Aerodynamics of a Lifting System

in Extreme Ground Effect

Springer-Verlag
Berlin
Heidelberg
GmbH

ONLINE LlBRARY
Physics and Astronomy
http://www.springer.de/phys/
Kirill V. Rozhdestvensky

Aerodynamics
of a lifting System
in Extreme Ground Effect
With 135 Figures

Springer
Professor Kirill V. Rozhdestvensky
Staint-Petersburg State Marine Technical University
Lotsmanskaya 3
190008 Saint-Petersburg
Russia
E-mail: xmastllinfopro.spb.su

Cover picture: Alexeyev's flying wing design concept as depicted by the artist Konstantin L. Vassiliev

Library ofCongress CataJoging-in-Publication Data

Rozhdestvensky, Kirill V., 1945-
Aerodynamics of a lifting system in extreme ground effect / Kirill V. Rozhdestvensky.
p.em.
Ineludes bibliographieal referenees.
ISBN 978-3-642-08556-7 ISBN 978-3-662-04240-3 (eBook)
DOI 10.1007/978-3-662-04240-3
1. Ground-effect machines-Dynamics. 2. Lift (Aerodynamies)

TL574.G7 R69 2000

629.3--dc21
00-026910

ISBN 978-3-642-08556-7

This work is subject to copyright. AlI rights are reserved, whether the whole or part of the material
is concerned, specifica1ly the rights of translation, reprinting, reuse of illustrations, recitation, broad-
casting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of
this publication or parts thereof is permitted only under the provisions of the German Copyright Law
of September 9, 1965, in its current version. and permission for use must aIways be obtained from
Springer-Verlag. Violations are liable for prosecution under the German Copyright Law.

© Springer-Verlag Berlin Heidelberg 2000

Qriginally published by Springer-Verlag Berlin Heidelberg New York in 2.000
Softco~ reprint of the hardcover 1st edition 2.000

The use of general descriptive names. registered names. trademarks. etc. in this publication does not
imply. even in the absence of a specific statement. that such names are exempt from the relevant pro-
tective laws and regulations and therefore free for general use.
Typesetting by the author.
Camera-ready by Steingraeber Satztechnik GmbH Heidelberg.
Cover design: Erich Kirchner. Heidelberg
Printed on acid-free paper SPIN 10694720 57/3144/mf 5 4 3 2 1 o
Preface

This book is dedicated to the memory of a distinguished Russian engineer,

Rostislav E. Alexeyev, who was the first in the world to develop the largest
ground effect machine - Ekranoplan. One of Alexeyev's design concepts with
the aerodynamic configuration of a jlying wing can be seen on the front page.
The book presents a description of a mathematical model of flow past a
lifting system, performing steady and unsteady motions in close proximity
to the underlying solid surface (ground). This case is interesting for practical
purposes because both the aerodynamic and the economic efficiency of the
system near the ground are most pronounced. Use of the method of matched
asymptotic expansions enables closed form solutions for the aerodynamic
characteristics of the wings-in-ground effect. These can be used for design,
identification, and processing of experimental data in the course of developing
ground effect vehicles. The term extreme ground effect, widely used through-
out the book, is associated with very small relative ground clearances of the
order of 10% or less. The theory of a lifting surface, moving in immediate
proximity to the ground, represents one of the few limiting cases that can be
treated analytically.
The author would like to acknowledge that this work has been influenced
by the ideas of Professor Sheila E. Widnall, who was the first to apply the
matched asymptotics techniques to treat lifting flows with the ground effect.
Saint Petersburg, Russia
February 2000 Kirill V. Rozhdestvensky
Contents

1. Introduction.............................................. 1
1.1 Some Definitions .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 ABrief Reference on Wing-in-Ground-Effect Craft. . . . . . . . . . 1
1.3 Asymptotic and Numerical Methods
in the Aerodynamics of the Wings-in-Ground Effect ........ 4
1.3.1 On the Mathematical Modelling of the Aerodynamics
of a Lifting System in the Ground Effect ............ 4
1.3.2 Asymptotics for Ground-Effect Lifting Flow Problems. 7
1.3.3 Numerics: Euler Codes for Steady Flows ............ 17
1.3.4 Numerics: Euler Codes for Unsteady Flows .. . . . . . . .. 19
1.3.5 Numerics: Viscous Flows, Navier-Stokes Solvers ..... , 20

2. Problem Formulation for the Flow Past a Lifting Surface

in Proximity to aSolid Boundary . . . . . . . . . . . . . . . . . . . . . . . .. 23
2.1 Formulation of the Three-Dimensional
Unsteady Flow Problem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 23
2.2 Flow Below the Lifting System and Its Wake (Channel Flow) 25
2.3 Flow Above the Lifting System and Its Wake. . . . . . . . . . . . .. 28
2.4 Flows in the Vicinity of Edges ........................... 29
2.5 Matching of Flow Descriptions in Different Regions . . . . . . . .. 36
2.6 A Uniformly Valid Flow Description
and Aerodynamic Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . .. 43
2.7 A Limiting Mathematical Model
of the Extreme Ground Effect . . . . . . . . . . . . . . . . . . . . . . . . . . .. 45

3. The Linear Theory of a Lifting System

Moving Close to the Ground ............................. 47
3.1 Features of a Linearized Formulation and an Algorithm
of the Solution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 47
3.2 Two-Dimensional Steady Flow Past a Slightly Curved Foil. .. 51
3.3 A Wing of Small Aspect Ratio in a Steady Linearized Flow .. 58
3.4 Steady Flow Past a Wing of Arbitrary Aspect Ratio . . . . . . .. 63
VIII Contents

3.5 Harmonie Oscillations of a Thin Fon

in a Two-Dimensional Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 73
3.6 Three-Dimensional Unsteady Linearized Lifting Flows. . . . . .. 79

4. Nonlinear Flow Problems for a Lifting System

in the Extreme Ground Effect . . . . . . . . . . . . . . . . . . . . . . . . . . .. 85
4.1 A Curved Thiek Fon in a Two-Dimensional Ground Effect . .. 85
4.2 A One-Dimensional Flow Model for a Flying Wing
with Endplates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 95
4.3 Steady-State Solutions for Flow Past a Wing with Endplates. 101
4.4 Unsteady Flow Solutions for a Wing with Endplates ........ 112
4.4.1 A Nonlinear Unsteady Solution for Small Gaps
Under Endplates ................................. 112
4.4.2 Unsteady Perturbation of a Steady Flow ............ 114

5. Compressible Flow Past a Wing

in the Extreme Ground Effect ............................ 121
5.1 Channel Flow in a Compressible Fluid .................... 121
5.2 Steady Linearized Compressible Flow Past a Wing
of Finite Span. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 124
5.3 Compressible Linearized Unsteady Flow Past a Wing
of Finite Span .......................................... 127
5.4 Nonlinear Steady Compressible Flow Problem for a Fon ..... 133

6. The Influence of Endplates, Flaps, and Slots .............. 137

6.1 An Estimate of the Infiuence of Endplates ................. 137
6.2 A Lifting System with a Jet Flap
in the Extreme Ground Effect ............................ 146
6.3 Nonlinear Steady Flow Problem for a Fon with a Jet Flap ... 155
6.4 Steady Flow Past a Jet-Flapped Wing of Finite Span ....... 158
6.5 A Wing with a Rotating Flap ............................ 166
6.6 Slotted Wings in the Ground Effect ....................... 171

7. The Aerodynamics of a Lifting System

Near Curved Ground ..................................... 183
7.1 The Infiuence of Waves on the Aerodynamics
of a Lifting Surface ..................................... 183
7.2 An Estimate of the Acceleration of a Lifting Surface
in Flight over a Wavy Wall .............................. 200
7.3 The Aerodynamie Response of a Lifting System
to Aperiodic Perturbations .............................. 205
Appendix: The Motion of a Wing Over a Wavy Wall
at an Arbitrary Course Angle: Derivation
of the Unsteady Lift and the Moment Coefficients .......... 214
 Contents IX

8. Schematized Flow Models

for a Power-Augmented Lifting System ................... 221
8.1 Local Flow Problems for Analysis of PAR Mode ............ 223
8.1.1 Reentrant Jet Flow Near the Leading Edge .......... 224
8.1.2 Reentrant Jet Flow Near the Leading Edge
with a Defiected Tip .............................. 227
8.1.3 Local Separated Flow Past a Flap (Endplate) ........ 229
8.2 Unseparated Flows Near Leading and Trailing Edges ........ 231
8.2.1 Unseparated Coanda Flow Past a Leading Edge ...... 231
8.2.2 Coanda Flow of Finite Width
Past a Defiected Leading Edge . . . . . . . . . . . . . . . . . . . . . 234
8.2.3 The Model of Flow Past a Leading Edge
with a Winglet ................................... 237
8.2.4 Unseparated Flow Near a Trailing Edge with a Flap .. 239
8.2.5 Flow of Infinite Width Past a Trailing Edge
with a Flap ...................................... 241
8.3 Envelopes of the Efficiency of Power Augmentation
for a Scheme with a Reentrant Jet ........................ 244
8.3.1 Estimate of the Efficiency and the Thrust Recovery
of PAR Based on a Reentrant Jet Scheme ........... 245
8.3.2 Margins of Efficiency of PAR Based
on a Reentrant Jet Scheme ........................ 250
8.4 A Discussion of a Mathematical Model of PAR
with the Coanda Effect .................................. 257

9. The Aerodynamic Efficiency of a Wing

in the Extreme Ground Effect ............................ 263
9.1 Optimal Wing-in-Ground Effect .......................... 263
9.2 The Lift-to-Drag Ratio of a Wing in the Ground Effect ...... 269

10. Integral Formulations for Lifting Surfaces

in the Extreme Ground Effect ............................ 281
10.1 A Slightly Curved Foil in a Two-Dimensional Ground Effect . 281
10.2 A Foil with a Jet Flap in Proximity to the Ground ......... 283
10.3 A Wing of Small Aspect Ratio ........................... 284
10.4 Lifting Line(s) in Close Proximity to the Ground ........... 285
10.4.1 A Single Lifting Line in the Extreme Ground Effect ... 285
10.4.2 A Comment on "Span-Dominated"
and "Chord-Dominated"Extreme Ground Effects ..... 288
10.4.3 A Tandem of Lifting Lines
in the Extreme Ground Effect ...................... 289
10.5 Quadruplication of the Integral Equation for a Wing
of Finite Span. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
X Contents

11. Equations and the Stability of Motion

of a Lifting System in the Extreme Ground Effect ........ 295
11.1 Linear Equations of Longitudinal Dynamies. . . . . . . . . . . . . . . . 296
11.1.1 The Quartie Characteristic Equation ................ 297
11.1.2 The Quintie Charaeteristic Equation ................ 299
11.2 The Equations of Motion in the Extreme Ground Effeet ..... 300
11.2.1 Order Estimates and Assumptions .................. 300
11.2.2 Asymptotic Form of the Equations of Motion
for h -+ o........................................ 301
11.3 Statie Stability and "Binding" Near the Ground ............ 302
11.3.1 A Single Wing in the Extreme Ground Effeet ........ 306
11.3.2 A Tandem in the Two-Dimensional
Extreme Ground Effeet ........................... 311
11.3.3 The Degree of Binding of the Vehicle to the Ground .. 315
Appendix: Formulas for the Coefficients Bi(i = 1 ... 5) .......... 318

12. Simple Mathematical Models of Elastic

and Flexible Wings in the Extreme Ground Effect ., ...... 319
12.1 Evaluation of the Speed of Flutter of a Foil
Close to the Ground . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
12.2 An Aeroelastie Wing in the Extreme Ground Effeet ......... 323
12.3 A Flexible Wing - a Simple Theory of a PARAWIG ........ 328

Bibliography .................................................. 337

Index ......................................................... 349

1. Introduction

1.1 Some Definitions

This book discusses the aerodynamics of vehicles, that utilize the favorable
effect of the proximity to an underlying surface upon their performance. Al-
though this underlying surface may be not only land, but also water, snow,
or ice, it will be called ground.
In what follows, the ground effect is understood as an increase in the
lift-to-drag ratio of a lifting surface (a wing) moving close to the ground.
For an appropriately designed lifting surface, this phenomenon exhibits it-
self at distances from the ground less than the chord of the wing, but the
most advantageous range of ground clearances normally lies below 25% of the
chord. The term extreme ground effect, widely used throughout the book, is
associated with very small relative ground clearances under 10%. This latter
range is characterized by a significant increase in the efficiency of the lifting
system and is expected to be operational for the lifting systems of the next
generation of the craft.
Ekranoplan (a wing-in-ground-effect vehicle) can be defined as a vehicle
with an engine and heavier than air that is designed to fly close to an un-
derlying surface for efficient utilization of the ground effect. At present many
terms exist to designate such a craft, namely, ekranoplan, wing-in-ground-
(WIG) effect vehicle, wing-in-surface-effect ship (WISES), flaircraft, ground
effect machine (GEM), etc.

1.2 ABrief Reference on Wing-in-Ground-Effect Craft

The earliest, albeit unintentional, use of the ground effect in technology is

probably due to the Wright brothers whose gliders covered longer distances
when f:l.ying close to the sands. Later, in 1910, some pilots reported a pecu-
liarly "different feel," when their aeroplanes f:l.ew closer to the ground. This
phenomenon, referred to in the literature as a "cushioning effect," sometimes
led to a sudden loss of lift. Nonetheless, the inf:l.uence of the ground was re-
vealed to be something more than just aerodynamic nuisance after a 56 t
Dornier DO-X seaplane increasing its payload and range when f:l.ying near

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
2 1. Introduction

the water during its transatlantic service in the early 1930s. It is probable
that this evidence triggered a purposeful engineering effort to conceive a craft
to take advantage of the ground proximity.
From 1935 a Finnish engineer, Kaario, had been working on aseries
of his "Aerosledge" vehicles, which were basically a low-aspect-ratio wing
mounted on skis. Kaario's craft showed pitch instability, which was later on
recognized as an inherent feature of wing-in-ground-effect vehicles. Following
Kaario's work, many projects of large ground effect machines and real, albeit
moderately sized, vehicles appeared; see Belavin [1], Ollila [2], Hooker [3],
Rozhdestvensky [4).
The list of projects of large craft includes, for example, Weiland's "Large
Weilandcraft," "Columbia" of the Vehicle Research Corporation, Boeing's
"Lowboy," Grumman's missile ekranoplan, SETOL by McDonnell Douglas,
Bertin's "Signe-14" and others.
Among the moderately sized vehicles built one can name Thoeng's "Aer-
oboat," Bertelson's GEMs, Kawasaki KAG series, the Lippisch Aerofoil Boat
and their derivative "Airfish" [5,6), Jörg's Tandem Flairboats [7, 8], China's
RAM and PAR vehicles [9, 10], Kubo's j.t-Sky Sliders developed with the
support of Mitsubishi [11, 12], more recent two-seater vehicles "Hoverwing
2-VT" of Fischer Flugmechanik [13, 14], and "Hydrowing VT-Ol"of Techno-
trans [15, 16).
Starting from the 1960s Russia took the world's lead in creating large
ekranoplans, [17)-[21). This became possible due to the outstanding effort of
Rostislav Alexeyev and his famous Hydrofoil Design and Construction Bureau
in collaboration with other leading institutions and enterprises. During the
following two decades a number of ekranoplans had been built with takeoff
weights of 120 through 550 t and a cruise speed in the range of 350-500 km/ho
The largest of these was and still is the KM (ship prototype), dubbed in the
West, the Caspian Sea Monster. A collection ofRussian wing-in-ground-effect
vehicles is shown in Fig. 1.1. It reflects the evolution of design concepts and
building technology for ekranoplans. Shown in Fig. 1.1 are

• 1 - the tandem scheme borrowed from the hydrofoil design and developed
by Alexeyev in the early 1960s.
• the first-generation vehicles, based on an aeroplane type aerodynamic
configuration with a fuselage, a highly elevated tail unit, and a power-
augmented takeoff feature: 2 - KM, 1967, designed by Alexeyev; 3 - Orly-
onok, 1972, designed by Alexeyev and Sokolov [22); 4 - LoonjSpasatel, 1987,
designed by Kirillovykh [20). One should note that the highly mounted tail
surface in the aeroplane scheme is intended to secure (pitch) stability to
longitudinal motion.
• the concepts and projects of a new generation of ekranoplans with ftying
wing and composite wing configurations: 5 - Alexeyev's flying wing con-
cept (late 1970s); 6 - composite wing configuration (Alexeyev, 1970s); 7 -
Bartini's contact free takeoff (landing) craft with a composite wing aero-
 1.2 ABrief Reference on Wing-in-Ground-Effect Craft 3

3 4

5 6

7 8

9 10

Fig. 1.1. A Collection of Russian Ekranoplans in Concepts and Vehicles.

4 Introduction

dynamie scheme [23]; 8 and 9 - Marine Passenger Ekranoplans MPE 400

and MPE 200, designed by Sinitsyn [24].
It should be added that pitch stability for the tailless fiying wing config-
uration 5 can be ensured by special profiling of wing sections and the use
of an automatie motion control system. The composite wing scheme (6-9)
employs the idea of a functional sub division of the vehicle's lifting area into
two parts. The central wing of the system enables taking full advantage
of power augmentation (blowing under the wing) at takeoff. Side wings
add efficiency and, if properly positioned, longitudinal stability in cruising.
Further elaboration of the composite wing scheme can be achieved by pro-
filing wing sections, whieh serves to reduce the tail unit and, consequently,
empty weight fraction .
• A smaller vehicle (10; 1996, Sinitsyn) features blowing under the wing
(power augmentation) both during takeoff (landing) and cruising. This
craft known by the name Amphistar is a six-seater with a cruising speed of
150 kmjh [25]. Other examples of the family of unconventional ekranoplans
with permanent power augmentation features are Volga-2 and Raketa-2
(project), developed in the Central Hydrofoil Design Bureau named after
R.E. Alexeyev.

1.3 Asymptotic and Numerical Methods

in the Aerodynamics of the Wings-in-Ground Effect

1.3.1 On the Mathematical Modelling of the Aerodynamics

of a Lifting System in the Ground Effect

A rational approach to the design of any unconventional vehicle, for which

restrieted prototype data exists, should be based on an appropriate mathe-
matieal model, refiecting the essential features of the craft under considera-
tion.
Ekranoplan can be viewed as such an unconventional type of superfast
water transport, utilizing the favorable infiuence of the underlying surface
(ground) upon its lift-to-drag ratio and, consequently, on its economic effi-
ciency, expressed in terms of fuel consumption and direct operating costs.
The ground effect is most pronounced when the main lifting system of
the ekranoplan operates at distances of the order of 25% of the chord or less
from the underlying surface. 1 From the viewpoint of mathematical modelling
this means that the corresponding fiow problem involves aleast one distinct
small parameter - relative ground clearance.

1 This means that a cost-effective ocean-going vehicle capable of handling high seas
should be sufficiently large. On the other hand, a vehicle of small dimensions
is expected to be efficient on inland waters, where high seaworthiness is not
required.
 Asymptotic and Numerical Methods 5

Ekranoplan operates in a wide range of modes, including subsonic cruise

flight in the presence of wind and wave perturbations, transient motions at
takeoff and landing accompanied by intensive blowing of air under the main
lifting surfaces, and floating on water, when the structure of the craft is
subject to action of hydrostatic and wave-induced bending moments.
It is worth noting that in the cruising mode of operation in proximity
to the ground, the qualities of such a transport in terms of the lift-to-drag
ratio, the static and dynamic stability, the steerability, and the ride comfort
are extremely sensitive to the design decisions adopted (aerodynamic config-
uration, geometry of lifting surfaces, takeoff and landing devices, cruise and
power augmentation engines, automatie control system).
The mathematical model employed for the purpose of conceptual and
preliminary design should meet certain requirements, in particular
• be sufficiently similar to a real system and include most essential factors;
• seeure the possibility of fast, inexpensive, and interactive evaluation of the
quality of the system;
• be tailored for the application of optimization procedures concerning the
economic vi ability of a transportation system, reserves of stability, etc.;
• be acceptable for prediction of the behavior of the system in extreme sit-
uations (risk assessment).
To work out an appropriate tool kit for the aerodynamic design of ekra-
noplans, one should be careful in choosing and applying existing mathemat-
ical technologies.
At present, two main approaches can be considered:
1. Numerical methods, based on advanced techniques of numerical solu-
tion of boundary problems and integral equations, describing a corresponding
physical process. In most cases, these approaches do not impose restrictions
on the geometry, kinematics or dynamics of the system under modelling. Al-
though many of numerical approaches and algorithms are essentially heuristic
and are not backed by proofs of existence, uniqueness, or convergence of the
solution, they often lead (when applied by an expert) to mathematically
reasonable and physicaHy adequate results. At the same time, from the view-
point of practical design, numerical approaches have certain inconveniences
and shortcomings, especiaHy when used at a preliminary stage of design.
These methods
• require powerful computers and considerable epu time;
• do not provide explicit representation of the structure of the system's re-
sponse, which often prompts designers to a path to successful project de-
cisions,
• are not weH suited for use in optimisation procedures at preliminary stages
of design, which is to a large extent related to the so-called "curse of di-
mension;"
6 Introduction

• in certain cases of practical interest (thin and slender bodies, flows in small
gaps, extreme variations of geometry, high Reynolds numbers, etc.) exhibit
unsatisfactory convergence and computational instability.
2. In cases when the flow problem is characterized by one or several small
parameters (which often happens in practical situations), one can employ, as
an alternative, asymptotic approaches, which also have pros and cons.
The disadvantages of asymptotic approaches include substantial restric-
tions on geometry and magnitudes of kinematic and dynamic parameters;
nonuniversal character, which entails the necessity to adapt methods for a
concrete problem; the need to use rather sophisticated mathematics to de-
velop solutions and obtain final closed form results.
Nevertheless, from designer's and system analyst's points of view, these
methods possess some useful properties.
• They are conveniently adapted to consideration of a concrete problem of
interest, which lays the ground for development of multifactor cost-effective
mathematical models of the system under design;
• They often provide closed analytical formulas and fast algorithms, which is
important for the preliminary design stage, as well as for feasibility studies
dedicated to evaluation of the quality of a large system;
• They serve to reveal similarity criteria, which are useful for the purpose of
design, the processing of experimental data, and identification procedures.
The method of matched asymptotic expansions used in this book to treat
the aerodynamics of the ekranoplan at very small relative ground clearances
implies separate consideration of the flow in subregions, characterized by
different length scales with subsequent "blending" of corresponding solutions.
This formalism not only leads to simplification of flow problems in these
subregions, but also makes the process of derivation of the solution similar
to that of design analysis and synthesis.
It is clear that none of the two approaches (numericalor asymptotic)
discussed ab ove , can be seen as a perfect tool for design. In this connection,
at least two strategies may be proposed for expedient use of these methods:
• Use of the asymptotic approach for the purpose of conceptual and prelim-
inary design of a craft (for example, for the selection of basic aerodynamic
configuration and assessment of expected operational behavior of the sys-
tem) with subsequent application of numerical techniques and computer
codes for the final elaboration of the vehicle;
• Rational synthesis of both numerical and asymptotic methods, implying
utilization of all useful properties and the results of the application of
asymptotics in computational procedures.
In what follows, abrief survey is presented of the use of asymptotic and
numerical methods in the aerodynamics of ekranoplans. This survey by no
means pretends to be complete and refers mostly to research work published
after 1970.
 Asymptotic and Numerical Methods 7

1.3.2 Asymptotics for Ground-Effect Lifting Flow Problems

The possibility of deriving solutions of the flow problem and corresponding
aerodynamic coefficients of a lifting system in a closed form is connected with
analytical and asymptotic approaches.
Analytics in a classical sense of the word is mostly related to two-
dimensional steady-flow problems, involving ground-effect phenomena. To-
motika et al. [26] obtained an exact solution for a problem of steady flow past
a flat plate at an arbitrary angle of attack in the presence of a ground plane.
Therewith, the formulas for the aerodynamic coefficients were expressed in
terms of elliptic theta functions. Later on these results were extended to a
Zhukovsky type aerofoil with consideration of the case, when the trailing edge
of the foil was touching the ground; see Tomotika and Imai [27]. The latter
case was also treated by Dätwyler [28].
Sedov [29] derived the same results as Tomotika et al. for a flat plate more
str aightforwardl y.
The first asymptotic approaches relevant to ground-effect phenomena em-
ployed Prandtl's lifting line model and its "mirror reflection," e.g., Wiesels-
berger [30], Serebriyskiy [31], etc. In this research, the distance of the wing
with a large aspect ratio from the ground was considered to be of the order
of the span, whereas the chord of the wing was assumed much smaller than
both the span and the ground clearance.
Because the ground effect is most pronounced at very sm all relative dis-
tances h from the underlying surface, 2 it is reasonable to seek an approximate
solution of the corresponding flow problem in the form of an asymptotic ex-
pansion in terms of a small parameter related to h.
Some of the earlier approaches were based upon asymptotic expansions
with respect to a small parameter, which is inversely proportional to the
ground clearance. Keldysh and Lavrent'ev [32] applied parameter l/h to treat
the flow past a hydrofoil, moving near a free surface. A similar expansion was
used by Plotkin and Kennel [33] to obtain the lift coefficient of an arbitrary
thin aerofoil in the presence of a ground plane and by Plotkin and Dodbele
[34] and Plotkin and Tan [35] for solving the flow problems for large-aspect-
ratio wings and slender wings in motion near a solid wall.
It is obvious that an expansion in l/h is appropriate for investigations of
the ground-effect phenomena at distances from the boundary, which are larger
that the chord of the wing. However, because wing-in-ground-effect vehicles
normally operate at distances below 25% of the chord of the main lifting
surface, it is practical to introduce a small parameter providing convergence of
the solution se ries at clearances less than the chord. Panchenkov [36] obtained
asymptotic solutions of a set of lifting flow problems involving interfaces (free
surface and solid wall) in terms of the so-called T-parameter, related to the
relative ground clearance by the following equation:
2 This parameter represents the ratio of the distance of a wing from the ground to
the chord of the wing and later on will be called the relative ground clearance.
8 Introduction

T = VI + 4 h2 - 2h. (1.1)
It is easy to see that (1.1) transforms a semi-infinite interval of the actual
clearance 0 ~ h ~ 00 to a unit interval ofvariation ofthe T-parameter 1 ~ T ~
O. In fact, these solutions were developed as perturbations of the unbounded
fluid case (h = 00) and yielded the results of Keldysh and Lavrent'ev for
large h. At the same time, the introduction of the parameter T 3 noticeably
enlarged the range of validity of the asymptotic solution.
Because wing-in-ground-effect vehicles have maximum aerodynamic effi-
ciency in very close proximity to the ground (Le., at distances essentially less
that the chord andjor the span), it is challenging to develop an asymptotic
expansion of the flow problem solution around the limiting case h = 0, rather
than h = 00. However, any attempt to develop a straightforward (outer or
pedes tri an; see Nayfeh [37]) asymptotic expansion of the solution for h -t 0
results in a degeneracy of the flow problem. In fact, from the viewpoint of
an outer observer, Le., at distances of the order of the chord from the lifting
surface, the gap between the wing and the ground vanishes. Consequently,
within the boundary problem formulation it be comes impossible to satisfy
the flow tangency condition on the lower surface of the wing and on part of
the ground plane under the wing. In other words, in the outer limit, the chan-
nel fiow between the lifting surface and the ground is lost. If one stretches
the vertical coordinate to retain the gap when h goes to zero, the governing
Laplace equation can be shown to lose one dimension, so that the description
of the flow above the wing and the ground becomes incomplete. Thus, the
problem under consideration definitely displays features of a singular pertur-
bation problem, for which neither an outer nor inner asymptotic expansion is
uniformly valid throughout the whole flow field; see Van-Dyke [38]. The flow
problem for a wing-in-ground effect is characterized by the "coexistence" of
two characteristic length scales (on the one hand, the ground clearance and,
on the other hand, the chord or the span of the wing), the ratio of which
vanishes when the small parameter h goes to zero. Such a problem can be
handled by the method of matched asymptotic expansions (MAE).
Apparently, Strand, Royce, and Fujita [39] were the first to indicate the
channel nature of highly constrained flow between the wing and the ground.
They stressed the point that in the two-dimensional case the channel flow
becomes one-dimensional. However, no method was presented then to deter-
mine the total amount of mass flow under the wing without solving the entire
flow problem.
The first MAE applications for lifting flows near the ground were intro-
duced by Widnall and Barrows [40] within the framework of linear theory.
Their basic linearizing assumption was that both geometrie (curvature and
thickness) and kinematic (angle of pitch) perturbations are small compared
with relative ground clearance h. It was shown that for a thin wing at an
3 It can be shown that an infinite number of regular parameter expansions can be
derived around h = 00, providing different ranges of validity with respect to h.
 Asymptotic and Numerical Methods 9

angle of pitch 0, the vertical perturbed flow velocities are of the order of
0(0), whereas in the confined region under the wing, the same vertical flow
ind uces a horizontal velo city of the order of 0 (0/ h). Therefore, the response
of the lifting system to perturbations of the same magnitude is amplified
in the ground effect compared to the out-of-ground-effect case. It was also
demonstrated that for a three-dimensionallifting flow in an extreme ground
effect, the corresponding channel flow problem becomes predominantly two-
dimensional in a horizontal plane parallel to the ground. Thus, as indicated by
Widnall and Barrows, the extreme-ground-eJJect theory forms an interesting
complement to Prandtl's lifting line theory and Jones's slender body theory,
in which fiow fields are basically two-dimensional in the transverse and lon-
gitudinal planes respectively. Detailed calculations were carried out with an
asymptotic error of the order 0(h 2 ) for a steady linearized flow around a
flat plate near a flat solid wall. As an example of three-dimensional flow,
the same authors obtained the leading order results for a flat plate near a
wall. A semielliptic flat wing elose to the ground was shown to have con-
stant induced downwash in the wake for all aspect ratios and, consequently,
to have minimal induced drag for a given lift. The spanwise distribution of
aerodynamic loading for this optimal wing turned out to be parabolic rather
than elliptic as in the unbounded fluid case. The problem of minimization of
induced drag for a range of ram wing transportation vehieles for a variety of
guideway configurations and small relative ground elearances was discussed
by Barrows and Widnall [160].
Extension of the MAE approach, advocated by Widnall and Barrows [40],
to a linear unsteady flow case accounting for flaps, endplates and compress-
ibility effects was carried out by Rozhdestvensky [41]-[48]. Due to the fact
that unsteady lifting flow exhibits a finite jump of both perturbed velo city
and velo city potential across the trailing edge, the matching of corresponding
local flow with the outer and channel flows was performed in terms of pres-
sure. Unsteady aerodynamic derivatives were found in a closed analytical form
for a rectangular wing of arbitrary aspect ratio for various unsteady motions
versus Strouhal and Mach numbers. It was shown that compressibility effects
give rise to some qualitatively new phenomena, which can be characterized as
acoustic resonance. On the basis of the asymptotic theory, it was also shown
how, to determine suction force, acting upon the rounded leading edge of the
wing. In Rozhdestvensky [44], elosed form expressions were presented for an
induced drag coefficient of a rectangular wing in steady and unsteady motion
near the ground with and without leading edge suction. It was demonstrated
that, depending on parameters of contributing vertical motions (heave, pitch,
deformations) drag or thrust force enhanced near the ground may act. Based
on these results, special propulsion systems can be designed, incorporating
oscillating wings and wing systems near solid boundaries. Essentially, the
idea was utilized of employing the MAE as a mathematical constructor by
way of matching appropriate asymptotic solutions, corresponding to different
10 Introduction

local geometry of the wing (forward and rear flaps, endplates, slots) andjor
physical peculiarities of performance of the edges (jet and rotor flaps and
endplates, shock or shock-free entry at the leading edge, vortex wake detach-
ment and roIl-up) to the main flow solution. A theory of the slotted wing
in the extreme ground effect was developed for both lateral and longitudinal
gaps on the wing's surface in steady and unsteady cases, featuring special
solutions for the flow in the immediate vicinity of the gap accounting for dif-
ferent possible flow patterns, as weIl as mass and vorticity transport through
the gap.
Kida and Miyai [49] applied the MAE approach to solve the flow prob-
lem of a nonplanar wing of finite span in motion very dose to the ground.
They provided a simple first-order analytical solution for a nonplanar (un-
cambered in the chordwise direction) wing of small aspect ratio in the form
of apower series of the ratio of semispan to the wing's root chord. The lift
and induced drag coefficients were discussed for the case when the chord dis-
tribution spanwise was semielliptic. The authors also calculated the lateral
(restoring) moment, acting on an indined wing of semielliptic planform.
Kida and Miyai [50] were the first to extend the approach based on the
method of matched asymptotic expansions to the case of a steady linearized
flow around jet-flapped wings in very dose proximity to the ground. They
assumed that both the angle of pitch and the angle of deflection of the jet,
emerging through a thin slot in the wing's trailing edge, are small compared
to the relative ground dearance. The Kutta-Zhukovsky condition of smooth
detachment of the flow at the trailing edge was replaced by a condition of jet
injection at a prescribed angle to the horizon. To describe the form of the jet,
the corresponding dynamic and kinematic conditions were introduced. These
require, respectively, that the pressure jump across the jet surface should be
proportional to its longitudinal curvature and that the said surface should
comply with the flow tangency condition. Concrete results were obtained
by these authors for a steady two-dimensional flow past a jet-flapped flat
plate and a semielliptic flat wing in the ground effect. Kida and Miyai also
found the optimal distribution of the jet momentum for which a semielliptic
wing in the extreme ground effect has the minimum induced drag for a given
lift. Rozhdestvensky [44] used the same technique to derive a solution for a
rectangular jet-flapped flat wing and calculated the optimal jet moment um
distributions for different aspect ratios.
It should be underlined that at very small relative ground dearances even
slight changes in geometry and kinematics of the lifting system may result
in considerable perturbations in the channel flow under the wing and, con-
sequently, in the aerodynamic response of the system. Therefore, the theory
should account for the nonlinearity of the ground-effect lifting flow with re-
spect to perturbations. Note that in linear theory it is assumed that vertical
displacements of points on the wing's surface are small compared with ground
dearance. For example, the angle of pitch (in radians ) is small with respect
 Asymptotic and Numerical Methods 11

to relative ground clearance. In nonlinear approaches this restrietion has to

be alleviated.
For the case of the extreme ground effect (leading order) in a compressible
isentropic lifting fiow, an unsteady, nonlinear three-dimensional treatment of
the problem was given by Rozhdestvensky [41, 42J. It was assumed that the
wing is advancing at constant speed in the vicinity of a curved ground and,
at the same time, is free to perform small vertical motions and deformations.
Displacements of points of both surfaces of the wing and the ground were
assumed comparable with the ground clearance. Mathematical model of the
fiow also included vertical gusts. A two-dimensional governing fiow equation
was derived from the mass conservation considerations for the constrained
fiow domain under the wing rather than from the formalism of the method
of matched asymptotic expansions. Boundary conditions at the edges of the
channel fiow region were obtained through application of simple physical
arguments such as continuity of pressure at the trailing edge and continuity
of the velo city potential at the leading and side edges. In the incompressible
fiow limit (Mach number equal to zero), the channel fiow equation was shown
to yield a quasi-harmonie equation, which differs from the well-known shallow
water equation only in the sense that the distribution of ground clearance -
both chord and spanwise - is prescribed.
Later, a leading-order nonlinear formulation of the problem was developed
by Thck 4 for a two-dimensional (1980, unsteady [52]) and three-dimensional
(1983, steady [53]) incompressible fiows and by Newman [54J for a lifting
surface of small aspect ratio. In Thck [52], a one-dimensional channel fiow
problem was reduced to a nonlinear first-order differential equation in time
variable as a variable. As an illustration of unsteady fiow theory, the problem
was solved for a fiat plate, driven toward the ground by its own weight and
advancing at the same time in a direction parallel to the ground at constant
speed.
Newman [54J was able to represent the channel fiow beneath the lifting
surface by a simple nonlinear solution in a cross-fiow plane with appropriate
conditions imposed at the trailing and leading (side) edges. A distinguished
feature of Newman's work was that the location of the point separating the
leading and trailing edge parts of the wing's planform contour had to be
determined as part of the solution. It should be mentioned that within the
linear formulation, the transition between the leading and trailing edges is
normally supposed to occur at the wing's tips. According to Newman, the
transition point for the steady fiow around a delta wing is located upstream
of the traling edge for sufficiently large angles of pitch. For high rates of heave
motions of the delta wing, the position of the transition point was shown to
be cyclic.

4 E.Tuck seems to have been the first to introduce the term extreme ground effect,
widely used in this bookj see also Read [51].
12 Introduction

Thek [53] eonsidered a nonlinear , extreme ground-effect lifting fiow for

thin wings of arbitrary aspect ratio in assoeiation with some problems of
aerodynamies of raeing ears. He performed numerical ealculations for a par-
tieular ease of a cireular planform with exponentially varying clearanee. For
this partieular class of clearanee distribution, it beeame possible to reduee
the ehannel fiow equation to a Helmholtz type equation and, using separation
of variables, to find eorresponding fundamental solutions in the form of mod-
ified Bessel functions of the first kind. Unknown eonstants of the solution,
as well as the position of the transition point, were determined numerieally
through the applieation of boundary eonditions at the leading and trailing
edges of the wing. The transition point loeation was found to shift upstream
from the point of maximum loeal span at positive angles of piteh.
Systematie analysis of nonlinear unsteady three-dimensional fiow past a
lifting system of finite thickness in a eurved ground effeet was earried out in
Rozhdestvensky [44, 48, 55]. The angle of piteh, maximum relative thiekness
and eurvature, amplitudes of oseillations of the wing, as asolid body and
preseribed deformations of its surfaee as well as the displaeements of points
of the underlying surfaee, were assumed to be small quantities of the order
of relative ground clearanee. In the general ease, the main translational mo-
tion of the wing parallel to the ground was supposed to be time-dependent.
It was formally demonstrated that, if longitudinal and lateral slopes of the
upper surfaee and slopes of the ground are small quantities of the order of
h, the outer fiow ean still be linearized and eonstrueted by a straightforward
distribution of sourees and sinks on the projeetions of the wing and the wake
on the unperturbed horizontal position of the ground plane with appropri-
ate admissible distributions of singularities along the planform eontour. The
strengths of singularities were determined direetly through a thin body the-
ory with the use of a known first-order downwash in the wake. The latter
was obtained through the leading-order ehannel fiow solution and Thom-
son theorem on eonservation of vortieity in the wake. The strength of the
souree-sink eontour distribution was obtained by matehing the outer fiow to
the ehannel fiow through edge fiows. Loeal (edge) fiow solutions were shown
to be two-dimensional in the planes normal to the wing's planform eontour
if the radius of eurvature of this eontour is large eompared with the ehar-
aeteristic ground clearanee. The latter cireumstanee signifieantly simplifies
the proeess of developing a loeal solution through the applieation of pow-
erful teehnique of eomplex variable analysis. Matehing was performed with
respeet to the veloeity potential at the leading and side edges and with re-
spect to the press ure at the trailing edge. Wherever possible, the ehannel
fiow equation was solved in a closed form. Otherwise, a numerical teehnique
is applied, based on a simplified mathematical model for the fiow under the
wing, or furt her perturbation parameter expansions were earried out to in-
clude nonlinear terms. The effects of thiekness and eurvature were studied. It
was shown, in partieular, that lifting and thickness effeets are not separable
 Asymptotic and Numerical Methods 13

in extreme proximity to the ground and that the dominating (channel flow)
contribution is determined by the configuration of the lower surface of the
wing-in-ground effect. The influence of waves on the underlying surface was
studied for both cases of translational motion of the wing in the direction
normal to the wave front and for an arbitrary course angle in Rozhdestven-
sky [56, 57], respectively. Another solution of the former problem was also
given by Efremov and Lukaschik [58].
A similar theory was published for a particular case of an incompressible
flow around a thin lifting surface in the curved ground effect by Qian-Xi
Wang [59]. In this work translational motion of the wing along the ground
was assumed to take place at constant speed. Based on the fact that the input
of the channel flow equations involved the instantaneous distribution of elear-
ance rather than ordinates of the wing's lower surface and that of the solid
boundary, the author formulated an equivalence rule between the extreme
curved ground effect and flat ground effect with an appropriate downwash on
the wing. No calculated results were presented.
As indicated previously, due to the dominating influence of the flow be-
tween the lower surface of the wing and the ground upon the aerodynamics
of the lifting surface in the extreme ground effect, the corresponding three-
dimensional flow problem can be reduced to that in two dimensions in the
planes parallel to the unperturbed position of the underlying surface. Fur-
ther simplification can be introduced for a wing with endplates moving in
elose proximity to the ground. In this case, the flow description can be shown
to be predominantly one-dimensional. A simple one-dimensional nonlinear
mathematical model of the flow past a rectangular wing with small relative
elearances under the tips of the endplates was derived and then validated
experimentally by Gallington et al. [60, 61]. This approach was based on an
assumption that (channel) flow parameters are independent of the chordwise
coordinate and on the observation that the leaking fiow escapes from under
the tips of the endplates into the external region with atmospheric pressure.
The author also assumed the occurrence of separation at the tips of the
endplates. Though very simple, Gallington's model agreed qualitatively with
experiments and provided interesting similarity criteria. An important conse-
quence of the introduction of this model from the theoretical standpoint was
that it helped to overcome a paradox of the infinite (logarithmic) increase of
the flow velo city at the gap, encountered by other researchers; see Ando [62].
One of the restrictions of Gallington's one-dimensional model ensues from
the assumption of the constancy of the loading along the chord. As a con-
sequence the model cannot be used for prediction of the longitudinal mo-
ment and the characteristics of stability. Secondly, it is confined to the case
of steady motion, whereas the analysis of the transient motion of wing-in-
ground-effect vehieles is of utmost importance.
14 Introduction

Rozhdestvensky [63] extended Gallington's one-dimensional nonlinear

mathematical model of channel flow, taking into account the chordwise vari-
ation of flow velocity (pressure) and introducing unsteady effects.
Certain attention should be attached to integral formulations. The re-
seach, referred to so far, was carried out by solving a boundary problem for
a wing near the ground. In principle, the same results (at least to the lead-
ing order) can be obtained by asymptotic treatment of the integral equation
of the lifting surface moving elose to asolid boundary. A singular integral
equation for the bound vorticity of a lifting surface in the extreme ground
effect can be readily derived by using the Biot-Savart formula and the mir7'or
image technique. The kernel of this equation reduces to the elassicallifting
surface kernel for an unbounded fluid as the elearance-to-chord ratio tends
to infinity. In this case, it is not difficult to develop a regular perturbation
expansion of the part of the kernel due to the image vortex system in terms
of an appropriate parameter, tending to zero when h --+ 00. As a result, one
would derive a corresponding regular perturbation solution around the infi-
nite fluid case (h = 00). On the other hand, if one lets the elearance-to-chord
ratio tend to zero, the regular part of the kernel can be shown to cancel
out with the singular part of the kernel. Thus, in the limit of the extreme
ground effect, the kernel of the integral equation vanishes. The limiting form
of the equation, however, can be shown to be nontrivial. If expansion of the
integrand in h is performed properly, the integral equation can be shown to
degenerate into a differential equation. It is not a surprise to discover that
this differential equation is identical to the channel flow equation derived in
the process of solving the boundary problem.
The integral formulation approach was applied to the investigation of a
wing-in-ground effect by Panchenkov [64], Kida and Miyai [65] and Rozhdest-
vensky [66]. Panchenkov [64] was the first to study the limiting degeneracies of
the integral equation of the lifting surface in the ground effect. Assuming that
the relative ground elearance h goes to zero, he constructed an asymptotic
representation of the solution of this equation. The corresponding theory was
dubbed the "quadruple theory." The term quadruple is quite representative
because in the limit of vanishing ground elearance the two double layers,
namely, the dipole sheets of the main lifting surface and its mirror image
with respect to the ground plane approach each other and, eventually, form
a nontriviallimiting "quadruple layer." To the leading order, the quadruple
theory gives results identical to the results of matched asymptotics. Higher
orders are incomplete due to the fact that no account was taken of asymptotic
expansions of the governing integral equation of the wing-in-ground effect
near the leading, side, and trailing edges. The contributions of these local
descriptions to the aerodynamic characteristics are negligible in the extreme
ground effect (leading-order solution), but become more substantial when the
clearance increases to 5% of the chord or more.
 Asymptotic and Numerical Methods 15

Kida and Miyai [65] developed what they called "an alternative analyti-
cal method for ground-effect aerofoils," based on asymptotic treatment of a
corresponding integral for a steady two-dimensional case. They included in
their treatment a flow around a foil near the ground in a somewhat more gen-
eral situation of a nonparallel stream, a flow problem for an airfoil between
parallel walls, the case of an airfoil in a free jet, and the case of an airfoil in a
slipstream. For a particular case of a foil near the ground in a uniform stream,
the authors applied a special technique of expansion of the integrand, which
becomes nonuniform when the difference (x - x') (where x is the abscissa
of a control point, x' the abscissa of the loading point) is comparable to the
ground clearance. In the vicinities of edges, the "local" integral equations on
a semi-infinite interval were derived. Finally, it was demonstrated that the
expression for the lift coefficient obtained by the authors is identical to that
of Widnall and Barrows [40] if truncated to the same order.
Rozhdestvensky [66] utilized the concept of "inner and outer" contribu-
tions to the integral equation and demonstrated for both two-dimensional
and three-dimensional steady flow cases that at small relative clearance the
integral equation degenerates into differential equations, identical to those
obtained through the matched asymptotics treatment of the boundary prob-
lem. He used this algorithm to determine the leading-order lift coefficient for a
flat plate in a two-dimensional flow, small-aspect-ratio wing, and jet-flapped
flat plate. The asymptotic description of the edge flow resulting from this
formulation was shown to reduce to an integral equation on a semi-infinite
interval.
Rozhdestvensky [67] also gave an analysis of the flow past a lifting line
and a tandem, comprising two lifting lines, in the immediate proximity to the
ground. In the former case, for a vanishing clearance-to-span ratio, he was
able to reduce Prandtl's integrodifferential equation to a simple ordinary dif-
ferential equation of the second order for the distribution of loading spanwise.
In the latter case, a system of two integrodifferential equations degenerates
for vanishing hll (l is the ratio of the span to the chord) into a corresponding
system of ordinary differential equations of the second order. In both cases,
the solutions of the resulting differential equations, subject to conditions of
zero loading at the ti ps of the wing, were obtained in analytical form.
To conclude the part of the survey dedicated to contributions from asymp-
totic analysis of the integral equation for a lifting surface in the ground effect,
one should mention that this technique was extensively used by Efremov et
al. [68]-[72] under the name asymptotics of small clearances to investigate
the effect of the compressibility of the flow, the flexibility and elasticity of the
foil, and the response of a lifting system in the extreme ground effect to the
unsteady periodic (oscillations) and aperiodic (abrupt variation of the angle
of pitch, the influence of a step-type vertical gust) perturbations.
The larger the wing-in-ground-effect vehicle, the more effective it becomes
in both economic performance and seaworthiness. At the same time, the
16 Introduction

elasticity of the vehicle can no longer be neglected, and both theoretical and
experimental evidence exists that the static (divergence) as well as dynamic
stability (flutter) of an elastic wing-in-ground effect worsens as the ground
clearance decreases.
There are very few works connected with mathematical modelling of the
flow around an aeroelastic wing moving in close proximity to the ground.
The limiting flow model of the ground effect based on matched asymptotics,
in conjunction with appropriate mathematical models of an elastic structure
subject to action of aerodynamic loads, facilitates evaluation ofthe aeroelastic
behavior of the lifting surface in motion near an underlying surface. To the
author's knowledge, Efremov [70]-[72] was the first to use the asymptotics
01 small clearances to provide so me qualitative and quantitative prediction
of aeroelastic phenomena for schematized lifting flows in the extreme ground
effect. To relate elastic displacements of the wing's surface with aerodynamic
loads, he used the well-known equation of the unsteady bending of an elastic
plate accounting for longitudinal and lateral stresses in its mean plane and
the boundary conditions, corresponding to the mode of fastening the plate
at the planform contour. Using a linearized version of the aerodynamics of a
lifting surface in the extreme ground effect, he was able to derive a system of
coupled equations for the aeroelastic lifting surface moving in close proximity
to asolid boundary and to reduce this system to one differential equation
with respect to deformations. In particular, Efremov studied the case of a
two-dimensional foil. It was shown that the speeds of divergence and the
flutter of the lifting foil in the ground effect decrease as the ground clearance
diminishes.
Lifenko and Rozhdestvensky [73, 74] used the same approach to study the
aeroelasticity of a lifting surface of finite aspect ratio in the extreme ground
effect. To solve the corresponding combined aeroelastic equation, they used
the Bubnov-Galerkin method. The same asymptotics (in the relative ground
clearance) for the critical speeds (divergence and flutter), as in Efremov [71],
was found. Quantitative results were presented to describe both the static
and dynamic stability domains for an elastic lifting surface in the ground
effect in parametric space.
This review featuring the applications of the matched asymptotics tech-
nique to the aerodynamics of lifting systems in the ground effect would be
incomplete without abrief discussion of a cycle of Japanese works dedicated
to optimal ground-effect lifting systems with average clearance of the order
of the span and small gaps between the ti ps of the wing and the ground
plane. Kida and Miyai [75] and Mamada and Ando [76, 77] applied the MAE
approach to develop a ground-effect theory for a wing, having arbitrary front-
view shapes and vanishing gaps under the tips. The problem was reduced to
consideration of flow in the Trefftz plane and, therefore, holds for the case
when both the angle of pitch and the longitudinal camber are reasonably
small and, in addition, the spanwise distribution of the lift is optimal in
 Asymptotic and Numerical Methods 17

Munk's sense. The main idea of the approach is based on an observation

that, in the outer limit, the tip gap disappears and the flows above the wing
and under the wing can be considered separately, accounting for a possible
point sour ce (outside) and sink (inside) at the intersection of the wing's con-
tour in the Trefftz plane with the ground. The inner flow in the vicinity of
the gap under the tip of the wing and the ground was considered in stretched
variables and contains the limiting geometry of the tip (a semi-infinite flat
slit tangent to the tip) and the ground plane. The matching procedure en-
ables determination of both the local flow pattern and the strengths of the
outer source-sink flows, which model the leakage of the flow through the gap
at distances of the order of the span. Ando and Ando and Yashiro [78, 79]
introduced certain simplifications into the solution of Kida and Miyai [75].
In particular, they found more simple procedures to develop the outer solu-
tions. In addition, they carried out their formulation explicitly for arbitrarily
curved (laterally) wings, whereas Kida and Miyai confined their analysis to
concrete cases.

1.3.3 Numerics: Euler Codes for Steady Flows

Belotserkovsky's monograph "Thin Wings in Subsonic Flow" published in

1965 [80], became a standard for vortex lattice methods applications. It con-
tains a set of results on rectangular wings flying over asolid boundary ob-
tained within linear theory. Later on, vortex lattice methods were successfully
applied to numerical solutions of nonlinear problems of aerodynamics of lift-
ing systems; see Belotserkovsky et al. [81, 82].
The linearized vortex lattice approach was used by Farberov and Plissov
[83], Plissov [84], and Konov [85, 86] to determine the characteristics of wings
of different planforms and lateral cross section near the interface, and by
Plissov and Latypov [87] to compose tables of rotational derivatives for wings
of different aspect ratios near the ground. Ermolenko [88] conceived an ap-
proximate nonlinear approach to the steady aerodynamics of wings near a
solid boundary based on iteration of position of trailing vortices in connec-
tion with the induced velocity field in a plane, normal to the wing and passing
through the trailing edge.
Treschevskiy and Yushin [89], Pavlovets et al. [90], Yushin [91], and Volkov
et al. [92] utilized various modifications of discrete and panel vortex methods
to compute the characteristics of wings and wing systems in the ground effect.
Katz [93] used a vortex lattice method incorporating a freely deforming
wake to investigate the performance of lifting surfaces dose to the ground
with application to the aerodynamics of racing cars.
Deese and Agarval [94] employed an Euler solver, based on the finite-
volume Runge-Kutta time-stepping scheme to predict the 3-D compressible
flow around airfoils and wings in the ground effect. They applied this tech-
nique to calculate the aerodynamics of Clark-Y airfoil (infinite aspect ratio)
and a low-aspect-ratio wing with a Clark-Y cross section.
18 Introduction

Kataoka et al. [95] extended the two-dimensional steady-state approach

to treat the aerodynamics of a foil moving in the presence of a water surface.
The foil was replaced by a sour ce distribution on its contour and a vortex
distribution on its camber line. The free surface effect was taken into account
by distributing wave sour ces on an unperturbed position of the boundary.
Mizutani and Suzuki [96] used an iterative approach based on panel meth-
ods for the airflow field and the Rankine method for the water flow field to
account for the free surface effects and calculated wave patterns generated
by a rectangular wing with endplates. In both of these works, the differences
in the aerodynamic predictions for a wing near a free surface and near a
corresponding solid boundary were found to be insignificant.
By using the method of continuous vortex layers (vortex panel approach),
Volkov [97] carried out some computational investigation into the influence
of the geometry of the foil upon its aerodynamic characteristics in proximity
to the ground.
Morishita and Tezuka [98] presented some numerical data on the two-
dimensional aerodynamics of the airfoil in the compressible ground-effect
flow.
Day and Doctors [99] applied the vortex lattice method to calculate the
steady aerodynamic characteristics of wings and wing systems in the ground
effect with the incorporation of the deformation of the wakes into the numer-
ical scheme.
Chun et al. [100] carried out computations for the aerodynamic charac-
teristics of both isolated wings and the entire craft of the flying wing configu-
ration in ground effect using the potential-based panel method of dipole and
source distributions. Standingford and Tuck [101] applied the high-resolution
approach developed within the steady lifting surface theory to investigate the
influence of endplates upon the characteristics of a thin flat rectangular wing
in the ground effect. The authors report this approach yields better accuracy
than the standard vortex lattice method near the edges of a junction of a
wing-plus-endplates system.
Hsiun and Chen [102] proposed a numerical procedure for the design of
two-dimensional airofoils in the ground effect. The corresponding inverse
problem was based upon the least square fitting of a prescribed press ure
distribution and a vortex panel method to obtain a direct solution of the
problem.
Design problems for airfoil sections near the ground have also been con-
sidered by Kühmstedt and Milbradt [103], whose objectives for optimization
were stability and maximum lift. These authors used and compared several
potential code methods in both 2-D and 3-D formulations.
 Asymptotic and Numerical Methods 19

1.3.4 Numerics: Euler Codes for Unsteady Flows

Efremov [104] used a collocation method to solve numerically the integral

equation describing oscillations of a flat plate of infinite aspect ratio in motion
near the interface of fluids of different densities.
Gur-Milner [105] used a continuous representation of loading with a
Birnbaum-Prantdtl double se ries to calculate the steady and unsteady char-
acteristics of wings of arbitrary planform near the ground within the frame-
work of linear theory.
Vasil'eva et al. [106] formulated a three-dimensionallinearized unsteady
theory of a lifting surface moving in the presence of a boundary of two media
with different densities. The authors were able to calculate the aerodynamic
derivatives for unsteady motions of the lifting surface of a given planform. In
particular, they performed calculations for an air-water interface, concluding
that, within the assumptions of the mathematical model adopted, the surface
of water under a wing behaves as if it were a solid wall.
Avvakumov [107] performed calculations of the aerodynamic derivatives
of a wing of finite aspect ratio in motion above the a wavy wall surface, using
the vortex lattice method to model the unsteady aerodynamics of the lifting
system and the distribution of three-dimensional sources on the underlying
surface.
A 2-D estimate ofwhat the authors designate "dynamic ground effect"was
made by ehen and Schweikhard [108] by using of the method of discrete
vortices of such transitional maneuvers of a flat plate airfoil as descent or
climb. In this work, a simplifying assumption was made that the foil unsteady
vortex wake is directed along the flight path.
A 3-D numerical simulation of the aerodynamics of wings of finite thick-
ness in the ground effect was carried out by Nuhait and Mook [109, 110],
Mook and Nuhait [111], Elzebda et al.[112], and Nuhait and Zedan [113]. In
these works, the authors employed the vortex lattice method and modelled
the shedding of the vorticity into the wake by imposing the Kutta-Zhukovsky
condition along the trailing edges and wing tips. The position and distribu-
tion of the vorticity in the wake were determined by requiring the wake to be
force free. The model is not restricted by plan form , camber, angle of pitch,
roll, or yaw as long as stall and vortex bursting do not occur. The method is
sufficiently general, and the authors gave calculated examples of the unsteady
ground effect related to the descent of the wing and performed computations
of the steady-state aerodynamics of wings and wing-tail combinations.
Ando and Ishikawa [114] investigated the aerodynamic response of a thin
airfoil at zero pitch angle moving in proximity to a wavy wall which advances
in the same direction but with different speed. The authors showed that the
wavelength and speed of advancement have considerable influence on what
they called "the second-order ground effect."
Kornev and Treshkov [115] developed an approximate method for calcu-
lating the aerodynamic derivatives of a complex lifting configuration based
20 Introduction

on the vortex lattice approach and the linearization of the unsteady flow com-
ponents in the vicinity of a nonlinear steady state of the system. They made
some numeric81 estimates of the contribution of different nonlinear factors to
the aerodynamics of wing-in-ground-effect craft.

1.3.5 Numeries: Viscous Flows, Navier-Stokes Solvers

With the increasing power of computers, it becomes possible to study the

aerodynamics of wing-in-ground-effect vehicles by direct numerical modelling
on the basis of appropriate viscous flow problem formulations. Similar to
computation81 methods based on Euler equations, this approach enables suf-
ficiently rigorous representation of the geometry of the vehicle. Besides, such
a technique is not constrained by technic81 difficuties related to simultaneous
physic81 modelling by using sever81 similarity criteria (say, Froude number
and Reynolds number). It is not restricted either in ranges of variation of
the said criteria, so that the future of full scale numerical modelling looks
promising.
However, practic81ly, the implementation of numerical approaches intro-
duces certain difficulties. First of all, the numeric81 treatment of viscous flows
for large magnitudes of the Reynolds number is associated with the necessity
of solving the Navier-Stokes equations with a very sm81l parameter in front
of the higher derivative. This fact gives birth to numerical instabilities due
to ill conditioning of corresponding matrices and considerably complicates
computations even in two dimensions. The difficulties multiply for the case
of unsteady, three-dimension81 flow.
An approximate method for computing ground-effect lifting flows with
rear separation was proposed by Jacob [116]. The author used an iterative
approach combining the three-dimension81 lifting surface theory for inviscid
incompressible flow with a two-dimensional flow model incorporating com-
pressibility and displacement effects. The author's analysis implicitly contains
an assumption that the aspect ratio of the wing is moderate or large.
Kawamura and Kubo [117] used a finite-difference method to solve a 3-D
incompressible viscous flow problem for a thin rectangular wing with end-
plates moving near the ground plane. They employed the standard MAC
method (implying the use of the Poisson equation for the pressure and the
Navier-Stokes equations) and a third-order upwind scheme. They did not use
any turbulence closure models and restricted their calculations to Re = 2000.
Akimoto et 81. [118] applied a finite-volume method to study the aerody-
namic characteristics of three foils in a steady two-dimensional viscous flow
on the basis of the Navier-Stokes equations with the Baldwin-Lomax turbu-
lence model. To provide modelling of the wake, the position of its centerline
was determined by a numerical streamer. The centerline ofthe wake was rep-
resented by a line of segments, extending from the trailing edge of the foil
to the boundary of the computational domain. The number of finite volumes
 Asymptotic and Numerical Methods 21

used in the calculation was 30x120 in vertical and longitudinal directions, re-
spectively. All calculations were carried out for a Reynolds number equal to
3 x 106 . The authors report that the typical CPU time for a set of parameters
was of the order of 150 minutes on an alpha-chip workstation.
Steinbach and Jacob [119] presented so me computational data for the air-
foils in a steady ground effect at a high Reynolds number. Their approach was
based upon an iterative procedure ineluding the potential panel, boundary
layer integral method and the rear separation displacement model.
In 1993 Hsiun and Chen solved the steady 2-D incompressible Navier-
Stokes equations for laminar flow past an airfoil in the ground effect. Later
on [120], the same authors developed a numerical scheme based on a stan-
dard k - € turbulence model, generalized body fixed coordinates, and the
finite volume method. They presented some numerical results concerning the
influence of Reynolds number, ground elearance, and angle of attack on the
aerodynamics of a NACA 4412 airfoil. The range of Reynolds numbers there-
with did not exceed 2 x 106 .
The Reynolds averaged Navier-Stokes (RANS) approach was also applied
by Kim and Shin [121] to treat a steady two-dimensional flow past different
foils, ineluding NACA 6409, NACA 0009, and an S-shaped foil, the latter
form providing static stability of longitudinal motion. Transformed momen-
turn transport equations were integrated in time using the Euler implicit
method. A third-order upwind-biased scheme was used for convection terms,
and diffusion terms were represented by using of a second-order central differ-
ence scheme. The pressure field was obtained by solving the pressure Poisson
equation. Since a nonstaggered grid was adopted in this method, the fourth-
order dissipation term was added in the Poisson equation to avoid oscillation
in the pressure field. Two-block H-grid topology was adopted both above and
below the foil surface. Two grid points from each block overlapped to ensure
flow continuity. For a Reynolds number of 2.37 x 10 5 adopted for calculations
and a 150x120 grid, employed to simulate turbulent flow, 300-500 seconds
were required to produce a calculation on Cray C-90 supercomputer.
Hirata and Kodama [122] performed a viscous flow computation for a
rectangular wing with endplates in the ground effect. For this purpose, they
used a Navier-Stokes solver, based on a third-order accurate upwing differ-
encing, finite-volume, pseudocompressibility scheme with an algebraic turbu-
lence model to elose the system of equations. To be able to treat complex
configuration of the flow the, authors used a multi block grid approach.
Hirata [123] extended the same technique to attack numerically the prob-
lem for a power-augmented ram wing (PARWIG)-in-ground effect. The thrust
of the propeller, ensuring power augmentation, was represented by prescribed
body-force distributions. However, the Reynolds number for which the cal-
culations were made was somewhat moderate (Re = 2.4 x 105 ). A similar
approach was used by Hirata and Hino [124] to treat the aerodynamics of a
ram wing of finite aspect ratio.
22 Introduction

Barber et al. [125] applied RANS equations with a k-c turbulence model
to investigate the influence of a boundary condition on the ground on the
resulting calculated aerodynamic characteristics of a foH in two-dimensional
viscous flow. The authors aimed at discerning differences in the existing mod-
elling technique of ground effect aerodynamics. Calculated results were pre-
sented for Re = 8.2 x 106 • In another paper by the same authors [126], the
RANS technique was used to analyze the deformation of an air-water inter-
face, caused by a wing flying above the water surface.
It should be noted that the application of computational fluid dynamics
(CFD) at very high Reynolds numbers is not straightforward for both numer-
ical and physical reasons; see Patel [127] and Larsson et al.[128]. The general
problem is that the ratio of the smallest to the largest scales of the flow
decreases with an increase in the Reynolds number. Numerically, it me ans
that more grid points are required to obtain a given resolution and physically
the nature of turbulence changes, which means that turbulence models de-
veloped at low Reynolds numbers might not be valid for high ones. Viscous
effects are comparable with inertial ones in the immediate vicinity of the wall.
Therefrom, for sufficient resolution, the number of points of the grid in the
direction normal to the surface of the body should be much larger than along
that surface. As a result, as a consequence of limited computer resources,
extremely elongated numerical cells appear near the body surface, causing ill
conditioning of corresponding systems of equations and breakdown of most
of the solvers; see Larsson [128]. In spite of the progress envisaged in numeri-
cal solution of Navier-Stokes equations with the use of large eddy simulation
(LES), and ultimately through direct numerical simulation (DNS), the ex-
perts do not expect that these methods would be realized earlier than 10 and
20 years, respectively.
2. Problem Formulation for the Flow
Past a Lifting Surface
in Proximity to aSolid Boundary

2.1 Formulation of the Three-Dimensional

Unsteady Flow Problem

Consider a wing of small thickness and curvature, performing an unsteady

motion above asolid nonplanar underlying surface in an ideal incompressible
fluid l (see Fig. 2.1). Assume that motion of the wing is the result of super-
position of the main translational motion with variable speed U(t) and small
vertical motions due to heave, pitch, and possible deformations of the lifting
surface. Introduce a moving coordinate system in which the axes x and z
are located upon an unperturbed position of the underlying boundary (the
ground). Axis x is directed forward in the plane of symmetry of the wing,
and axis y is directed upward and passes through the trailing edge of the root
chord.
In what follows, all quantities and functions will be rendered nondimen-
sional by using the root chord Co and a certain characteristic speed Uo. De-
fine the relative ground clearance h o as the ratio of the characteristic dis-
tance of the trailing edge of the wing's root section from the unperturbed
position of the ground to the length of the root chord. Introduce functions
Yu(x, z, t), Yl(X, z, t), Yg(x, z, t), and Yw(x, z, t), characterizing, respectively,
the instantaneous positions of the upper and lower surfaces of the wing, the
surfaces of the ground, and the wake from the plane Y = o.
Introduce a small parameter c, characterizing aperturbation (for example,
angle of pitch, curvature, thickness, amplitude of oscillations or deformations
of the wing, deformation of the ground surface, etc.).
With the intention of developing an asymptotic theory valid in close prox-
imity to the ground, suppose that the relative ground clearance is small, that
is, h o « 1. Assume that at any moment the instantaneous distances of points
of the wing, wake, and ground surfaces from the plane y = 0 are of the same
order as ho and are changing smoothly in longitudinal and lateral directions.
Thus, if y(x, z, t) describes either of these surfaces, then

By By)
( y, Bx' Bz = O(c) = O(ho) « 1. (2.1)

1 Extension to compressible flow will be considered in 5.1.

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
24 2. Problem Formulation

Vortex
--Y'-7''--- ----sh::O:ee"'t-_-~-

-w

Fig. 2.1. A lifting surface moving in proximity to the ground.

It should be noted that in the case of the extreme ground effect, the assump-
tion adopted 15 = O(ho ) does not mean that flow perturbations are necessarily
smalI. It will be shown later on that in the extreme ground effect, the input
of the order of 0(15) can result in the system's response of the order of 0(1).
The inviscid incompressible flow around a wing in the ground effect is
governed by the three-dimensional Laplace equation and is subject to
• the flow tangency condition on the surfaces of the wing and the ground,
• the dynamic and kinematic conditions on the wake surface (pressure and
normal velo city should be continuous across the wake), and
• the decay of perturbations at infinity.
The Kutta-Zhukovsky requirement of pressure continuity can be specified at
the trailing edge, although it is automatically satisfied through the boundary
conditions on the wake surface.
With this in mind, one can write the following flow problem formulation
with respect to the perturbation potential ep:

(2.2)

Oep = [oep _ U(t)] OYu,1 + oep OYu,1 + OYu,l,

oY OX OX oz OZ ot
Y = Yu,I(X, z, t), (x, z) E S; (2.3)
oep = [oep _ U(t)] oYg + oep oYg + OYg,
oY ox OX oz oz ot
Y = Yg(x, z, t), (x, z) E G; (2.4)
(V'epn)- = (V'epn)+, Y = Yw(x,z,t), (x, z) E W; (2.5)
 2.2 Flow Below the Lifting System and Its Wake 25

Fig. 2.2. Scheme of sub division of the fiow into characteristic zones and the se-
quence of asymptotic matching.

\l'P~O, (2.6)
where S, C, and Ware the areas of the wing, the ground, and the wake
related to the square of the root chord.
According to the technique of matched asymptotics, the flow domain will
be subdivided into the following subdomains with different characteristic
length scales (see Fig. 2.2, where, for simplicity, the sub division is illustrated
in two dimensions):
• the upper flow region Du above the wing, its wake and part of the ground
outside the projection of the wing and the wake upon the unperturbed
ground plane;
• the channel (lower) flow region D[ under the wing and the wake;
• the edge flow regions D e in the vicinity of the edges of the lifting surface
and the wake.
In each of the regions, asymptotic solutions of the problem (2.2)-(2.6) are
constructed for h o ~ 0 in appropriately scaled coordinates.
The asymptotic matching and additive composition of these solutions en-
able accounting for the interaction of different parts of the flow and obtaining
a uniformly valid solution for the entire flow domain. In what follows, con-
sideration is restricted to the asymptotic accuracy of the order of O(h o ).

2.2 Flow Below the Lifting System and Its Wake

(Channel Flow)

In the region D 1, where 'P rv 'PI, X = 0(1), z = 0(1), andy = O(ho), we

introduce stretching of the vertical coordinate

- Y (2.7)
Y=h'o
26 2. Problem Formulation

where h o = const. is a characteristic magnitude of the ground clearance. 2

After substitution of <PI, x, z and Y into the equations of the problem, we
obtain the following formulation for the channel flow perturbation poten-
tial <PI:
02<pI h2 (02<PI 02<PI) = 0 (
oy2 + 0 ox 2 + oz2 , x , y, z) E D I; (2.8)

0<PI = ho2 {[O<PI _ U(t)] 0YI + 0<PI °YI + OYI}, Y = YI = h

Yl ;
o (2.9)
oY OX ox oz OZ ot
0<PI = ho 2 { [O<PI _ U(t)] oYg + 0<PIOYg + Oyg} (2.10)
oY ox OX oz oz ot'
In the channel flow region, both the condition at infinity (decay of pertur-
bations in three-dimensional flow) and the Kutta-Zhukovsky condition at the
trailing edge are lost. The influence of these conditions will be transmitted
to the channel flow region by matching with the asymptotic solutions to be
obtained in regions De and Du.
We seek <PI in the form of the following asymptotic expansion:

<PI = <pj + h~ <pj* = <PlI + <P12 ho In ~o + <Pb ho+ h~ <pj*, (<pj, <pj*) = 0(1),
(2.11)
which can be shown to satisfy the requirement of matching of the asymptotic
representations of the velo city potential in the regions DI , Du, and De. Sub-
stituting (2.11) in (2.8), yields the following equations for the functions <pj
and <pj*:

(x, y, z) E D I ; (2.12)

02 rn
rl** _ 02 rn
rl* 02 rn
rl*
oy2 - ox 2 + oz2 ' (x, y, z) E DI. (2.13)

Then, using the same asymptotic expansion (2.11) in the flow tangency con-
ditions on the lower surface of the wing (2.9) and on the ground (2.10), we
obtain the following set of equations:
• on the lower surface of the wing

O<p*
_I =0 Y = YI(X, z, t); (2.14)
oY

o<pj* = [O<Pj _ U(t)] 0YI + o<pj 0YI + 0YI Y = YI(X, z, t). (2.15)
oY OX ox oz OZ 8t '

2 In the steady case, this parameter coincides with the relative ground clearance
h measured at the trailing edge.
 2.2 Flow Below the Lifting System and Its Wake 27

• on the ground
8epi _
8y - 0, Y = Yg(x, z, t); (2.16)

8ept = [8epi _ U(t)] 8Yg + 8epi 8Yg + 8Yg ' Y = Yg(x, z, t). (2.17)
8Y 8x 8x 8z 8z 8t
Integrating (2.12) two times with respect to Y and accounting for (2.14)
and (2.16), we obtain an important conclusion: with an asymptotie error
of the order of O(h o ), the description of the channel flow is two-
dimensional in the plane parallel to an unperturbed position of the
ground surface, Le., to the plane Y = 0,

epi = epi (x, z, t). (2.18)

Integrating (2.13) one time with respect to Y, gives

(2.19)

where f**(x, z) is an unknown function. Taking into account equations (2.15)

and (2.17), we obtain

( 82epi
8x2 +
82epi) _
8z2 Yl +
f**(
x, z
) = [8epi _ U( )] 8Yl
8x t 8x +
8epi 8Yl
8z 8z +
8Yl
8t'

Y = Yl(X, z, t); (2.20)

( 82epi 8 2epi ) _ f** ( ) = [8epi _ U( )] 8Yg 8epi 8Yg 8Yg

8x2 + 8z2 Yg + x, z 8x t 8x + 8z 8z + 8t '
Y = Yg(x, z, t); (2.21)
Substracting (2.21) from (2.20), we obtain the following channel flow equa-
tion:

~ (h* 8epi) + ~ (h* 8epi) = U(t) 8h* _ 8h*, (x, z) E 5, (2.22)

8x 8x 8z 8z 8x 8t
where h* = h*(x, z, t)/h o = Yl - Yg, h*(x, z) = Yl(X, z) - Yg(x, z) is the
instantaneous distribution of the gap between the wing and the ground.
Thus, it has been shown that for very small clearances (an extreme
ground effect), the flow field under the wing in the ground effect
has a two-dimensional description and its perturbation velocity po-
tential CPl r..J cpj satisfies quasi-harmonie equation (2.22) in a two-
dimensional domain S bounded by the wing planform contour.
The boundary conditions for epi at the leading hand trailing edges l2 of
the lifting surface will be obtained by matching.
From a physical viewpoint, equation (2.22) can be interpreted as the equa-
tion of mass conservation in a highly constrained channel flow region with
28 2. Problem Formulation

known distributed mass addition due to tangency conditions on the lower

surface of the wing and part of the ground situated under the wing.
For channel flow under the wake, the same procedures can be used to
relate the induced downwash a w = O(h o ) in the wake,

(2.23)

to the wake channel flow potential 'Pi and the corresponding instantaneous
gap distribution h~(x, z) = Yw(x, z, t) - Yg(x, z, t) by the following equation:

(x, z) E W, (2.24)

2.3 Flow Above the Lifting System and Its Wake

In the upper flow field Du, where (x, Y, z) = 0(1) for ho -+ 0 and e = O(h o),
both the wing and the wake approach the ground. In the limit, one comes in
the upper half space to the problem for the flow, generated by the tangency
conditions on the upper surfaces of the wing and the wake. Because e tends to
zero (which practically means that, e.g., the relative ground clearance, angle
of attack, curvature and thickness of the wing are small) the upper flow can
be linearized, so that tangency conditions are satisfied upon the projections
of the wing and wake sheet onto the plane Y = o.
In the region Du, we seek the upper flow perturbed velocity potential in
the form
'PUl = 0(1). (2.25)
Substitution of this expansion in the flow problem (2.2-2.6) leads to the
following equation and boundary conditions for the upper flow problem:

[j2'Pu1 + 02'PU1 + 02'PU1 _ 0

(x, Y, z) E S; (2.26)
ox2 oy2 oz2 - ,

0;;1 = O!u _ U(t) 0:: = -ä u, (x, z) E S, Y = 0 + 0; (2.27)

Y = 0 +0; (2.28)

oy -- 0,
0'PU1 (x, z) ~ S + W, y = 0; (2.29)

(2.30)
 2.4 Flows in the Vicinity of Edges 29

where ä u = Ci uI ho, ä W1 = CiwtI ho. The channel fiow and edge fiow descri p-
tions are lost in Du. Their infiuence will be recovered by asymptotic matching
of the upper fiow potential with that of the channel fiow through edge regions.
Note that the boundary condition on the upper surface of the wake vortex
sheet may be formulated as the fiow tangency condition if the downwash
CiW1 = hoä w1 in the wake is known.
The upper fiow potential 'PUl is constructed in the form

'P --~1 Q(l,t)dl+~JräudS+~J'r äWldW (2.31)

Ul - 47r h+13 rl 27r Js r 27r Jw r
The first term of (2.31) represents the induced velo city potential of the
source-sink distribution along the contours of the leading and side edges
of the wing hand the edges of the wake l3. This contour distribution charac-
terizes the infiuence of the channel fiow upon the upper fiow due to leakage of
air from under the wing. The strength ofthe contour sour ces (sinks) QI(l,t)
is determined as a result of matching the upper and channel fiow potential
through edges hand b. The second and third terms of (2.31) correspond to
the potential of the surface distributions of sour ces and the sinks of strength
-2Ci u upon Sand -2Ci wl upon W. The latter result is based on a thin body
theory.
Near the edges II and l2, the function 'PUl has the following asymptotic
representations:
• near the leading and side edges ll,

Q(l, t) 1 t) 1 A 2 (l, t) O( 2)
'PUl'" - -
27r
nlJ + -äu(l,
- - l J nlJ +
7r
AI(l, t)
7r
lJ +
7r
+ lJ; (2.32)

• ne ar the trailing edge 12 ,

(2.33)

where l is the arc coordinate measured along the planform contour, lJ is the
external normal to the planform, and ( Ci uw ) = Ciu(l, t) - Ciw(l, t) is the jump
of downwash on the upper surface of the wing across the trailing edge l2,
(ä uw ) = (Ci uw )Iho. Parameters Al, A 2 and BI, B 2 characterize the infiuence
of distant sources.

2.4 Flows in the Vicinity of Edges

In region De , i.e., in the vicinity of the edges of the wing (h and l2) and the
wake l3, one has to derive local asymptotic descriptions of the fiow.
Inspection of the Laplace equation shows that for sufficiently large radii
of curvature of planform contours (Pe» h o), the flow near the edge can be
30 2. Problem Formulation

considered two-dimensional in the plane normal to the planform contour. This

circumstance considerably facilitates the analysis of edge flows. In particular,
effective methods of the functions of the complex variable can be used to
account either for the specifics of the geometry of the edge (thin or thick,
with endplates, etc.) or for the physics of the local flow (shock or shock-free
entryat the leading edge, vortex roll-up at wing's tips, jet and rotating flaps,
etc.). Examples of various edge flows are shown in Fig. 2.3.
Here one considers an example of a sharp edge with a small vertex
angle. Later on, when examining particular effects, such as the jet blowing
from the trailing edge or the use of rotating flaps, one will have to consider
corresponding local flows.

Leading Edge

) I 7 J } ; ; ; ; J

Trailing Edge

~- Zi~~--~~~~L
;);;);;;777 ;:;'777))//; 1171>1))))1}

Side Edges

Longitudinal and Lateral Siots

I ~
~i~
; ) ; ;; 7 ~;};; ; ; ;;;;;
~~ >; ; j; ) ;;;;;-~

Jet Edges and Edges with Singularities

~
~
17;;;;;;;;>
~c
, ,. ,. ,. > > > > , , ;,. ,,,.>~>..,>.,>-,:>.-:>.-:>".>~>7'7"

Fig. 2.3. Possible local flows in the problem of a wing in the extreme ground effect.
 2.4 Flows in the Vicinity of Edges 31

Consider a flow near the leading edge l1. Introduee a loeal eoordinate
system v 0 y, where v is an external normal to the planform eontour, and
stretch independent variables

(2.34)

where h; = h;(l, t) is an instantaneous relative distanee of the edge vertex

from its projeetion upon the ground. Parameter l, as introdueed previously,
represents the are eoordinate measured along the planform eontour. Note
that due to assumption (2.1) both the wing and wake surfaees have spanwise
and ehordwise slopes of the order of O(h o ). Therefore, the edge flow problem
formulation ean be linearized. In fact, as preseribed by (2.1), in the stretehed
loeal region De the distanees of points on the wing's upper and lower surfaees
from the horizontalline y = 1, as weH as the distanees of points of the ground
surfaee from horizontal line y = 0 are of the order of O(h o ) (see Fig. 2.4).
Therefrom with an asymptotie error of the order of O(h;), the tangeney
eonditions on the wing and on the ground ean be transferred to horizontal
lines y = 0 and y = 1, respectively.
The eorresponding problem for the leading (side) edge flow velo city po-
tential 'Pe = 'PIe ean be formulated as foHows:

(2.35)

O'Ple = _h 2 d lvi< 00, y= 0 + 0; (2.36)

oy 0 g,

O'Ple = _h 2 d v ::; 0, y = 1 + 0; (2.37)

oy 0 u,

O'Ple = _h 2 d v ::; 0, y= 1- 0; (2.38)

oy 0 I,

Parameters du, dl, and dg depend on are eoordinate land time t and have
to be determined in the proeess of matching with the asymptotie solutions
eonstructed in the upper and ehannel flow regions. Note that without loss

i y=y/h* O(h)

0(1) O(h)

Fig. 2.4. Local flow in the vicinity of the edge with a finite vertex angle.
32 2. Problem Formulation

of generality, it is possible to set d g = O. In the problem far the flow near

a leading edge, the boundary eonditions at infinity on the wake and at the
trailing edge have been lost. Their influenee on the leading edge flow will
be reeovered by matehing with the asymptotie deseriptions of the veloeity
potential in the ehannel and upper flow regions.
Due to the linearity of the loeal problem, its general solution ean be
presented in the form

(2.39)

where ai are parameters to be determined through matehing. Function 'Pae

is a homogeneous solution, satisfying the eondition of no normal veloeity on
the wing and ground surfaees (du = d1 = dg = 0). This homogeneous solution
eorresponds to a eireulatory flow around the edge. Function 'Pbe represents a
nonhomogeneous solution, which is generated by a preseribed normal velo city
upon the wing and the ground. Both 'Pae and 'Pbe are of the order of 0(1).
Linear eombination a3hov + a4 automatically satisfies the Laplaee equation
and does not violate the flow tangeney eondition on solid boundaries.
We turn to determination of functions 'Pae and 'Pbe' To find solution of the
homogeneous problem for the flow past the leading (side) edge, we perform
a eonformal mapping of the loeal flow domain De onto the upper half plane
~( = Tl > 0 of the auxiliary eomplex plane ( = ~ + i Tl by the Christoffel-
Sehwartz integral; see Lavrent'ev and Shabat [129]. The eorrespondenee of
points in the physical plane f-L = v + i fj and auxiliary plane ( is illustrated in
Fig.2.5.
The mapping funetion has the form
1
f-L= -(l+(+ln(), (2.40)
7l'

where i = A is an imaginary unit. For purely cireulatory flow around the

edge in the plane of the eomplex potential fae = 'Pae + i 'ljJae, the flow with
unit velo city at the left infinity v = -00 is represented by a strip of unit
width. Conformal mapping of this strip onto the upper half plane ~( > 0 is
realized by the function
. 1
fae = 'Pae + 1 'ljJae = -ln
7l'
(. (2.41)

Im f ae = 1

Im f ae = 0 0 Fig. 2.5. Boundary conditions for the homo-

geneous component of the complex potential
of the flow near the leading edge.
 2.4 Flows in the Vicinity of Edges 33

Eventually, the solution of the homogeneous problem is given by the formulas

(2.40) and (2.41). Excluding the auxiliary variable (, we obtain the following
relationship between the "physical" complex plane p, = ii+i Y and the complex
potential fae = 'Pae + i '0ae, representing the homogeneous component of the
problem solution.
(2.42)
The flow pattern, corresponding to the homogeneous solution 'Pae is depicted
schematically in Fig. 2.6. At points on the wing surface near the leading (side)
edge, where fae = 'Pae + i, J.L = ii + i, ii::::; 0, the potential 'Pae can be
determined through the following implicit relationship:

(2.43)

It can be shown that the flow velo city has a standard square root singularity
at the leading edge. In fact, in the immediate vicinity of the edge vertex
ii -+ 0- 0, 'Pae -+ 0, and it follows from the expansion of equation (2.43) that

- ( 1 22 1 22
1rl1 c:::: 1 - 1 + 1':'Pae + 2"1': 'Pae + ... ) + 1':'Pae c:::: -2"1': 'Pae' (2.44)

wherefrom for ii -+ °- 0,

f-?§ O'Pae 1
'Pae c:::: ±y -;:-' ~C::::=f ~_'
uv v -,c,1':V
ii ::::; 0. (2.45)

To match the flow potentials in regions Du and DJ, one needs the asymp-
totic representations of the edge flow potential 'Pae far from the edge. Turning
to the variable v = hieii = hohieii and setting ho to zero for fixed v = 0(1),
one obtains
• on the upper surface of the wing (ii = v/hie -+ -00, Y= 1 + 0),

'Pae rv
1 I1':ii I = -1 I-*
-ln 1': V I = -ln
1 I-=-;-
1':V I, (2.46)
1': 1': h 1e 1': h oh 1e

...-"" ---
J J J J J 1 11 J J
...-""
'Pbe
Fig. 2.6. Flow patterns corresponding to the
homogeneous 'Pae and nonhomogeneous 'Pbe
components of the edge flow potential.
34 2. Problem Formulation

·1
Fig. 2.7. Boundary conditions for the conjugated complex velocity corresponding
to the nonhomogeneous component of the edge flow potential.

• on the lower surface of the wing (ii = /Jjhie -+ -00, Y= 1 - 0)

_ 1 /J 1 /J 1
CPae f"V /J - - = - - - = --- - -. (2.47)
n h~ n hh~ n

We turn to the determination of the nonhomogeneous solution CPbe. Using

the same mapping function (2.41), one comes to the following problem for a
complex conjugate velocity Wbe = Ube - i Vbe in the auxiliary plane (: Find
an analytic function Wbe«() in the upper half plane <;$( = TJ > 0 in terms of
its imaginary part <;$Wbe = -Vbe given on the axis ~ (see Fig. 2.7) .
It can be verified that the following expression satisfies the problem under
consideration:
1
Wbe = -[(du - d\) In(1 + () + d\ln (l. (2.48)
n
At points on the wing in the vicinity of the edge « = ~ < 0),

(2.49)

The velocity potential corresponding to the nonhomogeneous solution can

be derived by means by integrating(2.49) :

(2 .50)

At points on the wing surface, the auxiliary variable ~ = ~( is related to the

variable ii in the following way:

~ < o. (2 .51 )

The flow pattern corresponding to a nonhomogeneous solution is presented

schematically in Fig. 2.6. In the particular case of an infinitely thin edge (zero
vertex angle du = d\ = d 3), the following expression for CPbe can be derived
from the previous more general relationships:

3 Without loss of generality, constant d can be taken as d = 1.

 2.4 Flows in the Vicinity of Edges 35

To match the edge flow potential with the velo city potentials in regions Du
and D I , we need to know asymptotic representations of Ube and Cf'be far from
the leading (side) edge. These can be found in the form
• on the upper surface of the wing (v = v/hie -+ -00, iJ = 1 + 0, ~-+
-00)
du
Ube "" -;- In
17fV-I = du I 7fV I
-;- In hof/'ie ' (2.53)

Cf'be "" -duv (ln 1 -I

7fV - duv ( In I-=;-
1) = ----* 7fV I - 1) j (2.54)
7f 7fh oh le hohle
• on the lower surface of the wing (v -+ -00, iJ = 1 - 0, ~ -+ 0-0)

(2.55)

-dI ( -
Cf'be - 2
V ) -d - -
v2 - -
-
2
I 2h2h*2
-- (V
h-
h* .
V) (2.56)
7f oie 7f ole

Taking into account expressions (2.39) and (2.45) we obtain the following
estimate of the behavior of the velo city near the leading edge of the lifting
surface along the normal to the planform contour:

V< 0, (2.57)

where the "minus" sign corresponds to the upper surface of the wing, and
the "plus" sign corresponds to the lower surface of the wing. Formula (2.57)
is useful far calculating the suction force at the leading edge of the lifting
system in both steady and unsteady motion.
Note that the asymptotic solution of the problem for the flow near the
edge of a vortex sheet in the extreme ground effect has the same structure
as that for an infinitely thin leading (side) edge.
Near the trailing edge l2, the velocity potential Cf'e = Cf'te should sat-
isfy not only the flow tangency conditions on the wing and the ground but
also comply with the dynamic and kinematic conditions on the vortex sheet
emanating from the wing due to unsteady and three-dimensional effects. Ex-
cluding the homogeneous component of expression (2.39) which incorporates
a square root singularity for the flow velo city at the edge, we can write the
velo city potential Cf'te for the flow in the vicinity of the trailing edge in the
form
Cf'te = blh~Cf'be + b2h~v + b3 ho, (2.58)
where v= v / h * (l, t) = v / h;e and Cf'be is given by the formula
36 2. Problem Formulation

(2.59)

1
Ube = -[(eu - el) In 11 + ~I
+ edn 1~ll· (2.60)
7r
The auxiliary variable ~ is related to v byequation (2.51). Parameters eu , el
and bl, b2 , b3 are unknown and have to be determined by asymptotic match-
ing. It is essential to note that parameters b2 and b3 have different magnitudes
on the upper (Y = yjh;e = 1 + 0) and lower (Y = yjh;e = 1 - 0) surfaces of
the vortex sheet behind the trailing edge, Le., bt -=J b'2 and bj -=J b3'. This is
caused by the jump discontinuity of the velo city potential and the tangential
velo city at the trailing edge l2, the latter in the case of unsteady fiow. At
the same time, the solution satisfies the condition of the continuity of the
normal velo city component ac ross the vortex sheet. In the particular case of
an infinitely thin trailing edge (zero vertex angle eu = el = e), it follows from
(2.59) and (2.60) that

CPbe =;' [v(7rcpae -1) - ~7rCP;el. (2.61 )

The asymptotic representation of Ube and CPbe far from the trailing edge
can be found from expressions (2.53)-(2.56) by replacing du,l by eu,l and hie
by h;e' Note that the solutions of local fiow problems, presented above, lose
validity in the vicinity of the order of O(h o ) of the corner points of con-
tours h, h, where the fiow is essentially three-dimensional. Near such corner
points, additional solutions should be constructed, but this question will not
be discussed here.

2.5 Matching of Flow Descriptions in Different Regions

To determine the unknown constants that enter the asymptotic expansions of

the solutions in different parts of the fiow and also the boundary conditions
at the contour of the two-dimensional domain S for quasi-harmonie equation
(2.22) of the channel fiow, it is necessary to match these expansions in the
overlap regions. From a physical viewpoint, the matching procedure is sim-
Har to accounting for the inter action of different parts of the fiow. From a
mathematical viewpoint the process of asymptotic matching allows attaining
uniqueness of the composite uniformly valid solution of the problem.
We match of the fiow characteristics in the different regions in accor-
dance with the scheme shown in Fig. 2.2, where for simplicity the sequence
of matching is shown in a two-dimensional case .
• In the first stage, match the velocity potential CPu of the upper fiow with
the potential CPle of the fiow near the leading (side) edge lt (l2). This step
will result in the determination of parameters al, a2, a3, a4, and du.
 2.5 Matching of Flow Descriptions in Different Regions 37

• In the second stage, match the channel fiow potential <PI with potential <PIe
of the fiow near the leading (side) edge lt (l2). This step gives the possibility
of determining the boundary condition for the channel fiow quasi-harmonie
equation at contour lt, as well as the magnitudes of dl and the strength Q
of the contour distribution of sources (sinks) in expression (2.31).
• In the third stage, match the upper fiow potential <Pu with the trailing edge
fiow potential <Pte. This step will determine parameters bl , bt, bt, and eu .
Imposing the Kutta-Zhukovsky condition in the form of equation (2.5),
find the relationship of parameters b2, b3 with parameters bt, bt.
• Finally, in the fourth step, match the pressure coefficient in the fiow below
the trailing edge PIe with the channel fiow pressure coeffcient PI.
In what follows, we use the "asymptotic matching principle," as introduced
by Van-Dyke [38], namely: the " m-term inner expansion of the n-term outer
expansion is equal to the n-term outer expansion of the m-term inner ex-
pansion," where m and n are integers. Note that within the formalism of the
method of matched expansions, the term "outer" expansion stands for the
asymptotic expansion, obtained in variables, based on the primary charac-
teristic lengths of the problem. The "inner" expansion implies the use of
variables stretched with the help of the gauge function of a small parameter
so that they have the order of 0(1) in the regions of the nonuniform validity
of the outer expansion.
To match in the first stage, it is necessary to derive an asymptotic
representation of the upper fiow potential <Pu at points of a normal v in
the immediate vicinity of the leading (side) edge. This asymptotic expansion
has been obtained previously and is given by expressions (2.25) and (2.32).
Replacing the variable v in (2.32) by hiev, we obtain the following two-term
asymptotic expansion:

(2.62)

where hie' as earlier, is the instantaneous distance of the edge vertex from
the ground, which has the order of O(h o ), and (asw ) = au(l, t) - ag(l, t) is
the difference between the downwash upon the upper surface of the wing and
that upon the ground in the vicinity of the leading edge. The first term of
expression (2.62) is known from the theory of potential functions to refiect
the behavior of a potential near a li ne distribution of sourees. Other terms
describe the behavior of the surface distribution of sources (simple layer) near
the edge. Parameters Al(l, t) and A 2 (l, t) characterize the infiuence of distant
sourees. Expressions Al and A 2 are cumbersome in the general case, so they
are not presented here, but will be written for some concrete problems later
on.
38 2. Problem Formulation

On the other hand, collecting expressions (2.39), (2.46), and (2.54), we

obtain the following asymptotic expansion of the leading edge flow potential
far from the edge on the upper surface of the wing (fj = 1 + 0):

(2.63)

where hie = hie/ho. Equating expressions (2.62) and (2.63) in the same vari-
able (v or ii), we obtain

(2.64)

du = (öug)hie = [öu(l, t) - ög(l, t)Jhie, (2.65)

a3 = ~[hieAl +du(l-lnlho~iJ)], (2.66)

a4 = ~ (A 2 - a1lnl ho~ie 1). (2.67)

Note that after the first stage of matching, the quantity Q has been expressed
through a coefficient al, whieh will be determined later.
In the second stage of matching, it is necessary to rewrite expression
(2.39) for the leading edge flow potential in terms of the coordinate v = hieii
and pass over to the limit ho(hie) -+ 0 for y = 1 - 0 and fixed v. Taking into
account (2.39),(2.47), and (2.56), on the lower surface of the wing,

VI) a2h~dl
IPle -+ IPlel f'Valho ( h* - - + -,;;-
(12h*
v2 h~v
- -V) + a3V + a ho. 4 (2.68)
le 7r le le 7r le
This asymptotie representation should be equated to expression (2.11) for
the channel flow potential IPl evaluated for v = hieii -+ 0, that is,

IPl f'V IPi = IPh + IPl 2 ho In ~o + IPla ho

V
= alho ( --;;- - -
1) + --*
a2dlh~(1v2
- ---;;- -
V)
-
2V
+a3ho--;;- +a4 ho. (2.69)
h le 7r h le 2 h le 7r h le
Setting v to zero, we can determine parameters al, a2, and d l :

- ~Q - h* °IPh
al - 2 - le ov ' at v = 0, a2 = 1, (2.70)

at v = 0, (2.71)
and also the boundary condition for the quasi-harmonie equation (2.22) on
the line h whieh corresponds to the leading (side) edge. It follows from (2.69)
at v = 0 that
 2.5 Matching of Flow Descriptions in Different Regions 39

on h, (2.72)

or taking into account the expressions found for al and a4, as well as for the
asymptotics of the function 'Pi,

• horA h-· 8'PI 1 ( I 1 7r I)J 1/ = 0; (2.73)

'PI = -; 2 - le 81/ 1 + n hohie '

at 1/ = 0; (2.74)

(2.75)

(2.76)

The boundary conditions written above must be fulfilled on the line h, which
corresponds to the leading (side) edge. Note that one of the results of match-
ing in the second stage is the determination of the strength Q(l, t) of the
sources distributed along the lines hand b. This parameter enters the ex-
pression for the upper flow potential 'Pu. As follows from (2.70), the strength
of the sources Q is proportional to the velo city of flow escaping from the chan-
nel into the upper flow region. Simultaneously, it follows from comparison of
expressions (2.57) and (2.70) that the strength of the square root singularity
of the perturbed velocity at the leading (side) edge of the wing is directly
related to the intensity of the circulatory flow around the edge.
We turn to the third stage of matching. The asymptotic representation
of the upper flow potential 'Pu near the trailing edge can be written similarly
to (2.62) as

'Pu -+ 'Pute rv
h~(äuw) h-.tel/-1n (h 0 h-·tel/-) + -h~B (l )h-·tel/- + -hOB 2 (l , t,
I ,t
) (2.77 )
7r 7r 7r

where (ä uw ) = äu(l, t) - ä W1 (l, t) is the jump discontinuity of the düwnwash

upon the upper surface of the wing and the wake in the vicinity of the trailing
edge; BI (l, t) and B 2 (l, t) are known parameters, which have the same sense
as Al (l, t) and A 2 (l, t). On the other hand, in the expression für trailing edge
flow velo city potential 'Pte pass to the variable 1/ = hoh;ev and set h o (h;e) to
zero for fixed 1/ and fj = 1 + O. Accounting for (2.54) with du replaced byeu ,

'Pte -+ 'Pteu rv
h;b1e u ( In 1 h7r1/
---,;;-1/ h. 1
- 1) + b+2 h*
h; 1/ + b+h
3 0 (2.78)
7r te Ote te

Equating expressions (2.77) and (2.78) in the same variable (1/ or v), we
obtain parameters eu , bt, and bt:

(2.79)
40 2. Problem Formulation

(2.80)

(2.81)

One has to remember that h:e = hoh:e is an instantaneous distance of the

trailing edge from the ground and parameters b2 and b3 , entering solution
(2.58), take different magnitudes above and below the wing (btf:b2", bj ;t:b3").
To determine the relationship between bt, bj and b2", b3", it is necessary
to impose the Kutta-Zhukovsky at the trailing edge in the form (2.5), which
complies with the requirement for the continuity of pressure and normal
velocity across the wake in the immediate vicinity of the trailing edge. We will
show how to ensure realization of this condition for a thin edge (eu = el = e),
and then results will be presented for the trailing edge with finite vertex
angle.
The expression for pressure coefiicient in the adopted moving co ordinate
system x, y, z can be calculated with the asymptotic error of the order of
O(h~) as

8<p - 2-
P = 2-U(t) 8<p - (8<p)
- 2
- (8<p)
- 2
+ O(h 2 ). (2.82)
8x 8t 8x 8z 0

We introduce the local coordinate system vyl, where v is the external normal
to the planform contour, axis y passes through a given point of this contour
and is directed upwards, and l is a local tangent to the same contour. Tangent
l and normal v lie in the same plane, coinciding with the projection of the
wing upon an unperturbed position of the ground. For the points of the
trailing edge, follow the relationship between directional derivatives:

! =cos(v,x):v +sin(v,x):l'

:z = sin(v, x) :v - cos(v, x) :l·

It can be shown that in the case of an infinitely thin trailing edge, the po-
tential <Pte in the immediate vicinity of the edge (ii -t 0) has the form

(2.83)
where above the edge b2 = bt, b3 = bj, and below the edge b2 = b2" ,b3 = b3" .
Taking into account that at v = 0,

8<Pte _ b h~
8 V - 2 h *te '
(2.84)

and equating for ii = 0 the pressure coefiicients above and below the edge
(p+ = p_), we obtain the following relationship between the upper and lower
values of parameters b2 and b3 :
 2.5 Matching of Flow Descriptions in Different Regions 41

where
_ eos(v,x)U( )
J.Ll - h* t , J.L2 = sin(v, x)U(t). (2.86)
te
In the fourth stage of matching, we take into aceount that far from an
infinitely thin trailing edge under the wing (v = v /h;e ~ -00, jj = 1 - 0),
the edge flow potential <Pte has the form

<Pte ~ <Ptel
h 2e
= b1 h~
(1'2";;-
v 2 V) v
- - + b2h;.,;;- + b"3 ho• (2.87)
te te 7r te

Using formula (2.87) and expression (2.82) for the pressure eoefficient, it is
not diffieult to obtain the following boundary eondition for quasi-harmonie
equation (2.22) on the line l2, eorresponding to an infinitely thin trailing
edge:

PI = 2h o [J.Ll ho (b 2 - ;) + J.L2 ß;r -ß;f] , (2.88)

taking into aeeount equation (2.85),

(x, z) E l2. (2.89)

Noting that for an infinitely thin trailing edge, parameter e = eu is given by

formula (2.79) and parameters bt and bt are expressed by (2.80) and (2.81),
we finally obtain

2h o [ J-t1h-*te ( B1-(a_uw ) in 1hok

7r I)
te +J-t2 m -Tt,
PI = --:;r ßB2 ßB2 ] (x, z) E l2, (2.90)

where the relationship between the ehannel flow pressure eoefficient PI and
the potential <PI is given by formula (2.82). Representing the pressure in the
form of an asymptotic expansion

PI = PI 1 + PI 2 ho In ~o + PI 3 ho + O(h;), (2.91)

eorresponding to the adopted asymptotic expansion of the potential <pi, we

obtain the following boundary eonditions for magnitudes Ph ,Ph, and PI 3 on
the line l2:
(2.92)

(2.93)

(x, z) E l2. (2.94)

42 2. Problem Formulation

The relationship between the functions PI 1 ,PI2 , and PI 3 and the corresponding
terms of the asymptotic expansion of the channel flow potential is obvious:

PI = 2°<P11 U(t) _ 2 0 <P11 _ (0<P11)2 _ (0<P11 )2; (2.95)

1 OX ot OX OZ

PI - 2 [0<P12 U(t) _ 0<P12 _ 0<P11 0<P12 _ 0<P11 0<P12]. (2.96)

2 - OX ot OX OX OZ OZ '
PI. = 2 [O<Ph U(t) _
0<Ph _ 0<P11 0<Pb _ 0<P11 O<Pb]. (2.97)
OX 3 OX OX ot oz oz
In a simpler case of a wing with a straight trailing edge we obtain from
expression (2.90)

PI = -2h o [ I
( (ä uw ) In h 'Irh-* -
U(t) I BI )- oB ] .
~
2 (2.98)
'Ir ° h UL

Suppose that the trailing edge has a finite vertex angle (au =F al, eu =F
el). Then, as a result of transfer of the boundary conditions in the local region
Dte onto the line fj = 1 ± 0, the corresponding expression for the velo city
reveals a logarithmic singularity at the edge vertex. In principle, it is possible
to correct this deficiency of the local solution and fulfill the requirement of
the finiteness of the velo city at the trailing edge by introducing an additional
asymptotic expansion in the vicinity of the vertex ofthe order o [exp( -I/ho)].
However, some analysis shows that disregard of this nonuniformity does not
bring along any noticeable errors either in press ures or in integral lifting
characteristics.
For the trailing edge with a nonzero vertex angle, the matching procedures
result in the following expression for the pressure coefficient on the line l2:

PI = 2 [f.LIho ( b2+ - el)

-;
abt - h0ßt
+ f.L2 h oaz abt] ' (x, Z) E l2, (2.99)

where parameter el was found in the form

- *2 02<P11 0 l
= hte ov 2 ' v=
C
el lor on 2, (2.100)

or, finally, taking into account (2.80), (2.81), and (2.100),

(X,Z) E b. (2.101)
Thus, the application of the method of matched asymptotic expansions for
treating the problem of the flow past a lifting system in dose proximity of
the ground leads to the following algorithmic solution with an asymptotic
error of the order of O(h~):
 2.6 Uniformly Valid Flow Description and Aerodynamic Coefficients 43

1. The channel fiow potential <PI is found through solution of quasi-harmonie

equation (2.22) in the two-dimensional region S, bounded by the wing's
planform contour with boundary conditions (2.73) at the leading edge
(line h) and (2.98) or (2.101) at the trailing edge (line 12).
2. The potential <Pu of the upper fiow is constructed in a straightforward
manner with the help of the surface distribution of the sources on the
two-dimensional domain S + W + C with the addition of the admissible
contour distribution of the sources along the projections of the leading
(side) and trailing edges onto the plane y = o. The strength of surface
distribution of sources is equal to (-2a u ) upon S, (-2a g ) upon C, and
(-2a w J upon Wand is determined by using formulas (2.23), (2.24), and
(2.27). The strengths of the sources distributed along the contour h + 13
are determined by formula (2.70).
3. Local fiow solutions are constructed near the edges, depending upon the
geometry and particulars of the physical performance of the latter. For
example, expressions for the velo city potentials near a leading (side) and
trailing edges are given by expressions (2.39) and (2.58).

2.6 A Uniformly Valid Flow Description

and Aerodynamic Coefficients
U niformly valid asymptotic expansions for the velo city potentials on
the upper and lower surfaces of the wing can be determined by using the rule
of additive composition (see Van-Dyke [38]) in the form

<P+ = <Pu + (<p~ - <P~J + (<p~ - <P~Jj (2.102)

<P- = <PI + (<p~ - <P~I) + (<pu, - <P;'I) , (2.103)

where <Pu, <PIe, <Pte, <P~u' <P~u' <P~I' <P~I and <PI are given, correspondingly, by
formulas (2.25), (2.39), (2.58), (2.63), (2.78), (2.68), (2.87), and (2.11).
The aerodynamic coefficients of lift Gy, longitudinal moment m z
(around a point with abscissa x o ), and lateral moment m x can be calculated
by using the formulas

Gy = ~ Jfs (p- - p+) dS = ~ Jfs (p- - p+) dS, (2.104)

mz = ~ Jfs (p- - p+)(x - x o) dS = ~ Jfs (p- - p+)(x - x o) dS, (2.105)

(2.106)

where S is the reference area of the wing related to its root chord squaredj >.
and l are the aspect ratio and relative span of the wingj and p- and p+ are
44 2. Problem Formulation

uniformly valid asymptotic expressions for pressure coefficients on the lower

and upper surfaces of the wing, determined by the formulas

_ [{)<p-
P =2 - U ( t ) - - -
{)<p-] ({)<p-)2
-
({)<p-)2 +O(h),
- -
2 (2.107)
{)x {)t {)x {)z 0

+
p = 2
[{)<p+ {)<p+] 2
()x U(t) - 8t + O(h o ), (2.108)

where <p- and <p+ are uniformly valid asymptotic representations of the ve-
locity potential on the lower and upper surfaces of the wing, determined with
help of formulas (2.102) and (2.103). Analyzing expressions for pressure, it is
easy to see that for the upper surface of the wing, the corresponding formula
is linearized, whereas for the lower surface a nonlinear formula is used. These
representations of pressure are compatible with primary assumptions and
asymptotics obtained for the flow velo city potentials. In fact, for small geo-
metrie and kinematic perturbations the flow above the wing is only slightly
perturbed. On the other hand, in the flow beneath the wing, induced pertur-
bations can no longer be considered small.
To evaluate the lift-to-drag ratio (aerodynamic fineness) oflifting systems
in the ground effect, it is necessary to calculate part of the drag due to normal
loads. At the same time, one has to take into account the fact that due to
low pressure at the wing's leading edge, the so-called suction force may occur,
directed upstream and reducing the drag.
The induced drag of the wing incorporating the suction force can be
determined by investigating the flow far downstream in the 'frefftz plane.
The induced drag coefficient is given by the formula

CXt = S11 1 2
/
1/2
roo(z)awoo(z)dz,

where r oo(z), a woo are the spanwise distributions of circulation and downwash
in the 'frefftz plane, whereas S is the wing's reference area related to the
square of the root chord.
To the leading order, this expression can be represented in the form

C Xi = - h; 1 12
/ <Ph
1/2
(z)äW1 dz. (2.109)

Consideration of expression (2.109) shows that for a fixed lift coefficient the
induced drag coefficient is of the order of O(h o ) and, in extreme proximity to
the ground, diminishes in direct proportion to the relative ground elearance.
Sometimes it is practical to be able to evaluate the suction force
contribution to the induced drag. The magnitude of this force is defined by
the perturbation velocities near the leading edge. As demonstrated by Be-
lotserkovsky and Scripatch [130], the leading edge flow is elose to that in
two-dimensions in the plane, normal to the leading edge planform contour.
 2.7 A Limiting Mathematical Model of the Extreme Ground Effect 45

Near the edge, all components of the perturbation velo city can be shown to
be finite with the exception of component U y , which is normal to the planform
contour and lies in the plane tangent to the lifting surface. If the behavior of
U y is described by the relationship

A
Uy = r-:;' 11 < 0, (2.110)
V-li

where 11, as earlier, is an external normal to the planform contour, then the
suction force coefficient can be calculated by the formula

Cs = S
27l' 1
h
A 2 (l, t) COS(II, x) dl = S
27l' 1 11
A 2 (l, t) dz. (2.111)

Within the theory of the lifting surface in the extreme ground effect, the
perturbation velo city in immediate vicinity of the leading edge is given by
expression (2.57)
8'Ple hoa l
- - '" =F-;~=;=== 11 < O. (2.112)
811 v'
-27l'h ie ll '
Comparing expressions (2.110) and (2.112) and employing formula (2.111),
one can derive the following formula for the suction force coefficient on a wing
in the ground effect:
Cs(t) = hojl/2 al(z, t)2 dz. (2.113)
S -1/2 h*(z, t)
The expression for al was obtained by matching and is represented by formula
(2.70).

2.7 A Limiting Mathematical Model

of the Extreme Ground Effect
To the lowest order (the extreme ground effect) the fiow problem for a wing-
in-ground effect takes a simple form even for the case of curved ground:
• Asymptotic orders of the upper and lower surface velo city potentials:

(2.114)

• Channel fiow equation for the extreme ground effect:

~ (h* 8'Ph ) + ~ (h* 8'P11) = U(t) 8h* _ 8h* . (2.115)

8x 8x 8z 8z 8x 8t
• Boundary condition at the leading (side) edge h:
'Ph = 0, (x, z) E ll. (2.116)

• Boundary condition at the trailing edge l2:

2 [8'Ph U(t) _ 8'P11] _ (8'P11)2 _ (8'Ph)2 = 0, (x, z) E l2. (2.117)

8x 8t 8x 8z
46 2. Problem Formulation

Comment on the Definition of the Leading and Trailing Edges in a

Nonlinear Formulation. So far the notions of the leading (side) edge and
trailing edges have been dealt with without attention to the questions: Where
does the transition lie between these and how can we determine the position
of the corresponding transition point'? It should be said that such transition
points may be fixed apriori by the geometry of the planform, especially if its
tips are sharp; see Tuck [53]. Otherwise, when the planform contour is suffi-
ciently smooth, the position of the transition point is not known in advance,
and its determination becomes part of the fiow problem. In the nonlinear case,
this transition does not necessarily occur at the wing tips, corresponding to
the maximum local span.
Following Newman [54J, one can define the transition point as that where
both leading edge and trailing edge channel fiow boundary conditions are ful-
filled simultaneously. To the lowest order of 0(1), it implies that at the point
(XT, ZT) E II n l2, where subscript "T" denotes the transition point, both
(2.115) and (2.116) hold simultaneously. Introducing the normal-tangent lo-
cal coordinates (/I, r) and designating by X the (least) angle between a tan-
gent to the wing 's planform contour and x axis, one can rewrite expressions
(2.115) and (2.116) at the transition point, i.e., for (x,z) = (XT,ZT) as

rpll = 0, (2.118)

(2.119)

0'Ph . 0'Ph
2 [( ---smXT+--cOSXT )U() O'Ph]
t -- - - (8'Ph)2
- - - (8'Ph)2
- - =0.
0/1 or ot 0/1 or
(2.120)
Substituting (2.118) and (2.11 g) in (2.120), we finally obtain the following
equation for the determining the position of the transition point as part of the
lowest order problem solution:

Orph
0/1 = - 2U()'
t smXT· (2.121)

If the perturbations are sm all (linearized problem), the squares of the pertur-
bation: velocity components in (2.120) can be neglected, which results in the
following equation:
Orph. 0 (2.122)
0/1 smXT = .
Therefrom
sinXT = 0 at the transition point. (2.123)
It is easy to conclude from (2.123) that for a linearized problem and a smooth
wing planform contour, the transition points coincide with the tips of the wing.
3. The Linear Theory of a Lifting System
Moving Close to the Ground

To reduce a nonlinear formulation to the linear theory, one has to assume that
the deflections of the surfaces of the wing, its vortex wake, and the ground,
respectively, from horizontal planes Y = h 1 and y = 0 are small compared to
the ground clearance h, Le.,

IYu,l,w - hl «: h, (3.1)

These assumptions lead to the possibility of imposing boundary conditions

for the wing and the wake on the plane Y = hand boundary conditions for the
ground on the plane Y = O. Note that, upon linearization, the quasi-harmonie
equation (2.22) is reduced to the Poisson equation. It is obvious that within
the linear theory, the pressure is related to the velo city potential through
linear differential operators for both the upper and channel flows.

3.1 Features of a Linearized Formulation

and an Algorithm of the Solution
A linear formulation for the perturbed velocity potential of absolute motion
can be easily derived from the nonlinear formulation in the form
• Equation:
(3.2)

• Flow tangency conditions on the upper and lower surfaces of the wing:

öcp = -U(t) ÖYu,1 + ÖYu,l, Y = h ± 0, (x, z) E S; (3.3)

öY öx ät
Signs "plus" and "minus" correspond to the upper and lower surfaces.
• Flow tangency condition on the ground:

(3.4)

1 Here notation h stands for the characteristic relative ground clearance.

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
48 3. Linear Theory of a Lifting System

• Dynamic condition of pressure continuity across the wake:

(x,z) E W (3.5)

or

U(t) 8rp- _ 8rp- = U(t) 8rp+ _ 8rp+ , y = h, (x, z) E W. (3.6)

8x 8t 8x 8t
We introduce a new time variable

i= lt·
t
U(t) dt, (3.7)

so that rp = rp(x, y, z, i) and rewrite (3.6) as

y = h, (x, z) E W. (3.8)

Equation (3.8) can be demonstrated to be identical to the following set of

equations:

(x,z) E W, x - Xte(z) +i = 0, (3.9)

where Xte = Xte (z) is the equation of the trailing edge in the adopted
coordinate system. 2 In terms of "physical" time t,

(x,z) E W, x - Xte(z) + l t
t·
U(t) dt = O. (3.10)

In fact, the alternative forms (3.9),(3.10) of the dynamic boundary condi-

tion (3.5) in the wake correspond to the Kelvin (Thomson) theorem and
express the jump of the velo city potential across the vortex sheet at an ar-
bitrary point of the latter, if this jump is known at any time at the trailing
edge of the wing. The Kutta-Zhukovsky condition at the trailing edge can
be viewed as incorporated into the above dynamic condition in the wake.
• The continuity of the vertical component of the flow velo city across the
wake:
Linearizing the kinematic condition (2.5), we derive

8rp+ 8rp-
--=--, y = h, (x, z) E W. (3.11)
8y 8y

2 For a wing with a straight trailing edge, Xte = o.

 3.1 Features of Linearized Formulation and Algorithm of Solution 49

• Decay of perturbations at infinity:

\1cp --+ 0, (3.12)

In channel flow, the same asymptotic expansion (2.11), as earlier, can be

utilized, although in the linear case

(3.13)

Using the linearized version of the procedures, demonstrated in 2.2, one can
show that the channel flow equation is identical to the Poisson equation

8 2cp* 8 2cp* 8h* 8h*

8x i + 8z21 = U(t) 8x - at ' (x, z) E S, (3.14)

where h* = h* /h = fli - fjg = 1 + O(e:/h), whereas

( 8h* 8h*) = O(~) (3.15)

8x' 8t h .
For channel flow under the wake a similar equation holds, but in this case it
has to be solved with respect to the induced downwash

(3.16)

The components of the induced downwash in the wake can be determined by

a linearized version of (2.24), namely,

(x,z) E W. (3.17)

Using the dynamic condition in the wake in the Kelvin (Thomson) form
(3.10), one can calculate aw as

2
_ 1 8 [ 1 8 *( *)] 8 *( *) (3.18)
Qw = U(t*) 8t* U(t*) 8t* CPI 0, z, t + 8z2 CPI 0, z, t ,

where t* is related to x and t byequation (3.10).

For a steady flow with asymptotic error O(h), the downwash in the wake is
not dependent on x. With this in mind, determination of aw '" a W1 becomes
still simpler:
(3.19)

The flow above the lifting system and its wake (upper flow) is identical to
what was considered previously. The upper flow potential CPu is of the order
of O(e:) and, to the lowest order, is described by an expansion
50 3. Linear Theory of a Lifting System

(3.20)

in which <PUl is represented by (2.31) and has the "edge" asymptotics given
by (2.32) and (2.33).
Local fiows are linearized consistent with the overalllinearization scheme.
Stretching of local coordinates is performed by the ground clearance h rather
than the local instantaneous distance h:
of the edge from the ground as in
the nonlinear case. Otherwise, all previous results of section 2 hold for the
linear case.
The boundary condition for the channel fiow equation (3.14) at the leading
(side) edge II for an infinitely thin edge can be derived from (2.73) setting
h; = 1. Then
*
<PI =:;h[A 2 - 0<PII
ov ( 1 + In h7r)] ' v = 0 (3.21 )

taking into account the asymptotics of the channel fiow (2.11)

<PlI = 0, v = 0, (3.22)

1 0<Ph
<PI
2
= ----
7r ov ' V = 0, (3.23)

1 0<Ph
<Pb = -[A 2
7r
- ~(1
uV
+ In 7r)], v= o. (3.24)

The boundary condition for the channel fiow equation (3.14) at the trailing
edge l2 for a sharp straight edge can be derived from (2.98) setting = 1. h:
Then
p* = 2 [o<Pj U(t) _ o<pj]
I ox ot

= ~[U(t)((äuw)ln~ -BI) _ 0!2] , v =0. (3.25)

or, with reference to the asymptotics of channel fiow,

PlI = 0, v = 0, (3.26)

v = 0, (3.27)

Pb = ~[U(t)((äuw)ln7r-Bl) _ 0: 2 ].
The coefficients ofthe lift and moments are given by formulas (2.104)-(2.106),
but the press ures on the upper and lower surfaces of the wing are both cal-
culated by using linear differential operators

p+'- = 2 [0<P+'- U(t) _ 0<P+'-]. (3.28)

ox ot
 3.2 Two-Dimensional Steady Flow Past a Slightly Curved Foil 51

Within the linearized formulation, the extreme ground effect case has a still
simpler mathematical description than in the nonlinear theory. Correspond-
ing relationships have the form
2 2 --
8 <Pll +8 <Ph _ U(t)8h* _ 8h* ( ) S (3.29)
8x2 8z2 - 8x 8t' x, z E ;

<Ph = 0, (x,z) E h; (3.30)

8<Ph U(t) _ 8<Pll = 0 () I (3.31)
8x 8t' x, z E 2·

Transition points, separating the leading (side) and trailing edges, are coin-
cident with the wing tips.

3.2 Two-Dimensional Steady Flow

Past a Slightly Curved Foil
Consider a slightly curved infinitely thin foil, moving steadily near asolid
flat ground plane with an angle of pitch (). In this case with asymptotic
error of the order of O(h 2 ), the flow under the foil does not differ from the
one-dimensional, and the channel flow potential is governed by the following
elementary equations:

(3.32)
with the following boundary conditions at the end points of the segment
o~ x ~ 1:
• At the leading edge (x = 1),

<Pll (1) = 0, <P12(1) = _.!. 8<Pll (1);

7r 8x

<P13 (1) = 1[A 2 - 8:~l (1)(1 + ln 7r)]; (3.33)

• At the trailing edge (x = 0),

(3.34)
52 3. Linear Theory of a Lifting System

where Ys (x) = h + E f (x) is a distribution of the gap between the foil and the
ground and f (x) = O( 1) characterizes the position of the foil with respect to
the line y = h. Parameters A 2 and BI have been described earlier.
We construct the upper flow potential by a distribution along the segment
o ::; x ::; 1 of two-dimensional sources (sinks) whose strength is determined
on the basis of the thin body theory as [-2y~(x)l. Additionally, one should
account for the admissibility of a point source solution at the leading edge,
i.e., at x = 1, y = 0 + o. Note that this latter solution does not violate the
flow tangency condition on the upper surface of the foil.
Thus, the values of the upper flow potential at the points of a foil on its
upper surface (0::; x ::; 1) are given by the expression

<PUl Q
= -ln(l
27r
- x) - - 11
7r 0
1
y~(~) ln(x -~) d~. (3.35)

The strength of the point source in accordance with (2.70) is equal to

(3.36)

Asymptotic expansions of a potential <Pu near the edges are obtained in the
following form:
• Near the leading edge,
hQ h hAI hA 2
<Pu ~ - l n lJ
27r
+ _y'
7r s
(1) lJ In lJ + - - l J + --,
7r 7r
(3.37)

where lJ = x-I;

Al = -y~(l) -11 [y~(~) - y~(l)ll ~ ~;

A2 = -11 y~(~) In(l -~) d~;
• Near the trailing edge,

h hB 1 hB2
<Pu ~ - y~(O) lJ In lJ + - - l J + --, (3.38)
7r 7r 7r

where lJ = -x;

BI = ~Q - y~(O) -11 [y~(e) y~(O)l ~~; -

B2 = -11 y~(~)ln~d~.
 3.2 Two-Dimensional Steady Flow Past a Slightly Curved Foil 53

The solution of equations (3.32) with boundary conditions (3.33) and (3.34)
can be written as follows:

CPIl = 1y~(~) d~ =
1
ys(x) - Ys(1) = ys(x) - 1;

dcpIl _ -'( ) dCPh = Q = - (1) -1'

dx - Ys x , dx 2 Ys ,
CPh = Xcp[2 (0) + CPl2 (1) - CP[2 (0);
CPb = XCP[3(0) + CPI3(1) - CP[3(0).
Uniformly valid additive expressions for the velo city potential on the upper
and lower surfaces of the foH are

cP+ = CPUl + h~ [cp~(v) - ~ ln(7l'v)] + h{ CPte(v) - ~Y~(1) hi/[ln(7l'v) -1]}

+h{ CPte(V1) - ~Y~(O) hV1[ln(7l'V1) - 1]}; (3.39)

cP- = CPI + h~ [CP;;:e(V) - v + ~] + h [CPhe(V) - hy~(1) Gv 2 - ~v)]

(3.40)

where v = (x - 1)/h, V1 = -x/ho Functions CPte' CP;;:e, CPte' CPhe are calculated
by formulas (2.43) and (2.52), in which ii has to be replaced by v.
Differentiating (3.39) and (3.40) with respect to x, one can find uniformly
valid expressions for the pressures on the upper and lower surfaces of the foH

+ _ 2 dcp+
p - dx'

It is convenient to take into account the following relationships:

dcpae 1
dv 1 - exp(7l'cpae) ,
We calculate the lift coefficient

Cy = l\p- - p+) dx = 21 1 (d:x- - d:x+) dx = 2[cp+(0) - cp-(O)]

= 2[cpu(0) - CPI(O)] = -2[cpI(0) - hcpUl (0)] + 0(h 2 ).

Setting x = 0 in expressions (3.39) and (3.40) and representing the foH ordi-
nates as Ys = h+ef(x), where f(x) = 0(1), we obtain the following formula
für the lift coefficient of a slightly curved foH in the ground effect:

(3.41)
54 3. Linear Theory of a Lifting System

21 f(~)
where
1
Cl = dx;

C2 = ~[f(l)
7r
+ 1'(0)];

Here are some examples:

• For a Hat plate at an angle of piteh (), the form function f (x) = x and
formula (3.41) is redueed to that obtained by Widnall and Barrows [40]:

C =- () ( l +4h
- l n7r- +2h)
-. (3.42)
y h 7r h 7r

Note that the same result ean be obtained from the asymptotie analysis of
the exaet solution of the nonlinear How problem for a Hat plate in the ground
effect derived by Tomotika et al. [27] for () -+ 0, h -+ 0 and () jh -+ O. Figure
3.1 illustrates the distribution of the pressure eoeffieient along the upper and
lower surfaees of the Hat plate in eomparison with the results obtained by
eolloeation .

• Parabolie are
Let the foil have the form of a parabolie are with relative eurvature De . In
this ease, f(x) = 4x(1 - x), a

2.0 - - - - - - , - - - - - - - - - - , - - -_

_ h pe

1.0 ----+----i------------;~

! Suctioni Side
I I
0.0 00--:-::::="'--"'==+:-==--~----i
i

- -1-
-1.0 -·--+1------'~~-,;----'---___i
I ,
Pressurei Side
I
-2.0:-,---'---+-------c----~-"""

-3.0 --~----------
0.0 0.2 0.4 0.6 0.8 x 1.0

Fig. 3.1. Distribution of the pressure coefficient on the upper and lower surfaces
of a flat plate [40], h = 0.1. The dashed line corresponds to a one-term asymptotic
solution. The difference between the three-term asymptotic solution and the results
of the collocation method (solid lines) is indistinguishable.
 3.2 Two-Dimensional Steady Flow Past a Slightly Curved Foil 55

(3.43)

• Flat plate with a flap

Let the flap have a chord equal to bf and a deflection angle ()f. In this case
the form function is described by the equation f(x) = x, for O:S: x :s: bf,
and f(x) = bf, for bf:S: x. Generally speaking, the foil can be oriented with
respect to the ground at an angle of pitch (), but in linear theory it is sufficient
to study the case () = 0, {)f -# O.
Note that in the vicinity of the hinge of the flap, the channel flow can-
not be considered one-dimensional no matter how small the relative ground
clearance h. Therefore, one has to analyze a loeal flow near the hinge. First
of all, use the local coordinate system Xf = bf - x, Yf = y. Introduce stretched
coordinates Yf = yr/h, Xf = xr/h. After stretching, the local region near the
hinge transforms into a strip (0 :s: Yf :s: 1,lxfl < (0), on the boundary of
which a normal derivative of the flow potential is known. Mapping the strip
onto a half plane and using the Schwartz formula (see Fuks and Shabat (131]),
one can write the expression for the flow perturbation veloeity on the lower
surface of the foil near the hinge as

dCPf ()
dx = - 7Th [ln 11 - exp( -7TXt) I + 7TXtl + R,
where R is a eonstant. In the immediate vicinity of the hinge (Xf -t 0),

dcpf ()f_
-dx '"
- --lnxf
7Th .
Far from the hinge:
• To the leJt,

dCPf
Xf -t -00, dx ~ R + O[exp(7Txt)], (3.44)

• To the right,
dCPf
Xf -t +00, dx ~ -Xf+ R , (3.45)

Turning to the variable x = bf - hXf and setting x to bf, we obtain the

following result:

dCPh (b f + 0) = dCPh (b f - 0), CPll (b f + 0) = CPh (b f - 0) (3.46)

dx dx
It follows from (3.46) that for h -t 0, the magnitudes of both the potential
CPh and the perturbation velocity dcpll / dx in the channel flow are equal when
approaching the hinge from the left and right-hand sides.
56 3. Linear Theory of a Lifting System

Here are uniformly valid formulas for the pressure coefficient, obtained
by additive composition of asymptotic solutions, derived in different parts of
the Row:

1 2: x 2: bf ;
bf 2: x 2: o.
where

_ 2Bf [ 7 r bf d'Pae _]
Pi = -h
7r
hIn hl ( - )1
1 - exp -7rXf
+ h(l - br) + 7rbf - d v- - 7rXf ,

_ 2Bf [ 7 r bf _ ]
P2 = -h
7r hIn hl 1 - exp ( - )1
-7rXf - h(bf + 7r'Pae(vt)) ,

where 'Pae is given by (2.43), v = (x - l)/h, Vi = -x/ho Representative

distributions of the press ure coefficient on a foil with a Rap are plotted in
Fig.3.2.
Due to the fact that the perturbations under the foil induced by the
hinge decay exponentially, the lift coefficient can be derived from the general
formula (3.41) as

() Bf { ( bf 2h 1
Cy=C/Bf=h 2b f l-"2)+-:;;:-ln h (l+b f )
+ 2: [1 + (1 + br) In 7r + (1 - br) In 1 ~ bf] + O(h 2 ) }. (3.47)

1.2 1 I Pressure Side : bf =0.3

h pe f I I h =0.10:
0.8 ii~---t--jf--<~=====~=-'7-:
! ! h = 0.05:
! I ·
I

h = 0.05:
~~~~===:;:========~
I,'

~;
0.0 j :

I
·0.4 '-----r---~--.:-------'----\r'
, i

I Suctionl Side
I !
·0.8 ,--'- - - ' - - - - - - ' - - - ' - - - - - - -
0.0 0.2 0.4 0.6 0.8 1.0
x
Fig. 3.2. Distribution of the pressure coefficient on the upper and lower surfaces
of a flat plate with a flap for different magnitudes of the relative ground clearance,
bf = 0.3.
 3.2 Two-Dimensional Steady Flow Past a Slightly Curved Foil 57

Parameter hc~r versus ground clearance and the length of the flap is
plotted in Fig. 3.3. In Fig. 3.4 a comparison is presented of the results of the
present theory with calculated data of Shadrin [132J based on the method of
the 7-parameter.

2.0 r -- ---,-- - - . - - - -- - - - - - - - - ,

0.2
0.4 r=-i --::=:::+:====-r----l
0.0 L--_ _...l.-_ _--L._ _ _ --'~ _____'
0.00 0.04 0.08 0.12 h 0 .16

Fig. 3.3. Parameter hCZ r versus the relative ground clearance and the length of
the flap , .A = 00 .

0.8 !--rr-f---+-+----r- ----=---+---i

0.4 "'
, -.j.--+-+--+-+-~--!-__t--1

0.0 I'---'----'---'-----'--~-----'--'
0.1 0.2 0.3 0 ." 0 .5 0 .6 0 .7 0 .8 0 .9 1 .0
bt
Fig. 3.4. Comparison of the calculated results for a flat plate of infinite aspect
ratio, h=O.15 . Solid !ine: MAE; circles: from [132] .
58 3. Linear Theory of a Lifting System

3.3 A Wing of Small Aspect Ratio

in a Steady Linearized Flow

Consider another limiting case of a wing of small aspect ratio >. « 1. In

this case, due to elongation of the wing in the direction of the flow, instead
of a three-dimensional Laplace equation, a two-dimensional equation can be
considered:
[Ptp {)2tp
{)y2 + {)z2 = O.
For a wing with a slow increase in local span chordwise, it is sufficient to
investigate a flow in a transverse plane passing through the trailing edge of the
wing. Here it is convenient to render the problem nondimensional by a semi-
span ll2 of the wing and speed Uo and introduce relative ground clearance,
based on the span of the wing at the trailing edge hj = hll. For a flat wing, the
corresponding velocity potential should satisfy the flow tangency condition
on the wing and on the ground:

{)tp =0 y = o.
()y ,

Following the general algorithm, one can demonstrate the main stages of the
solution. For h « >. « 1, the channel flow is described by the following
relationships:

1
<PI = <PI! + 2h l <p12 In -h
2 j
+ 2h l <p13 + O(h 2l ),
{)2tpj ()
(3.48)
{)z2 2h j '
In accordance with the general scheme, the potential of the upper flow tpu
is constructed by placing along the segment -1 ::::: x ::::: 1 a distribution of
sources with strength () I h j with the addition of admissible point sour ces at
the points z = ±1. These point sources model (in the upper flow) the leakage
of air from beneath the wing around its edges. Then

tpUl Q
= -ln(l
2n
- z 2 ) - - ()
2h j n
1 1

-1
ln(z - () d(

= !{ In(l -
2n
z2) - -h() [(1
2 jn
+ z) In(l + z) + (1 - z) In(l - z) - 2].

For a vanishingly thin and flat side edges, the structure of the edge potential
remains the same as earlier:
 3.3 Wing of Small Aspect Ratio in a Steady Linearized Flow 59

where iI = (z - 1) /2h l at the right tip and iI = (-1 - z) /2hl at the left tip of
the wing.
Asymptotic matching of the velo city potentials <PI, <Pu, and <Pe leads to the
following results
• Constants ai

z = ±1;

a3 = ~(al-
27r
~ln~);
hl hl

a4 = .!.7r [-al In ~
2h l
+~
2h l
(2 -ln 2)] ;

• Boundary conditions for the channel flow potential under the wing

<PI(±l) = 2hl ( a4 - ~)
or
al
<Ph (±1) = 0, <Ph(±l) = --;
7r

<Pb (±1) = ~ [-al(l + In 7r) + 2~1 (2 -ln2)].

For the problem of the leading order,

() 2
<PIl = 4h l (z -1),

wherefrom it follows that in the extreme ground effect, the span-

wise distribution of circulation for a wing of small aspect ratio
is parabolic. It is easy to see that the corresponding downwash distribution
is uniform. In accordance with Munk's theorem, such a wing is optimal, Le.,
has minimum induced drag for a given lift coefficient.
Including approximations of the order of O(h l ), the following expression
for the lift coefficient of a wing of small aspect ratio can be obtained:

(3.49)

To the leading order, the induced drag coefficient can be derived by using
formula (2.109) in the form

or in Prandtl's representation,

(3.50)
60 3. Linear Theory of a Lifting System

0.5

Cy II
0." I

I
0.3 ,
I
I
0'1
0.1
1
I

0.0 '
0 2 3
" 9°
5 6

Fig. 3.5. Comparison of calculated results for a Hat plate for different ground
clearances with experimental data (A = 1, solid lines: theory; circles: experiment).

Formulas (3.49) and (3.50) hold for a small-aspect-ratio wing of arbitrary

planform. 3 Thrning to a rectangular wing I = A, hl = h).. = h/\ one can
derive from (3.49), (3.50) that

BA ( 6h).. 7r ) (}A 2 ( 6h 7rA) (3.51 )

Gy = 6h).. 1 + -:;;:-ln h).. = 6h 1 + 7rA In h '

A
p, = 37rh' (3.52)

where h is the relative ground clearance based on the root chord, and Ae is
the effective aspect ratio of the wing. Figure 3.5 presents a comparison of
calculated results for A = 1, obtained by (3.51), with experimental results of
Ermolenko et al. [133].
Within the assumption of a small aspect ratio, it is easy to consider the
steady flow problem for a wing with curvilinear lateral curvature. Let the gap
distribution be described as

h*(x, z) = 2h 1H(z) + Bx,

where h1 is the ground clearance based on the span, () is a pitch angle, and
H(z) is a function of the order of 0(1) characterizing the form of the wing in
the lateral direction, H(O) = 1. In this case, the linearization of (2.22) results
in the following channel flow equation for a small-aspect-ratio wing with
a curvilinear lateral axis:
3 Strictly speaking, the theory of a wing of small aspect ratio implies that the local
span increases monotonically in the downstream direction.
 3.3 Wing of Small Aspect Ratio in a Steady Linearized Flow 61

~ [H(z) d'Pl] = ~, Izl ::; 1. (3.53)

dz dz 2hl

This equation should be solved with boundary conditions

In the general case of a wing with a nonzero angle of heel a1 (1) i= a1 (-1)

The potential of the upper flow 'Pu for a wing with a curvilinear lateral axis
has the form
'Pu ~ 2h l 'PUl'
1 1
'PUl = -a1(1) ln(l - z) + -a1( -1) ln(l + z)
1r 1r
e
- - h [(1 + z) ln(l + z) + (1 - z) ln(l - z) - 2].
21r 1

To the leading order, the expressions for the coefficients of the lift and
lateral moment of wings of a small aspect ratio with a curvilinear lateral axis
are
_ ~(J _ J 1 J 2 ) (3.54)
m x - 8h l 3 J '
where
1 d( 1 (d(
/ /
J= -1 H(()' J1 = -1 H(()'
Taking into account that, to the leading order, the lift coefficient of a fiat
plate of small aspect ratio is equal to GyfP = B>'/6h l , we obtain the following
formula, which allows estimating the relative influence of the lateral curvature
of a wing in the ground effect

Gy =
GyfP
~ (J2
2
_ Jr).
J
(3.55)

Consider some simple lateral forms of the lifting surface. Lateral configu-
ration in the form of a "hat" is described by the function

z E [-1,1].

In addition, we introduce parabolic and elliptic lateral configurations, repre-

sented, respectively, by the functions
62 3. Linear Theory of a Lifting System

'I
4 r-~~~--~-----T-----+----~

2 ~
. ----+-~~~--~----_+----~

0.2 0.4 0.6 0.8 1.0

Fig. 3.6. Relative influence of the form of the lateral cross section of the small-
aspect-ratio wing upon the lift coefficient in the extreme ground effect (1 : "hat;" 2:
parabolic; 3: elliptic).

5 1-----------------~--------_,

I
4 1r----+--------~r=--~~~

,1C=::l:===--=--=y-----ri
IG:)
1
I
I
II
:1
0.0 0.1 0.2 0.3 0.4 0.5
~
hA

Fig. 3.7. Relative lateral moment coefficient versus heel angle for small-aspect-ratio
wings with different spanwise configurations (1: "hat;" 2 : parabolic; 3: elliptic) .

Figure 3.6 illustrates the relative influence of the lateral curvature of the
wing upon its lift coefficient for different magnitudes of the gap Ct between
the tips and the ground plane. Figure 3.7 shows the relative influence of
the angle of heel upon the lateral moment coefficient m x for small-aspect-
 3.4 Steady Flow Past a Wing of Arbitrary Aspect Ratio 63

ratio wings of different spanwise configurations. This parameter characterizes

the static lateral stability of the wing, Le., the capacity to restore its upright
position after the action of heeling perturbations. Note that Fig. 3.7 compares
the behavior of the ratio of the lateral moment coefficient m x for a given
configuration to that of a fiat plate m Xfp versus heel angle ß.
The induced drag coefficient for a wing with a curvilinear lateral axis can
be found by using (2.109) in the form
C2
C - Y (3.56)
Xi - 7rA./L'

3.4 Steady Flow Past a Wing of Arbitrary Aspect Ratio

To avoid cumbersome algebra, consideration will be confined in this section
to the simplest case of the extreme ground effect for a wing with a straight
trailing edge. In the steady fiow problem, the velocity potential does not
depend on time. Consequently, there are no unsteady free vortices in the
wake, and the corresponding fiow model is substantially simplified.
Recalling the formulas for induced downwash Qw at the trailing edge, one
should mention that to the lowest order of approximation Qw ~ QW1' and
downwash does not depend on the abscissa x of the point in the wake, Le.,
d 2 <p11
QW1 = h dz 2 (0, z). (3.57)

Therefore, if the wing is in steady motion and its relative clearance is small,
then at all points of the wake with the same z coordinate, the downwash
is "almost,,4 the same as that behind the trailing edge. This important cir-
cumstance allows writing formula (2.109) for determining the induced drag
coefficient (accounting for suction force) for the case of small relative ground
clearances in the following way:

C Xi = -sh 1 1/ 2

-1/2
<pdO, z)
()2<p
()
Z
;1 (0, z) dz, (3.58)

where, as earlier, S is the reference area of the wing related to the square of
the root chord. In addition, the condition of optimality of the wing (in the
extreme ground effect) can be written as a requirement that the downwash
distribution at the trailing edge should be uniform in the spanwise direction.
As indicated in 3.1, in the case of the extreme ground effect, it is sufficient
to solve the following problem:
• Equation:
{)2<p11 {)2<ph {)Yl
{)x 2 + {)z2 = ()x'
(x,z) E S; (3.59)

4 With an asymptotic error of the order of O(h).

64 3. Linear Theory of a Lifting System

• Boundary conditions:
<Pi} = 0 at the leading edge h,
PI
1
= 2 °<Ph
OX = 0 at the trailing edge. (3.60)

For a fiat wing, y{ = Ö = 0/ h, where 0 is a pitch angle.

The Hft and moment coefficients can be calculated by using the formula

Gy = -821 1 2
/
-1/2
<Ph (0, z) dz; (3.61)

(3.62)

We consider the example of a rectangular wing, for which the domain

S is a rectangle 0 ::; x ::; 1, Izl ::; >"/2, where >.. is the aspect ratio. At first,
suppose that the wing is fiat, so that the right-hand side of equation (3.59)
is equal to Ö = 0/ h, where 0 is the angle of pitch. We seek <Ph in the form

ÖL Xn(x)cosqn z ,
00

<Pli = (3.63)
n=O
where qn = 7r(2n + 1)/>... Expression (3.63) should vanish for z = ±>"/2.

f
Noting that
1 =.i (_l)ncosqnz
>.. n=O qn
and substituting (3.63) in Poisson equation (3.59), we obtain an ordinary
differential equation for functions Xn(x):

(3.64)

A general solution of (3.64) can be sought in the form

( )-
Xnx-an sinhqnx b coshqnx 4(-1)n
cosh qn +n cosh qn \3·
/\qn

Bearing in mind that the boundary conditions for the nmction <Pli at the
ends of the interval 0 ::; x ::; 1 are

<Pli (1, z) = 0, 0<PIi ( )

8x O,Z = 0,

we obtain the conditions for the functions Xn(x) at the ends of the segment
o ::; x ::; 1:
X~(O) = 0,
 3.4 Steady Flow Past a Wing of Arbitrary Aspect Ratio 65

0.372 1.698
~ <p >

__ ~O

A-O.5
Fig. 3.8. Distribution of the aerodynamic loading on a rectangular Hat wing in the
extreme ground effect.

wherefrom

Finally, we find the following expression for the channel flow velocity
potential:

Cflt = '"\
4e ~ (_l)n (coshqnx
~ --3- h
)
-1 cosqn Z .
A n=O qn COS qn
The distribution of aerodynamic loading to this order of approximation is

_ + 0Cflt
p - p = (p) c:: Plt = 28
~ (-1 sinh qn x
= ~ ~ --2- h
se
cosqn Z .
t
x n=O qn COS qn

In Fig. 3.S examples are given of the distribution of parameter h(p) /(), which
characterizes the aerodynamic loading, for rectangular wings of aspect ratios
A = 1 and A = 3. The loading distributions have been ca!culated on the basis
of the leading-order solution.
The lift coefficient can be calculated as

G = G O() = eije = _3. />./2 (0) d = 16 () ~ tanh qn tanh(qn/ 2)

y Y Y A Cflt ,Z Z hA2 ~ 4 .
->'/2 n=O qn
(3.65)
The longitudinal moment coefficient with respect to the leading edge is

mz=mß=m z()=,
o ö- 2 />./2
A ->'/2 0
1 1
(x-1)~dxdz
OCflt
uX

16 () ~ 1 (tanh qn qn )
= - hA 2 ~ 4" + tanh qn tanh 2 - 1 . (3.66)
n=O qn qn
The induced drag coefficient is

GXi -- _~
A
J>' ( ) 02Cfl1 ( ) d _ S ()2 ~ tanh 2 qn tanh 2 (qn/ 2)
Cflt 0, Z 0 2 0, Z Z - hA 2 ~ 4 .
->'/2 Z n=O qn
(3.67)
66 3. Linear Theory of a Lifting System

As h « 1, the formulas obtained above are valid for practical arbitrary aspect
ratio h « >. :s 00. The distance of the center of pressure from the leading
edge is defined by the ratio x p = -mz/Cy and m z and Cy are determined
by (3.65) and (3.66).
If >. -+ 00, we obtain the formulas for a fiat plate of an infinite aspect
ratio:
o 1
Cy = h' xp = 3' (3.68)

It follows from (3.68) that for a plate of an infinite aspect ratio moving
in very dose proximity to the ground plane, the center of pressure
is located at a distance one-third of the chord from the leading
edge. Recall that in unbounded fiow the center of pressure of a fiat plate
(>' = 00) is located at one-fourth ofthe chord from the leading edge. 5
For a small aspect ratio >. -+ 0, the general formulas yield the following
results:
o>.2 0>' (3.69)
Cy = 6h = 6h>.'
160>.3
mz = - h-rr5 '

(3.70)

Some results of calculations of the aerodynamic coefficients of a rectan-

gular wing versus the aspect ratio>. are plotted in Fig. 3.9 with ten terms
retained in the corresponding series. The terms of the series in expressions
(3.65)-(3.67) decrease with the increment of n as 1/(2n + 1)4. For not very
large >., it is practically sufficient to retain one term, so that
160>.2 -rr -rr
Cy ~ h-rr4 tanh ~ tanh 2>' '

m rv - 160>.2 [tanh ~ (~ + tanh.!!...) - 1]

z - h-rr4 >. -rr 2>' '
80 2 >.22 -rr 2 -rr
CX1 ~ h-rr4 tanh ~ tanh 2>..
It is interesting to compare the infiuence of the aspect ratio in the ex-
treme ground effect and out of ground effect. Such an analysis is presented

5 Note that the case under consideration corresponds to the order relationship
o« Cl! « h « 1, complying with the assumptions of linear theory: both the
relative ground clearance and the angle of attack are small, and the latter is
always smaller than the former.
 3.4 Steady Flow Past a Wing of Arbitrary Aspeet Ratio 67

1.0 I
0.8 1
.~
~
~
eGI 0.6
.c:
()

'e
0

GI
c:
0."
>-
-g
.i

2 .. i.. 6
Aspect Ratio

Fig. 3.9. The aerodynamie eharaeteristies of a reet angular Rat wing in the extreme
ground effeet versus the aspeet ratio.

!!?
-g ~ 0.6 \---+-~--~~-+!.:..---+----i
GI!!?
-0-
.~'ö
~Cii
()~"E
c3 0...
~CII
...J=
CII_
> 0
~5
CII:o=
0:: 'in
o
0-

2 3 .. 5 6
Aspect Ratio i..

Fig. 3.10. The inRuenee of the aspeet ratio of a Rat rectangular wing upon the lift
eoeffieient and the distanee of the center of press ure from the leading edge (solid
lines: extreme ground effect; dashed lines correspond to h = 00).

in Fig. 3.10, where the results of the present theory for h -+ are compared
with those obtained by Belotserkovsky and Scripatch [130] for h = 00.
°
It is easy to determine the suction force coefficient es. Calculating the
factor al
68 3. Linear Theory of a Lifting System

0CP1, 4B ~ (_l)n
al = ~(l,z) = h)'" ~ --2-tanhqncosqnz,
uX n=Ü qn
and substituting this expression in formula (2.113) with € = Band h* ~ h,
we obtain the suction force coefficient in the form

(3.71)

In the limiting cases,

h>
B2
for )... -+ 00,
Gs =
B2)...2
for )... -+ 0, Gs = 12h'
Comparing the latter formula with (3.70), one can see that for the case of
the extreme ground effect, quite similarly to the case of unbounded flow,
realization of suction force on a wing of small aspect ratio reduces the induced
drag by half.
Now, assume a spanwise distribution of the pitch angle of the wing
sections in the form
B(z) = Bo8(z).
The solution of problem (3.59)-(3.60) can be readily constructed in the fol-
lowing form:
• Flow potential:

1T(2n + 1)
qn = )... . (3.72)

• Corresponding lift coefficient:

(3.73)

• Induced drag coefficient:

(3.74)

• Coefficient of the moment around the leading edge:

__ 4Bo ~ Bn(-l)n (tanh qn

mz - h)'" ~ 3 + tan h qn tan h qn _ )
1 , (3.75)
n=ü qn qn 2
where Bn are determined by the following formula
 3.4 Steady Flow Past a Wing of Arbitrary Aspect Ratio 69

On =:x21>../2 8(z)cosqn zdz .

->"/2

Following Widnall and Barrows [40], consider the case of the wing of
semielliptic planform with a straight trailing edge. In this case the domain
S is represented by a semiellipse with axes equal to 1 and l/2, where l is the
relative span ofthe wing. Equation (3.59) and the boundary conditions (3.60)
are satisfied, if the function CP1l is sought in the form

CPh = d (x 2 + 4l~2 - 1), (3.76)

where dis a constant. Substituting (3.76) in Poisson equation (3.59), in the

case of a Hat semielliptic wing,

Ol2
d = --:-=--...,.. (3.77)
2h(l2 + 4)'
wherefrom
Ol2 ( 4z 2
CP1l = 2h(l2 + 4) x 2 + r - 1) .
Taking into account that for vanishing ground clearances the distribution of
circulation spanwise is equal to

one can see that the loading of a semielliptic wing has a parabolic distribution
spanwise.
The lift coefficient becomes
881 2
(3.78)
Gy = 31l"h(l2 + 4)·
For a semicircular wing (l = 2) the lift coefficient is two times less than for an
semielliptic wing of an infinite aspect ratio (l -+ (0). It is interesting that the
limiting results for l -+ 00 do not coincide with those for the two-dimensional
problem because in the limit the wing retains a semielliptic planform. In fact,
in the two-dimensional case,
o
Gy = h'
whereas here for l -+ 00,
80
Gy = 31l"h.
It is worthwhile to mention that for a semielliptic wing in the extreme
ground effect, the induced downwash in the wake is constant spanwise:
 70 3. Linear Theory of a Lifting System

40
a W1 = l2 + 4' (3.79)

so that such a wing is optimal for any aspect ratio. 6 The induced drag coef-
ficient is given by

_ 20Cy _ C; 41 2 )..
(3.80)
C Xi - l2 + 4 - 7r)..p,' P, = 37r2)"h = 12h·

Calculating the factor al taking into account the equation of an ellipse, and
using expression (2.113), one finds the suction force coefficient of a semielliptic
wing in the form
(3.81)

Use has been made ofthe formula).. = 4l/7r for the aspect ratio of a semiellip-
tic wing. It is easy to check that expressions (3.78), (3.80) and (3.81) comply
with the evident requirement CXi = OCy - Cs for a flat wing. The influence
of the aspect ratio on the lift coefficient and the indueed drag eoefficient for
semielliptie flat wing in the extreme ground effeet is illustrated in Fig. 3.11
in eomparison with a reet angular wing.
Now, we can turn over to the case of a reet angular wing with a flap.
Let the wing of aspect ratio ).. have a flap whose ehord is equal to bf and
defleetion angle is Of. The domain 8 bounded by the wing planform con-
tour eonsists of two rectangular subdomains 8 1 (b f ~ X ~ 1, Izl ~ ),,/2) and
8 2 (0 ~ X ~ bf , Izl ~ ),,/2). Due to the linear formulation, the effect ofthe flap
deflection angle can be studied separately from that of the pitch. Therefore,

.l!l
c:
CD
~ I

~ I
c; 0.6 1-' ---+----,t=---i-7"'~+-__+-__I
!!!
o
~
-5 0.4
1
Fig. 3.11. The infiuence of the

~
:5
0.2
I,
>--I-r-tf--t---+---j-----::-:+----j
hC~91
planform of the wing on the lift
and induced drag coeflicients for
different magnitudes of the as-
I:' , - ----_'1 ___ _
0.0 '-"""'---'---_-'-_--L_----1_ _..:.....::....:...::...=!
pect ratio (dashed lines corre-
spond to a semielliptic flat wingj
o 2 3 4 5 6 solid lines to a rectangular fiat
Aspect Ratio 1.. wing).
6 This conclusion correlates well with an exact solution of the problem of the
optimal wingj see de Haller [134].
 3.4 Steady Flow Past a Wing of Arbitrary Aspect Ratio 71

it is possible to set {} = O. Consequently, on part 8 1 of the wing area, Le.,

between the leading edge and the hinge axis of the flap, the right-hand side
of the Poisson equation should be equal to zero. On part 8 2, i.e., between
the hinge of the flap and the trailing edge of the wing, oYs/ox = iJf , where
iJf = {}r/h. Within the mathematical model of the linearized extreme ground
effect, we find the following formulation:

epl (X Z) - { epll(X, z) for (x, z) E 8 1 ,

1 , - epI2(X, z) for (x, z) E 8 2;

02epll + 02epll _
ox2 oz2 -
°, (x, z) E 8 1 ;

epll(l,z) = epll(x,±>'/2) = 0;
02ep12 02ep12 _ {}-
ox 2 + oz2 - 1> (x, z) E 82,

epI2(X, ±>'/2) = 0, O;~2 (0, z) = 0;

Accounting for the result of the solution of the local flow problem in the
vicinity of the hinge (3.46), one can formulate the additional conditions at
the hinge axis:

Applying the same method as for the wing without a flap, we derive the
solution in the form

- ~[ coshqn x 4(-1)n]
ep12 = {}f ~ bn h b - >. 3 cosqn z ,
n=O cos qn f qn
where the coefficients an and bn are equal to

The lift coefficient can be found as

C - (}fGOf _ 16{}f ~ tanhqn bftanh(qn br/2) +tanh[qn(l- bf))tanhqnbf

y - y - h>.2 ~ q~{l + tanh[qn(l - br)] tanhqnbf} .
(3.82)
The coefficient of the longitudinal moment with respect to the hinge axis
of the flap is
 72 3. Linear Theory of a Lifting System

1.0 -----,-- - 'I- -.....,2-::"

o-::"ea-se- -::;:::::::::;
h C~f i
0.8 hc:'=2bt (1- bt /2 )'

I
0.<1
Ii----1.-f--r--p.4--+--::= : : - - i

Fig. 3.12. The lift eoeffieient

of a reet angular wing in the ex-
treme ground effeet versus the
relative ehord of the fiap for dif-
0.0 'E::::._.J..._ _.!---'-:-~0~8:------:'1 0 ferent magnitudes of the aspeet
0.0 0.2 0.<1 0.6 . bt .
ratio.

2 0 .16 ,,....- - - -- - - - - ; - - -- ---;

hC .9f
XI 0.'<1
Ii---i-- --'-- ---'--.;----;-- --:

0.12
I bt=1
~--i----,.,L.~....---+---t----1

i
0.10 r--+-I--:----;----':....::-T---+--~

; 0<1

Fig. 3.13. The indueed drag

eoeffieient of a rectangular
wing in the extreme ground
effeet versus the relative ehord
2 3 .. 5 6 of the fiap for different magni-
Ä tudes of the aspeet ratio.

tanh[qn(l - br)] h [qn(1- bf )]}

an tan . (3.83)
qn 2
The coefficient of the induced drag

C
x;
= 80f f
h)...2 n=O
{tanh(qn bf/2) + tanh[qn(1- br)]}2 tanh 2 qnbf
qH1 + tanh[qn(l - br)] tanhqnbfl2 .
(3.84)

Figures 3.12 and 3.13 illustrate the dependence of the parameters hCt!, and
hCx;/Of upon the aspect ratio of the wing for different chords of the fiap.
 3.5 Harmonie Oseillations of a Thin Foil in a Two-Dimensional Flow 73

3.5 Harmonie Oseillations of a Thin Foil

in a Two-Dimensional Flow
Consider a thin foil advancing near the ground with a constant speed Uo and
simultaneously performing small vertieal motions (heave, pitch or deforma-
tions ). An unsteady wake, consisting of the shed vortieity, is formed behind
the foil in the process of such a motion. According to the general scheme
discussed previously, the channel flow in this case will be described by the
following relationships:

(3.85)
where
IPI c::: IPi(x, t) = IPh (x, t) + hIn ~IPb(X, t) + hIPh(x, t),
ä s is a nmction of the order of 0 (e / h), characterizing the vertieal com ponent
of the flow velo city given on the foil.
The solution of equation (3.85) can be written in the form

(3.86)

where C 1 (t) and C2 (t) are the functions of time to be determined with the
help of the boundary conditions for the channel flow potential at the leading
and trailing edges.
The lowest order induced unsteady downwash in the wake can be deter-
mined by using the formula

where t* = t + x. In the case of harmonie motions,

IPl 1 (0, t*) = a cos kt* + b sin kt*,

where a and bare constants determined in the course of solving the lowest
order problem, k = wCo/Uo is the Strouhal number, and w is the circular
frequency of the unsteady process.
In the upper flow, the corresponding asymptotic expansion for the velo city
potential has the form

Solution IPh is constructed by distributing along a semi axis x ::; 1 sour ces
(sinks) whose strength is equal to the doubled local downwash, which is
74 3. Linear Theory of a Lifting System

given on the foH and determined in the wake. At the leading edge, we place
an admissible point source (sink) solution that does not violate the flow
tangency conditions either on the foH or in the wake. Thus, at points on the
upper surfaces of the foH and the wake

c,oUl = Q
211" In(l - x) + 211"
1 1
-00
1
q(e, t) ln(x - e) de, (3.87)

where
q(x, t) = { -2_äs (x, t),
-2aW1 (x, t),
°: :;<x x:::;:::; 0.
-00
1;

Note that the strength of the point source (sink) at the leading edge has
to be determined by matching. Note that when calculating c,oUl with the help
of (3.87), the following integrals are encountered:

h = 1 00
cos k(ln (d(, h = 1 00
sink( ln(d(;

These integrals are divergent in the conventional sense. Nonetheless, they can
be calculated in a generalized (Abel-Poisson) sense; see Fikhtengoltz [135J.

11 = lim (OO exp( -8() cos k(ln (d(

~-?o Jo

= l~{ - k2 ~ 82 [~8In(k2 + 82) + karctan + 8'Y]} = - ~. i (3.88)

12 = lim (OO exp( -8() sin k( In ( d( = lim [8 arctan ~ - k'Y

~-?o Jo ~-?o "I

+2
!k I (k 2 82)] = { (ln k - 'Y)/k for k > 0, }
n + for k = 0, '° (3.89)

where "I !::: 0.5772 - the Euler constant.

Prom a physical viewpoint, the generalized integration means that we
consider slowly amplifying oscillations, turning in the limit into oscillations
with constant amplitude. An analogous approach was used by Theodorsen in
calculating integrals of the type

1 00
sin k( d(,

in the work on harmonie oscillations of a thin foH in an unbounded flow.

Taking into account (3.88) and (3.89), the expression for c,oUl can be rewritten
as
Q ln(l - x) - -
c,oUl = -2
11"
1
11
äs(e, t) ln(x - e) de
11" 0
 3.5 Harmonie Oseillations of a Thin Foil in a Two-Dimensional Flow 75

where
GI = Lcoskx + M sinkx, G2 = Lsinkx - M eoskx;
L = si(kx) - sin kx lnx, M = 2ln k + eos kx lnx - ci(kx).
Functions si(kx) and ci(kx) are integral sine and eosine, defined by the ex-
pressions
{CO sin ( ci(kx) = _ {CO eos( d(.
si(kx) = - Jkx -(- d(;
Jkx (
To match, one needs to know the asymptotic expansions of !PUl near the
edges. These were found in the form
• Near the leading edge (v = x-I --t 0)

Q ä s (l, t) Alv A 2 2
'Pu ~-lnv+ vlnv+-+-+O(v);
1 27T 7T 7T 7T

Al = -ä s(l, t) - (I [äs(~, t) - ä s(l, t)] d~ c - k 2[(aG 2l - bGu) eos kt

Jo 1- '>
+(aG u + bG 2l ) sin kt],

A2 = -11 äs(~,t)ln(1-~)d~+k[(aGu+bG2l)Coskt+(bGu-aG2l)sinkt],
where G u = G l (l) and G2l = G2(1) .
• Near the trailing edge (v = -x --t 0),
(ä uw )
!Pu = - - v In v
1 7T
+ -vB!
7T
+ -B7T2 + 0 ( v 2) ;
(ä uw ) = äs(O, t) - ä Wl (0, t),
1 (I ~
BI = "2 Q- äs (O,t)- Jo [äs(~,t)-äs(0,t)]T+k{[a(kG2o-1)-bkGlO]eoskt

+[b(kG20 - 1) + akG lO ]sin kt};

B2 = -11 äs(~, t) ln~ d~ + k[(aG lO + bG20 ) eos kt + (bG lO - aG2o ) sinkt],

7T 'Y -ln k
GlO =Gl (0)=-2k' G20 =G2(0) = k·
The following results ean be obtained from matching the solutions determined
in the upper, ehannel, and edge regions.
76 3. Linear Theory of a Lifting System

• Constants of the loeal solution at the leading edge (x = 1):

a3 = ~[Al +ös (l,t)(l-lnX))' a4 = ~(A2 -allnX).

• Constants of the loeal solution near the trailing edge (x = 0):

bl = 1,

• Applieation of the Kutta-Zhukovsky eondition at the trailing edge (0 =

-x/h = 0) gives a relationship between b2", b3" and bt, bt:

_ 8b3" _ + 8bt _ 1 [ _ ( 11") 8B2 )

b2 + 8t - b2 + at -;: B l + (auw ) I-ln h + 8t .
• Boundary eonditions for the equation (3.85):
At the leading edge,

CPl(I, t) = ~ [A 2 - al (1 + In X))·
At the trailing edge (x = 0),

8B2 11")
8CPl 8CPl
8x - 8t =;:h ( -Bl - Tt _
+ (auw)lnh: .
As a result of matehing, additional information was obtained that provides
uniqueness of the asymptotie solutions, determined in different regions of
the How. Forming uniformly valid (additive) eomposite expressions for the
pressure eoeffieient on the upper p+ and the lower p- surfaees of the foil and
integrating these expressions, we obtain formulas for the lift and moment
eoefficients

where
P+ = 2 (8cp+ _ 8cp+) p-_2(8cp-
- --- 8cp-).
8x 8t' 8x 8t
As a first example, eonsider heave oscillations of a Hat plate, for which in-
stantaneous positions with respeet to the ground are deseribed by the equa-
tion
Ys = h + ho sin kt,
whereas the vertieal downwash on a plate
_ ho
a s = -hkeos kt.
 3.5 Harmonie Oseillations of a Thin Foil in a Two-Dimensional Flow 77

The lift coefficient was obtained with an asymptotic error of the order of
O(h 2 ) in the following form:

(3.90)

h 2 + k2 h 1 2(1 + k 2 )2 + 2 - k2
hCy = - 2(1 + k 2 ) - ;: in h (1 + k2)2

_~ {3k 2 (1 + k 2 ) + 2(1 + in 7l')[k 2 + (2 + k 2 )(1 + k 2 )]_ k3 [(1 - k 2 )G l1 + 2kG 21 ] }.

7l' 2(1 + k2)2 '
(3.91 )

;;, 2-k 2 h 1(I-k 2 )(2+k 3 )

hCy = 6(1 + k2) + ;: in h (1 + k2)2
~ {2(1 + in 7l') - 2k 2 1n 7l'(2 + k 2 ) - k 2 [2kG l1 + (k 2 - I)G 21 ]}
(3.92)
+ 7l' 2(1 + k2)2 '
where
G l1 = G 1 (1) = si(k) cosk + [21nk - ci(k)] sink;
G 21 = G 2 (1) = si(k) sin k - [21n k - ci(k)] cos k.
The solution obtained here is valid for a small clearance (h « 1) and a
large range of Strouhal numbers k « l/h. Some calculated results based on
formulas (3.91) and (3.92) are presented in Figs. 3.14 and 3.15.
On the same graphs are plot ted the numerical results of Efremov [104],
obtained by the discrete vortex method for h = 0.1. In a similar manner

-0.4 ~-~--.,--------,---------,

Ii
I
i
-0.8 [ - -i,
-

-1.2 1----+'--:"L-------;..--..L---"""1

-1.6 L-_--'--_ _'---_--'-_ _. L - _ - - - . J

0.0 0.4 0.8 1.2 1.6 k 2.0

Fig. 3.14. Aerodynamie derivative hC; for a foil of infinite aspeet ratio in
heave versus Strouhal number (solid lines: formula (3.91); dashed lines: eolloca-
tion method).
78 3. Linear Theory of a Lifting System

0.8 1'-- - . , . . - - - - -- ---.-------,

h c~ 0.6 l r-~-+--_,_--'---+--_;

0.4 r----1<'----+----- -+- ---i

0.0 0.4 0.8 1.2 1.6 k 2.0

Fig. 3.15. Aerodynamic derivative hC~ far a foil of infinite aspect ratio in
heave versus Strouhal number (solid lines: formula (3.91); dashed !ines: coIIoca-
tion method).

it is possible to derive formulas for the aerodynamie derivatives for pitch

oscillations around a point with abscissa a o and oscillations of a flap
around a hinge point. For briefness, the eorresponding expressions for
aerodynamic eoeffieients will be written only for the extreme ground effect
(leading order):
• For pitch oscillations,

(3.93)
where

• For oscillations of the flap,

Cy = fhCZf + BrCif, (3.94)

hellt = 2b r - b; + k 2 br[1 - br + b;(b - 4)/12]- k4bt(1 + 3bd/6 O(hl~)
y 1 + k2 + n h '

heih = br (2b r - 2bl/3 - 1) + k 2 bl(1 - bd2) O(h I .!.)

y 1 + k2 + n h .
The solution obtained above for the general ease of harmonie perturbations
is valid both for oseillations of the foi! as asolid body and for deformations
of the foi!.
 3.6 Three-Dimensional Unsteady Linearized Lifting Flows 79

3.6 Three-Dimensional Unsteady Linearized

Lifting Flows
For a wing of finite aspect ratio, all derivations can be carried out to the
order of O(h), which is practically enough to provide reliable information
for relative ground clearances up to 0.1-0.15. However, here the analysis will
be restricted to the least cumbersome (from the viewpoint of the algebra
involved) case of the extreme ground effect. Moreover, the concrete case of
an oseillating fiat plate with a rectangular planform will be eonsidered. Due
to linearization, the effects of eamber ean be superimposed.
Let a finite wing perform heave oseillations with the following deseription
of its instantaneous position with respect to the ground:

Ys = h + h o sin kt,
where h o is the amplitude of the heave and k = wCo/Uo is the Strouhal
number. The downwash at points on the wing is given by the function
ays ays -
as = ax - 7ft = -hok eos kt,
For a rectangular wing of aspeet ratio A, the following boundary problem has
to be solved:

(3.95)

(3.96)

(3.97)
For harmonie oseillations, the solution ean be represented as

ipl, = ipll eos kt + ip12 sin kt. (3.98)

Substituting (3.98) in (3.95), we obtain the following (time-independent)

problems for functions ipll and ip12:

ipll(l,t) = ip12(1,z) = 0, ipll(X, ±A/2) = ip12(X, ±A/2) = 0,

aipll a<p12
8x = kip12; 8x = -kipll, X = o.
The solution of these problems was found in the form
80 3. Linear Theory of a Lifting System

where
X coshqn(x -1) b coshqn x 4(-1)n
in = aln h
COS qn
+ in COSh qn '3'
/\qn

X 2n = a2n coshqn(x -1) b coshqn x 11"( )

h
cos qn
+ 2n h '
COS qn
qn = \" 2n + 1 ,
/\
and the solution series coefficients are
4( _1)nk 2 tanh 2 qn tanh(qn/2)
aln = - Aq~(q~ + k2 tanh 2 qn) ,

The lift coefficient for heave oscillations is obtained in the form

.
hC; = 'k
2
/\
1>./21 (-a- - kCP12)
->'/2 0
1 acpll
X
dx dz;

h 2
hCy = - 'k 2
/\
1>./21 (aCP12
->'/2 0
-a +
1

X
)
kCPll dxdz,

where
h= kh o cos kt,

C~ and C; are the aerodynamic derivatives of the lift coefficient with respect
to the heave rate and heave acceleration:

(3.100)

The results of calculating the aerodynamic coefficients for heave oscilla-

tions are presented in Figs. 3.16 and 3.17. Ten terms were retained in the
series. One does not have to retain many terms because the series under con-
sideration converges quickly. For not very large A, it is sufficient to leave just
one term, as in the steady-state case.
For the limiting case of an infinite aspect ratio, the above formulas yield
the two-dimensional results of Barrows and Widnall [136]:

A -t hCi. = _ 1 + O.5k 2
1 + k2 '
00,
y
 3.6 Three-Dimensiünal Unsteady Linearized Lifting Flüws 81

· 0.0
hCh I
Y
GI
>
-0.1 I
<'0
GI
~

.
.50
'"~
:;; 4
t
.~ I
k=5
c!i
'e
(.)

<'0
c
>- -0.6
"0
eGI
<{

234 5.. 6
Aspect Ratio

Fig. 3.16. The aerüdynamic derivative hC; für a rectangular wing in the extreme
ground effect versus the Strouhal number.

~
<'0
GI
~
.50
~ '"
~
>
~u
'e<'0
-0.05
1
c
>-
~GI
-0.10 I
<{ I
-0.15

-0.20 .
I 14 .0 3.0

0 234 5 .. 6
Aspect Ratio

Fig. 3.17. The aerüdynamic derivative hC~ für a rectangular wing in the extreme
ground effect versus the Strüuhal number.

For small aspect ratios, the corresponding limiting formulas become

hCy
h
= - {5'
,\2 hCii = _ ,\2
y 6
(1 _ 192,\).
7[5

The suction force coefficient of the wing performing heave oscillations can
be found with the help of formula (2.113) in the form

C5 = C sh it 2 + Chhith
y + c shh h 2 '
82 3. Linear Theory of a Lifting System

where
00
hh 1 ~ 2 ( al n ) 2
Cs = 2h L qn --h- - bIn tanh qn ,
n=O cos qn

. ..
hh
Cs = -h
1
L
00 q2n 2n
k2 h
a (aIn
--h- - bIn tanhqn ,
)
n=O COS qn COS qn

.... 1 00q2 a 2
Chh =_ ~ n 2n
s 2h ~ k4cosh2qn'
with coefficients aIn, a2n and bIn, b2n determined previously. In the limiting
cases of wings of very large and small aspect ratios, the above formulas can
be considerably simplified.
For ). -+ 00,

.. 1 (2
C hh = _ -+- k 2)2 C hh _ --:-.,----1--:-~
s 4h 1 + k 2 ' s - 4h(1 + P)2 .
For ). -+ 0,

Note that in the case of heave, the suction force coefficient averaged through-
out the period of oscillation is not zero. This means that the wing experiences
a thrust. It is easy to see from the formulas for the suction force coefficient
that this thrust increases when the reference relative ground clearance h be-
comes smaller.
If the wing performs pitch oscillations, the instantaneous position of
the wing's surface is described by the equation

y = h + Bo(x - ao) cos kt,

where Bo is the amplitude of pitch oscillations, and a o is the abscissa of the

center of rotation. The lift coefficient for pitch oscillations is obtained in the
form

hCy() = "\
4L ~- (_l)n [( kb2n )
- ain - - - tanhqn + (bInk+a
--2n ) qn
tanhqn tanh-
/\ n=O qn qn qn 2

k 2bIn - 2a o) ] ,
+-2-(1 (3.101)

hCyo = - 'k
4 ~ (-l)n [( kbIn) In
L - - a2n + - - tanh qn + (b2n
k- a
- - ) tanh qn tanh -qn
/\ n=O qn qn qn 2

-2kb In ] , (3.102)
 3.6 T hree-Dimensiünal Unsteady Linearized Lifting Flüws 83

1.0

h C 9y
.c:.
!::! 0.8
ii:
...
,5
41
.~
'ii 0.6
>
'e::
c!
.!:! 0.4
E
'"
c:
>-
~ 0.2
~ Fig. 3.18. The aerodynamie
derivative hC~ für a reetangular
wing in the extreme ground ef-
0.0 feet versus the Strouhal number,
0 2 3 4
Aspee1 Ratio ao = 0.5.

0.3
9
hC y
.:g .· 3 4 k=5
Q.
0.2 "
...
.E
41
>
~
>
'e::
0.1
c!
c.J
'E
'"c:>-
0.0
~
~ Fig. 3.19. The aerüdynamie
k=O
derivative hci for a reetangular
wing in the extreme ground ef-
-0.1
0 2 3 4 5 6 feet versus the Strouhal number,
Aspect Ratio
A.
ao = 0.5.

where B = Bo cos kt, iJ = -Bok sin kt,

7r
qn = ~(2n + 1) ;

W 2 + kW1 tanhqn / kn
a2n = 2;
1 + k 2 tanh qn/q~

k2bln
W1 = - ( 1 - ao
- --- + a o) , 1n ( 1 + tanh qn tanh -qn) .
kb -
W2 = -
qn cosh qn qn 2
Some results are presented in Figs. 3.18 and 3.19 that illustrate the behavior
of the aerodynamic coefficients hCZ a nd hci versus the aspect ratio for
84 3. Linear Theory of a Lifting System

different magnitudes of the Strouhal number. For >. -t 00, formulas (3.101)
and (3.102) yield results of the corresponding unsteady two-dimensional fiow
problem. For >. -t 0, the following expressions are obtained:

hCy() =
>.2 [1 + k 2(1 -
'6 2a o ) (12 -;:>.)] , . >.2
hC(}y = -6 (a 0
+1-
3>'
- )
7f'
4. Nonlinear Flow Problems for a Lifting
System in the Extreme Ground Effect

4.1 A Curved Thick Foil

in a Two-Dimensional Ground Effect

First, we consider an example of a fiow problem for a moderately curved

thin foil in the ground effect, l and then present some results for thick foils.
Essentially, as discussed at length in section 2, the procedure for the solution
uses the assumption that, for h « 1 and € = O(h), nonlinear effects exhibit
themselves mainly in the narrow channel under the foil. The foil and the
corresponding coordinate system are shown in Fig. 4.1.

h
x

Coordinate System for a Foil in Ground Effect

The Foil with Trailing Edge Touching the Ground

Fig. 4.1. A foil of moderate thickness and curvature near the ground.

1 The term "moderately curved" implies that the distances of points on the foil
contour from the horizontalline y = h are of the order of O(h).

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
86 4. Nonlinear Flow Problems

The complete problem for the perturbation velo city potential has the form

(4.1)

ßcp
ßn = cos( n, x) on the foil; (4.2)

ßcp = 0 on the ground; (4.3)

ßy
'Vcp --t 0 at infinity. (4.4)
The Kutta-Zhukovsky condition can be applied in the form of the require-
ment that the velo city (pressure ) should be continuous across the trailing
edge.
The application of the general algorithm to the problem under consider-
ation gives the following:
• The upper flow region Du
We describe the local gap as Ys = hys, where Ys = Ys(€, x) is a function of the
order of 0(1) and € = c/h, c = O(h) is the parameter of perturbation (e.g.,
maximum relative curvature, thickness, angle of pitch). We seek the upper
flow potential CPu as
CPu = hcpUl + 0(h 2), (4.5)
where 'PUl satisfies the following problem:

(x,y) E Du; (4.6)

y = 0 + 0, 0 ~ x ~ 1; (4.7)

ßUl =0 y = 0 + 0, x > 1 and x < 0; (4.8)

ßy ,

'V CPUl --t 0, (4.9)

The solution of the upper flow problem can be written in a straightforward
manner. The velo city potential at points on the upper surface of the foil is
defined by the expression

(4.10)

Asymptotic expansions of 'PUl near the edges are obtained in the following
form:
 4.1 A Curved Thick Foil in a Two-Dimensional Ground Effect 87

• Near the leading edge,

Ql 1 Al A2
!.pu c:::- -lnv + -y~(1) vlnv + v - + - , V -t 0, V = x -1,
I 27r 7r 7r 7r

Al = -y~(1) -11 [y~(~) - y~(1)11 ~ (

A2 = -11 y~ ln(1 - ~) d~;
• Near the trailing edge,
1 _I BI B2
!.pUl C:::--Ys(O)vlnv+v-+-, v-tO, v= -x, (4.11)
7r 7r 7r

BI = ;1 -y~(O) -11 [y~(~) - y~(O)l ~~,

• The channel fiow region D I ,
The nonlinearity of the ground effect exhibits itself mainly in DI. After
introduction of a stretched vertical coordinate y = yjh into the full problem,
we obtain the following channel fiow problem with respect to the perturbation
potential !.pI:
{)2!.p1 1 {)2!.pl _ 0.

:1
{)x2 + h2 {)y2 - , (4.13)

= h2y'(X) (~:I - 1), Y= Ys(~, x); (4.14)

{)!.pl =
{)y °, y = 0. (4.15)

Note that in the upper fiow limit, the infiuence of the condition at infinity is
lost.
We seek !.pI as an expansion

(!.pj,!.pj*) = 0(1), (4.16)

where
1
!.pI = !.plI + hIn h!.p12 + h!.pl. (4.17)

Substituting (4.16) in (4.13)-(4.15), we obtain

!.pI = !.pj(t,x); (4.18)

d (_ d'PI) _I
0::; x::; 1, (4.19)
dx Ys dx = Ys'
88 4. Nonlinear Flow Problems

Le., for h « 1 and t = 0(1), the flow under the foil is "almost" one-
dimensional 2 and is governed by the elementary equation (4.19). The solu-
tion of this equation can be derived in the form

(4.20)

where the constants Cl and C 2 are found from the boundary conditions at
the ends of the interval x E [0, 1J. Because the function Ys can be represented
as
N
ys(t,x) = 1 + L€.iJj(x), (4.21)
j=l

the function 'Pi depends nonlinearlyon t .

• The edge regions D e
Near the edges of the foH, we introduce isotropie stretching of coordinates
- y
y= h'
As previously discussed in a more general case, the distances of the points
on the foH from the horizontal line passing through the trailing edge are of
the order of O(h). So, with an asymptotic error of 0(h 2 ), the flow tangency
condition can be imposed on the lines Ys = y(l) (near the leading edge) and
Ys = Ys(O) (near the trailing edge). This simplification enables us to formally
utilize the edge fiow solutions obtained within the linear theory.
The leading edge flow velo city potential

(4.22)

where
1
'Pae = - Je, 'Pbe -
_ y~(1)Ys(l) [-(I
V Je
-1) _ J;]
2' (4.23)
11" 11" 11"

v= V/ Ys (1), and Je is determined through the equation

1I"V = 1 - exp Je + Je. (4.24)

Near the trailing edge,

_ _ V
V = V = h. (4.25)

The solution for 'Pbe is found from (4.23), substituting ys(l) and y~(l) by 1
and j}~ (0), respectively.

2 To the order of O[h, h In(l/h)J.

 4.1 A Curved Thick Foil in a Two-Dimensional Ground Effect 89

Matching gives the following results:

(4.26)

(4.27)

a4 = ~ [A 2 - alYs(1) in hY~(1)]; (4.28)

b1 =1, b2=~[Bl+Y~(0)(1-ln~)], b3+=~2. (4.29)

The boundary conditions to determine the constants of the solution (4.20)
were found in the form

(4.30)

The lift coefficient is obtained by integrating the composite express ions for
the pressure coefficients on the upper and the lower surfaces of the foil, i.e.,

(4.31 )

where the upper and lower surface contributions to the press ure coefficients
are defined by the formulas

p+ = 2 d<p+ on the upper surface, (4.32)

dx

:x -(:x)
and
d - d - 2
p- =2 on the lower surface. (4.33)
Neglecting the terms of the order O(h 2 ) and higher, we arrive at the following
expression:

where
(4.34)

(4.35)
90 4. Nonlinear Flow Problems

We consider so me examples:
• Flat plate (Ys = 1 + Bx, B= () j h, () - pitch angle)

G = B _ + 2h B(2 + ~) In.!. + 2h B(ln ~ + 1 + In 1T). (4.36)

Y 1 + () 1T (1 + ()) h 1T 1 + () 1 + ()
For vanishing B ~ 0, formula (4.36) yields the linear result of Widnall and
Barrows [40]; see formula (3.42). For the extreme ground effect (first term),

(4.37)

It follows from (4.37) that

for B~ 0, Gy ~ B; for B~ 00, Gy ~ 1.

These results indicate that when both incidence and ground clearance
tend to zero, permutation of the limits (() j h ~ 0 or () j h ~ (0) yields different
results.
The limit () j h ~ 00 implies that the trailing edge of the foil touches the
ground before the pitch angle becomes equal to zero. It can be concluded
from observation of (4.37) that, if one measures the ground clearance from
the leading rather than the trailing edge, i.e.,

h1e = h+(), (4.38)

then
B () ()
Gy = 1+ B= h + () = h 1e '
(4.39)

that is, the lift coefficient of a Rat plate found from the nonlinear solution
becomes linear in ().3 This conclusion holds exactly to the lowest order.
Figure 4.2 shows so me results, obtained by using formula (4.36), in com-
parison with calculated data of Grebeshov et al. [137].
The next example is related to a parabolic foH, for which Ys = 1 + Bx +
48c x(l - x),8c = Jjh, B= ()jh,

(4.40)

qc = V(B + 48c )2 + 168c ,

2 - - -
G2 = -[() + (() + 4Jc )(1- GI)],
1T

3 Panchenkov (1974) was the first to notice this feature of a nonlinear solution for
a flat plate.
 4.1 A Curved Thick Foil in a Two-Dimensional Ground Effect 91

1.2
i I,
I

~, I I I
\\\ '- K
1.0

', I1

\ ~~K-
0.8
a~3

""
0.6
, "-:
~ ~II
i
0.4
,\, 2" ~~
-..::...~
~ '-"-
~~
0.2
1"
- -- - ....... I

0.0
I
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
h
Fig. 4.2. A comparison of the asymptotic theory (solid lines) with the calculated
data of [137] (method of discrete vortices: dashed lines) for a fiat plate of infinite
aspect ratio.

C3 = ~ [Ö(ln ~ - 1 (}_) + 48e + [(Ö + 48e ) In 7r - 48e ](1 - Cd].

7r 1+(} 1+
For Ö -t 0 and 8e -t 0 we obtain the linear result (3.43).
Another example is a flat plate with a flap at the trailing edge. For
this case, Ys = 1 + (Ö + 8r) x, 0 ::; x ::; bf 4 and Ys = 1 + 8f bf + Öx, bf ::; X ::; 1,
where bf is the chord of the flap and 8f represents the flap deflection angle,
8f = 8f/h,
C =1- 1+bf (Ö+bf8r) . (4.41)
1 (1 + () + öfbr )[1 + bf «() + Öf)] ,
2 - - - -
C 2 = -[obf + () + «() + 8r)(1- Cl];
7r
C -
3 -
~7r {(8f bf +
Ö) (ln 7r _ 1
1 + 8f bf + () 1 + () + 8f bf
)

-8d(1 - br) In(1 - br) + brln bf ]

+[8f (lnbf - bf + 1) + (Ö + 8r) In 7r](1 - Cl)}'

If Ö = 0 and 8f -t 0, expression (4.41) yields the corresponding linear

results; see formula (3.47). Some of the calculated results for the lift force co-
efficient of a moderately curved foil-in-ground effect are presented in Figs. 4.3
and 4.4.

4 Note that in this formulation the ground clearance is measured from the trailing
edge of the fiap.
92 4. Nonlinear Flow Problems

2.0,,.-------...,------:---,--,

Cy I
1.5 ~-t.-_i_--L-f_-+-+-+_---1

1.0 I

0.5 f--+--I--"..~7'""""'fC=...:,,4=--_+"~~'-'-1

0.0 hL-i-/'-""'/<"---+.--+--i---j--+--j

-1.0 L-...-l_-.L_.-:..._-L-_-'----lL---L_---'
-2 -1 0 2 3 4 5 6
9°
Fig. 4.3. The lift coefficient of a parabolic foil-in-ground effect versus pitch angle
and relative curvature, h = 0.1.

2.5
Cy I
1

2.0 I

1.5

1.0

0°
0.5

-1 o 2 3 5 6
9
Fig. 4.4. The lift coefficient of a flat foil with a flap in the ground effect versus
pitch angle and flap deflection angle, h = 0.05, bf = 0.3.

In what follows, so me results are presented for a moderately thick foil 5

in a steady ground effect. For a foil with maximum relative thickness c5t , the
ordinates of the upper and lower surfaces can be defined as

5 As previously, this terminology implies that the maximum relative thickness is

of the order of the relative ground clearance h. This assumption is practical and
allows analyzing the influence of the thickness up to 10-15% of the chord.
 4.1 A Curved Thick Foil in a Two-Dimensional Ground Effect 93

1.2.
Cy
1.0 I
0.8

0.8

0.4

0.2

0.0
-2 -1 0 2 3 4 5 6
e"
Fig. 4.5. The influence of thickness on the lift coefficient of a foil with a flat lower
surface and a parabolic thickness distribution, h = 0.05.

Yu(x) = h + (}x + 8du(x), Yl(X) = h + (}x - 8dl(X),

where (} is the angle of pitch, and functions fu(x) and fl(X) describe the
positions of the upper and lower surfaces of the foH with respect to the chord
line.
The lift coefficient is found in the form

(4.42)

1
where
1 dx
Cl = 1 - -2()' y = y/h;
o Yl x
2
C2 = -[Yu(l) - 1 + y~(O)(l - Cl)];
'Fr

C3 = ~{B2-A2+a1 [I-al +In Yu~I)] +(l-CyJ[y~(O)(1n 'Fr-I)+y{(O)-B1]};

a1 = 2'
Q1
B 2 - A2 = Jo
r y~(e) In -e-ede,
1 I -

B1 = ~1 - y~(O) -l\jj~(e) - y~(O)] ~e.

For the limiting case of zero clearance (h = 0), one can deduce from (4.42)
the formula that corresponds to the situation when the trailing edge of the
foH slides along a flat ground surface:
94 4. Nonlinear Flow Problems

yt
1.5
I
I,

y U
1.4 I

1.3

/) , ' 1I.· ' ' I II

/
I

I I
1.2

J!V
,
,

1.1
/'
1.0 V
o 2 4 6 8
I
I
10

Fig. 4.6. A comparison ofthe asymptotic theory (solid line: formula (4.44); dashed
line: formula (4.44a)) with the exact solution [28] (circles) for the case when the
trailing edge of a flat plate touches the ground.

(4.43)

In the particular case of a Bat plate for which Yu = Yl = 1 + Bx, B= ()/h,

the third term of (4.43), vanishes, and the formula is reduced to

Gy = 1 + -;-ln
2()
o.
7r
(4.44)

Formula (4.44) is in satisfactory agreement with the exact theory of Tomotika

and Imai [27J and Dätwyler [28J up to angles of pitch of 5-6°. We can extend
the range of validity of (4.44) by constructing an asymptotically equivalent
expression
(4.44a)

which is in agreement with the exact theory up to angles 12-15°.

Consider a moderately thick foil with a Bat lower surface and the upper
surface in the form of a parabolic arc. In this case Yl = 1 + Bx, iJu = 1 + Bx +
4Jt x(1 - x), Jt = ot/h, where Ot is the maximum relative thickness. The lift
coefficient is found in the form

+ 7r(18J+hB) (-B + In h
t 7r )
Gy = GYPlate - 1 , (4.45)

where GYPlate is the lift coefficient of a Bat plate, as given by (4.36). If the
trailing edge of the foil touches the ground, B= 00, and it follows from (4.45)
that
 4.2 One-Dimensional Flow Model for a Flying Wing with Endplates 95

2B 'Fr 8c5t
C =l+-ln-+-.
Y 'Fr B 'Fr

4.2 A One-Dimensional Flow Model for a Flying Wing

with Endplates

The general formulation of the ftow problem for a wing in elose proximity
to the ground, presented in section 2, covers a wide range of aerodynamic
modes of the operation of ground-effect machines for h « 1.
In what follows, attention will be attached to the particular case of a
jlying wing with endplates in the extreme ground effect, when relative gaps
under the tips of the endplates are small.
In this case, the leakage of air from under the lifting surface is hampered,
resulting in a considerable improvement in performance. Whereas for h ---t 0,
the description of the ftow under the wing is independent of the vertical
coordinate (see section 2), for vanishing gaps between the endplates and the
ground, the channel ftow becomes almost one-dimensional. Accounting for the
fact that the upper ftow contribution can be shown to be ofthe order of O(h),
one can conelude that for a lifting system in the extreme ground effect
and small clearances under the tips of the endplates, the dominant
nature of the flow is one-dimensional.
A simple one-dimensional model of channel ftow with leakage was first
introduced by Gallington et al. and will be called the G-theory herein. It was
assumed therein that the ftow parameters are independent of the chordwise
coordinate and that the leaking ftow escapes into the external region of at-
mospheric pressure. To account for the intensive generation of vortex sheets
emanating from wing's side edges, the G-theory implies that separation oc-
curs at the tips of the endplates. Though simple, the G-theory of channel flow
agrees qualitatively with experiments and provides useful similarity criteria,
convenient from the viewpoint of processing test data and designing vehieles.
As pointed out by Ando [62], Gallington's ftow model does not exhibit the in-
finite (logarithmic) increase of velo city at the gap encountered in other ftow
models. However, due to the assumption that the ftow parameters (veloc-
ity, pressure) are not dependent on the chordwise coordinate, the G-theory
cannot be used to predict the moments and characteristics of longitudinal
stability. Secondly, the model under discussion does not account for edge ef-
fects, thus preventing determination of such characteristics as, for example,
the suction force at the leading edge.
In what follows, an extended one-dimensional jlow model is introduced
for a wing with small gaps under the endplates; see Rozhdestvensky [63].
This new model accounts for chordwise variation of the channel ftow velocity
and incorporates unsteady effects. It can be used for evaluating the efficiency
and stability of a simple ftying wing configuration in the extreme ground
effect. It also produces formulas, useful for processing of experimental data,
96 4. Nonlinear Flow Problems

identifieation of parameters of the lifting system, and eventually, ean serve

as a tool of eoneeptual and preliminary design.
Below, follows a derivation of the governing equation for unsteady
flow past a lifting surface with small gaps between the tips of the endplates
and the ground.
We reeall a general limiting problem, formulated earlier for the velocity
potential of the absolute motion of a fluid when h -+ 0; see paragraph 2.7.
This problem is governed by the quasi-harmonie equation

:x [h*(x, z, t) ~!] + :z [h*(x, z, t) ~~] = -! h*(x, z, t), (4.46)

whieh ean be obtained from (2.115) by replacing the absolute potential 'Pli
with the potential of relative motion cjJ and h* = h* jh o with h* = h*(x, z, t).
As previously, the latter quantity represents the loeal clearanee under the
wing. All funetions and parameters are rendered nondimensional by using
the root ehord Co and a eharaeteristie velo city Uo.
Restricting the analysis to a reet angular wing (see Fig. 4.7), we average
equation (4.46) spanwise by using the integral operator
11>./2
C =).. {}dz,
->'/2
where >. is the aspeet ratio of the wing.
Assume additionally that the clearanee distribution funetion depends only
on the longitudinal eoordinate and time, Le., h* = h*(x, t). Then

o { * 0 [1 1>./2 ]} 1 * [ocjJ >. ocjJ

ox h (x, t) ox ).. ->'/2 cjJ(x, z, t)dz +)..h (x, t) oz (x, 2' t)- oz (x,
>.]
-2' t)

a [11>'/2
= - ot ).. ->'/2 h*(x, t)dz
] (4.47)

y
WING
FLOW

Fig. 4.7. A rectangular wing

with endplates in the extreme
ground effect.
 4.2 One-Dimensional Flow Model for a Flying Wing with Endplates 97

and, introducing the spanwise averaged potential ~, we obtain

o [ * O~] 1 * [ocjJ,\ ocjJ ,\] oh*(x,t)

ox h (x,t)ox +X h (x,t) oz(x'2,t)- oz(x'-2,t) =- ot .
(4.48)
Equation (4.48) includes a function ocjJ(x, ±,\/2, t)joz that represents the
transversal velo city component under the wing in the close vicinity of the
endplate. For a practical case of symmetricalleakage, when ocjJ/ox(x, ,\/2) =
-ocjJ/oz(x, -,\/2), we obtain from (4.48)

!..-
ox
[h*( ) O~] 2h*(x, t) ocjJ ( ~) =
x, t ox + ,\ oz x, 2
oh*(x, t)
(4.49)
ot
We relate the velocity vep of transversalleakage to the spanwise averaged
longitudinal velo city o~/ox(x, t), assuming that the perturbed pressure out-
side of the endplate is equal to zero. Let 8ep (x, t) represent an effective gap
under the endplate (nondimensionalized with respect to the chord). Then the
velo city v ep of the leakage is given by

Vep(x, t) = h*(x, t) ~~ (x, t, ~)/8ep(x, t), (4.50)

and the corresponding dynamic condition just outside of the endplate takes
the form

p(x, t) = U 2 (t) - vep(x,

2
t) - [O~]2
ox (x, t) - 2 o~
ot (x, t) = 0, (4.51)

where U(t) is a function, describing the time variation of the speed of the
vehicle cruising. It follows from (4.50) and (4.51) that

ocjJ(x t ~) = ± 8ep (x, t)

oz "2 h*(x,t)

8ep (x,t). A( )y'IA( )1 (4.52)

= h*(x, t) slgnp x, t p x, t ,
where ß(x, t) is the span-averaged pressure coefficient under the wing. Taking
into account (4.51), equation (4.49) takes the form

Equation (4.53) is a one-dimensional description of the flow in a

highly contracted channel under the lifting surface with endplates
that accounts for the lateral leakage of air through the gaps under
the endplates. Essentially, such a description corresponds to a small ground
clearance and small (with respect to the characteristic height h above the
98 4. Nonlinear Flow Problems

ground) gaps under the tips of the endplates. Note that the plus in front of
the square root term in (4.53) corresponds to the outward leakage, whereas
the minus corresponds to the inward leakage. As accounted for by the signum
function in (4.53), the direction of leakage depends on whether at a given
moment and for a given station x along the chord, the pressure coefficient
under the wing is positive (pressure) or negative (suction). In the former
case, the leaking flow is directed from under the wing into the external area,
whereas in the latter case, it is directed inward.
Note that the definition of the effective gap 8ep under endplate depends
on the choice of the flow model of the leakage from under the endplate (or
flap).6
Two boundary conditions necessary to solve equation (4.53) ought to be
determined by smooth blending (asymptotic matching) of the channel flow
with local flows near the leading and trailing edges. It has been shown in
section 2 by using local flow solutions that for the extreme ground effect case
(h « 1), an appropriate boundary condition at x = 1 (leading edge), has the
form
~(1, t) = o. (4.54)
From the physical viewpoint condition (4.54) signifies that in the extreme
ground effect, the vorticity in the channel under the wing is accumulated
(due to the development of the boundary layer), counting from the leading
edge.
The boundary condition at x = 0 can be obtained from the requirement
of pressure continuity at the trailing edge and by using asymptotic estimates
for h --+ 0 of the press ure coefficient above the wing ß = 0 (h) and in the
channel beneath the wing ß = 0(1). Introducing an effective gap 8f(t) under
the trailing edge, we can write the corresponding dynamic condition as

2 2 8~
p(x, t) = U (t) - Vf (t) - 2 8t (0, t) = 0, (4.55)

where Vf(t) is the velo city of the flow leaking out from under the trailing
edge. This velocity is related to that of the channel flow in the immediate
proximity of the trailing edge by the flow continuity relationship

(4.56)

wherefrom taking into account (4.55), we deduce the second boundary con-
dition for equation (4.53) at x = 0:

(4.57)

6 It may, for example, incorporate the factor of the contraction of the flow leaking
from under the endplate.
 4.2 One-Dimensional Flow Model for a Flying Wing with Endplates 99

In addition to boundary conditions (4.54) and (4.57), an approprate initial

condition should be imposed. For example,

:: (x, 0) = vo(x). (4.58)

The lift and moment (with respect to the trailing edge) coefficients can
be calculated by using the formulas

Cy(t) = 1 1
ß(x, t) dx, mz = 1 1
xß(x, t) dx,

where the pressure coefficient is given by the expression

On the basis of the solution of a one-dimensional unsteady nonlinear flow

problem past a rectangular wing with endplates in the extreme ground efIect,
we can derive the induced drag coefficient. As previously in the general
case, the induced drag is defined as the difIerence between the pressure drag
in inviscid flow and the suction force acting upon the leading edge. Hence,
the corresponding coefficient can be written as

where Cxp is a drag coefficient due to the action of the normal aerodynamic
loading in the longitudinal direction (ideal pressure drag) and C s is the suc-
tion force coefficient. To obtain Cxp , we project the spanwise averaged pres-
sure forces, obtained within the present theory, onto the longitudinal direc-
tion. For a small rear flap, we can determine separate contributions of the
wing CXpw and of the flap C Xpf to the pressure drag of the lifting system.
Accounting for the coordinate system adopted in this book, one can derive
the following expression for C xpw :

CXpw(t) = 1
1
ß(x,t)cos(n,x)dx = -ho 1ß(x,t)~~dx,
1
(4.59)

where ili = yl/h o , Yl = Yl(X, t) are the ordinates of the points on the lower sur-
face of the wing, measured from the unperturbed position of the underlying
surface. In the case of a flat wing, it follows from (4.59) that Cxpw = -Ci}.
The ideal pressure drag of the rear flap can be obtained by using the
solution of Gurevich [138] for the drag of a wedge with streamline separation.
Renormalizing the expression of the pressure drag coefficient obtained by
Gurevich with respect to the velo city on the (jet) free boundary and the
area of the wing planform, we obtain a simple formula for the pressure drag
100 4. Nonlinear Flow Problems

coefficient of a rigid flap in separated flow 7

(4.60)

Now, we can turn to the derivation ofthe coefficient ofthe suction force act-
ing upon the leading edge. This force presents an integrated effect of suction,
occurring due to the large curvature of streamlines in the vicinity of the lead-
ing edge. In principle, the suction force contribution should be determined by
integrating the projection of the local suction force over the contour of the
leading edge, but it is rather difficult to single out this force numerically. At
the same time, calculations show that, as the foH becomes thinner, the local
suction increases whereas the radius of the leading edge decreases. These two
factors vary so that their product remains finite and almost constant up to
the limiting case of zero foH thickness. That is why it is often assumed that
the theoretical value of the suction force determined at the leading edge of
zero thickness can be utilized for practical rounded edges with a finite radius
of curvature. By using formula (2.39), the asymptotics of the flow velo city
of the relative motion at the points of the leading edge of the wing in the
extreme ground effect can be written as

ßl/l1e = -U(t)
ßx
+h 0 al
ß<pae
ßx
+ O(h 2 )
0 ,
(4.61)

where <Pae is determined by formula (2.42). Matching (4.61) with the channel
flow velo city v(x, t) and accounting for the asymptotics of (2.47) leads to the
following expression for al:

al = h*(1, t)[U(t) + v(1, t)J. (4.62)

Substituting (4.62) in formula (2.113) to determine the suction force on a
wing in the extreme ground effect, we obtain

(4.63)

Eventually, the induced drag coefficient can be written as

Examples of calculation of the induced drag coefficient for particular cases

will be presented later.
Thus, a one-dimensional nonlinear formulation has been found for the
aerodynamics of the longitudinal motion of a rectangular flying wing with
7 The assumption of the ideal fluid separation scheme on the flap is adequate for the
power augmentation mode at rest. It can also be considered practically adequate
for the cruise mode in a pronounced ground effect due to the domination of the
channel flow contribution.
 4.3 Steady-State Solutions for Flow Past a Wing with Endplates 101

endplates in the extreme ground effect. This formulation indudes equation

(4.53), boundary conditions (4.54) and (4.57), and initial condition (4.58).
This mathematical model accounts for unsteady effects, which can be caused
by motions of the wing as a rigid or deformable lifting body, the action of
control devices and wind-wave perturbations. Though sufficiently simple, it
retains the inherent nonlinearity of the aerodynamics of the extreme ground
effect with respect to the geometry and kinematics of the lifting system. The
common sense basis of simplifying the flow model consists of the observation
that when the endplate tip clearances are very small, the channel flow, already
"squeezed " vertically for h -+ 0, becomes almost one-dimensional.
This formulation can be used to evaluate the aerodynamic characteristics
and preliminary design of a jlying wing configuration for both cruise and
power augmentation performance. Note that equation (4.53) describes the
main contribution to the aerodynamics of the wing-in-ground-effect vehide,
namely, that of the channel under the main lifting surface. The contribution
of the upper surface of the wing can be added, using the general asymptotic
approach presented in section 2. Alternatively, due to experimental evidence
that, in dose proximity to the ground, the upper surface pressure distribution
varies insignificantly with variation of h, the upper surface characteristics can
be, with a certain degree of approximation, borrowed from the unbounded
fluid case.

4.3 Steady-State Solutions for Flow

Past a Wing with Endplates

In the steady flow case, the one-dimensional formulation for a wing with
endplates in the extreme ground effect can be simplified. We rewrite equation
(4.53) for steady flow as

(4.65)

or, in a more compact form,

d~ [h*(x)v(x)] + 21S~~X) signß(x)vIß(x)1 = 0, (4.66)

where v(x) = dJ/dx is the spanwise averaged flow velo city under the wing,
h*(x) = h* /h, h*(x) is the chordwise distribution of the dearance between
the wing and the ground, h = h*(O) is the relative ground dearance at the
trailing edge, and lSep and ISf are the effective gaps under the endplates and
rear flap. Equation (4.65) is an ordinary differential equation of the first order
with respect to the function v(x). The boundary condition follows from (4.57)
for U (t) = 1 and takes the form
102 4. Nonlinear Flow Problems

(4.67)

Equation (4.66) with boundary condition (4.67) can be easily solved numer-
icaIly. In particular cases discussed later, it can be integrated analytically.
Writing Öep as Öep(x) = ö~pLl(x), where Ö~p is the effective gap under the
endplate at the trailing edge of the wing and function Ll(x) = 0(1) charac-
terizes the form of distribution of local gap in longitudinal direction, we can
rewrite equation (4.65) in the form

d~ [h(x)v(x)] + CLl(x) sign [1- v(x)2lJI1 - v2(x)1 = 0, (4.68)

which incorporates a similarity criterion C,

(4.69)

This criterion, which can be called a generalized gap parameter, refiects

the combined infiuence on the leakage from under the wing of three important
quantities, namely, the gap under the tips of endplates Ö~P' the aspect ratio of
the wing A and the characteristic relative ground clearance h. It follows from
(4.68) and (4.69) that within the mathematical model under consideration,
the aerodynamics of the wing with endplates depends on C, rather than on
Ö~P' A, and h separately.
For C = 0, the fiow under the wing in the extreme ground effect can be
viewed as almost one-dimensional. Note that function h*(x) in a sufficiently
general case can be presented in the form
N
h*(x) = 1 + Bx + LEj fj(x), (4.70)
j=1

where B = ()/h, () is the angle of pitch, €j = O(h) and fJ(x) = 0(1) are pa-
rameters and functions characterizing deformation of the lower surface of the
wing, Ej = €j/h = 0(1), B = 0(1), and Ej = 0(1). For example, if the lower
surface of the wing has the form of a parabolic arc with a relative curvature
of öc , the corresponding contribution to the sum in formula (4.70) is equal to
48cx(1 - x), where 8c = öc/h. It follows from (4.68)-(4.70) that the channel
fiow velo city v(x) and, consequently, the other aerodynamic characteristics
(pressure coefficient, forces, and moment) should depend upon C and the set
of parameters B, 8f , Ej, (j = 1, ... , N).
The lift and moment coefficients, as weIl as the abscissa of the center of
pressure, are represented by the expressions

(4.71)
 4.3 Steady-State Solutions for Flow Past a Wing with Endplates 103

mz
xp = C' (4.73)
y

As diseussed previously, the leakage of the flow from under the endplates
leads to generation of vorticity and, eonsequently, indueed drag. For a steady
lifting flow in the extreme ground effeet, it follows from (4.64) that

(4.74)

We eonsider some analytical solutions of the main equation for

steady flow. In some practieally interesting eases, equation (4.65) with eon-
dition (4.67) can be integrated in closed form.
Let the effective gap under endplates 8ep be eonstant ehordwise. Note that,
if one defines the effective gap as a geometrie gap under the endplate, the
above assumption implies ehordwise uniformity of the latter. If one introduces
the outflow contraction model, the effective gap would depend on the ratio
of the geometrie gap to the loeal ground clearanee h*(x), and, eonsequently,
would vary chordwise even when the geometrie gap is eonstant. However, for
small magnitudes of the relative geometrie gap, the effeetive (eontraeted) gap
can be eonsidered praetieally eonstant for a given setting of the endplate (or
flap) with respect to the ground plane. We set Ll = 1 in equation (4.68), so
that 8ep (x) = 8~p = eonst.
We consider some closed form results for the flow with leakage.
We turn to consideration of the simplest ease of a flat reet angular wing at
zero pitch angle B = 0 with endplates and a rear flap. In this ease the lift
of the wing is due to deflection of the flap (8f i 0). Because both the loeal
ground clearance and the distanee from the tips of endplates to the ground
are eonstant ehordwise (h*(x) = 1 and 8ep (x) = 8~p), equation (4.68) takes
the form
d~~) + GJ1 _ v(x)2 = O. (4.75)

Note that both signum function in front of the square root and the absolute
value sign under the square root were omitted beeause in this ease one expects
no suction under the wing, so that v(x) ::; 1.
The integral of (4.75), eomplying with boundary eondition (4.67), is given
by
v(x) = -sin(Gx+arcsin8r). (4.76)
The distribution of the pressure eoefficient ß(x) along the ehannel under the
wing ean be obtained by the formula

ß(x) = 1- v(x)2 = cos 2(Gx + arcsin8r). (4.77)

104 4. Nonlinear Flow Problems

The lift and moment (with respect to the trailing edge) are given by the
expressions

Ja(ß(x)dx = ~2 + 2D cos(G + 2 arcsinJf ) sinG,

Cy = 1 (4.78)

mz = 1 1
xß(x) dx = ~{1 + ~sin(2G + 2 arcsinJf )
+ 2~ [cos(2G + 2 arcsin Jr) - cos(2 arcsin Jr)] }. (4.79)

With reference to (4.74), the induced drag coefficient may be found as

G = 2b"ep (4.80)
>'h .
As seen from these formulas, in the example under consideration the aero-
dynamic characteristics of the wing depend on only two parameters, namely
G and J f . At the same time, the original problem contained four parameters,
including the relative ground clearance h, the aspect ratio >., the effective
gap under the endplates b"~p and the effective gap under the flap b"f. Thus, in
this example, the use of similarity criteria reduces the number of independent
parameters of the problem twofold! Generally, for a uniform distribution of
the gap under the endplates along the chord, the number of parameters that
characterize the flow problem for h -+ 0 will be n - 2, where n is the initial
number of parameters.
The lift coefficient C y and the abscissa of the center of pressure x p =
mz/Cy versus the similarity criteria Jf and G = 2b"~p/>'h are plotted in
Figs. 4.8 and 4.9. Plotted in Fig. 4.10 against the similarity parameter G
for different flap settings is the induced drag coefficient C x ; related to h.
Another integrable case is that of a Hat plate at pitch angle B with
endplates at a constant gap b"ep(x) = oep and a short rear flap. In this case,
the ground clearance function h*(x) = 1 + 11x, 11 = B/h, and equation (4.68)
yields
d~ [(1 + 11 x)v(x)] + G \/1- v(x)2 = O. (4.81)

Rewriting (4.81) as

- dv - y'
(1 + Bx) dx + Bv(x) + G1 - v(x)2 = 0, (4.82)

we can separate the variables in the following way:

dv dx
(4.83)
Bv + G y'1 - -0 2 -1 + 11x·
 4.3 Steady-State Solutions for Flow Past a Wing with Endplates 105

1.0
-....... !
Cy
.............
....... '-.....1 Gi O I !

'" "" "'"1""1 ~

........

0.8
~Jo.11

0.6 """"><
0.5 ~ ."'l )C·'

0.4
""I,""J""
~""!\ \
I\

G= 0.7~~ \ \

0.2 I '\ ~\
Ii .~ ~
i
I
"
I
0.0
0.0 0.2 0.4 0.6 0.8 - 1.0
Öt
Fig. 4.8. The lift coefficient of a wing with endplates at zero pitch angle versus
the trailing edge fiap setting for different magnitudes of similarity criterion G =
28~p/>'h.

0.6 I

-------- - r-H--'
G= 0,
I / 0.1 I
-
0.5 r-.....
::::::
"-.. .1. 3
:--- r-- I

r-r-..I

Ii:""
0.4 ~

I'--- "'
0.5
0.3 \
G~0.7 ~
0.2 I I i'-

0.1 I

0.0 I I
0.0 0.2 0.4 0.6 0.8 - 1.0
Öt
Fig. 4.9. The abscissa of the center of pressure of a wing with endplates at zero
pitch angle versus the trailing edge fiap setting for different magnitudes of similarity
criterion G = 28~p/ >'h.

Introducing the alternative similarity parameter Go,

G 28~p
Go = ~ = --, (4.84)
() A()
we obtain from (4.84)
106 4. Nonlinear Flow Problems

1.0
I
I I 1----'-1
.vr
CXi
I
h 5f~0 I
0.8
V I
VV
0;"
1

-------
0.6
VVL

~V
0.3
0.4 V V- I
// V V

0.2 IY / ~
I, V/ .,/'"

0.0 ~
f-- - 5,= 0.7 i II
0.0 0.2 0.4 0.6 0.8 1.0
25ep
G=)Jl

Fig. 4.10. The induced drag coefficient of a wing with endplates at zero pitch angle
versus the endplate similarity criterion G = 28~p/)"h and the flap settings.

dv Ödx
----==== (4.85)
v + Gov'1 - v 2 - 1 + Öx·
It follows from the preceding calculations that the solution of the basic
equation is dependent on a pair of parameters, in particular, G = 28~p/)"h
and Ö= () / h, or on an alternative pair of parameters Go = 28~p/)..() and Ö. It
is worthwhile mentioning that parameter Go, discussed herein, is identical to
parameter H, introduced by Gallington et al. [61). The integral of equation
(4.85) can be obtained in closed form. The result in the form of an implicit
relationship x = x(v) can be written as

()- x-
_ C* ( -Go arcsin v - In Isin( arcsin v + arctan Go) I) _ 1
exp G2 . (4.86)
1+ 0
Applying the boundary condition (4.67), we can determine the constant
C*as

C* [-Go arcsin 8r + In Isin( - arcsin 8r + arctan Go) I]

=exp I+G~ . (4.87)

So, finally,
()- x -_ exp [-Go(arcsinv+arcsin8r)-L(v,Go,8r)]
G2
1
- , (4.88)
1+ 0

where

=lnl.sin(arcsi~v+arctanGo) I =lnlv+Go~l·
L(v,Go,8r)
sm( - arcsm 8r + arctan Go) 8r - Go J1 - 81
 4.3 Steady-State Solutions for Flow Past a Wing with Endplates 107

1.0 r--_-_,-_-_-J-_- .-_-_--:.,-.-_--:-1-.-_--:-..-.- _-:-,_-.-.-.-.-_-_'_

Cy i I i ' .\ . ' . -.•• . •
. l ~ ~ :: ~ .. i- . ·i__ ~ . - - :. - . . ...
-
0.8 I'

0.4 1--+--+-+---+~r_;;_.....:..;_--.,.,--4-_:::I

0.2 ~.L-..J.........<~~~-I_-'--
' -_~~.....,

~·=,~~·:==:=:
0.0 -
0.0 0.2 0.4
.. =.====~
I ~I.=
0.6 0.8 _ 1.0
e
Fig. 4.11. The lift coefficient of a wing with endplates versus pitch angle for
different magnitudes of similarity criterion G = 2fJ2p/ >"h and Rap settings.

o
0.8

0.4

I
0.0 +1- , - - , --.,..---,--.----r---,--,-----,- +
0.0 0.4 0.8 1.2 1.6 2.0
Ge

Fig. 4.12. The lift and drag coefficients of a rectangular wing with endplates versus
parameter Go = 2fJ2p/ >"B: theory and experiment (>.. = 0.5; circles: experiment [61] ;
solid lines: present theory; dashed lines: G-theory).

The structure of (4.88) shows that in the ex am pIe under discussion the
solution (span-averaged velo city and pressure coefficient) depends on a new
independent variable Xl = Ox and the similarity criteria G() and 8r, i.e.,
108 4. Nonlinear Flow Problems

LID
40

30

20

10

O+--r--r-~-r--r-~-r--r-~-+
0.0 0.4 0.8 1.2 1.6 2.0
Ga
Fig. 4.13. The lift-to-drag ratio of rectangular wings with endplates: theory and
experiment (empty circles: experiment for ,\ = 2/3 [61J; black circles: experiment
for ,\ = 0.5 [61J; solid lines: present theory; dashed lines: G-theory).

The coefficients of Hft, longitudinal moment (with respect to the traiHng

edge) and the abscissa of the center of pressure are calculated by the formulas
(4.71)-( 4.73).
The fiow model, described by equation (4.82) with boundary condition
(4.67), contains as a particular case the previously mentioned G-theory, pro-
posed by Gallington et al. [61J. The latter theory, based on elementary conti-
nuity considerations, implies that the pressure does not vary along the chord
of the wing.
To retrieve the governing equation of the G-theory from equation (4.82),
suppose that a solution of this equation exists, for which the span-averaged
velo city (pressure) is constant along the chord, Le., v(x) = const. = V. Then,
for a fiat rectangular wing with a constant gap under the tips of the endplates
and pitch angle 0, equation (4.82) can be rewritten as

(4.89)

Equation (4.89) can be easily solved with respect to v to give

v=
J1+Gf
_
p=
1 -2 1
-v =-----2'
1 +Go
Gy = 1 1

o
ßdx = ~G2'
1+ 0
(4.90)

It follows from (4.90) that the both pressure and Hft coefficients in the G-
theory depend only on parameter Go, Le., are defined in terms of a certain
combination of the gap under the endplates, the aspect ratio and the adjusted
 4.3 Steady-State Solutions for Flow Past a Wing with Endplates 109

pitch angle of the wing. On the other hand, (4.89) provides a simple tool for
designing a ram wing vehiele for a given pressure in the channel. At the
v
same time, the assumption of constant together with prescribed boundary
condition (4.67) necessitates the following equalities:
- bf Go ~
bf = h = 1 + G~ = V 1 - p,

indicating that the rear Rap should be "tuned up" to ensure the prescribed
pressure in dynamic air cushion. These relationships show, in particular, that
an increase in the design ground elearance for the same magnitude of loading
should be followed by an opening of the gap.
The foregoing approach to determining the aerodynamic characteristics of
rectangular wings with endplates in the extreme ground effect remains valid
for foils with loeal suetion. One example of such a foil is illustrated in
Fig. 4.14, which shows the lift coefficient versus the generalized gap param-
eter G for a foil with a parabolic lower surface and zero pitch angle. Other
examples of calculation of the aerodynamic parameters of foils with local suc-
tion are discussed in the next section in connection with the problem of the
static stability of longitudinal motion.
Now, we turn to the ease when the generalized gap parameter is
equal to zero, i.e., G = O. Note that this can occur either for zero elearance
under the ti ps of the endplates bep = 1 or when the wing has an infinite aspect
ratio (A -+ 00). For Jf = 0, the latter case corresponds to the descri ption of
the order of 0(1) of the two-dimensional Row problem for a foil moving elose
to the ground, as discussed in paragraph 4.1.

0.0
Cy
f-- ~
öc= 0.1
---...----
/'" I--
~
L.....-
-0.2
V /1 /
/'
----
/'
V
V /'
-0.4
u.~
/ /
/ n'l· /
-0.6 / / I ,
/
1

I I

-0.8 / i
Öc=, 0.4 I
/, ! I I
:
/ I I
I i i
-1.0
0.0 0.2 0.4 0.6 0.8 1.0
2liep
G=-
Ah
Fig. 4.14. The lift coefficient of a wing with endplates and a symmetrical parabolic
thickness distribution versus the similarity parameter G = 28~p/ >"h.
110 4. Nonlinear Flow Problems

In this case, the equations (4.65) and (4.67) can be rewritten as

v(O) = -8f . (4.91)

The solution of problem (4.90) is elementary,

.
p(x) = 1 -
[h*(x)
~ ]2 .

The lift and moment coefficients are calculated by the formulas

mz
xp = C' (4.92)
y

Here are some simple analytical expressions for Gy and m z for some par-
ticular cases:
• Flat plate with a rigid fiap at zero pitch:
1 -2 1
m Z =2(1-bf ), xp = 2' (4.93)

• Flat plate without a fiap 8f = 1 at a given incidence iJ:

(4.94)

mz = 21 + (12(1 1+ B) [-8 - (1 + 8)- ln(l + 8)- ] . (4.95)

Note that for small perturbations these formulas yield corresponding ex-
pressions of the linear theory, whereas for moderate and large perturba-
tions, they refiect the inherent nonlinearity of the aerodynamics of the
extreme ground effect.
• Flat plate at pitch angle B with a rigid fiap

mz = 21 + 82 (18f+ 8) [-B - (1
- + 8).
+ B)ln(l -] (4.96)

Employing formula (4.63), it is easy to find the suction force coefficient as

Gs - ( 1 - -8f- .
= h(l + 8) )2 (4.97)
1+8
Because for G = 0 (no gap und er the endplates or an infinite aspect ratio)
there is no lateral leakage, the overall drag force acting on the wing in
potential fiow should be zero. Taking into consideration the magnitude of
the pressure drag coefficient for this case,
 4.3 Steady-State Solutions for Flow Past a Wing with Endplates 111

-2
C = -C 0 + C = -h 0-( 1 - - ~)
- - - h(l - 6f)
- 2 (4.98)
xp Y X pf 1+0 '

we obtain the following result:

~ -
- 6
f - 6f
= -M(l - -- ) - h(l - 6r)
2
+ h(l + 0)(1-
-
(4.99)
-~)
2
=0
1+0 ' 1+0
which confirms the correctness of the determination of the suction force
coefficient .
• Flat plate with incidence and a jet Rap at the trailing edge:
In this case, the corresponding effective gap und er the jet Rap entering the
problem can be determined from local analysis of the Row near the jet Rap
(see section 6) as

6jf
- =- Yjoo = 1- TVrc;
2/;,' ( 4.100)

where q is a coefficient of the total momentum of the jet and T is the

angle of blowing.
Substituting;Sf with ;Sjf in expressions (4.96), as given by (4.100), we obtain
the following formulas for the aerodynamic coefficients of a a foil with a
jet Rap at the trailing edge:

C = 1- (1- TJCJ2h)2 (4.101)

Y 1+0 '

1 (1 - TJCj /2h)2 [- - -]
mz = "2 + 02 (1 + 0) 0 - (1 + 0) ln(l + 0) . (4.102)

The induced drag coefficient of a jet-Rapped rectangular wing with end-

plates at pitch angle 0 can be found from (4.74) with ;Sf replaced with the
effective gap under the jet Rap ;Sjf; see (4.100). In the particular case of
zero incidence, we obtain the following expression:

(4.103)

It is interesting that introduction of the effective gap under the jet Rap ;Sjf
renders identical the structure of the formulas for predicting the aerody-
namic coefficients in the cases of a wing with a rigid Rap and wing with a
jet Rap.
112 4. Nonlinear Flow Problems

4.4 Unsteady Flow Solutions

for a Wing with Endplates
To analyze the transient modes of motion of a wing-in-ground-effeet vehicle
(takeoff, landing, variation of ground clearanee, ete.), as weH as to evaluate
the infiuenee of wind-wave perturbations, it is useful to have the unsteady
eharaeteristies of the main lifting surfaee. To obtain the eorresponding solu-
tions for the ease of a sehematized ftying wing configuration in the form of
a reet angular wing with endplates in the extreme ground effeet, one has to
solve equation (4.53) with boundary eonditions (4.54) and (4.57) and initial
eondition (4.58).

4.4.1 A Nonlinear Unsteady Solution for Small Gaps

U nder Endplates

If one assumes that the relative gap under the endplates is sufficiently
small at any moment, i.e.,

28ep (x, t) = O(G) 1

A h(t) « ,
where h(t) is the instantaneous distanee of the trailing edge from a eorre-
sponding point on the ground and G is the generalized gap parameter intro-
dueed earlier, it is possible to derive an approximate solution of the nonlinear
unsteady problem of the flow past a wing with endplates, by iterating on a
leading-order solution for G = o. If one assumes that

~(x, t, G) = ~o(x, t) + ~l(X, t, G) + O(G 2 ), ~l(X,t,G) = O(G),

then the equation for the first-order eontribution of the gap under the end-
plates to the velo city potential will be

8 [h* (x, t )8x8 ;, (x, t )] + 28

8x '1'1 A t).
ep (x,
slgnpo A viPo x, t I = 0,
A ( )
(4.104)

where
2
Po(x, t) = U(t) -
A (8~O)2
8x
8~o
- 2 8t . (4.105)

The boundary eonditions for ~1 at the edges will be

~1(1, t) = 0, 8~1 8~o + 8; 8~1 I = o. (4.106)

8x Bx 8t x=O

The appropriate initial eondition for equation (4.104) is

8~11
8x t=O -
_ 0
•
 4.4 Unsteady Flow Solutions for a Wing with Endplates 113

As seen from the preceding equation, to find the unsteady characteristics of

the lifting system for the case of a nonzero gap under the endplates, it is
necessary to begin with the solution of the unsteady aerodynamics with no
leakage from under the endplates (zero gap), i.e., find Jo(x, t). Such a problem
in itself represents a certain practical interest for considering the takeoff of
a vehicle from water, because the gaps under the endplates are very small
from the begining of motion until almost the moment of detachment from the
water surface. In the case of zero gap under the endplates, the problem
of determining the potential Jo(x, t) of the relative motion of the fluid takes
the form

:x [h*(x, t) :x Jo(x, t)] = - :t h* (x, t); (4.107)

J1U(t)2 _ (7;) 2 - 2J18~o = 0, x = 0; (4.108)

J(l, t) = 0, x=l. (4.109)

Integrating (4.107) twice, we derive the following expressions for the channel
flow velo city and potential:

(4.110)

4Y(X, t) = -
JX
1
dX1
h*(X1,t)
rXI 8t(~'
Jo
8h*
t) d~ + Cl
JX d~
h*(~, t) + C2(t),
1 (4.111)

where, as previously, h* = h*(x, t) is a distribution of the instantaneous gap

between the lower surface of the wing and the ground. Functions of time
Cl (t), C2 (t) are to be determined by using the boundary conditions (4.108)
and (4.109). Imposing the leading edge condition, we obtain

(4.112)

Satisfying the trailing edge condition (4.109) and taking into account ex-
pressions (4.110) and (4.111), we can obtain the following nonlinear ordinary
differential equation for determining the function of time Cl (t):

( 4.113)

where the "dot" indicates differentiation with respect to time and functions
a(t), b(t), and c(t) are given by

r 1 h*(~, t) t d~ -2 * r 1 d~
Jo h*(~,t)]
-1
a(t) = - Jo h*2(~,t) d~/Jo h*(~,t)' b(t)=-[26[h (O,t) ,
114 4. Nonlinear Flow Problems

c(t) = {U(t)2 - 2 Jor1 Jor 1 •• • • d~ dX1 }

[h*(X1, t)h*(~, t) - h*(~, t)h*(X1, t)] h*(xl, t)2

Differential equation (4.113) is known as a Rieatti equation and eannot be

integrated analytically. However, it ean be integrated numerically with an
appropriate initial eondition. To formulate the initial eondition with respeet
to the time function C1 (t), suppose that at t = 0 the vehicle is at rest and the
rate of variation of the ground clearanee is equal to zero, Le., h*(x,O) = O.
Then, the initial eondition for Cl (t), eorresponding to zero relative velo city
of the fluid in the ehannel, should be

C1 (0) = O.

Having solved the Rieatti equation (4.113) numerieally, we ean determine the
aerodynamie eoefficients. For example, the lift eoefficient will be expressed
by the following formula:

r 1
Cy(t) = Jo Po(x, t) dx = U(t)2 - Jo [C1(t) - Jo
r1 r.h*(~, t) d~] 2 ~
h*2(X, t)

-2 11{Xl11xl
0 0
..
[h*(x!, t)h*(~, t) -
. . dcd
.. Xl
h*(~, t)h*(x!, t)] h*(xl, t)2
. r
+C1(t) J1
d~
h*(~, t) - C1(t) J1
r h*(e,t)de
h*(~, t)2 } dx. (4.114)

This solution permits ealeulating the unsteady nonlinear aerodynamies for

the different laws of motion and the deformation of a simple ftying wing
eonfiguration for a zero gap under the endplates. For example, for a flat wing
advancing along the ground and, at the same time, performing unsteady
vertical and angular motions, the instantaneous distribution of the relative
ground clearanee ehordwise ean be represented by the expression

h*(x, t) = h(t) + B(t) X,

where h(t) is the time-dependent relative ground clearanee at the trailing
edge and B(t) is the eurrent piteh angle.

4.4.2 Unsteady Perturbation of a Steady Flow

To analyze the linear dynamics of a lifting eonfiguration in the ground ef-

feet one needs derivatives of the aerodynamic eoefficients with respeet to the
relative ground clearanee h, the piteh angle B, and their rates hand iJ. To
determine these derivatives, we eonsider small unsteady perturbations of a
 4.4 Unsteady Flow Solutions far a Wing with Endplates 115

nonlinear steady-state fiow. Such a perturbation analysis enables us to re-

tain the nonlinear dependence of the aerodynamic derivatives upon basic
steady-state parameters, e.g., adjusted pitch angle, relative ground clearance
in cruise, etc. As an ex am pie of an application of such an approach, take the
case of a rectangular wing with endplates at h -+ O. For simplicity, consider
full opening of the fiap at the trailing edge, that is, ;Sf = 1. Represent the
velo city potential of the relative motion, the local ground clearance distribu-
tion, and the gap under the endplates in the following way:8
rP(x, t) = rPs(x) + ~(x, t), h*(x, t) = hs(x) + j~(x, t),
oep(x, t) = oep(x) + h(x, t),
where subscript "s" designates steady-state parameters, whereas the second
term in each equation represents unsteady contributions.
Substituting the perturbed quantities in (4.53) and accounting for the
description of the steady-state fiow problem,

(4.115)

rPs(l) = 0, drPs I --1 (4.116)

dx x=ü - ,

where

we obtain the following equations for the unsteady fiow potential ~(x, t):

(4.117)

(4.118)
When deriving the trailing edge condition for unsteady fiow potential in
(4.118), it was taken into account that drPs/dx = -1 at x = O. Using the
steady fiow equation, we obtain an alternative equation for the perturbed
unsteady velocity potential:

~[hs(X)ä~s] +~[h(x,t)drPs] _ h(x,t)~[hs(x)drPs] =_äh. (4.119)

äx äx äx dx oep dx dx ät
Having fulfilled the linearization of unsteady fiow with respect to non-
linear steady fiow, we can consider separately two practical cases of height
perturbation and pitch perturbation.
8 The "hats" are taken off.
116 4. Nonlinear Flow Problems

Unsteady Height Perturbation. In the case of height perturbation,

h(x, t) = h(t). The perturbation potential can be represented as

~(x, t) = ~h(x) h(t) + ~t.(x) h(t). (4.120)

Assume that the gap under the tips of the endplates in cruise is constant
chordwise 8ep (x) = 8ep = const. Corresponding equations for the components
of the perturbed solution can be derived in the form

(4.121)

(4.122)

Imposing conditions at the leading edges, we arrive at the following boundary

conditions at x = 1 for equations (4.121) and (4.122):

-t.
cjJ (1)=0. (4.123)

Applying the Kutta-Zhukovsky condition,

(4.124)

Assuming an oscillatory character of pert~rbations,9 we relate heave accel-

eration and heave displacement, namely, h = -k 2 h, where k = wCo/Uo is
the Strouhal number based on the chord of the wing. Therewith, we obtain
the following boundary conditio~ for the derivatives of the unsteady velocity
potential with respect to li and li at the trailing edge:
-t.
(~
dx
_cjJh)
x=o
=0. (4.125)

Integrating equations (4.121) and (4.122) and accounting for the boundary
conditions (4.123) at the leading edge gives the following expressions for the
corresponding perturbation velocities and potentials:

h d~h = c3 - x, (4.127)
dx hs(x)

9 Aperiodic perturbations can be analyzed on the basis of the Fourier integral.

 4.4 Unsteady Flow Solutions for a Wing with Endplates 117

To determine the constants Cl and C2, we apply the boundary conditions

(4.124) at the trailing edge. This gives

a + k 2 ß'Y
Cl = 1 + k 2 'Y2 '

where

1 1
ß= o Vs(x)[h()-r]dx,
s X ep
1 1
'Y
r l dx
= Jo hs(x)·

In these expressions, h is the relative ground clearance measured from

the trailing edge, c5ep = 8ep /h = 0(1), hs(x) = hs(x)/h = 0(1), and vs(x) =
d~s/ dx is the span-averaged channel flow velo city.
The derivatives of the unsteady pressure coefficient with respect to hand
h can be derived in the form

(4.128)

(4.129)

It. can be seen from equations (4.126)-(4.129) that the derivatives ph(x) and
ph(x) are inversely proportional to the relative ground clearance in cruise,
Le., are of the order of O(l/h).
Unsteady Pitch Perturbation. In the case of pitch motions around the
center of gravity h(x, t) = O(t)(x - x cg ), the perturbation potential of un-
steady flow can be represented as

(4.130)

The derivatives of the unsteady perturbed velo city I?otential with respect
to perturbation in the pitch 0 and the rate of pitch 0 are governed by the
following equations:

~
d
[hs ()d~Ö]
x d +~
tPs
d [( x _ x cg )dd ] _ (x -J< x cg ) ~
dX [h s (x )ddtP s] = 0 (4131)
,.
x x x X U ep X

(4.132)

The boundary conditions at the leading and trailing edges for equations
(4.131) and (4.132) are

(4.133)
118 4. Nonlinear Flow Problems

-Ö
=0, ( d~Ö _ ~Ö)
c

( ~+k2~())
dx x=o dx x=o
= 0. (4.134)

Solution of equations (4.131) and (4.132) is straightforward and leads to the

following formulas:

h4> =
-9 JX [12(~ -
1 C4 - x cg )
2] hs(~)·
d~
(4.136)

The constants C3 and C4 can be calculated by using the boundary conditions

at the trailing edge.
The derivatives of the pressure distribution, induced by the unsteady
variation of pitch, can be determined by the following expressions:

(4.137)

As in the case of unsteady height perturbations, these derivatives are inversely

proportional to the relative ground clearance h in cruise.
Derivatives of lift and moment (around the center of gravity x = x cg )
coefficients can be obtained by integrating the derivatives of the correspond-
ing pressure distributions:

C;,h,Ö,9 = 1 1
ßh,h,Ö,9(x) dx (4.138)

m~,h,Ö,9 = 1 1
(x - X cg ) ph,h,Ö,9 (x) dx. (4.139)

Note that in both unsteady height and pitch pertu~bations, ~he derivatives
of lift and the moment coefficients with respect to h, h, 0, and 0 are inversely
proportional to the relative ground clearance in cruise.
Now, we turn to determining the drag coefficient Cx and its derivatives
with respect to height and pitch perturbations. The drag coefficient can be
written as
(4.140)
with C Xi and C Xf representing, respectively, induced drag and viscous drag.
Later on, it will be assumed that the viscous part of the drag does not vary
with small perturbations of ground clearance and pitch angle.
As discussed earlier in this chapter, the induced drag coefficient can be
obtained by the formula
 4.4 Unsteady Flow Solutions for a Wing with Endplates 119

C Xi = C + Cs Xp (4.141)
where C Xp is part of the induced drag coefficient due to the longitudinal
component of press ure forces, and C s is a suction force contribution. For Rat
ground, C Xp can be written as

1 dh*(x)
Cxp(x,t) = h Ja p(x,t)~dx.
r (4.142)

Substituting h*(x, t) = hs(x) + h(t) + iJ(t)(x - x cg ) and the perturbation

expansion of pressure p(x, t) in (4.143), we obtain the following express ions
for the derivatives of C Xp in height and pitch:

Ch,h,fi,o = h
Xp Ja
r fjh,h,fi,O(x)dhs(x)
1

dx
dx. (4.143)

The steady-state pressure contribution to the induced drag is given by

Cx
Ps
= h 1 a
1 dhs(x)
Ps(x)-d- dx.
x
(4.144)

We consider the suction force contribution to the induced drag coefficient.

According to (4.63), it can be determined as

Cs = -h h*(l, t)[l + v(l, t)]2

= -h [h s + h(l, t)][l + vs(1) + v(x, t)]2. (4.145)
where v(x, t) = 8J/8x. Expanding (4.145) to the first order,
e ss = -2hh s (1)[1 + v s (1)]2, (4.146)

C:,h,fi,o = -2hh s(1)[1 +vs(l)]d J ::,fi,fi (1). ( 4.147)

It can be seen from (4.144) and (4.148) that for h --+ 0, the induced drag
coefficient is ofthe order of O(h). Because hfjl = 0(1), it follows from (4.144)
and (4.148) that derivatives of the induced drag coefficient in the extreme
ground effect are of the order of 0(1).
5. Compressible Flow Past a Wing
in the Extreme Ground Effect

It is practical to extend the analysis of the aerodynamics of a wing in the

ground effect to account for the dynamic compressibility of the air. In fact,
the cruise speed of ground-effect vehicles can amount to half or more of the
speed of sound. At the same time, it is known that the problem of unsteady
subsonic flow is one of the most challenging in lifting surface theory; see
Belotserkovskyet al. [139]. The complexity of the problem partly sterns from
the fact that in a compressible fluid the perturbations propagate with finite
speeds.
The problem of compressible flow past a wing in the extreme ground
effect can be treated on the basis of the approach, similarly to that applied in
section 2 for an incompressible fluid. Here one adopts the same assumption as
previously, stating that both deviations and slopes of the surfaces of the wing,
vortex wake, and the ground should be small, i.e., of the order of the relative
ground clearance; see (2.1). In this case, it becomes possible to linearize the
flows above the wing and the wake and introduce linearizing simplifications
into formulations for edge flows. As earlier, with an asymptotic error of the
order of O(h;), channel flow can be shown to retain almost a two-dimensional
nature and incorporate nonlinearity. In what follows, the derivation of the
solution will be confined to the leading order of 0(1) and the case of constant
speed U(t) = 1. Then, some examples are considered, including linearized
steady and unsteady compressible flows past a rectangular wing and the
nonlinear flow problem for a two-dimensional foil in the extreme ground effect.

5.1 Channel Flow in a Compressible Fluid

In three-dimensional, compressible, isentropic flow, the following relation-

ships can be shown to hold (see Ashley and Landahl [161]):
• Equation of fluid motion

• Energy equation

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
122 5. Compressible Flow

where cfJ is the velo city potential of relative fluid motion related to the per-
turbed velo city potential cp through the equation

cfJ = -x + cp, (5.3)

as and aso are, respectively, the local velo city of sound and the velo city of
sound at upstream infinity, Mo = Uo/ aso is the Mach number in the unper-
turbed oncoming flow, and 'Y is the ratio of the specific heat of gas (isentropic
parameter); for air, 'Y = 1.4;

n . 8 .8 k 8 n2 n n
= Z 8x + J 8y + 8z'
A
v .u = v = v . v.

Excluding (a s/a so )2 from (5.1), we can derive the following equation to de-
termine the relative velo city potential:

(5.4)
In the channel flow region D 1, we introduce stretching of the vertical coordi-
nate y = y/h and seek cfJ in the form of an asymptotic expansion

(5.5)
where
(cfJi, cfJi*) = 0(1), cfJi = cfJh +hln~cfJI2 +hcfJI 3• (5.6)
Passing over to the channel flow variables and accounting for the adopted
asymptotics (5.5) of the potential in the gap between the lifting surface and
the ground, we obtain the following relationships with respect to the potential
function cfJi:
• Continuity equation:
2
8 cfJi . (8 cfJi )2 = o·, (5.7)
8y2 8y
• Boundary conditions:

8y = 0 Elor Y = Yl
8cfJi an d Y = Yg· (5.8)

A solution, satisfying both (5.7) and (5.8), has the form

cfJi = cfJi(x, z). (5.9)

 5.1 Channel Flow in a Compressible Fluid 123

The eorresponding equation for cPi* ean be obtained by taking into aeeount
(5.6) and (5.9):
(5.10)

where N1,2 are nonlinear differential operators in the two variables x and z:

(5.11)

(5.12)

(5.13)

Integrating (5.10) onee with respeet to y and using the flow tangeney eondi-
tions for cPi* identical to (2.15) and (2.17), where 'Pi* - x should be replaeed
by cPi*, we obtain the following ehannel flow equation for eompressible isen-
tropie flow (Rozhdestvensky [41]):

{I + ~b -1) M; [1- (\12cP*)2 - 28~i]} [\12(h*. \12cPj) + 8!*]

-h* M; [8;bi + !(\12cPj)2+~\12cPi·\12(\12cPi)2] =0, (5.14)
where
\1 2 ' (h*\1 2cP*) = !...-(h*8cPi ) + ~(h*8cPi);
8x 8x 8z 8z
h* = h* jh, h* = h*(x, z, t) is a preseribed instantaneous gap distribution.
To solve the lowest order problem (the extreme ground effeet), one has to
replace cPi by cPlt and apply the following boundary conditions at the planform
contour:
'P1 1 = X + cPl 1 = 0 at the leading edge, (5.15)
Pit = 0 at the trailing edge. (5.16)
Inspecting the expression for the pressure coefficient in compressible flow,
(see Ashley and Landahl [161]),

(5.17)

To satisfy (5.16) for compressible case, it is sufficient to require that

2 8:: +
1 (\1 2cP)2 - 1 = 0 at the trailing edge. (5.18)

For small perturbations, linearization of equation (5.14) leads to the following

lowest order problem with respect to the perturbed velocity potential 'P11 :
124 5. Compressible Flow

• Equation:

( 1- M2) {j2CPh + 8 2cph _ M2 (8 2CP h _ 2 82CPh ) _ 8h* _ 8h*. (5.19)

o 8x2 8z2 0 8t2 8x8t - 8x 8t '
• Boundary conditions at the planform contour:

CP l 1 = 0 at the leading edge, (5.20)

Ph = 2 (8;~1 -8;;1 ) = 0 at the trailing edge. (5.21)

5.2 Steady Linearized Compressible Flow Past a Wing

of Finite Span
A linearized version of a steady compressible flow around a wing at arbitrary
relative ground clearances is described by the following boundary problem:
• Equation:
2 8 2cp 8 2cp 8 2cp_
(1 - Mo ) 8x 2 + 8 y 2 + 8z 2 - O. (5.22)

• Flow tangency conditions on the wing and the ground:

8cp 8yu I
8y = - 8x' at (x,z) E S, Y = h±O; (5.23)

8cp
8y = 0 at (x, z) E G, y = 0 + 0, (5.24)

where Mo is the Mach number of the oncoming stream. Using the Prandtl-
Glauert transformation for the y and z coordinates, y' = YV1 - M;,
z' = zV1 - M;, x' = x, we obtain the following problem for an equivalent
incompressible flow:
• Equation:
(5.25)

Note that the original Prandtl-Glauert rule implies that only the x coor-
dinate is transformed into x' = x / V1 -
M;. Here we prefer to retain the
same chord length. In fact, both transformations lead to identical results.
• Boundary conditions
8cp 1 8Yu,1
8y' = - ßM 8x at y' = hßM ± 0, (5.26)

8cp = 0 at (x', z') E G', y' = 0 + 0, (5.27)

8y'
 5.2 Steady Linearized Compressible Flow Past a Wing of Finite Span 125

where ßM = VI - Mt. As seen from (5.25)-(5.27), in a linearized steady

state approach we account for compressibility by utilizing the results obtained
previously for incompressible flow, but in aspace "squeezed" both vertically
and laterally. In particular, we consider an equivalent incompressible flow for
a wing of smaller aspect ratio)..' = A . ßM. In addition, the downwash on the
surface of the equivalent wing is 1/ ßM times larger than in a corresponding
compressible flow problem.! The aforementioned two factors are the same as
in unbounded flow. A distinctive feature of the flow in ground effect is that
the equivalent wing moves at a sm aller ground clearance h' = h· ßM. To
compare the influence of compressibility upon a wing in an unbounded fluid
and in the extreme ground effect, we consider the simplest cases of a flat
plate of infinite and small aspect ratio. For a two-dimensional incompressible
flow the following are the expressions for the lift coefficient:
• In an unbounded fluid,
(5.28)
• In the extreme ground effect,

(5.29)

Turning to the compressible flow case by the of Prandtl-Grauert correction,

we can obtain the following formulas for the lift coefficient:
• In an unbounded fluid,

C' = 21l"u = Cy (5.30)

y ßM VI-Mt
• In the extreme ground effect,

C' U Cy (5.31)
y= ß'tr.·h = I-Mt·
Now, it is easy to see that for a wing oflarge aspect ratio in both unbounded
and bounded flow, there is an increase of the lift coefficient due to compress-
ibility. However, in the ground effect, the influence of compressibility
is more pronounced. For example, for a Mach number equal to a 0.8, an
increase of the lift in the extreme ground effect is almost twofold compared
to unbounded flow. It should be noted that Efremov [71], studying a thin foH,
also came to the conclusion that, in proximity to the ground, the influence
of compressibility leads to a noticeable increase in the lift coefficient. Such a
conclusion can be easHy interpreted in terms of the Prandtl-Glauert trans-
formation, if one accounts for the fact that the equivalent wing flies closer to
the ground, h' = hVl - Mt.
1 If, for example, a wing in compressible flow has an angle of attack a, then the
angle of attack of the equivalent wing becomes a' = a/ VI - M;t.
126 5. Compressible Flow

C 2.8
Y I
--2.6
C
YM=Q 2.4 I

2.2
:/.
2.0
Ä.=\C // /
r;//
5"",,'

Vv
1.8
4_ }'j
1.6

1.4
h ~ V·
~ ~:...---- ~
~V_-
1.2
.&i!! ~ ~
1.0

0.8
I A=1

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Mo
Fig. 5.1. The compressibility correction to the lift coefficient of a flat rectangular
wing in the extreme ground effect versus the cruising Mach number.

Figure 5.1 features a compressibility correction in the form of the ratio

of the lift coefficients in compressible and incompressible flows versus the
Mach number for wings of different aspect ratios in the extreme ground effect
(Rozhdestvensky [41]). This correction can be shown to hold for a practical
range of ground clearances (up to 0.15) with an asymptotic error of the order
of O(h). For comparison, the curve corresponding to the unbounded two-
dimensional flow compressibility correction factor ßM = M~, is plotted VI -
in the same figure (dashed line).
Now, we can proceed to the case of the small aspect ratio, for which in
incompressible flow
• In unbounded flow, as predicted by the Jones theory,
1l'>'
Gy = 20:. (5.32)

• In the extreme ground effect, see formula (3.69),2

0:>.2
Gy = 6h . (5.33)

Using the Prandtl-Glauert correction, Le., replacing respectively >., h,

and 0: by >.' = ßM)" h' = ßMh, and 0:' = o:/ßM, we obtain a result, which
reads similarly für both unbüunded and bounded flow, namely, for wings of
a small aspect ratio, the influence of compressibility upon the lift
coefficient becomes negligible.
2 Here instead of the notation for pitch 0, we use a.
 5.3 Compressible Linearized Unsteady Flow Past a Wing of Finite Span 127

5.3 Compressible Linearized Unsteady Flow

Past a Wing of Finite Span
In paragraphs 5.1. and 5.2., both nonlinear and linear eompressible flows
around a wing in the extreme ground effeet were diseussed, and eorresponding
approximate mathematical models were proposed. In this paragraph based on
a linear formulation, the flow problem for a reet angular wing in the extreme
ground effeet will be treated for harmonie dependenee of perturbations on
time. Reeall equation (5.19) for linear eompressible flow past a wing in the
extreme ground effeet, omitting subseript "lI":

(5.34)

With the intention of investigating a representative example of he ave oseil-

lations of a reetangular wing, we represent the instantaneous loeal gap and
velo city potential as

h*(x, z, t) = h - hoiexp(ikt), <p(x, z, t) = <jJ(x, z) exp(ikt), (5.35)

where k = weo/uo is the Strouhal number, i = p. Taking into aeeount of

(5.35) equation (5.34), yields

2 8 2 <jJ 8 2 <jJ 2( 2 8<jJ) hok

+ 8z 2 + Mo k <p + 21k 8x = -ho (5.36)
A •

(1- Mo) 8x 2

We express the eomplex amplitude of the ehannel flow veloeity potential <jJ
in terms of aseries that satisfies the eondition of zero loading at the tips of
the wing:
<Xl
7r
<jJ(x,z) = LXn(x)eosqz, qn = >:(2n + 1), (5.37)
n=O
where Xn(x) is a eomplex function of real argument x. Simultaneously, we
expand the right-hand side of (5.36) into aseries aeeounting for the following
expression:

1=
<Xl 4(-1)n

n=O )..qn
L --':---'-- (5.38)

The resulting equation with respect to functions Xn(x) ean be written as

The eharaeteristie equation for (5.39) is

2 2ikM; k 2 M; - q~ _ 0
f-t + 1_ M2 f-t + 1 _ M2
o 0
- , (5.40)
128 5. Compressible Flow

5.0
h 1 I I , !I

1/\
-h C y 4.5
I I I MOd O.4
I
4.0

3.5 \
3.0 /
" \ // /
0.5
2.5

2.0
/
V
O.~ r"\/ O} '\
(\V \
1.5

1.0 ~ / V \\
0.5
~
\I \ /
Mo =0 1'--'./' Y
0.0
o 2 3 4 5 6 7 8
k
Fig. 5.2. The aerodynamic derivative hC~ of a flat foil heaving in the extreme
ground effect versus Strouhal number for different Mach numbers and >. = 00.

with roots
-ikM; ± i..(l5
{Lln,2n = 1 _ M2 ' (5.41)
o

We write the solution of (5.39) in the form

Xn(x) = An exp({LlnX) + B n exp({L2n X ) + Xnpart(x), (5.42)

where Xnpart (x) is a particular solution equal in this case to Qj(k 2 M; - q~).
Recalling that the perturbed flow potential to this order should satisfy
two boundary conditions, namely, cp = 0 at the leading edge (x = 1) and

ocp _ ocp = 0
ox ot
at the trailing edge (x = 0), we can write, respectively,

<t3(1) = 0, (5.43)

and

o<t3
ox -1
·kcp= 0 ,
A
X~ -ikXn = 0 x=o, zE (-~2' ~)
2 .
(5.44)

Applying the requirements (5.43) and (5.44), we can obtain the following
expressions for the coefficients of the solution series

A =X A* =X ikexp({L2n) + {L2n - ik ()
n npart n npart exp({L2n)({Lln _ ik) _ exp({Lln)({L2n _ ik)' 5.45
 5.3 Compressible Linearized Unsteady Flow Past a Wing of Finite Span 129

h~ 1.0\
0.8 '~--t\--+--+--+---+-+-+--i----j

0.6 ~\ \
"\ \ \ Mo F 0.6 0.5 0.4 I

1/ Mn=O
-0.2

-0.4 L---L.._.l.---..l..._...l.----.l_-L_L--l
o 2345678
k
Fig. 5.3. The aerodynarnic derivative hC; of a flat foH heaving in the extreme
ground effect versus the Strouhal number for different Mach numbers and .A = 00.

B =X B* = -X ikexp(J.Lln) + J.Lln - ik ()
n npart n npart exp(J.L2n)(J.Lln _ ik) _ exp(J.Lln)(J.L2n _ ik)' 5.46
The lift coefficient is obtained by integrating the loading

/>./2
2 r1 8cp 8cp 2 />./2 r1 8cp
Cy(t) = >: ->'/2 Jo (8x - 8t) dx dz = >: exp(ikt) ->'/2 Jo (8x -ikcp) dx dz

(_l)n 1
= Xexp(ikt) L --
4 N 1
(5.47)
A

(X' - ikX) dx = Gyexp(ikt).

n=O qn 0

The final expression for the complex amplitude of the Hft coefficient is

Cy = _ fXnpart{l+ik+A~[l+ikexP(J.Lln -1)]+B~[1+ikexP(J.L2n -I)]}.

n=O J.Lln J.L2n
(5.48)
Introducing derivatives with respect to the rate and the acceleration of
the heave, we write the Hft coefficient as

(5.49)

where
(5.50)

where ~ and s:s are real and imaginary parts of the expressions.
130 5. Compressible Flow

5.0
!
,
I II i 111 I
. 4.5
h
- h C y 4.0 I
I i
i

I I
I I
3.5
1 I
3.0 i I
Mo= 0.6
I

2.5
I ! I

2.0
10.5 !
1.5 I 11
1.0 J 11 K
0.5
I .- v_ V
) 1VJ:-.? ,....~.1\J
0.,3
0.0
o 2 3 4 5 6 7 k 8

Fig. 5.4. The aerodynamic derivative hC~ of a rectangular wing heaving in the
extreme ground effect versus the Strouhal number for different Mach numbers and
>. = 1.

0.4 , I
I
.. 0.3
I !(\
h I
hCy
0.2
I
i ~ I \
! \
0.1
I I

0.0
Mo=C,
\ / ~

I
-0.1
--...... r---
-0.2

-0.3
M0 = ~~
\ '"l\j'~ U ...

!
-0.4
I I
o 2 3 4 5 8

Fig. 5.5. The aerodynamic derivative hC~ of a rectangular wing heaving in the
extreme ground effect versus the Strouhal number for different Mach numbers and
>. = 1.

Some results for heave derivatives of the lift coefficient, multiplied by the
steady state ground clearance, i.e., hCt and hC~ are presented in Figs. 5.2-
5.7 versus the Strouhal number and for different Mach numbers for a rect-
angular wing of aspect ratio ..\ = 1,2,3,00. The characteristic feature of the
curves of (-hC;) as a function of the Strouhal number consists of the oc-
 5.3 Compressible Linearized Unsteady Flow Past a Wing of Finite Span 131

5.0
!
- h <fy
4.5 :
I
4.0 I

3.5 I
1

3.0
I
2.5
i1
2.0 1

1.5 !

I
1.0 !
0.5
I
0.0
0 2 3 4 5 7 8
k
Fig. 5.6. The aerodynamic derivative hCt of a rectangular wing heaving in the
extreme ground effect versus the Strouhal number for different Mach numbers and
>. = 2.

-0.1

-0.2

-0.3 i
I
I
-0.4 .
0 2 3 4 5 6 7 8
k

Fig. 5.7. The aerodynamic derivative hC~ of a rectangular wing heaving in the
extreme ground effect versus the Strouhal number for different Mach numbers and
>. = 2.

currence of pronounced maxima that te nd to decrease and shift to smaller

Strouhal numbers with increases in the Mach number and the aspect ratio.
The analysis of the curves representing hC~ versus the Strouhal number
shows that far the range of Strouhal numbers, corresponding to the max-
ima of (-hC~), points of "loss of vortex damping" appear, i.e., zeros of the
 132 5. Cümpressible Flüw

5.0
I !
, I i
ti 4.5
-h Cy
4.0
I I
3.5
I ,0·11\
3.0
I I \
2.5
/ \
2.0 I 0'1\ \
/\ 1/ \/ \
1.5
Mo" O.S; X Y\ B1"\1
1.0 V___
0.5
~VL-\ V ~

0.0
o
I
2 3 4 5 ° 6 k 7 8

Fig. 5.8. The aerodynamic derivative hC; üf a rectangular wing heaving in extreme
grüund effect versus the Strüuhal number für different Mach numbers and A = 3.

0.4~.--,---,--~----r--r--r-...,-----,

.. 0.3 ~-"---+-----.::t---r'<f--+",+--7+--',I
h Ch
Y
0.21---+---+---1-+-I-+-+-+-+--1

-0.2 f-----I-_+--+--+-~~-\_7">~-__i

-0.3 f-----I~_+--+--+--+_-r_--t-__i

2 3 4 5 6 7 8
k

Fig. 5.9. The aerüdynamic derivative hC~ üf a rectangular wing heaving in the
extreme ground effect versus the Strüuhal number für different Mach numbers and
A=3.

quantity hC~. These features of the unsteady aerodynamics of a wing in an

unsteady compressible ground-effect flow can be interpreted as an "acoustic
resonance." The theoretical possibility of the occurrence of acoustic resonance
was discussed by Söhngen and Quick [140] in connection with unsteady gas
flows through axial compressors and within a more general formulation of
flat plate cascade oscillations in compressible flow by Gorelov et al. [141].
 5.4 Nonlinear Steady Compressible Flow Problem for a Foil 133

The same conclusions were drawn by Efremov and Unov [142J, who studied
oscillations of a foH in a two-dimensional flow in the presence of the ground.
Acoustic resonance may take place when the frequency of oscillations of a
wing in a restricted compressible flow 3 coincides with some of the fundamen-
tal frequencies of oscillations of compressible flow with the same boundaries.

5.4 Nonlinear Steady Compressible Flow Problem

for a Foil
Consider a moderately curved foH in a two-dimensional steady compressible
flow near a flat ground. To the lowest order, the corresponding channel flow
equation, formulated with respect to the relative motion velocity potential
cPh, can be derived from the more general three-dimensional equations (5.14)-
(5.16) in the form

[1 + 21 b - ]d 1 d
l)M;(l - u 2 ) dx [Ys(x)uJ - 2Ys(x)M;u dx u2 = 0, (5.51)

where Ys(x) = ys(x)/h,

( ) _ dcPh (5.52)
u x - dx '

where CPh is the perturbed velo city potential of the fluid. Equation (5.51) has
to be solved subject to the following boundary conditions:
• At the leading edge (x = 1),

(5.53)
• At the trailing edge (x = 0),

PI! (0) = o. (5.54)

We introduce P = 1- u 2 , J.L = (1' -1)M;/2. Multiplying (5.51) by u, we

obtain
2 dys 1 cl 2) 1 _ 2 2 d 2
(1 + J.Lp) ( u dx + 2Ys dx u - 2YsMo u dx u = 0 (5.55)
A

or, in terms of p,

(1 + J.Lp) [(1 - p) dys _ ~Ys dp ] + ~M2(1 _ p) dp = O. (5.56)

dx 2dx 2 0 dx
We divide both parts of (5.56) by Ys. Hence,

(1 + J.Lp) [(1- p) dlnys _ ~ dp ] + ~M2(1- p) dp = O. (5.57)

dx 2dx 2 0 dx
3 In this case, in a highly constrained channel under the wing.
134 5. Compressible Flow

Grouping with respect to d In Ys and dp,

1 1
(1 + I1P)(l - p) d In Ys - 2(1 + I1P)dp + 2M;(1- p)dp = 0,

dl - _~[l+J1P-(l-P)M;]d (5.58)
nys - 2 (1 + I1P)(l - p) p.
Integrating (5.58),

lnys = -~ ln(l- p) - ~ M; In(l + I1P) +lnG*. (5.59)

2 2 11
Applying the trailing edge condition Ys(O) = l,p = 0, we find that G* = 1 so
that
:2 = (1- p)(l + I1P)M5/J1-.
Ys
(5.60)

We introduce the pressure coefficient

2 {[ 1 2 2 ]/,/(1-1) }
P = 'YM~ 1 - 2Mo b -l)(u -1) -1 ,

and express in terms of function p:

(5.61 )

The lift coefficient of a foil can be determined by using the following

1
formula:
1+0 _ dys
Gy = P(Ys) _'(_ )' (5.62)
o Ys Ys
where B= B/ h, and B is pitch. For a fiat plate

ys(x) = 1 + Bx, y~ =B xE [-1,1],

so that the lift coefficient is given by

- 11
Gy(B,Mo) = =
-P1 dys dp,
P(p,B,Mo)-d
Bop
(5.63)

where PI = p(B) is determined by using formula (5.60).

For a fiat plate in incompressible fiow, the expression for the lift coefficient
to the leading order is
(5.64)

So me calculated results, corresponding to the case of a fiat plate, are pre-

sented versus the relative pitch angle Bfor the Mach numbers Mo = 0.5 and
Mo = 0 (incompressible fiow case) in Fig. 5.10.
 5.4 Nonlinear Steady Compressible Flow Problem for a Foil 135

0.8

-1
I

~
, , , Incompressible
0.4
I
t

~
I
0.4 0.8 1.2 1.6 2.0
9th
Fig. 5.10. The lift coefficient of a flat plate of infinite aspect ratio in a compressible
extreme ground effect versus the relative pitch angle (} / h (solid line: compressible
Mo = 0.5; dashed line: incompressible Mo = 0).
6. The Influence of Endplates, Flaps, and Slots

6.1 An Estimate of the Influence of Endplates

Adefinite pecularity of the wing-in-ground-effect vehicle compared to the

airplane consists of the presence of endplates. Endplates are mounted at the
wing's tips and are intended to decrease leakage of air from under the lifting
system. Consequenly, the mounting of endplates results in the augmentation
of lift or in a decrease of the induced drag for a given lift. In practice, the
configuration of the endplates can vary. Figure 6.1 illustrates schematically
some of the possibilities.
According to the approach adopted herein, to account for the infiuence of
the endplates upon the fiow past a wing in the ground effect, one has to solve
the corresponding local fiow problem in the vicinity of the order of O(h) of
the tip of a wing equipped with an endplate. Strictly speaking, the local fiow
solution includes both homogeneous and nonhomogeneous parts, as is the
case for a wing without endplates. However, it can be shown that the ratio of
the nonhomogeneous to the homogeneous component is of the order of O(h).
Therefore, to the leading order, the analysis can be restricted to constructing
only a homogeneous (circulatory) component of the fiow around the wing tip
of a given geometry. Such a solution will be reduced to the conformal mapping
ofthe local flow domain onto the interior ofthe unit strip 0::; ';Sfae = 7/Jae ::; 1
in the plane of the complex velocity potential f ae = 'Pae + i7/Jae.
The solution procedure for the problem will be illustrated for the fiow
past a rectangular wing with vanishingly thin endplates.

r-
b c

I
d

I
Fig. 6.1. Various schematized configurations of endplates.

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
138 6. Endplates, Flaps, and Slots

At first, eonsider the loeal flow problem for an edge with a thin lower
endplate of height hep. Introduee a eomplex variable ( = ii+iy in the physical
plane. Map the domain of flow in the (-plane onto the upper half plane C:Sg =
g2 > 0 by the Christoffel-Sehwartz transformation; see e.g. Lavrient'ev and
Shabat [129J . The point-to-point eorrespondenee for the eonformal mapping
is shown in Fig. 6.2. The mapping is obtained in the form
1 (t
., - - 1n
I' + l)hep - t) + 2(1- , ep)t
(6.1)
(1 - t)(t + ,ep)
--''---c,-'....!....:;;..,-
- 7r , ep 7r(l - t 2 )'
where
(1 + ,ep)g + 2,;p 2 ß
t= ß' > 1.
(1 + ,ep)g + 2 ' 'ep =

The parameter ,ep is related to the height of the endplate through the
following equation:

h-ep -- hep -_ ~ [aret an hep - 1)J2,ep + 1 + (1 - ,ep)J2,ep + 1] (62)

, .
h* 7r 3,ep + 1 2,ep(1 + ,ep)
where h* is the loeal ground clearanee, Le., the distanee from the wing's
surfaee to the ground near the endplate. The mapping of the upper half
plane CJg > 0 upon the interior of the unit strip 0 ::; CJfae ::; 1 of the plane of
the eomplex potential is performed by the function

9 = gl + ig2 = exp(7rfae). (6.3)

For points on the wing,
• on the upper surfaee,

( = ii + i, 9 = gl < -ß, fae = 'Pae + i, 'Pae > 0; (6.4)

y
l;=v+iy
C B v
C' B' hep
A h
C' C

C B A
IIliillli
S'
~' IIIII! 1111
C

ß 1 ß'
Fig. 6.2. The correspondence of points for the conformal mapping of the fiow field
in the vicinity of an endplate onto the auxiliary upper half plane.
 6.1 An Estimate of the Infiuence of Endplates 139

• on the lower surface,

(=v+i, O>g=gl>-ß', fae='Pae+i, 'Pae<O. (6.5)
At the points on the endplate for v = 0,
• on the exterior surface of the endplate,

- -hh* h-
- Y> -ep
(= iy, 0> - =- ep, -ß < gl < -I,

fae = 'Pae + i, 'Pae ;?: 0; (6.6)

• on interior surface of the endplate,

(= iy, hep
- h*
- = -
h-
< Y- <0 -1< gl < -ß';
ep - - ,

fae = 'Pae + i, 'Pae ::; O. (6.7)

To match the local flow solution for the endplate with the asymptotic
descriptions of the upper flow and channel flow, we need to know the far-
field behavior of the edge flow potential at y --+ -00, Y = 1 ± o. These
estimates are readily obtained in the following form:
• Far from the endplate on the upper surface of the wing (v --+ -00, Y =
1 + 0),
(6.8)

• Far from the endplate on the lower surface ofthe wing (v --+ -00, Y = 1-0),
_ 1 (1 + I'ep) 2
'Pae c::: v - - in 2 - ( ) • (6.9)
1f 81' 1f 1 + I'ep

The general form of the solution valid near the wing tip with an endplate has
the following asymptotics:

(6.10)

Recalling that the upper flow potential in the immediate vicinity of the wing's
edge has the following asymptotic behavior,

(6.11)

and matching (6.11) with the asymptotic representation of (6.10) far from
the edge, one obtains taking into account (6.9),

(6.12)
140 6. Endplates, Flaps, and Slots

In similar fashion, matehing in the region of the flow below the wing, we find
al and the boundary eonditions for the ehannel flow equation (2.22) for a
wing equipped with endplates:

_ ~Q _ h* 8!.ph h* = h* /h, v = 0; (6.13)

al - 2 I - 8v'

1
!.pI ~ !.ph + !.p12 h ln h + h!.p13
= haI [11"zI -ln (1 + -;ep )2 _ 2 ] + ha4, v = 0; (6.14)
11" 8'Yep 1 + 'Yep

(6.15)

+ 'Ye p)3]
!.p13 - [2
= -al
11" 1 + 'Yep
+ 1n 11"(1 8'Yep + 11"-1 A 2, v =0. (6.16)

Thus, in the problem under eonsideration, the ehannel flow potential ean
be found, as previously, by solving the quasi-harmonie equation (2.22) for
the nonlinear ease and the Poisson equation (3.14) in the linearized ease.
However, here, the boundary eonditions to be applied at the wing's planform
eontour, ineorporate the influenee of the endplates. Note that in the linear
ease, the loeal clearanee of the wing near the endplate should be substituted
by h.
A relatively simple solution ean be derived for a reet angular wing of a
small aspeet ratio A; see Rozhdestvensky [44]. In this problem, the upper
flow potential outside of the tips is eonstructed in the same way as for the
small-aspeet-ratio wing without endplates. The ehannel flow is determined
by using the equation (3.48) with boundary eonditions

* _ 2h ,x [2'Ye p
!.pI - - -
1 1n (1 + 'Ye p)311"] z = ±1, (6.17)
11"(l+'Yep) 11" 8h,x'Yep ,

where h,x = h/ A. The lift eoefficient for the lower endplates was obtained in
the form

Gyep = BA {I + 6h,x [ln 11"(1 + 'Yep) _ 'Yep -1] + O(h2)}, (6.18)

6h,x 11" 8'Yep h,x 'Yep + 1

where h,x = h/ A and parameter 'Yep is linked to the endplate height hep by
equation (6.2). For h ep -+ 0, the lift eoefficient beeomes equal to that for the
small-aspeet-ratio wing without endplates, Le.,

BA ( 6h 11" )
Gy = 6h,x 1 + -:;-ln h,x . (6.19)

In Fig. 6.3, some ealculated results are eompared with experimental data for
A = 1, h,x = h = 0.057, hep/h = 0.875.
 6.1 An Estimate of the Influenee of Endplates 141

1.0

Cy
0.8

with endplates
0.6

0.4

0.2

0.0
0 2 4 6 8
9·
Fig. 6.3. A eomparison of theoretieal and experimental results for a reet angular
flat wing of small aspeet ratio (A = 1, h = 0.057, hep/h = 0.875; solid lines: formula
(6.18); dots: experiment).

The augmentation of the Hft coefficient of the wing resulting from the
installation of endplates ean be charaeterized by the eoeffieient

K ep = G = {I + 6h).. [ln 11"(1 + 1'ep)3 _ 1'ep - I]}

yep
Gy 11" 81'ep h).. 1'ep + 1

/ (1 + 6~).. In ~). (6.20)

Some ealeulated eurves, showing the behavior of this eoefficient versus the
relative height kep of the endplate for different clearances k).. of the wing-in-
ground effeet are presented in Fig. 6.4. It ean be observed from Fig. 6.4 that
the utilization of endplates may result in a eonsiderable inerease in the Hft.
In other terms, for a wing with endplates, the effeetive aspeet ratio may quite
notieeably exeeed the geometrieal aspeet ratio.
We turn to the eonsideration of endplates of a more complex eonfiguration;
see Fig. 6.1 b. Suppose that the upper part of the endplate has a height h 1ep '
whereas the lower part has the height h 2ep • The loeal flow velo city potential
ean be obtained by the same technique as for the simpler lower endplate.
The Hft coefficient for the wing of small aspeet ratio with endplates under
eonsideration was found in the form

G
Yep
= () ..
6h )..
{I + 6h).. [1n 11"(14ß'+ 1'd21'~
2h
+ ß'1'? _ 1] + O(h )2} (6.21 )
11" 1'1 ).. 1'22 )..,

where parameters 1'1 > 1'2 > 1/ß' are related to the dimensions of the end-
plate by the following relationships:
142 6. Endplates, Flaps, and Slots

1.8,----.-----,---,.----,-,--,
Cyep
Kep=-C-y-
1.6 i---+---f---j----f+----1

1.4 f..----+---I---j--f---t----j

1.2
1

I
1.0 ~§:::t:::=-L_...L_l_-.J
QO 02 OA Oß Oß 1~
hep I h;..

Fig. 6.4. The dependence of the relative increment of the lift of a small-aspect-ratio
wing on the ground clearance and the relative height of the endplates.

hlep _ 1 [ß'B'Yl('Y~-'Y~) +2 t B(1-'Yl)]

h- 1 = -h (6.22)
ep
- -7r 2
"12
are an 1 + "11 B2'

h2 = h2ep = .!. [A'Y1 (ß' -1) + 2 arctan A('Y1 -1)], (6.23)

ep h 7r "12 'Y~ + A

ß''Y~ - 1
A= 1-ß' .
(6.24)

Some results illustrating the influence of the lower and upper parts of
the endplate upon the Hft coefficient of the wing are presented in Fig. 6.5.
It is easy to see from the analysis of the graph that the upper parts of the
endplates have an insignificant effect upon the increase of the Hft. The same
conclusion was drawn by Ermolenko et al. [133J.
Now, return to a more general problem of the influence of endplates upon
the aerodynamics of a wing of arbitrary aspect ratio. As shown above, when
the wing tips are equipped with endplates, a change occurs in the bound-
ary conditions, starting from an approximation of the order of O(h). Conse-
quently, functions <PlI and <Ph, which characterize, respectively, the first and
the second approximations for the channel flow potential, are not dependent
upon the parameters of the endplates. On the other hand, the influence of
channel flow upon the upper flow is defined by the strength Q1 of the source
(sink) singularities, distributed along the wing's planform boundary contour.
Due to the fact that

1/ = 0, (6.25)
 6.1 An Estimate of the Influence of Endplates 143

1.4 ,;----,-------,,-;::::==='11
Cyep I 'li h = 0.1
"ep=--
Cy I ! h28p l h 1 0.5
13~1--~-~~~~~-+--~
. i

1.2, I
1
1
'I I Ih 10.25 I
I I I
l'l!J?ft1I,
I h2oJ/h=0 I
1.0
I ;
'
I
'
I
0.0 0.2 0.4 0.6 0.8 1.0
h,op' h

Fig. 6.5. The relative increment of the lift of a small-aspect-ratio wing versus the
heights of the lower and upper parts of the endplates, h A = 0.1.

within the approximation considered, the endplates practically do not affect

the upper flow.
The previously mentioned considerations lead to the following conclusion:
to account for the variation of flow past a wing in the ground effect
due to the presence of endplates, it is sufficient to solve the problem
just for the corresponding increment of the channel flow potential.
The corresponding boundary problem with respect to this increment <PIep
for unsteady linearized 1 flow around a lifting surface with endplates can be
written as follows:

(6.26)

hal
<PIep (1, z) = 0, <PI ep (x, ±>'/2) = - -7rG

- =f=~ a<plep G z = ±>'/2j (6.27)

- 7r az '
a<plep _ a<plep = 0
x =0, (6.28)
ax
8t '
where for thin lower endplates the function G is given by

G( rep ) _- 1 - rep + 1n (1 + rep)3 , (6.29)

1 + rep 8rep

1 Nonlinear version can be handled in a similar fashion.

144 6. Endplates, Flaps, and Slots

and parameter rep is related, as earlier, to the height of the endplate by

equation (6.2). Having at our disposal the lowest order flow problem solution,
we ean readily obtain information on the influenee of the endplates for both
the steady and unsteady motions of a wing near the ground.
Consider, as an example, the steady motion of a flat reet angular wing with
lower thin vertical endplates at given angle of piteh O. The relative inerement
of the lift due to the endplates is found in the form

/'bep = ~ep = 1 + hG(rep)A(>') + O(h 2 ), (6.30)

y

where A is a function of the aspeet ratio

7r
qn = X(2n+ 1). (6.31)

Some ealeulated results for the eoeffieient /'bep versus the aspect ratio >. and
for different hep/h are presented in Fig. 6.6. In the limiting eases of large and
small aspeet ratios of the wing and for hep :s 0.5h, the expression for /'b ep is
simplified:
• For >. -+ 0,
128) hep -
/'b ep ~ 1 + \6h [In ( 4 4 - 1] , eep = 1 - h = 1 - hep. (6.32)
7r A 7r eep

• For >. -+ 00,

2h
/'bep~l+\ (128) ]
[In 4 4 -1. (6.33)
7rA 7r eep

2.0 ~--,-----r--~--,-----,

ICep= Cyep I
C
y 1.8 t----++---'---+--+----+---~

1.61---t-+--+---t--t----+---j

1.41---t---\j----t--t-----+----j

1.2 f-, ---+-"<--t---"Ioc--~----,----I

1.0 L---.l-=::::±::===b~;;;;;;~;;;;;;;;J
o 2 3 4 5 6

Fig. 6.6. The relative increment of the lift of a rectangular flat wing with thin
lower endplates in extreme ground effect versus the aspect ratio, h = 0.1.
 6.1 An Estimate of the Infiuence of Endplates 145

It is not difficult to verify that with an asymptotic error O(h)..) (where h).. =
h/>') the first of these formulas is compatible with expression (6.18), which
was derived from a straightforward solution of the flow problem for a small-
aspect-ratio wing.
It follows from the analysis of the formulas presented and Fig. 6.4 and 6.6
that, for a wing near the ground, the efficiency of endplates increases with a
decrease in the aspect ratio and/or a decrease in the gap between the lower
tip of the endplate and the ground.
In the case considered before, the endplates were assumed to be vanish-
ingly thin. Inclined and/or thick endplates can be handled in a similar fashion.
Consider an endplate of a more general polygonal configuration (Fig. 6.7).
We map the flow domain around the endplate onto an upper half plane <.S > 0
by the Christoffel- Schwartz transformation so that the point A 3 corresponds
to a3 = -1 (the point-to-point correspondence is shown in Fig. 6.7). The
mapping function is

- .- 1
( = x + lY = 11'F(O)
1 9
a2
F(g) dg
9 , (6.34)

where

(6.35)

where 6:k = (Xk/7r are the external angles of the polygon.

The Christoffel-Schwartz constants a2, a4, ... ,ak-l are determined taking
into account the point-to-point correspondence in the planes ( and 9 and
depend on the parameters that characterize the geometry of the endplate. The

Z =v+ i Y

Fig. 6.7. The correspondence of points for conformal mapping of the fiow domain
in the vicinity of the endplate of a polygonal cross section onto the auxiliary upper
half plane.
146 6. Endplates, Flaps, and Slots

lift coefficient for a small-aspect-ratio wing-in-ground effect with endplates

of polygonal cross section was obtained in the form

(6.36)

where
T= 1° F(~)
a2 ~
d~.
- F(O) (6.37)

The endplate configurations presented in Fig. 6.la,b,c,d can be derived as

particular cases from the polygonal shape considered.

6.2 A Lifting System with a Jet Flap

in the Extreme Ground Effect
From the viewpoint of use in the transitional modes of motion (takeoff and
touchdown) the so-called jet flaps have a certain practical interest. These
flaps are formed by high-speed jets, utilizing the reserve of the air exhaust
of turbojet engines blown through narrow slots of the trailing edges of the
lifting system. Whereas the increment of lift due to the deflection of rigid
flap is always accompanied by a certain increase of drag, jet flaps do not
have this deficiency because the major part of the jet momentum gives rise
to the thrust. Consequently, jet flaps favorably combine propulsive and lifting
properties. One should also point out another reason that justifies the use of
jet flaps. When the vehicle is operating in transient modes above rough seas,
rigid flaps can touch the sea surface, thereby experiencing considerable hy-
drodynamic loads. The latter circumstance can lead to failures in operation
of the flaps and the corresponding flap control systems. The use of nonrigid
devices, based on jet blowing, gives the possibility of controlling the lifting
properties of the lifting system, including the cases when the jets touch the
water surface. Assuming that the relative ground clearance is small the prob-
lem of determining the aerodynamic characteristics of a wing with a jet flap
can be effectively solved by the method of matched asymptotic expansions.
We start with the formulation of a nonlinear flow problem for a
jet-flapped wing in the ground effect.
We consider a wing in steady motion with a jet flap along the trailing
edge; see Fig. 6.8a.
In what follows, it is assumed that the jet is vanishingly thin. The latter
assumption was introduced for the first time by Spence [143], who studied
the two-dimensional problem for the flow past a foil with a jet flap in an
unbounded fluid. No account was taken in Spence [143] of the ejection effect
of the jet upon the surrounding fluid. However, both the results of Spence
and other works, see, for example, Maskell and Spence [144], utilizing the
 6.2 A Lifting System with a Jet Flap in the Extreme Ground Effect 147

a.

FLOW SUBDMSION NEAR THE JET FLAP

b.
Fig. 6.8. Scheme of the flow past a wing with a jet flap in the ground effect:
(a) general view; (b) near the jet flap .

hypothesis of the thinness of the jet, are in fair agreement with the experi-
mental data even for large deflection angles up to 60°. These considerations
justify the validity of the adopted mathematical model of a thin jet, in which
expansion of the jet due to involvement into the motion of particles of the
surrounding fluid is neglected.
The velocity potential of the flow past a wing with a jet ftap should
satisfy the Laplace equation, the ftow tangency conditions on the wing and
the jet sheet, the condition of the decay of perturbations at infinity, and
the dynamic condition upon the jet surface. The Kutta-Zhukovsky condition
at the trailing edge should be replaced with the requirement that the jet is
blown at a given angle to the chord of the wing.
As in the general algorithm of the solution of the flow problem discussed in
section 2, the ftow field is conditionally subdivided into characteristic regions:
the channel ftow 'D\ under the wing and the jet sheet; the upper ftow, including
the region 'Du above the wing; the jet sheet and part of the ground outside of
the "shadow" of the wing and the jet on the ground; the regions of local ftows
near the leading and side edges 'De; and the local ftow region in the vicinity
of the trailing edge with a jet flap V j . Below, based on the assumption of
148 6. Endplates, Flaps, and Slots

the smallness of the relative ground clearance h « 1, the main stages of the
asymptotic solution of a steady flow problem for a wing with a jet flap will
be shown. In accordance with the general hypothesis, the deflections of the
jet sheet from the plane Y = h are assumed to be comparable to the ground
clearance, that is,
Ih - Yjl = O(h). (6.38)
As indicated in section 2, such an assumption enables us to account for non-
linear effects, at least in channel flow, with an asymptotic error of O(h 2 ).
Special consideration is required of the local flow near the trailing edge
with a jet flap. As for the rest, the solution procedure does not differ signifi-
cantly from the approach discussed in section 2 for a wing without a jet flap.
Therefore, corresponding modifications of the velo city potentials in regions
VI. V u , and V e will be discussed very briefly.
At first, consideration is restricted to a vanishingly thin flat wing with
a straight trailing edge moving at zero incidence. The fuH problem for the
perturbed velo city potential cp is described by the Laplace equation
[J2 cp [J2 cp [J2 cp
ox2 + oy2 + oz2 = 0, (6.39)

and the foHowing boundary conditions:

• Flow tangency conditions on the wing, ground plane and the jet sheet:

ocp = 0 Y = h±O, (x,z) E S; (6.40)

oy ,

ocp = 0 Y = 0 ±O; (6.41 )

oY ,
ocp = (oCP _ 1) oYj + ocp oYj , Y = Yj(x, z), (6.42)
oY OX OX oz oz
• Dynamic condition on the jet:
Interacting with the oncoming stream, the jet experiences deformations.
As a consequence, centrifugal forces occur proportional to the local lon-
gitudinal curvature. These forces are balanced by the pressure difference
across the jet surface

P- -p+ -- KC}-L' X< 0, (6.43)

where the pressure can be calculated by the formula

p = 2 ocp _ (oCP) 2 _ (oCP) 2.

OX OX oz
K is the curvature defined though the coordinates Yj of the jet surface as

02 y. [ ( oy. ) 2] 3/2
i
K(x, z) = Ox (x, z)j 1 + o~ , (6.44)
 6.2 A Lifting System with a Jet Flap in the Extreme Ground Effect 149

CI-' = CI-' (z) = 21 (z) / pU;;·l is the sectional coefficient of the jet moment um,
I( z) is the sectional jet momentum, plus and minus correspond to the upper
and lower surfaces of the jet sheet, and Sj is the area of the jet sheet. The
total jet momentum coefficient can be calculated by the formula

Cj 11
=S
1 2
/
-1/2
CI-'(Z) dz, (6.45)

where l is the relative span and S is the area of the wing related to the
square of the root chord.
Rewriting the dynamic condition (6.43) taking into account (6.44) with
asymptotic error O(h 2 ), we obtain

_ +_ 8 2 Yj
P - P - CI-' 8x 2 ' (x, z) E Sj.

In what follows, it will be shown, that for small relative ground elearances
h, the main modification of the jet surface takes place in the downstream
direction within a distance of the order of O( v'h). 2 Farther on, the jet sheet
loses longitudinal curvature, the pressure difference aCross it vanishes, and
(in the case of a finite aspect ratio of the wing) generation of a vortex sheet
begins. It can be easily seen that the dynamic and kinematic conditions
work automatically in the case of a vortex sheet.
• The requirement of a jet blowing at a given angle r with respect to the
chord of the wing:
As noted before in the problem under consideration, the Kutta-Zhukovsky
condition at the trailing edge is replaced by the requirement that the jet
should be blown at a given angle r = r(z) with respect to the chord at
cross section z = const., Le.,
8y·
arctan 8~ = r(z), for Xj = 0, (6.46)

where Xj is a local x coordinate, directed upstream .

• The condition of the decay of perturbations at infinity:
\lcp ~ 0, (6.47)
Now, we turn to the asymptotic solution of the loeal problem of
flow near a trailing edge with a jet flap. Consider a local flow in elose
proximity to a trailing edge equipped with a jet flap. Introduce local coordi-
nates stretched in vertical and longitudinal directions, Le.,

-
X·J - Yj
Zj = z,
Xj = al[h*(O, z)]' Yj = a2[h*(0, z)]'

2 The ground clearance is measured between the hinge of the fiap and the unper-
turbed position of the ground surface.
150 6. Endplates, Flaps, and Slots

where h*(O, z) is the local distance of a point on the trailing edge from the
ground at a given cross section z = const., in our case of a wing of zero lateral
curvature one can set h*(O, z) = h.
The stretching functions 0"1 and 0"2 are to be determined by the least
degeneracy principle. Note that, depending on the choice of the stretching
functions and the lowest order asymptotics of the local flow potential in the
region Dj , it is possible to distinguish different local subdomains for which
the local flow descriptions would have corresponding distinct limiting forms.
The sub division of the jet flow domain Dj into different subdomains, as weH
as respective orders of coordinates, are shown in Fig. 6.8.
In the subdomain Dh , independent variables are of the order Xj = O(h)
and Yj = O(h), so that one can set 0"1(h) = hand 0"2(h) = h. Substitution of
stretched variables in the fuH flow problem leads to the local flow problem in
the immediate vicinity of the hinge of the jet flap. The solution of the latter
problem was obtained in the foHowing form:
• On the upper surface of the wing-jet,
{)<p+ hT _ +
- - = --lnx·
{)x. 7r 11
+R.11 (6.48)
J1

• On the upper surface of the wing-jet,

{)<Ph
-{)_ hT [
= -ln 1-
-
exp( 7rxh
1+ Rh'
- (6.49)
Xjr 7r

where <pjr is the flow velocity potential in subdomain Djr> xh = Xjr /h, Rh'
and R j1 are unknown parameters to be determined by matching.
It follows from expressions (6.48) and (6.49) that the flow velocity near
the point of blowing has a logarithmic singularity with different signs on the
upper (acceleration) and lower (deccelaration) surfaces of the wing-jet. In
subdomain Djr, the jet degenerates into a segment of a straight line, so that
the kinematic boundary condition coincides with the requirement that the jet
is blown at a certain angle. The solution obtained in Dh is two-dimensional
and describes the local flow at distances of the order of O(h) from the hinge
of the jet flap. The most complete description of the jet can be obtained in
subdomain D h . Under the wing and jet Dj~ C Dh , it is convenient to choose
the stretching function in the vertical direction as 0"2(h) = h. Longitudinal
stretching 0"1 (h) should be selected so that the mathematical description
of the jet is the least degenerate. It means that to the lowest order, one
has to retain both the dynamic and kinematic boundary conditions on the
jet. We analyze in more detail the procedure of constructing the solution in
subdomain Dh . To retain the channel flow for h -* 0, it is logical to assurne
later on that for h -*°
(6.50)
Equation (6.50) implies that in the limit h -* 0, we will obtain a one-
dimensional description of the flow under the wing and jet in the vicinity
 6.2 A Lifting System with a Jet Flap in the Extreme Ground Effect 151

of the fiap. In this case, the dynamic boundary condition on the jet acquires
the form
~ {)cpj _ ~({)cpj)2 _ C (z)~ {)2 Yj (6.51 )
al {)xh ar {)xh - J1- ar {)xtl .
The governing Laplace equation, it can be shown as previously, reduces to

(6.52)

The distinct limit for the system (6.51) and (6.52) is secured by the following
choice of the stretching function al (h) and the lowest order asymptotics of
the local fiow potential:

(6.53)

With this in mind, equations (6.51) and (6.52) can be rewritten as

2 {)cpj _ ( ()cpj ) 2 = C (z) {)2cpj (6.54)

() - ()- J1- {)-2'
xh xh xh

(6.55)

To obtain an equation describing the configuration of the jet sheet, we first

integrate (6.55). We determine the constant of integration by requiring that
at downstream infinity (xh -+ -00), the jet becomes horizontal at any given
cross section z = const. Therewith Yj -+ Yjoo (z), and the perturbation ve-
locity in the channel fiow between the jet and the ground vanishes, Le., for

{)ipj 0
---+ (6.56)
{)xh '
where Yjoo (z) has to be determined by matching. Therefore,

(6.57)

Using (6.57) to exclude {)CPj/{)xh from (6.55), we derive the following differ-
ential equation for determining the jet configuration:

{)2Yj
CJ1- (z ) {)-2 _
-
1_ yL
-2 . (6.58)
xh Yj

Multiplying both parts of (6.58) by Yj, we can rewrite this equation as

(6.59)
152 6. Endplates, Flaps, and Slots

and
ß- 2
(-Jl-) (ih + y~",,)
2 -2
= - + C*. (6.60)
ßXh CI-' Yj
Consistent with the previous requirement that for Yj = Yj (Xj}) for any given
z = const. the jet becomes horizontal, its slope must vanish too, Le., for

ßYj- t
-
ßXh
° . (6.61)

Hence, we can determine the constant

C* = _ 4Yjoo . (6.62)
CI-'

Thus, the differential equation governing the form of the jet becomes

Byj
( --- )2 -_- 2 (_Yj + yj2"" - 2-Yjoo ) -_ - 2 (Yj - -Yjoo)2 .
~ (6.63)
ßXh CI-' Yj CI-' Yj

Selecting the appropriate sign of the square root, we can reduce (6.63) to

(6.64)

It is not difficult to integrate (6.63) to obtain the following implicit equation

for fh = Yj(xh):

c;. = 2(y'Yj -1) + v'Yj"" ln[(.jYj

Yf2"Xj} - v'YC)(1 + v'YC)].
(v'YJ + v'Yjoo)(l - v'Yjoo)
(6.65)

Now, we can apply the requirement that the air should be blown from the
trailing edge at a prescribed angle 1" = 1"(z). Using (6.46) and accounting for
the order of magnitude of the jet coordinates, that is,

(6.66)

we obtain that for xh = 0,

ßy·
arctan -ß~ = arctan
(h fL
ßY·)
ß-: ':::!
/J ßy·
v h ß _.J = 1"(z) (6.67)
~ vh ~l ~}
or
ß!j (0, z) = r(z). (6.68)
ßXh .;Ji
We apply (6.68) to (6.64), keeping in mind that for Xj -+ 0,
(trailing edge). Then
 6.2 A Lifting System with a Jet Flap in the Extreme Ground Effect 153

T(Z)
Vii = Yf2 -
c;: (1 - Yjoo), (6.69)

and the distance of the jet from the ground far downstream is

(6.70)

Setting '[hoo = 0 in (6.70), we can determine for which combination of sectional

magnitudes of the parameters Op" T, and h the jet would touch the ground
at a given cross section Z = const.:

T(Z)V°P,(Z) = 1. (6.71)
2h
For a uniform spanwise distribution of the jet deflection angle, the ground
clearance and the jet momentum coefficient, equation (6.71) can be inter-
preted as a condition of blockage, Le., the situation when the jet touches
the ground everywhere spanwise

rc; rc;
TV 2h = TV 2h = 1, (6.72)

where for uniform jet momentum distribution, 0 p, = q, where q is the total

jet momentum coefficient.
The equation of the jet sheet in the case of blockage follows from (6.65)
for Yjoo = 0:
Yj = (y'~~p, + 1)2 = (2~Xj + 1)2, (6.73)

and the blockage occurs at a distance

2Vii 2h
Xj = - - = -y'20p" Xj = - - = -y'2hOp, (6.74)
T T

from the trailing edge.

The longitudinal velo city in a narrow channel under the jet sheet is given
by equation (6.57), where Yjoo is described by (6.70). The spanwise distri-
bution of the longitudinal perturbation velo city at the trailing edge can be
obtained by setting Xj = 0, Yj = 1, wherefrom

(6.75)

Due to the conservation of mass, the magnitude of ocpj /OXj is practically

the same as the perturbation velo city under the wing in the vicinity of the
trailing edge. Therefore, to the lowest order, the boundary condition for the
154 6. Endplates, Flaps, and Slots

ehannel flow equation at a trailing edge equipped with a jet flap ean be
written identieally as

°'Ph (0
oX ,Z
)= ()V G/1(Z)
TZ 2h· (6.76)

This ean also be shown through the matching proeess.

The flow potential in the upper part of subdomain V h (xh) was found in
the form
(6.77)

where 'Pt and 'Pt (xh) are unknown functions to be determined by matching.
The next eharaeteristie subdomain of the flow near the jet flap is V h , whieh
is loeated above the wing and jet; see Fig. 6.8. In this subdomain, Yj = O( v'h)
and Xj = O( v'h). The expression for the flow potential in Vj, was found in
the form
+ 9(h)jOO - -2 2 -
(6.78)
'Ph = 27r -00 qj(~)ln[(xj3 -~) +YjJd~+ ... ,

where g(h) is an unknown gauge function of h, q(~) is an unknown funetion,

and xh = Xj2 = Xj / v'h. Thus, as a result of the asymptotie analysis of
the flow field in V j near the point of jet blowing, we ean determine the
eharaeteristics of the flow in subdomains VjI , V h , and V h with the help of
expressions (6.57), (6.77), and (6.78). The unknown parameters and functions
are determined by matching.
Continuing the diseussion of particular features of the asymptotic
algorithm for a wing with a jet flap in the ground effect, some eor-
reet ions will be shown briefly, whieh have to be introdueed into a general
algorithm of the solution in the partieular ease of a wing with a jet flap. In
the upper flow region V u , the expression for the potential 'Pu (2.31) has to
be supplemented by the term

-~1 Qj dl (6.79)
47r h r '

whieh represents the potential of the distribution of sinks with strength Qj

along the trailing edge 12 and models the influenee of the jet Rap upon the
upper flow. In sub region 'Oe near the leading edge and side edges, the velocity
potential is given byexpression (2.39), in whieh the eoeffieients ai depend on
eharaeteristies of blowing. If the wing has endplates, their influenee upon the
aerodynamies of the lifting system ean be determined in the same fashion as
for a wing without a jet flap.
In the ehannel flow VI the equation for determining the potential 'PI has
the same form as that for a wing without a jet flap. For zero piteh angle, the
equation for the lower flow potential is
 6.3 Nünlinear Steady Flüw Problem für a Foil with a Jet Flap 155

8 2 <pl 8 2 <pl
8x 2 + 8z 2 = 0, (x,Z) E S, (6.80)

with boundary conditions at the leading edge and side edges to be determined
by matching. Below, without going into details, some results are presented
of the matching needed for further calculations. Matching the upper flow
potential <Pu, valid in region Du, with solutions <P~ and <P~ in subdomains
Dh and Dh enables us to find the quantities g(h), qj, <Pt and Qj in the form

g(h) = h, (6.81 )

(6.82)

(6.83)

(6.84)

where fhoo describes ordinates of the jet for large xh -t -00. Note that
Yjoo depends on z, Le., varies along the span, and Qj(z) is negative, Le.,
represents the productivity of sinks, modelling local effects in the upper flow
subdomain, connected with deformation of the jet surface from the trailing
edge downstream.

6.3 Nonlinear Steady Flow Problem for a Foil

with a Jet Flap

Consider a jet-flapped foil with the upper and lower surfaces described by Yu
and Yl moving above the flat ground at incidence (). The jet deflection angle
is designated T, and the jet momentum coefficient Cj . In the simplest case
of the extreme ground effect, that is to the lowest order, we have to solve a
simple differential equation

d (_ d<Pll) _ dYl
dx Yl dx - dx' O:Sx:Sl. (6.85)

Equation (6.85) has to be solved with the following trailing edge condition:

(6.86)
156 6. Endplates, Flaps, and Slots

Solution of the fl.ow problem in an approximation of the extreme ground effect

is straightforward. The Hft coefficient was obtained in the form

(6.87)

where GyO ' the Hft coefficient of the same foH without a jet fl.ap is given by

1
the formula
1 dx
GyO = 1 - -2() , fit = Yl/h. (6.88)
o YI X
In the concrete case of a Hat plate of infinite aspect ratio with a jet fl.ap in
the extreme ground effect, expression (6.88) yields

C = 1 _ (1 -~j)2 = 1 _ (1 - T ..jCjfiJi)2
(6.89)
Yj 1+0 1+0'
where iJ = 0/ hand h is the relative ground clearance. The general algorithm
of the fl.ow problem solution enables us to proceed to higher approximations.
Omitting intermediate calculations, one can present the following formulas
for the Hft coefficient of a moderately thick foH with a jet fl.ap:3

1
Gy =:F1 + .r2 hln h + :F3 h + O(h ),
2
(6.90)

1
where
Vrc:
1 dx
:F1 = 1 - (1 - "'j)2 -2()'
o Yl x
"'j = T 21;,; (6.91)

2 [- 3 -/)( l-~j)
.r2=-0+-2"'j+Yu(0
7r
10
1
dx 1
-2()'
YI X
(6.92)

:F3 = ~{B2 - A2 + (iJ + ~j) C~ :~ + lnIYI~l) I) + ~j ln ß, - J("'j)

-(1-~.)[B1-Y{(0)(1-~.)+y{(0)(1-ln7r)1 hr ~(~
1 _dx } (6.93)
J J

1
with
1
J("'j) = ln'dYj, (6.94)
yjoo

,(-.) = 2( r.;. -1) + ~l

YJ V Yj
I
(yYj - vYJ:)(1 + vYJ:)
V YJoo n (yYj + v'Yjoo) (1 - v'Yjoo) .
I (6.95)

Für I\;j = 0, these fürmulas for Gy yield cürrespünding fürmulas für the foil
without a jet fl.ap, obtained in paragraph 4.1. In the case of a jet-fl.apped Hat
plate in the extreme grüund effect, when Yu = YI = 1 + Öx,
3 That is, with thickness of the order of the ground clearance.
 6.3 Nonlinear Steady Flow Problem for a Foil with a Jet Flap 157

1.6
1

Cy
I

0.0 '--"-'-'---L....L.:--'---_ _- ' -_ _- ' -_ _....J

-2 o 2 4 6

Fig. 6.9. The lift coefficient of a Rat foil with a jet Rap versus the angle of pitch
for different magnitudes of the jet moment um coefficient, T O = 30°, h = 0.1.

(1 - h:Y
F 1 = 1- 1+ ~ ; (6.96)

F 2 = -2 [-8 + -h:o 0
3 + ---(1 - h:') ] . (6.97)
7r 2 J 1+8 J,

F3 = ~ [0 (1 - h:j ) (1 + In 7r - h:j) + (0 + h:j) In I~ I

7r 1+ h:j 1+ 8

+h:j In {l; - J(h:j)]. (6.98)

Figure 6.9 presents the lift coefficient Gy of a fiat plate versus the angle
of pitch 8° for different magnitudes of the moment um coefficient q and the
angle of blowing T O = 30° for h = O.l.
Figure 6.10 shows the dependence of the lift coefficient of a jet-fiapped
plate upon the relative ground clearance for T O = 15°, 8° = 2° for different
magnitudes of the jet moment um coefficient Gj • Linearizing the expression
for the lift coefficient Gy of the fiat plate, we obtain the formula

(6.99)

where 'Ye = 0.5772 is Euler's constant. This formula is identical to the ex-
pression derived within assumptions of the linear theory in Kida and Miyai
[50].
158 6. Endplates, Flaps, and Slots

1.2,-------,---,-------,-----,

0.2 ~---'-----'-----'--------'
0.00 0.04 0.08 0.12 0.16
h
Fig. 6.10. The lift coefficient of a flat foil with ajet flap versus the ground clearance
for different magnitudes of the jet momentum coefficient, T O = 15°, BO = 2°.

6.4 Steady Flow Past a Jet-Flapped Wing of Finite Span

If perturbations in the fiow are smalI, we can apply the linear theory to obtain
some useful practical results. Linearization admits separate investigation of
the effects of angle of pitch, curvature, and the jet fiap. Therefore, when
considering the fiow problem for a fiat wing with a jet fiap, it is sufficient to
treat a fiat jet-fiapped plate at zero incidence. Within the approximation of
very small ground clearances, the fiow problem is described by relationships
derived from the general formulas of the theory developed in section 2. The
equation is
[P<pI a2 <pI _ 0
ax 2 + az 2 - , (x,Z) ES. (6.100)
The boundary conditions are

<PI = 0 at the leading and side edges (6.101)

and
a<pI = jCJ.L(Z) = at the trailing edge. (6.102)
ax T
2h
J),'
J

It follows from the more general nonlinear jet description that for small per-
turbations, the equation of the jet becomes

(6.103)

The jet becomes horizontal far from the trailing edge, and its ordinate far
downstream is given by
 6.4 Steady Flow Past a Jet-Flapped Wing of Finite Span 159

y.
Joo
= Yjoo = 1 _
h
K,'
J
= 1- T rc;
V?J;' (6.104)

It is easily seen from the latter expression that the condition of blockage
in the linear case is identical to that obtained from nonlinear theory; see
formula (6.71). Thrning from stretched to natural coordinates Yj = hYj and
xh = VhXh, we can obtain

GY'
GX~ = T exp
(/2)
Vhi'f"; ,
Xj2 (6.105)

YJ ~h -7 Jh;" [1- exP(Xhj h~Jl (6.106)

It follows from expressions (6.105) and (6.106) that within the small pertur-
bation theory, the form of the jet behind the trailing edge is exponential. The
jet becomes horizontal at distances of the order of O( Vh) from the point
of blowing. Note that due to the dependence of T and Cp. on the spanwise
coordinate the magnitude of Yjoo also depends on z. We consider a partic-
ular case of a rectangular wing of aspect ratio >. with a jet flap in the
extreme ground effect. We let the jet have an arbitrary spanwise distribu-
tion of the momentum coefficient Cp.(z) and jet blowing angle T(Z). Note
that we can rather speak about a given spanwise distribution of the quan-
tity K,j(z) = T(Z) JCp.(z)/2h = K,joKj(z). The solution of the lowest order
problem (6.100)-(6.102) can be obtained in the form

7r
K,n = :x(2n + 1). (6.107)

The lift coefficient, the moment coefficient around the trailing edge, and the
coefficients of the induced drag and suction force for a rectangular wing with
a jet flap were found in the form

(6.108)

(6.109)

(6.110)
160 6. Endplates, Flaps, and Slots

1 00 2 2
C - _ "'" anqn
s - 2 20
cosh 2 qn '
(6.111)

where the coeffi(:ients an are determined by the formula

an = ~
2 1>./2 ICj(z)cosqnzdz. (6.112)
Aqn ->'/2

Note that these formulas do not include the reactive vertical and horizontal
components of the force due to the jet momentum. For example, the reactive
component of the lift coefficient can be calculated by using the formula

Cy! = ~
11>./2 CI'(z)r(z) dz. (6.113)
->'/2

When considering the horizontal projection of the balance of forces acting

upon the wing, we have to account for the reactive thrust of the jet. Within
the frame of linear theory, the latter is equal to the coefficient Cj of the
total jet momentum. In the particular case of uniform distribution of the jet
momentum and the jet blowing angle CI' = q = const., r(z) = r = const.,
the formulas (6.108)-(6.111) yield

(6.114)

(6.115)

2C h2
~ "'" tan
00
C = qn (6.116)
Xi ),2 ~ q2 '
n=O n
c _ 4r 2 q ~ 1
(6.117)
s - q2 cosh2q .
),2 ~
n=O n n
For not very large aspect ratios, one can truncate the series to one term, that
is,

(6.118)

(6.119)

(6.120)

(6.121)
 6.4 Steady Flow Past a Jet-Flapped Wing of Finite Span 161

es = 2
4T Gj
7f 2 eosh(7f/>')
(
6.12
2)

An important eonclusion ab out the approximate equality of effeetive aspeet

ratios for a wing with and without a jet flap follows immediately from eom-
parison of expressions (6.121) and (3.70). It means that the influenee of the
relative ground clearance on the effective aspect ratio is the same no matter
how the lift is generated, by a jet flap at zero angle of pitch or by the angle
of pitch without jet flap. Passing to the limit>. --t 00 in (6.114)-(6.117), one
obtains corresponding results for two-dimensional extreme ground effect:

rc;
mz =TV2k' (6.123)

For a wing of small aspect ratio>. --t 0,

Gs =0. (6.124)

Consideration of the expressions (6.114)-(6.117) shows that the aerodynamic

coefficients Gy, m z , Gxjh, and Gs/h depend on the aspect ratio>. and the
parameter
(6.125)

whereas the quantities Gy/K,j,mz/K,j, Gxi/hK,J and Gs/hK,J depend only on

the aspect ratio of the wing. Thus, for a wing of given aspect ratio it is suffi-
cient, onee and for all , to calculate the quantities Gy / K,j, m z / K" Gxj hK,J and
Gs/hK,J. Based on the data ealculated in this way, it is easy to determine the
aerodynamic characteristics, corresponding to different magnitudes of the jet
momentum coefficient q and the relative ground clearance h. The parameter
Kj can be viewed as a similarity parameter that characterizes the aero-
dynamics of the wing with a jet flap in the extreme ground effect. Figure 6.11
illustrates the influenee of the aspect ratio of a jet-flapped reet angular wing
in the extreme ground effect upon the lift coefficient, related to the similar-
ity parameter K,j.4 The ealculation was performed by using formula (6.114)
with ten terms retained in the series, which converges very quickly. In the
same figure there are plotted corresponding results for a wing of semielliptic
planform; see Kida and Miyai [50].
We let the jet momentum coefficient distribution vary as

(6.126)

whieh is identical to the variation of the velo city of blowing along the span
in proportion to cos(7fz/>'). Then, the following formulas hold for the lift and
the moment coefficients:
4 Here, the reactive component of the lift coefficient is substracted from the total
lift coefficient.
162 6. Endplates, Flaps, and Slots

1.2 ,---r-----r----,--,--,------,

1.0 I---+---+----I-~+--.<.;----i

0.0 '--.--'----'------'---'----'----'
o 2 3 4 5 A 6

Fig. 6.11. The influence of the aspect ratio on the lift coefficient of a jet-flapped
flat wing of reet angular and semielliptic planform (solid line: rectangular planform,
see formula (6.114); dashed line: semielliptic planform [50]).

Cy
4>.
= 7r27 Vh
rc; tanh ~ = 4.12>.
7r
--;Z/'i,j
7r
tanh~,

4>.2 7 rc; 7r 7r
mz = ~ Vh tanh ~ tanh 2>' .
Comparison of the lifting properties of a rectangular wing with a uniform
jet momentum distribution and the jet momentum distribution, given by for-
mula (6.125), shows that for the same angle of blowing, ground clearance,
and the total jet momentum coefficients q, nonuniform blowing results in
somewhat larger magnitudes of the lift coefficient than that of uniform blow-
ing.
In section 9 a law of blowing is discussed, which leads to a minimum
induced drag for a given lift coefficient and also the influence of optimization
upon the lift-to-drag ratio of the wing with a jet flap in the ground effect.
When the jet-flapped wing has endplates, the calculation of the lift co-
efficient and other aerodynamic coefficients can be performed taking into
account the results presented in paragraph 6.1. Application of the approach
set forth in paragraph 6.1 leads to the following formula for the relative lift
coefficient of a jet-flapped wing with endplates:
 6.4 Steady Flow Past a Jet-Flapped Wing of Finite Span 163

where the function G( 'Yep) accounts for the configuration of the endplate and
the function Aj , depending upon the aspect ratio, was found in the form

A j (>') = 7r~ [f c~
n=O
tanh(qn>.2 /4)jqn /f tanhqn/q~],
n=O

where

7r
qn = :x(2n + 1),

Leaving one term (l = n = 0) in this series, we can obtain the following

approximate formula for Aj ().):

It is remarkable that the magnitude of Aj from these formulas does not depend
on the jet moment um.
Now some results will be discussed of the nonlinear theory of a jet-
flapped wing of a finite aspect ratio in the extreme ground effect. For
a wing of rectangular planform, the aerodynamic coefficients can be derived
in analytical form. Taking into account the trailing edge condition obtained
earlier,
ß<PI
ßx
= 1 _ y. = T
Joo
rc:: = ",.J'
VV;
the nonlinear problem is formally described by the same set of equations
except for the fact that the pressure distribution should be calculated by
using the following nonlinear differential operator:

(6.127)

Integrating the pressure difference (p) = p- - p+ on the wing surface taking

into account the expression (6.127) for <PI, we can derive the following expres-
sions for the lift coefficient of a rectangular wing with a jet fiap in the case
of moderately large fiow perturbations:

Gy = \" 11 1
A 0
1 A 2
/
-A/2
{p)dxdz

(6.128)

where
164 6. Endplates, Flaps, and Slots

The moment coefficient calculated around the trailing edge. has been found
in the form

(6.129)

For constant K,j = T JG!-,/2h = T Jq/2h, the following expressions were

obtained for the coefficients of the lift, moment, and induced drag:

(6.130)

16K,j ~ tanh qn ( qn 1 )
mz = \2 ~ 4 tanh-
2 - -K,jtanhqn , (6.131)
/\ n=O qn 4

(6.132)

As earlier, retaining just one term in the series (6.130)-(6.132), we obtain the
following simple approximate formulas:

(6.133)

(6.134)

(6.135)

Passage to the limiting cases of wings of large and small aspect ratios results
in the following formulas:

,x -t 00, Gy = 2K,j ( 1 - i), m z = K,j ( 1 - i), GXt = 0; (6.136)

Cy = 16,xK,j (1 _ K,j) = K,j,x2 ( _ K,j)

,x -t 0, 71"3 2 , mz 6 1 4'

GXi = hK,J(l - K,j). (6.137)

For small magnitudes of K,j these formulas yield the corresponding formulas of
linear theory. Of practical interest is an estimate of the maximum magnitudes
of the lift coefficient of the a jet-flapped wing in the extreme ground effect.
This maximum is achieved when the jet touches the ground, Le., in the case
 6.4 Steady Flow Past a Jet-Flapped Wing of Finite Span 165

of blockage. In accordance with the blockage condition /'l,j = 1 and equation

(6.130)
C = ~ ~ tanh qn . (6.138)
Ymax >,2 6 q3
n=O n
The magnitudes of the lift coefficient for other "nonblocking" magnitudes of
parameter /'l,j can be determined by taking into account (6.130) and (6.138)
in the form
(6.139)

It is interesting to compare the magnitudes of C Ymax ' predicted by the non-

linear (6.138) and linear (6.114) theories. Due to the fact that the condition
of blockage is identical for both the linear and nonlinear theories it follows
immediately from (6.114) that for /'l,j = 1,

C lin _ 16 ~ tanhqn
Ymax - >.26 q3 . (6.140)
n=O n

Comparing expressions (6.130) and (6.138), we see that the linear theory
predicts magnitudes of the maximum lift coefficient which are twice as large
as those predicted by the nonlinear theory. It can be readily shown that the
nonlinear description of the blockage phenomenon is closer to reality than
the linear one. For instance, in the case of the wing of infinite aspect ratio,
it follows from equations (6.138) and (6.140) that
cnonlin =1 C Ymax
lin =2.
Ymax '

At the same time, it is clear from the physical viewpoint that for zero or
vanishing incidence the blockage of flow near the trailing edge of a wing of
infinite aspect ratio moving near a wall, results in complete stagnation of the
flow under the wing. In very close proximity to the ground, this situation
corresponds to magnitudes of the lift coefficient close to unity.
To conclude the consideration of the jet-flapped wing in the extreme
ground effect, a comparison is presented in Fig. 6.12 of the results of the
asymptotic theory with the experimental data of V.P. Shadrin, obtained in a
wind tunnel for a rectangular wing with endplates. Note that, when conduct-
ing calculations by using the asymptotic theory, account was taken of the
jet re action force and the presence of the endplates. For large magnitudes
of the blowing angle, the reactive components of the jet in the vertical and
horizontal directions have to be predicted by the formulas

11).,/2
CYj = ~ CJ.L(z) sin r(z) dz,
-),,/2

11).,/2
Ctj = ~ CJ.L(z)cosr(z)dz. (6.141 )
-),,/2
166 6. Endplates, Flaps, and Slots

0.20

Cy-CYo 0.18
I
I i I I
I i
I .1 I
!
0.16 I I
1 ! I
I
I I I
0.14

0.12
1 ~. I I
0.10
1\
L\- I
I
0.08

0.06
~
....... _\....
0.04
I
0.02

0.00
i
0.00 0.05 0.10 0.15 0.20 h 0.25

Fig. 6.12. The additional lift coefficient of a rectangular wing with endplates due
to a jet Rap (7° = 30°, q = 0.01, B = 0, solid line: asymptotic theory, circles:
experiment, Shadrin (1975».

It is worthwhHe to mention that in the experiments conducted by Shadrin

a model of a rectangular wing of aspect ratio>. = 1 with a "Göttingen"
type foH section with an almost flat lower surface was tested. The relative
thickness of the foH was equal to 11%, and the width of the jet injection slot
with respect to the chord was 0.0053. Specially designed changeable trailing
edge elements provided variation of the angle of blowing 7, measured with
respect to the flat lower surface of the wing. The injection of air was provided
by special fans buHt into the model.

6.5 A Wing with a Rotating Flap

It is known that rotor devices in the form of rotating cylinders and rotating
flaps attached to the main foH can be used as one of the possibilities to
enhance and control the aerodynamic efficiency of lifting systems.
In connection with the importance of edge effects for the aerodynamic
behavior of wings operating in the ground effect, rotor devices can be fitted
on the leading, trailing and side edges to form rotating forward and rear flaps
and rotating endplates.
In what follows, we discuss some simplified mathematical models of ro-
tating devices fitted near the edges of a lifting surface in the ground effect.
It is assumed, in particular, that when a rotating device is operating near a
sharp edge, the Kutta-Zhukovsky type condition is satisfied at this edge at
all times.
 6.5 A Wing with a Rotating Flap 167

aamaamuamtmmmmmm
HHffifiMiiMffiMMBfimmmmammm iiii" i

Fig. 6.13. A wing with a rotating flap at the trailing edge in the extreme ground
effect.

In the case of a rotating ftap mounted in the vicinity of the trailing edge,
this assumption implies that a supercirculation arises around the wing, which
should eliminate large velocities induced at the trailing edge. It should be
noted that the lift resulting from this supercirculation is much larger than a
simple Magnus lift acting upon a rotating cylinder or a ftap. The lift-to-drag
ratio is significantly larger that for an isolated rotating ftap due to a large
chord to thickness ratio.
In the case of a rotor endplate, the mechanism of enhancing the lift is
somewhat different. Here the requirement of finite velocities at the side edge
of the wing (or ti ps of conventional endplates used for wing-in-ground-effect
vehicles) slows down the leakage from und er the side tips. Hence, we expect
an increase in lifting capacity.
It should be noted that for both rotating ftaps and rotating endplates, the
resulting lift-to-drag ratio should be estimated accounting for the additional
energy required for the rotation of the devices.
From the viewpoint of applying matched asymptotics techniques to the
problem offlow past a wing-in-ground effect with rotating devices, the scheme
of the solution should incorporate the analysis of the corresponding local ftow
in the vicinity of the wing tip with a rotor (rotating ftap or rotating endplate);
see Fig. 6.13 .
We turn to the derivation of the boundary condition for channel
flow at the trailing edge with a rotating flap. Consider the case when
the main lifting surface of a wing-in-ground-effect vehicle has a rear ftap that
can rotate around an axis located at its midchord. Replace the rotating ftap
by a vortex of circulation r. In a local stretched trailing edge region, this
vortex has a complex coordinate ("1 = ii'Y +iy'Y' where ii = -x; see Fig. 6.14a.
We define the nondimensional circulation r as

(6.142)
168 6. Endplates, Flaps, and Slots

y
(3)
B A
,:J I~' 0
A
0 o
B
C
C B 9,
v
(a) (b) ·r
Fig. 6 .14. (a) The stretehed loeal region near the trailing edge with a rotating
flap, (b) The auxiliary plane in the loeal problem of flow past a trailing edge with
a rotating flap .

where r o is the nondimensional radius of an equivalent rotating cylinder equal

to the semichord of a rotating fiap , w is the circular frequency of rotation,
Co is the chord of the main wing, Uo is the cruise speed of the vehicle, and
k = wCo/Uo is the Strouhal number based on the chord. The local fiow
potential 'P'Y of the relative motion of the fiuid can be sought in the form

The local fiow near the trailing edge without a rotating fiap was considered
in section 2. Therefore, we have to determine only the additional term due to
the rotor, that is, 'P'Y' To solve for 'P'Y' we map the fiow field onto an auxiliary
complex plane 9 = gl +ig2 shown in Fig. 6.14 b. This mapping is realized by
the function
1
( = -(g + lng + 1). (6.143)
7r

We find 'P'Y = ~{F'Y( ()}, where the complex potential F'Y of the vortex in
the g-plane can be written as

(6.144)

where g is the complex conjugate with respect to g. The complex coordinate

g'Y of the vortex in an auxiliary plane can be found from the equation

(6.145)

Assume further that the rotation of the fiap results in supercirculation around
the wing. The magnitude and direction of this supercirculation is governed
by the Kutta-Zhukovsky condition at the trailing edge. With this in mind,
after complementing the right-hand side of (6.144) with a supercirculation
term, we obtain
(6.146)
 6.5 A Wing with a Rotating Flap 169

Aeeepting the previously mentioned meehanism of the operation of a ro-

tating flap, we find the eonstant C"I from the eondition of no flow aeross the
trailing edge of the main wing. Aeeounting for violation of the eonformity of
the mapping at point 9 = -1 (image of the trailing edge), we ean satisfy the
Kutta-Zhukovsky eondition by requiring that for 9 = -1,

(6.147)

wherefrom
C __
1 ( 1 _ 1 ) (6.148)
"1- 2ni -1-g1"1-ig2"1 -1-g1"1+ i g2"1'

or, finally
C g2"1
(6.149)
"I - n [( 1 + gl"1 )2 2 l'
+ g2"1
We consider the outer asymptotics of the flow velo city due to the rotor at
some distance from the trailing edge along the upper and the lower surfaces
of the lifting surfaee. We calculate the complex velo city in the physical plane

dF"I dF"I dg [ g2"1 C"I] ng

(6.150)
d( = dg d( = n(g - g"l)(g - 9"1) +9 (1 + g)"
On the surface of the wing in the vicinity of the trailing edge, we have to
replace 9 with gl, so that
dF"I dF"I g2"1g1 nC"I
-=-- +-- (6.151)
d( dD - [(gI - gl"1)2 + g~"Il(l + gd 1 + gl·

The upper surface asymptotics follows for gl ---* -00:

D ~ gl + 0(1). (6.152)
n
It follows from (6.151), (6.152) that for D ---* -00, y = 1 +0 (gI ---* -00, g2 =
0+ 0), the flow velo city behaves as

dF"I = ocp"l = C"I O(~). (6.153)

d( oD D + D2

We turn to eonsideration of the outer asymptotics of the loeal flow velocity

on the lower surface of the wing inside the ehannel, Le., when gl ---* 0 - o. In
this ease we ean obtain from the expression for the mapping function

(6.154)
170 6. Endplates, Flaps, and Slots

wherefrom
gl ::::i -exp(1I"iI), (6.155)
so that far upstream in the channel, gl vanishes exponentially. Hence

dF1' 8'P1'
d/" = 8- = 2
g21'gl
2 + 11"Cl' = 11"Cl' + O[exp (-)]
1I"V • (6.156)
'> V gl1' + g21'

Accounting for the obvious relationships iI = -x = -x/h,8'P1'/8i1 =

-h8'P1'/8x, we can represent the contribution of 'P1' to the channel flow
boundary condition at the trailing edge as

(6.157)

The corresponding boundary condition for the velo city of the relative motion
of the fluid accounting for the structure of the local flow potential 'Pe can be
written as folIows:

(6.158)

where

(6.159)

In the previous expression, 1'0 = rolh is an equivalent radius of a rotating

flap (identical to a half-chord of the flap) related to the ground clearance and
k ro = wCorolUo is the ratio ofthe flap edge linear velocity to the characteristic
flow velo city Uo. The parameter k ro can be interpreted as a Strouhal number
based on the radius of an equivalent rotating cylinder. Formula (6.158) should
be used instead of formula (4.67) of section 4 as a channel flow boundary
value of the spanwise-averaged velo city at a point x = 0, when solving flow
equation (4.66) for a wing with a rotating flap. Observing that the structure
of the flow problem (hence the structure of its solution) for the rotating flap
is identical to that of a traditional rigid flap, we can interpret the quan-
tity 8r ,
f g21' (6.160)
l' - (1 + gl1')2 + g~1"

as a relative effective gap behind the rotating flap. For example, for a wing
with zero pitch angle, a rotating flap at the trailing edge, and nonzero clear-
ances under the endplates, we can determine the lift and moment and induced
drag coefficients by formulas (4.71), (4.72), and (4.74), replacing 8f with 8r
given by expression (6.160). It is easy to see from (6.160) that to diminish
the effective gap 8r , that is, to increase the lift, the flap should be rotated in
the clock-wise direction (f < 0).
 6.6 Slotted Wings in the Ground Effect 171

6.6 Slotted Wings in the Ground Effect

When the lifting surface is slotted, the same formalism can be applied to
account for the influence of gaps upon its aerodynamics in the ground effect.
As indicated in Durand (1934), the problem of the slotted wing is of practical
importance for several reasons. In the case of lateral slots with a proper choice
of the adjusted angles of attack of portions of the wing, separated by the slots,
a better lifting capacity can be achieved. On the other hand, introduction of
longitudinal slots especially near the root section of the lifting surface, can
result in an increase in induced drag.
It should be mentioned that numerical solutions of the flow problem for a
wing with narrow slots can bring about computational difficulties and even
divergence of the computational processes. In any case the number of control
points should be considerably increased in the immediate vicinity of the gap
to avoid numerical errors.
An exact solution for a two-dimensional foil with one small gap in an
unbounded fluid was developed by Chaplygin. Bleviss and Strubble [145] ob-
tained some exact solutions for longitudinal gaps on a wing of small aspect
ratio in unbounded flow. Application of the method of matched asymptotic
expansions opens certain possibilities for a simplified analysis of the flow
problems for slotted wings both in and out of the ground effect. White and
Landahl [146] handled corresponding 2-D problems by the matched asymp-
totics technique. In what follows, we use the asymptotic theory set forth in
previous sections of this book to evaluate the influence of a lateral gap of
small width 8g = 8g h, 8g = 0(1), located at a distance 19 = 0(1) from the
trailing edge, upon the aerodynamic characteristics of a thin rectangular wing
of aspect ratio .A, moving at very small distances from the ground h ~ 1; see
Rozhdestvensky [45, 46].
As discussed previously, to account for the influence of the gap, we have
to analyze the local flow near it, find the corresponding asymptotic solution,
and subsequently, match it to the main flow. This procedure can be shown
to be applicable in nonlinear case, but here analysis will be limited to a
the linearized case. In compliance with the linear theory it will be assumed
that the vortex sheet emanating from the leading edge of the lateral gap
stays in the same horizontal plane. Sufficiently far from the intersection of
the slot with the side edges of the wing, we will consider the local flow as
a two-dimensional flow in the planes z = const. which are normal to the
gap's axis. Near the gap, the local stretched coordinates Yg = y/h,x g =
xg/h = (x -l)/h, Zg = Z are introduced. In the plane of the complex variable
Z = xg + iYg, the complex velo city potential Fg and the complex conjugate
velo city
dFg •
wg = dZ = ug - lVg
of the flow near the gap are analytic functions, satisfying the following bound-
ary conditions:
172 6. Endplates, Flaps , and Slots

• on the ground U/g = 0 + 0),

a) ~Fg = 1/Jg = 0, b) ~Wg = -vg = 0; (6.161)

• On the wing behind the gap (ijg = 1 ± 0, xg < -8g /2),

a) ~Fg = 1/Jg = 1, b) ~Wg = -vg = hO:s(lg + 0, z, t); (6.162)

• On the wing in front of the gap (yg = 1 ± 0, xg > 8g/2),

(6.163)

Here I'g is an unknown constant associated with the rate of the fluid passing
through the gap;
ays ays
O:s = ax - ßt
is a prescribed downwash near the leading and trailing edges of the gap. The
superseripts plus and minus correspond to points On the upper and lower
surfaces of the wing, respectively. According to the scheme already discussed
above, the gap flow solution will be decomposed into (a) homogeneous and
(b) nonhomogeneous solutions. To proceed with the solution of the local
problem, formulated above, we map the flow region onto an upper half plane
of the auxiliary complex plane. The point-to-point correspondence is shown
in Fig. 6.15.
The mapping function was obtained by application of the Christoffel-
Schwartz integral

(6.164)

Z=X +iy
__-------1~--~~ g

A' 0'
- 69 6g 1-
A" 0 0" )(g
-

~,,'"
0"0' ~ A'A"
_Hili lIIIiIIlilliliif ~
-1 -p 0 P 1

Fig. 6.15. The correspondence of points of the flow field near the slot and auxiliary
half plane.
 6.6 Slotted Wings in the Ground Effect 173

where parameter ß is related to the width of the gap by the relationship

-
8g =
8g
h =
4 [2
71"(1- ß2) ß
1
+ 2(1 -
2 1+ ß]
ß ) In 1 - ß . (6.165)

Note that for sufficiently small 5g < 0.5, we can approximately set ß = 7I"8g /8.
The solution of problem (a) was found in the form

1( (-1 ( )
Fag = -; In - ( - + 'Yg In (+ 1 . (6.166)

It is easy to verify that this solution does not violate the fiow tangency
condition at points on the wing. The velo city potential on the surface of the
wing is calculated by taking the real part of the complex potential and setting
(=~. Then,

4'ag
1(
= -; In -~-
1-~
+ 'Yg In 1 +~ ~ ). (6.167)

The solution of problem (b) was found with help of the Schwartz integral for
a half plane, which enables us to construct an analytic function by using its
imaginary part, given on the real axis; see Fuks and Shabat [131]:

Wbg = Ubg - . = -;11

lVbg
00

-00
8'W bg
(1 _ ( d(l +C

= -;h[as(l- 0) In -
(-1 ( ]
( - + as(l + 0) In ( + 1 + C. (6.168)

Omitting the real constant C and separating the real part of expression
(6.168), for points on the surface of the wing near the gap,

h[as(lg - 1-( + as(lg + 0) In -(- ] ,

Jor
g
Ubg = - 0) In -,.- 'Pbg = Ubg dx g.
71" ., 1-(
(6.169)
The general solution for the perturbed velo city potential in the vicinity of
the gap, that is, in region 'Dg , can be written in the form

(6.170)

where parameters li must be determined by matching procedures. Parameters

l3 and l4 take different values above (yg = 1 + 0) and below (yg = 1 - 0)
the gap, that is lt -=!= l3 and lt -=!= l4' The real constant 'Yg was found by
accounting for the requirement that the perturbation velo city should be finite
at the leading edge of the gap (point C in Fig. 6.15 ) as

1-ß (6.171)
'Yg = 1 + ß'
It can be readily demonstrated that for such a choice of 'Yg, the perturbation
velocity at the trailing edge of the gap (point B in Fig. 6.15) has a square
174 6. Endplates, Flaps, and Slots

root singularity, Le., behaves as 8- 1/ 2 (where 8 is the distance of the point

from the trailing edge of the gap in the downstream direction). Asymptotic
representations of the gap velo city potential (6.167) far from the gap, which
are needed for matching with the main flow solution, can be written as follows:
• far from the gap on the upper surface of the wing (y = 1 + 0),

2ßh 1 7T'(1 - ß2)xg

cpg ~ 7T'(1 + ß) n 2ß2

(a s ) [7T'(1 - ß2)xg ]
+-:;;:-hxg In 2ß2 - 1 + l3 + xgh + l4, (6.172)

(a s) = as(lg - 0) - as(lg + 0) .
• Far from the gap upon the lower surface of the wing (Yg = 1 - 0),

Xg -t -00, Yg = 1 - O. (6.174)
The velo city potentials of the upper flow, channel flow, and leading and
trailing edge flows can be constructed in a manner explained in detail in
previous sections. For briefness we discuss only the changes to be introduced
into the procedure for solution when the wing has a lateral gap.
The upper flow potential in region 'Du is represented by a straightforward
distribution of sources (sinks) over the projection of the wing plus the wake
onto the ground with the addition of the admissible contour distribution of
the sources (sinks) along the edges of the wing (wake) and along the axis of
the gap. Expression (2.31) of section 2 for the upper flow potential should be
supplemented by the term

(6.175)

which represents the potential of a linear distribution of the sources (sinks)

with a strength Qg and models (in the upper flow region) a leakage ofthe fluid
through the gap from beneath the wing. The strength Qg of the singularity
 6.6 Slotted Wings in the Ground Effect 175

distribution has to be determined by matching with the gap fl.ow poten-

tial cpg.
Taking into account expression (6.175), the asymptotic representation of
the upper fl.ow potential near the gap is found in the form

CPu I
x-tlg±O
- Qg(z, t) 1n x g
rv

211'
+ (a )s x g In11' x g + L11'l x g + L11'2 + O( x g2) • (6 . 176)

Here, parameters L l and L 2 characterize the infl.uence of distant sources

(sinks). Corresponding expressions for L l ,2 are not presented herein for brief-
ness.
The channel fl.ow potential satisfies the same Poisson equation as that
for the wing without slots. However, this equation, strictly speaking, has to
be solved separately in two portions of the wing planform divided by the
gap axis. The boundary conditions at the leading (side) and trailing edges of
the wing are given by the same formulas as previously discussed. The only
difference is that the constants A2 , B il and B 2 comprise contributions from
the additional term (6.175) in the expression of the upper potential.
A special feature of the channel fl.ow for a slotted wing is that both values
of the potential CPI and the corresponding pressures are, generally speaking,
different when one approaches the line x = 19, Izl ~ >'/2 from the left and
from the right. Physically, this is connected with mass and vorticity transport
through the gap. Rewriting expressions (6.174) and (6.176) in terms ofthe gap
variable x g = hX g enables us to find asymptotic structure of the parameter
h when x g tends to 0 + 0 and 0 - 0:

(6.177)

as weH as the behavior of the channel flow potential and press ure in the
vicinity of the line 19 ± 0:

(6.178)

- ß)
( 11 + 4ß { 8l;
Ph (l + 0) = ß Pl 1 (lg - 0) + 11'(1 + ß) l3 - 8t

2ß 81 11 ß [
+ 1 + ß(1-ln2)"ßt + (as ) In2 - 1- ß as(lg + 0)

+as(lg-O)G~~)]}, Ph =2(8:~1 _8;;1) (6.179)

and a parameter
(6.180)
176 6. Endplates, Flaps, and Slots

Matching the upper flow <Pu and the gap flow potentials is identical to com-
paring of the expressions (6.172) and (6.176) in the same variables and leads
to the following results:

Q = 4ßl l1
(6.181)
g 1+ß'

(6.182)

Finally, applying the Kutta-Zhukovsky condition at the gap trailing edge

(point C) in the form of the pressure continuity across the gap, we can find

(6.183)

Taking into account (6.181)-(6.183), the relationship between Ph (lg + 0) and

PI! (lg - 0) becomes completely defined. The solution in region is constructed
similarly to that of a wing without the gap. To the leading order (the extreme
ground effect), the problem acquires the sufficiently simple form

<P+ = 0(1), (6.184)

where <Ph should satisfy the following equation and boundary conditions

A
Izl:S 2";
(6.185)
A A
<Ph = 0, x = 1, Izl ::; 2 and 1 2:: x 2:: 0, z= ±2; (6.186)

8<Ph _ 8<pl! =
8x 8t '
° A
x = 0, Izl::; 2' (6.187)

with the following boundary relationships at the gap centerline:

<Ph (lg + 0) = <PI! (lg - 0), (6.188)

- ß)
Ph (lg + 0) = Ph (lg - 0) ( 11 + ß ' PI! = 2 (8<Ph 8<Ph)
8x - fit . (6.189)

Now, consider some examples .

• For a slightly curved thin foH (infinite aspect ratio),

Ys = h + ef(x), 1 ::; x ::; 0,

we can write the lift and moment (around trailing edge of the wing) coef-
ficients in the following way:
 6.6 Slotted Wings in the Ground EfIect 177

(6.190)

Cl = 21 f(~) d~
1
- 4~:(~) (1 -lg),
C = ~ [1 ~ ßlg/'(O) -
2 f(lg)] + C~ ~)f'(O) + 1'(1)],

C3 = ~{B2 - 1I"(lg - 1) ß:~3 (lg + 0) - A2 + (1 + In 11") [1(1) - 1 ~ ßf(lg)]

4ß2f(lg) [ '( ) ]}
+ (1 + ß)(l _ ß2) + 19 f 0 In 11" - BI ,

mz = X[mI + m2 h1n ~ + hm3 + O(h2 )], (6.191)

ml = 21 U(~) d~
1
- 2~:(~) (1 -l~),
m2 = ~{1 ~ ß[lg/'(O) - 2f(lg)] + C~ ~)f'(O) + 2f(1)},

m3 = (1 -l~) d:~3 (lg + 0) + ~{l~[J'(O) In 11" - B 1] - 2A2 + Qlln 11"

+Qg[lg In 19 + (1 -lg) In(l -lg) - 1]

(, [ ( () 1 8ß2 f(lg)lg }
-2 Ja f(~Hln~+ l-~)ln 1-~ -1 d~- (1+ß)(1-ß2) ,
where
Qg = 1 ~ ßf(lg), Ql = 2 [f(l) - 1 ~ ßf(lg)],

- ß) ,
dcp l 3
dx (lg + 0) = 1 2ß
+ ßl3 + ;:1 { (11 + ß [f (0) In 11" - BI]

2ß(f')ln2_~[(1-ß)f'(l -0) f'(l O)]}

+ 1+ß 1 - ß2 1 + ß g + g+ ,
where (f') = f'(lg + 0) - f'(lg - 0).
Parameters l3, BI, and A 2 are calculated by the formulas

l = .!.{l - (f')
3 11" 1
[1n 11"(12ß- hß2) -I]} '
2

BI = Ql + Qg - 1'(0) - ([f'(~) - f'(O)]d~,

2 2l g Ja ~

A2 = ~g In(l -lg) -11 f'(~) In(l-~) d~,

178 6. Endplates, Flaps, and Slots

LI = f' (lg + 0)[1 + In(l - 19)] - f'(lg - 0)(1 + In 19) - 2(1~ 19)

-11 [f'(~) - f'(lg - O)]lg ~~ -11 [f'(~) - f'(lg + O)]lg ~(

To the leading order (the extreme ground effect) ,

Cy = ~ [11 f(~) d~ 2ßf(~g~~

_ -lg)], (6.192)

mz = 2: [11 U(~) d~ ~f~l~ -l~)].

- (1 (6.193)

The relative infiuence of the gap upon the aerodynamics of the wing can be
conveniently described by ratios of the lift and moment coefficients for a wing
with a gap to those for a wing without a gap, that is, for Jg = O. These ratios
are
/'i,g = Cy/CyO and /'i,m = mz/m zo '
To the lowest order, that is, with an asymptotic error of O(hln(l/h), the
magnitudes of /'i,g and /'i,m are not dependent upon the relative clearance of
the foil hand have the following form:

_ 1 2ßf(lg)(1 -lg)
(6.194)
/'i,g - - (1 + ß) Jo1 f(~) d(

ßf(lg)(l -li)
(6.195)
/'i,m = 1 - (1 + ß) Jo1 U(~) d(
In the particular case of a fiat plate at an angle of pitch e, f(x) = X,c = e.
It follows from the general results that

hCy _ hCB _ _ 4ßlg(1 - 19) 4h I ~

e - y - 1 1+ß + 7T'(1 + ß) nh

2h { 1 + -2ß- [(1 -
+- 1--
2l ) In - 19 - 2]
7T' 1+ ß g 19
4ß2 ß
+ (1 + ß)2 [lg(lg -1) - 1 - ß]}' (6.196)

(6.197)

3 + ß[2(1 -li) - 1]
m == ~ - 2ßlg (1 _ l2)
1 3 l+ß g' m2 = 1 +ß '
3+ ß 2ß [ 1 - 19
m3 = -1 { 1 + - - l n 7T' + - - (1 - 3l ) In - - - 2 - 2l
2
7T' 1+ß 1+ß g 19 g
 6.6 Slotted Wings in the Ground Effect 179

0.8 ~~:,--=-+--",::::..--+---t---+--j

0.6 L-_...L...-_..-L._-':"'_-'---"---::-'
0.00 0.02 0.04 0.06 0.08 0.10 h 0.12

Fig. 6.16. The lift coefficient of a fiat plate of infinite aspect ratio versus the
ground clearance for different magnitudes of the relative width of the slot Og/h,
19 = 0.5.

2 2ß(1 + 19) 2] 4ß2(1 + Li) }

-2lg In 7r + 19 In 7r + 1 + ß [lg + (1 -lg) J - (1 + ß)(1 _ ß2) .
Some results, calculated by using formulas (6.196) when the gap is located
at the midchord of the plate (lg = 0.5) for different hand 8g are given in
Fig.6.16.
As seen from this figure, within the range of clearances under considera-
tion, the presence of the gap brings along a certain reduction of the lift and
moment coefficients. 5
Analysis of calculations of the factors "'g and "'m for different 0 :::; h :::;
0.12 shows that these quantities are only weakly dependent on the relative
clearance. The latter circumstance provides a sound basis for using much
simpler formulas, using leading order contributions (6.192) and (6.193). In
particular cases, these expressions take the following forms:
• Hat plate:
3ßlg(1 - l~)
/'i,m = 1- 1+ß . (6.198)

• parabolic arc, f(x) = 4x(1 - x):

12ßlg(1 -lg)(1 -li)

/'i,m = 1- 1+ ß . (6.199)

5 These coefficients are based on the length of the chord without a gap.
180 6. Endplates, Flaps, and Slots

• Rat plate with a Rap J(x) = x for 0 :::; x :::; bf , J(x) = 0 for bf :::; x :::; 1,
where bf is the ehord of the fiap. When the gap is loeated near the hinge
of the Rap, 19 = bf ,

4ß(1 - br) 6ß(1 - bf)

= 1 - (1 + ß)(2 - br) , (6.200)
K: g K: m = 1 - (1 + ß)(3 - bf)"
It follows from the expressions for K: g and K: m that for a foH with a given
eamberline J(x), a position of the gap exists for whieh the deerease of the
aerodynamic eoefficients due to the gap is most pronouneed. For example,
for a Rat plate, the minimum of K: g is reached at 19 = 0.5, whereas for an are
foH, this minimum is aehieved at l = 1/3. Figures 6.17-6.19 illustrate some
results for K:g . It ean be seen from analysis of the figures that the inRuenee
of the gap for a given width 8g and loeation 19 depends on the shape of the
eamber line. For example, at 19 = 0.5 and 8g = 0.3 for a Rat plate, K: g = 0.9,
whereas for a plate with a Rap and 19 = bf = 0.5, K: g = 0.85 .
• Rat plate of finite aspeet ratio A and reet angular planform:
2ßR l
K: m = 1- (1 + ß)Ro ' (6.201)

L gn(lgß) tanh(qnlg),
00
C = ~ tanhqn tanh(qn/ 2)
Cl = o ~ 4 '
(6.202)
n=O n=O qn
where

1.5
I
0.6 '--_---'-_ _.....ll_ _-'--_ _.l.-_----'
0.0 0.2 0.4 0.6 0.8 1.0
Ig
Fig. 6.17. The relative influence of the width and the position of the slot in a flat
plate upon the lift coefficient.
 6.6 Slotted Wings in the Ground Effeet 181

C 1.0 ,- ,
Yg ! 1\=0.1 !
K=
9
Cy 0.9 i
1-\\\\,-".---l---4---f-,.L,---/-'-H-,~----I

I
0.8 -+\-\~+---+--"LL-H+-+---
f-'

0.6 f----t-4---+--+-f1'-----t------I

0.5:-
1
----''i-------I---I--+-----+---J
11

0.4
I
L i_ _ -'---_---.L_ _-'--_ _-'---_--'
0.0 0.2 0.4 0.6 0.8 Ig 1.0

Fig. 6.18. The relative inRuenee of the width and the position of the slot in a
parabolie are foil upon the lift eoeffieient.

1.2~,--~-----,----r----,

K = CYg
9 C
I 1-
!09=0.1
Ii
y 1.0 LI ==r==t==t===:1~~

0.6 IJ
I i
. I
0.4:-,---"t'--~---+---+--i

0.2 1L _ _-1..-_ _'---_--1--_ _-'--_ _I

0.0 0.2 0.4 0.6 0.8 1.0
Ig
Fig. 6.19. The relative inRuenee of the width and the position of the slot in a Rat
plate with a Rap upon the lift eoeffieient.

Ro = f
n=Ü
~ (tanh qn
qn qn
- 1) ,
tanh[qn(1 - 19)]
~=
q~ coshqn[l + (1 - ß) tanh(qn1g) tanh(l -lg)j(l + ß)] ,
7r
qn = ::\(2n + 1).
182 6. Endplates, Flaps, and Slots

Some calculated results using formulas (6.201) and (6.202) are given in
Table 6.1 for 19 = 0.5 and 8g = 0.3.
These data demonstrate that the influence of the lateral gap upon the
aerodynamic coefficients diminishes quite noticeably with a decrease in the
aspect ratio.
Table 6.1 The influence of a lateral slot upon the lift and moment
of a rectangular wing in the extreme ground effect.

>. 1 2 3 4 5 00
/'i,g 0.991 0.959 0.938 0.926 0.920 0.901
/'i,m 0.985 0.949 0.927 0.915 0.908 0.888
7. The Aerodynamics of a Lifting System
Near Curved Ground

7.1 The Influence of Waves on the Aerodynamics

of a Lifting Surface

A large seagoing ground-effect vehicle has to combine sufficient seaworthiness

with acceptable magnitude of the lift-to-drag ratio, when fiying above rough
seas. The approach of matched asymptotic expansions furnishes a simplified
mathematical model of the unsteady aerodynamics of the wing-in-ground-
effect vehicle, based on the idea of domination of the channel fiow in the
extreme ground effect. This concept seems to be promising for wave per-
turbations, because the infiuence of waves upon the aerodynamics of such
a vehicle is in fact predominantly due to the corresponding variation of the
geometry of the gap between the wing and the sea surface.
Following Rozhdestvensky [56, 57], we consider the schematized problem
of a thin fiat rectangular wing of aspect ratio .x, advancing with constant
speed Uo and an angle of pitch () above a still, solid, wavy boundary.l
The wing and the coordinate system are presented in Fig. 7.1, where a w
is the amplitude of the wave, Lw is the length of the wave, h is the clearance,
rneasured frorn the trailing edge of the root section to the unperturbed level
of ground surface, and ß is the course angle measured in the horizontal plane
y = 0 as the least angle between the direction of the wing's motion and
anormal to the front of a sinusoidal cylindrical wave. 2 As previously, one
assumes that at any time the gap between the wing and the wavy ground is
small compared to the chord. With these assumptions in mind, we can write
the gap distribution between the wing and the underlying surface as

h*(x, t) = h + (}x - aw cos[k(x cos ß + z sinß + t cos ß)]

= hh*(x, t) = 1 + Bx - ä w cos(K-X + pz + K-t), (7.1)

1 i) It is common to assume that due to the very small ratio of densities of air
and water, the surface under a lifting system in the ground effect behaves as if it
were solid. ii) The assumption of stillness of the wavy boundary implies that the
speeds of propagation of sea waves are small compared to the cruising speed of
a wing-in-ground-effect vehicle. It is not difficult to introduce the effects of wave
speed into the model.
2 As earlier, all lengths are related to the root chord of the wing.

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
184 7. Aerodynamics of a Lifting System Near Curved Ground

I---Lw--l
Fig. 7.1. A lifting surface moving in proximity to the wavy ground.

where iJ = (J/h, ä w = aw/h, K, = kcosß, p = ksinß, and k = 27r/L w is the

Strouhal number based on the ratio Lw of the wavelength to the root chord
of the wing.
Using the lowest order model of the flow, we have to solve the problem
described by relationships (2.114)-(2.117) with the gap distribution input
given by (7.1). One ofthe approaches to the solution ofthis nonlinear problem
for the motion of a wing over a wavy wall may be based on a consistent
expansion of the velocity potential in terms of the two small parameters () / h
and aw/h; see Barrows and Widnall [136] and Rozhdestvensky [42]. In li ne
with this approach, we seek the solution of the problem in the form of aseries
and leave only terms that are linear in the angle of pitch
(7.2)
Function <Po characterizes the linear (in angle of attack) part of the potential
for steady motion above an unperturbed underlying surface. As shown in
3.4, the expression for <Po can be readily derived for the uniform spanwise
distribution of the angle of pitch in the form

4 ~ (_l)n (coshqnx ) 7r
<Po = \' ~ - - 3 - h -1 cosqn z , qn = );'(2n + 1). (7.3)
/\ n=O qn cos qn

Writing h*(x, z, t) for convenience in a complex form,

h* (x, z, t) = 1 + iJx - ä w exp[i(K,x + pz + K,t)], i = yCI, (7.4)
and substituting (7.2) in (2.114)-(2.117), we obtain the following problem for
the "wave-induced" component <POä w :
 7.1 Infiuence of Waves on Aerodynamics of a Lifting Surface 185

+ :Z [exP(iIl:X + pz + II:t) aJzÖ]; (7.5)

<PÖäw (1, z, t) = 0, ZE(-~,~); (7.6)

a<p-8ä w a<p-8ä w = 0
ax
_

at '
x=O, ZE(-~,~), (7.7)
where
a<PÖ ~ 4( _l)n sinh(qnx) ( )
~ = ~ \ 2 h() cos qn z ,
uX n=O I\qn cos qn

a<pö __ ~ 4( _l)n [COsh(qnX) _ ] . ( )

aZ - n=O
~ A2
qn COS
h()
qn
1 sm qn z . (7.8)

Writing the wave-induced contribution <PÖäw = cj;öäwexp(ill:t) and differenti-

ating, we obtain the following problem for the complex amplitude cj;Öä w of
the potential:

IpZ az ;
.
+lpexp (') a<PÖ] (7.9)

ZE(-~~)'
2' 2 '
(7.10)

x=O, ZE(-~,~). (7.11)

The procedure for the solution for the arbitrary course angle ß is some-
what cumbersome and is presented in more detail in the Appendix. Here
the description of the solution is restricted to a simpler case when the wing
moves normally to the wave front. In this case the equation for the complex
amplitude of the wave-induced contribution of the order of ()a w /h 2 can be
derived from (7.9) in the form
a 2cj;-_ a 2cj;-_ a<p8
a:2aw + a:2aw = (1 + ik ax ) exp(ikx). (7.12)

We write cj;Öä w in the following form, satisfying the requirement of zero loading
at the side edges of the wing:

2: Xn(x) cosqn
00

cj;ÖaJx, z) = Z' (7.13)

n=O
Substituting (7.13) in (7.12) yields

X'''-
n qn2X'n =4(-1)n(1_iksinhqnX)
\ h exp ('k)
1 x . (7.14)
I\qn qn cos qn
186 7. Aerodynamics of a Lifting System Near Curved Ground

We rewrite the second term at the right-hand side of (7.14) in the alternative
form
4( -l)nik sinhqn x ('k) _ 2( -l)nik { [( 'k)]
, 2 h exp 1 x - ' 2 h exp qn + 1 X
/\qn cos qn /\qn COS qn

- exp[( -qn + ik)x]}. (7.15)

We seek a particular solution, corresponding to the right-hand side of the
equation, in the form

Xnpart (x) = A 1n exp(ikx) + A 2n exp[(qn + ik)x] + A3n exp[( -qn + ik)x].

Parameters A 1n , A 2n , and A 3n were found in the form

A 2(-1)n
3n = Aq; coshqn(2qn - ik)'
The general solution for Xn(x) is

Xn(x) = an exp(qnx) + bn exp( -qnx) + A 1n exp(ikx)

+A2n exp[(qn + ik)x] + A 3n exp[( -qn + ik)x]. (7.16)
Applying the boundary conditions (7.10) and (7.11) with K = k, we obtain
the requirements

(7.17)

wherefrom

Xnpart (l)(qn + ik) + exp( -qn)[X~part (0) - ikXnpart (0)]

(7.18)
exp(qn)(qn + ik) + exp( -qn)(qn - ik)

b _ exp(qn)[X~part (0) - ikXnpart (0)]- X npar.(l)(qn - ik)

(7.19)
n- exp(qn)(qn + ik) + exp( -qn)(qn - ik) ,
where

Xnpart (1) = A 1n exp(ik) + A2n exp(qn + ik) + A 3n exp( -qn + ik),

Xnpart (0) = A1n + A2n
+ A3n ,
X~part (0) = ikA1n + (qn + ik)A2n + (-qn + ik)A3n .
The complex amplitude of the wave-induced lift coefficient can be determined
by the formula
 7.1 Infiuence of Waves on Aerodynamics of a Lifting Surface 187

A 2Ba
CYw = h2 ;'
A
1>../2
->"/2 0
1(8cp-
1
86aw
X
- ikCPija w) dx dz

(7.20)

or, finally,

exp( -q ) - 1]
+bn [1 - ik n + A 1n + A 2n + A 3n
qn
'k[A exp(ik)-l A exp(qn+ ik )-l
+1 In
z'k + 2n
qn + 1'k
A ex p (-qn+ ik )-l]} (7.21)
+ 3n
-qn +'k
1
.
This formula for Gyw = Gywo corresponds to the case when at time t = 0,
the crest of the wave passes under the trailing edge of the wing; see equation
(7.1). To pass over to any other reference point with the abscissa a (for,
example the center of gravity of the vehicle), we have to use the formula

(7.22)

where Gywo corresponds to the case when the reference point coincides with
the trailing edge, Le., a = O. Eventually, the lift coefficent for a wing in
motion over waves can be represented by taking into account the expansion
(7.2) in the form
(7.23)
where CYO = (jct is a lift coefficient in cruise and CYw(t) is an additional
time-dependent lift coefficient of the order of O( (ja w ) due to the presence of
waves on the underlying surface. As follows from the preceding analysis, the
latter can be represented as the real part of the expression

(7.24)

where Gw is a complex quantity of the order of 0(1). The wave-induced lift

coefficient, expressed by (7.24), can be written alternatively in a amplitude-
phase form
188 7. Aerodynamics of a Lifting System Near Curved Ground

Oa w ' Oa w
CYw(t) = h2"lR[Cw exp (kt)] = h2"Feos(kt - tJt), (7.25)

where F = F(k,)..) = IGwl and tJt = tJt(k,).., a) = arg Gw is the phase angle,
which eharaeterizes the shift in time between the moment when the erest of
the wave passes under the referenee point with the abscissa a and the moment
of action of the maximum wave-indueed lift.
For an arbitrary aspeet ratio and very long waves k -+ 0, the general
expressions derived above yield the following result:

F = F. = 16 ~ tanhqn tanh(qn/ 2 ) tJt = o.

o )..2 L..t 4 '
n=O qn

Taking into aeeount the preeeding equation and reealling formula (3.65) for
the lift eoefficient of a fl.at reet angular wing in steady motion near a ground
plane, we ean rewrite expression (7.23) for the total lift eoefficient as

Cy (t) = CyO + CyO -h Gw = CyO [1 a

aw -,- + -hw F.F eos(kt - tJt) ] , (7.26)
ICwl 0

where CyO = OFo/h is the lift eoefficient in eruise. Based on (7.26), we ean
define the relative wave-indueed lift, Le., the ratio of the wave-indueed lift to
the lift in eruise in the form

CYw (t) = a w !.- eos(kt _ tJt). (7.27)

C yO h Fo

Henee, the maximum of the relative wave-indueed lift is given by

(7.28)

Expressions (7.27) and (7.28) provide the basis for some simple praetieal
eonclusions. In particular, within the approximation considered here
the magnitudes of both the instantaneous and the maximum wave-
induced lift coefficents are direct1y proportional to the wave am-
plitude (wave height) and the cruise lift coefficient and inversely
proportional to the relative ground clearance.
We eonsider some limiting eases for a situation when the referenee point
is loeated at the leading edge (a = 1). Moving from general results to the
limit of a wing of an infinite aspect ratio ().. -+ 00), we obtain the
known result of Barrows and Widnall [136]

6 = Oa w [i(k 2 - 2) + 4k] exp( -ik) + i(2 + P) + k(k 2 - 2)

Yw h2 k2(k - i) ).. = 00.
(7.29)
 7.1 Influence of Waves on Aerodynamics of a Lifting Surface 189

For the case of a wing of small aspect ratio, we obtain the following
formulas for the amplitude and the phase angle:
16 00 1
F rv ..\2 ~ q;(P + q;)
= ~ ~(~ _ 1 ) = 2..\2 [ _ 2tanh(k>../2)]
..\2k 2 f='o q; P + q; k~ 1 k)., ,
(7.30)

(7.31 )

where k)., = k..\. In deriving formula (7.30), it was taken into account that
(see Gradshtein and Ryzhik [147])

1 ..\ ..\k
L P + q2 = 4k tanh 2'
00

n=ü n

For waves, which are long in comparison with the span of the wing, i.e., for
k)., -+ 0, it follows from (7.30) and (7.31) that

(7.32)

The maximum relative wave-induced lift for wings of a small aspect ratio can
be determined by the formula

CYw = aw F = 12a; [1 _ 2tanh(k)"/2)].

C yO h Fo hk)., k).,
For k).. -+ 00 (waves are short in comparison with span of the wing),

Figures 7.2 and 7.3 present the amplitude-phase characteristics of the wave-
induced lift coefficient CYw (t) versus the relative wavelengths (Strouhal num-
bers) for different aspect ratios of a rectangular Bat wing. The dashed lines
correspond to calculations performed by using the formulas (7.30) and (7.31)
for wings of a small aspect ratio. It follows from Fig. 7.2 that at a certain
relative wavelength for a wing of a given aspect ratio the amplitude of the
wave induced unsteady lift is minimal. With a decrease in the aspect ratio,
this minimum shifts toward the range of shorter waves. For example, for a
wing of an infinite aspect ratio, the minimum amplitude of the unsteady lift
is reached at k ~ 2.4, Lw ~ 2.6.
For a rectangular wing of aspect ratio ..\ = 2, this minimum is reached
at k ~ 3.1, Lw ~ 2,3 and for a square wing ..\ = 1 the minimum occurs at
3 For Lw = 2, the wave is two times longer than chord of the wing.
190 7. Aerodynamics of a Lifting System Near Curved Ground

.E
Fo
«:: 1.0 ~~=+-I--+-+-r-t--:::i=:::!
:::i
"8
~

~
j
'Ci
~
:ec.
ci 0.2 1----.4--+---I--+----+--+-.-....--!---1

0.0 L---L_...l...-_L---L_....I.-_'----'-_-'
o 234 5 6 7 8
Strouhal Number k= h
Lw
Fig. 7.2. The amplitude of the wave-induced unsteady lift coefficient of a rectan-
gular wing versus the Strouhal number for different magnitudes of the aspect ratio
(dashed lines: small aspect ratio approximation (7.30)).

0.8
'I'
0.6
~
'"0
0.4
~::J
'"0
.E
g!

-
~
0
CD
g, -0.2
c
ce
CD
UI
-0.4
IV
.&:
Q.
-0.6

-0.8
0 2 3 4 5 6 7 8
Strouhal Number
k=~
Lw
Fig. 7.3. The phase of the wave-induced unsteady lift coefficient of a rectangu-
lar wing versus the Strouhal number for different magnitudes of the aspect ratio
(dashed lines: small-aspect-ratio approximation (7.31)).

k ::::::' 4.5, Lw ::::::' 1.4. These results correlate with the calculated and experimen-
tal data of other researchers. According to calculations of Avvakumov [107],
obtained by using the method of vortex segments and distribution of sources
on wavy ground, the magnitudes of the Strouhal number, corresponding to
 7.1 Infiuence of Waves on Aerodynamics of a Lifting Surface 191

the aforementioned minima, were found as k ~ 3 for A = 2 and k ~ 4.3 for

A = 1. In the work by Shumsky [148], dedicated to the nonlinear flow problem
for a two-dimensional foH near a wavy wall (the method of vortex singular-
ities in combination with step-by-step linearization in time), the minimum
of the unsteady lift was estimated to occur at k ~ 2. In the experimental
investigations of Belinsky et al. [149], carried out by using the method of an
underwater screen (ground plane), it was shown that for a rectangular wing
with an aspect ratio A = 2, the minimum of the amplitude of unsteady lift
occurs at the Strouhal number k ~ 3.
The solution of problem (7.5)-(7.7) in the general case of an arbitrary
angle between the direction of motion of a wing and the wave front was
obtained in Rozhdestvensky [57]. A somewhat cumbersome derivation of the
solution of this problem and some final closed form expressions for the lift and
the moment contributions are given in the Appendix. Note that for oblique
waves, the wing is subject to the action of a lateral moment.
In accordance with the general structure of the solution, the wave-induced
lateral moment coefficient can be represented in the form

where K, = k cos ß, Fm,.(A, k, ß) and lJi(A, k, ß) characterize the amplitude and

the phase of the lateral moment due to the presence of waves. Some typical
results of calculations of the aerodynamic characteristics for oblique waves
and a reference point coinciding with the leading edge (a = 1) are plotted in
Figs.7.4-7.7.

1.0
F
-
Fa
0.8

0.6

0.4

0.2

13=45°
0.0
0 2 3 4 5
k
Fig. 1.4. The amplitude of the wave-induced unsteady lift coefficient of a rectan-
gular wing versus the Strouhal number for different magnitudes of the course angle
ß, >. = 3.
192 7. Aerodynamics of a Lifting System Near Curved Ground

1.0

111
0.5

0.0

-0.5

-1.0

-1.5

-2.0
0 2 3 4 k 5

Fig. 1.5. The phase of the wave-induced unsteady lift coefficient of a rectangular
wing versus the Strouhal number for different magnitudes of the course angle ß,
>. = 3.

Analysis of the results calculated for the wave-induced lift and the lateral
moment coefficients CYw and m xw leads to the following conclusions:
• With an increase in the course angle ß, the minimum of the amplitude of
the unsteady lift for a wing of a given aspect ratio decreases and shifts
toward the range of longer waves; this shift is more pronounced for wings
of a larger aspect ratio;
• Up to the Strouhal numbers corresponding to the minima of the amplitude
of the unsteady lift, a phase lag of the lift is observed with respect to the
moment of time when the leading edge passes the wave crest. This phase
lag vanishes when angle ß approaches 7r /2 (in the latter case the problem
becomes steady, which corresponds to wm ", equal to zero);
• In the range of Strouhal numbers considered, the amplitude curve of the
lateral moment coefficient contains two maxima. The magnitude of the first
maximum increases when the course angle tends to 7r /2. The position of
the maximum shifts toward longer waves with an increase in the aspect
ratio.
F'rom the general formulas, corresponding to arbitrary course angles, we
can derive limiting expressions for a wing of small aspect ratio and wave-
lengths much larger than the span:

The amplitude-phase characteristics, corresponding to an unsteady incre-

ment of the lateral moment, in the case of wings of small aspect ratio and
 7.1 Influence of Waves on Aerodynamics of a Lifting Surface 193

0.16 r---r--~---..,.---,-----,

0.12 f--+t----t--'t---+--+--+-J

0.08 f--+--t----t--~+---y+---j

2 3 5
k
Fig. 7.6. The amplitude of the wave-induced unsteady coefficient of the lateral
moment of a rectangular wing versus the Strouhal number for different magnitudes
of the course angle ß, >. = 3.

2.5 ,---.,.-----,---...---.,.-----,

"'mx ß = 15°
2.0 t-----1r-:".=--t-~..-+--........- - - I

13=45°
1.5 f - - - + - - - + - - - \ f - - \ - - ' : - - - l

1.0 f---j---li---++--+..::---I

0.5 f - - - j - - - - t - - - + - ' t - - i - - - - I

0.0 '--_---'_ _- - L_ _- ' -_ _" - - _ - - '

o 2 3 k 5

Fig. 7.7. The phase of the wave-induced unsteady coefficient of the lateral moment
of a rectangular wing versus the Strouhal number for different magnitudes of the
course angle ß, >. = 3.

long waves take the form

F, "" )..4k sinß 7!')..r;, 7!' r;,>.

m", - 192 ' Wm
'"
~ --
2
+ -27!'= -- + -.
2 27!'
Turning to the relative wavelength Lw, we obtain the following formula
for the amplitude of the lateral moment for >. -+ 0 and k -+ 0:
194 7. Aerodynamics of a Lifting System Near Curved Ground

wherefrom it follows that for wings of a small-aspect ratio over long waves, the
amplitude of the lateral moment is directly proportional to the aspect ratio to
the fourth power and the sine of the course angle and inversely proportional
to the relative wavelength.
Because the solutions presented above are linear with respect to the wave
amplitude and are valid in practically an unlimited range of Strouhal num-
bers, it is possible to apply the Fourier integral to investigate the following:
• motion of a rectangular wing over a cylindrical deformation of otherwise
Hat ground or over a solitary wave, and inHuence of a step variation of the
ground surfacej
• motion of a rectangular wing over a sea surface with irregular waves on it.
The solution of the latter problem in conjunction with consideration of the
equations of the dynamics of the lifting system in ground the effect opens
a possibility of developing the basics of normalization of the seaworthiness
and the ride quality of wing-in-ground-effect vehicles.
The asymptotic representation (7.2) of the velo city potential 'PI can be
continued to include the term ofthe order of O(Öä~), which accounts for non-
linear effects in the wave amplitude. In this case, the corresponding expansion
--2
for the lift coefficient should be supplemented by the term Öä~ C~aw .
In the simplest case of a wing of infinite aspect ratio). = 00, the expression
for this additional term can be obtained in closed form:
_ _ _

= A cos 2kt + B sin 2kt + Dj

~2

Gy w

.4=1+ cOS::2-1 + k(1:k 2 )[q(coSk-1)+f(sink-2k)]

2[Q(1 + 2k 2 ) - kR].
1 + 4k 2 '

- 1 (Sin2k)
B = 2k ~ -1 + k(1 +1 k 2 ) [q(sink - 2k) - f(cosk -1)]

2[R(1 + 2k 2 ) + kQ].
1 + 4k 2 '

- 1 1 [2 (2) cosk-2k+(k
D=2+ k(1+k 2 ) k+ 3k- k
2.
-2)smk] j (7.33)

_ cosk-1 . _ sink
q = k +k-smk, r = 2 - - - - cos kj
k
- 1 sin 2k
Q = 2 - ----:ik - 2(1 +1 k2) [_. k
qsm -
-( k 2)]
r cos - j
 7.1 Infiuence of Waves on Aerodynamics of a Lifting Surface 195

-0.2
1

k=5
-0.4 ---1_....L..--L_"----'---.l_....L..---L_"---"
L!

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 t 0.9 1.0

Fig. 7.8. The unsteady wave-induced lift coefficient of a fiat plate of infinite aspect
ratio versus time for different magnitudes of the Strouhal number k, aw/h = 0.5.

- cos 2k 1 1
R= ---:tk + 2(1+k 2 )[rsink+i}(cosk-2)]- 4k'
The magnitude of fJ is constant for a given wavelength and characterizes the
time-averaged contribution to the wave-induced lift coefficient. For a positive
angle of pitch fJ > o.
Calculation by using formula (3.33) shows that for the considered case
of a fl.at plate, a wing fl.ying over sinusoidal waves is subject to action of a
constant additional lift increment proportional to the magnitude of the cruise
lift coefficient and the square of the ratio of the wave amplitude to the ground
clearance. 4
Figure 7.8 presents a history of the wave-induced unsteady lift coefficient
related to the cruise lift coefficient CyO for fl.ying near fl.at ground with the
same relative clearance h.
Calculations were performed by the formula
C -
~ '" - c(}a w
C - aw y + a-2w C(}a
--2
y
w
.
Yo

The calculated data correspond to a wing of infinite aspect ratio (CyO =

Bjh, h -t 0) and relative wave amplitude ii w = awjh = 0.5. The dashed line
in Fig. 7.8 describes the wave profile, enabling us to determine the position
of the maximum (minimum) of the unsteady lift with respect to the position
of the crest or hollow of the wave.

4 The same is true for any wing with a fiat lower surface moving in the extreme
ground effect.
196 7. Aerodynamics of a Lifting System Near Curved Ground

I
\ i

I I

\0:-1
0.8

j.=2

0.6

~
0.4 ~
~
0.2
I 0 ~
u (
~ I----
y
00
D o 0 o n (
o
0.1 0.15 0.2 0.25 0.3
h
Fig. 1.9. A comparison of theory with experiment for a rectangular wing over
waves ( solid lines: theory; circles: experiment [137J; the abscissa of the reference
point a = 0.5, A = 2, k = 1.05).

Prom a practical viewpoint, for example, for using the results of this
solution in the dynamics of fiight of a vehicle over waves, it is convenient
to represent the aerodynamic coefficients by aerodynamic derivatives of the
lift and moment coefficients with respect to the wave coordinate Yw and its
rate Yw'
In this format, the lift and moment coefficients can be calculated with the
help of the following formulas:

(7.34)

mz = m za + m~w Yw + m~wyw, (7.35)

m x = m xa + m~w Yw + m~wyw, (7.36)
where CYa ' m za , m xa are the cruise values of the aerodynamic coefficients
Yw = a w cos K,t, Yw = -K, a w sin K,t,

.
CYw = -~C
e ' . (7.37)
Y h 2 K, Yw'

.
m Yw = --~m
e . (7.38)
z h2 K, Zw'

Yw _- h K,
m x 2
e "".;smxw , (7.39)

where in all cases, the lift and moment coefficients are defined with respect
to the same reference point a. When analyzing the dynamics of the vehicle,
this point can be chosen to coincide with the center of gravity. Formulas for
 7.1 Infiuence of Waves on Aerodynamics of a Lifting Surface 197

0.25

C -C 0.23
ymax ymin
0.20 ,

-
0.18

0.15 \' ,A.=2

,
/'
0.13 \" /

- -\ ,
/

--
0.10
-- \" /

---
A.=1
0.08
-~ /
0.05
~
-------- ~
0.03
---- - - - - - - -
0.00
o 2 3 4 5 6 7 8
k
Fig. 7.10. The doubled amplitude of the wave-induced lift coefficient versus the
Strouhal number (A = 2, (}O = 5°, h = 0.2, a w = 0.1; solid lines: present theory;
dashed lines: numerical calculation, [107]).

the above coefficients in the case when the reference point coincides with the
trailing edge, are presented in Appendix, 5 Recalculation from the trailing
edge to the arbitrary position of the reference point a i- 1 can be done
by multiplying the complex amplitude of the corresponding coefficient by
exp( -ka), see formula (7.22).
In Fig. 7.9 a comparison is presented of some results calculated for the
aerodynamic derivatives with experimental data for motion of a rectangular
wing near a wavy solid wall (a = 0.5,'\ = 2,()O = 3.3°,ß = O,/\: = k = 1.05).
The experiment was conducted by Grebeshov [137) by the method of an
underwater screen.
Figure 7.10 presents a comparison ofthe dependence ofthe doubled ampli-
tude (CYmax - C ym1n ) of the unsteady lift coefficient versus the Strouhal num-
ber with the calculated data of Avvakumov [107](128 panels on the wing, 720
panels on the ground). In Fig. 7.11 instantaneous magnitudes of the wave-
induced lift coefficient of the present theory are compared with calculated
data of [107]. Figure 7.12 presents a comparison of results calculated by the
present theory with the results of Shumsky [148], where a complete numerical
analysis of a two-dimensional inviscid flow past a flat plate above a wavy wall
was performed by the discrete vortex technique accounting for the structure
ofthe wake.
Figures 7.13-7.16 illustrate the influence of the Strouhal number and the
course angle upon the aerodynamic derivatives of the lift and the lateral

5 Note that in the Appendix the longitudinal moment coeffcient m zw was also
calculated around the trailing edge.
198 7. Aerodynamics of a Lifting System Near Curved Ground

0.04
,
Cy(t) ,
I
I'J,
0.03
,
,
~ ,
,V
0.02

0.01 r\'\'- I
-',
,
II .( f-'
0.00 .'

-0.01
wa\ eco tou r 1\~ ", - - .
,'/
,

-0.02 r\ !'--V
- ~ --~
'

-0.03

-0.04
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
t/Tw

Fig. 7.11. The instantaneous wave-induced lift coefficient of a rectangular wing

(>. = 1, a w = 0.1, h = 0.2, k = 2.1; solid line: present theory; dashed line: numerical
calculation [107]).

1.6
I I
I
i
I I - - -,
1.2
--
- /V '~
I ,
1.0

0.8
\, ,
,
/
/
,
,
\, ,
,
" I
,!'
/
\' 1/I
0.6

0.4
i'---"
0.2

0.0
o 2 3 4 5 6 7 k 8

Fig. 7.12. The relative infiuence of the Strouhal number upon the doubled ampli-
tude of the wave-induced lift for a fiat plate of infinite aspect ratio (>. = 00, BO =
6°, h = 0.2, a w = 0.15; solid line: present theory; dashed line: numerical calculation
[148]).
 7.1 Influence of Waves on Aerodynamics of a Lifting Surface 199

0.4 , . . - - - - , - - - r - - - - , - - - - ,

h2 CYw
Y
e
0.2 f--~<+---f-~~+-----1

0.0 f----t---'.:--f----t-7"""-----:::;

-0.21------'----1------'-----'
o 2 4 6 k 8
Fig. 7.13. The aerodynamic derivative C~w of the lift coefficient of a rectangular
wing in flight over waves versus the Strouhal number for different course angles ß,
>. = 2.
0.2

h2 C Yw
-_Y
ß = 75°
e 0.1

0.0 f-----H7L,.c....-f----!-""""""'-.-;:::::=!I

~.1r--+-_+-----1---~r_--~

-0.21------'----1------'-----'
o 2 4 6 8
k
Fig. 7.14. The aerodynamic derivative C:w of the lift coefficient of a rectangular
wing in ßight over waves versus the Strouhal number for different course angles ß,
>. = 2.

moment coefficients of a rectangular wing of aspect ratio A = 2 in fiight

over wavy ground. Note that the quantities h2Ctw je, h2cew je, h2m~w je
and h2m~w je do not depend either on the relative ground clearance or on
the adjusted pitch angle. This feature makes the data representation more
compact.
200 7. Aerodynamics of a Lifting System Near Curved Ground

a
0.00 -4'"~---+-~"'cth:----+-----;;J

-0.04 -+----~--_+---\---+----"-__'_1

_0.08.....l.---....-l'------'------'----'
o 2 4 6 8
k
Fig. 7.15. The aerodynamic derivative m~w of lateral moment coefficient of a
rectangular wing in fiight over waves versus the Strouhal number for different course
angles ß, >. = 2.

0.3 r - - - - r - - - - , - - - - r - - - - - - - ,

a
0.2 r----'\rl-----+----+----'

0.1 f----+----''';----j----f----i

-0.1 '--_ _--'-_ _ _....1--_ _ _" - -_ _---'

o 2 4 6 8
k
Fig. 7.16. The aerodynamic derivative m~w of the lateral moment coefficient of
a rectangular wing in flight over waves versus the Strouhal number for different
course angles ß, >. = 2.

7.2 An Estimate of the Acceleration of a Lifting Surface

in Flight over a Wavy Wall
To evaluate the ride quality of a wing-in-ground-effect vehicle versus the de-
sign parameters and the characteristics of waviness of the underlying surface,
qualitative analysis will be performed in this paragraph of the acceleration
 7.2 An Estimate of the Acceleration of a Lifting Surface 201

due to fiight of a lifting surface elose to wavy solid wall. Assuming for simplic-
ity that in cruise all perturbations are sufficiently small, we consider both the
linear aerodynamics and the equations of motion. In addition, the analysis
will be conflned to the pure heave motions of the vehiele under the action of
the wave-induced lift. 6 In dimensional form, the instantaneous vertical posi-
tion of the vehiele's center of gravity is described by the following ordinary
differential equation

(7.40)

where M and >'22 are the mass and added mass of the vehiele, the latter
corresponding to the inertial force in the direction of y axis in motion along
the same axis, H cg represents a dimensional distance of the center of gravity
from the unperturbed ground surface, and T is dimensional time. The right-
hand side of equation (7.40) contains external aerodynamic forces, acting
in the vertical direction. We render (7.40) nondimensional, using the root
chord Co and the cruise speed Uo of the vehiele as characteristic quantities.
Substituting the nondimensional parameters in (7.40), we obtain

(7.41)

In (7.41), CYn is the contribution of the lift force R Yn to the total lift coeffi-
cient, S stands for the reference area of the wing, and p is the density of the
air. Neglecting the added mass >'22 with respect to the mass of the vehiele,
we obtain the following nondimensional equation:

d 2 h cg _ ~C (7.42)
J.l dt 2 - L......- Yn'
n=l

where J.l = 2M / pSCo is the doubled relative density of the vehiele based on
the reference volume v;. = SCo . Introducing the perturbation of the ground
elearance h(t) = h cg - h, where his the time-averaged relative ground elear-
ance, we rewrite (7.42) as

(7.43)

We represent the right-hand side of (7.43) as the sum of three lift coefficients
responsible for contributions of the aerodynamic action of the waves and the
heave response of the vehiele,
6 Practically, this mode of motion takes place when the vehicle's pitch is kept fixed
by controls.
202 7. Aerodynamics of a Lifting System Near Curved Ground

(7.44)

where Gyw is the wave-induced lift coefficient, and the second term represents
the "spring" feature of the dynamic air cushion under the lifting surface in
cruise, Le., the response of the vehicle to the perturbation of the ground
clearance. 7 The last term is due to the wave-induced he ave rate and the ac-
celeration. By using the results obtained in paragraphs 3.6 and 7.1, dedicated
correspondingly to heave oscillations and the motion of a wing above wavy
ground, we can write formula (7.44) in more detail:

(7.45)

where a w is the amplitude ofthe wave as a fraction ofthe root chord, i = A,

k = 27r / Lw is the Strouhal number, and Lw is the wavelength related to the
root chord. In all cases we assume as meaningful only real part of the complex
expressions. The parameters Gw and Gh represent the complex amplitudes
of the wave-induced lift eoefficient and the lift eoefficient due to the heave.
Assuming that the wave-indueed oseillations take plaee for a long time so
that free oscillations are damped, we restrict the analysis to "foreed " motions
~ith the nondimensional frequeney of the waves. For harmonie oseillations,
h = -k2 h, so that the formula for he ave motions induced by waves on the
ground becomes
aw(}Gw(>', k) exp(ikt)
h(t) = (7.46)

The maximum nondimensional aeeeleration due to heave is given by

(7.47)

Taking into aecount formula (3.65) for the lift coefficient Gy of a flat plate in
a steady ground efIect, we can derive an approximate relationship between
the eruise lift eoeficient and its derivative in height:

aGyO GyO
7ih=-T' (7.48)

One can also link the (doubled) density j.L of the vehicle and its eruise lift
coefficient in the following way:

GYO Fr 2 = j.L,

1 The minus derivative of the cruise lift coefficient with respect to the ground
clearance can be interpreted as the "stiffness" of the spring in a simple mass-
spring model of a vehicle on a dynamic aircushion.
 7.2 An Estimate of the Acceleration of a Lifting Surface 203

0.• 0 -7-=======--- - -- 12
Ah I I
- 0.35 At, Near Resonance I '
g :
0.30 ,- 't.-:----,-- -- --::::i=-- - - - ' g

0,25 :;""4~-'-..;..-_,~.k---=--­
a
I
I
O~ L!~~~-~~~---~~ I
0.15
! I
i
1
.:
0.10 I,-.;-~~;,.a..~"'----''-----
1
0.05 1
:--~--+_!...-'"----:-----:----
I
0.00 'L --'-_ _L - _' -_ __ --:_
o I 2 3 • 5 6 7 a 9 10
O ~~~~~~~;=~
0.00 0.0. 0.16 0.20
2n
k=-
Lw
Fig. 7.17. The wave-induced heave accelerations of a rectangular wing versus the
Strouhal number for different magnitudes of the aspect ratio (h = 0.2, J-t = 75,
CYO = 0.65, a w = 0.03) .

where Fr = Uo/.;gc;, is the Froude number based on the root chord. With
this in mind, the maximum (dimensional) wave-induced heave acceleration
Ah as a fraction of the gravity acceleration 9 can be shown as

awJ-tICw(\ k)/Cw(A, 0)lk 2

(7.49)
1- C yO + hJ-tk 2 + iCh ( \ k)l'
In so me cases it may be more convenient to use the parameters Fr and CyO
than the relative density of the vehicle /-l and the lift coeffcient C yo ' In this
case (7.49) can be written in the alternative form

a wFr 2 C yo ICw(A, k)/Cw(A, 0)lk 2

(7.50)
1- CYO + k2hFr 2 CyO + iCh(A, k)l'
Figures 7.17 and 7.18 illustrate the dependence of the acceleration level
on the Strouhal number (the ratio of the wavelength to the chord) for dif-
ferent aspect ratios and ground clearances. In particular, Fig. 7.17 shows the
behavior of the acceleration level versus the Strouhal number for different as-
pect ratios of the lifting surface. Figure 7.18 demonstrates the dependence of
the wave-induced heave accelerations upon the Strouhal number for different
magnitudes of the relative ground clearance.
An analysis of the calculations shows that for certain combinations of
the cruise lift coefficent, the relative density of the vehicle, and the relative
ground clearance, the wave-induced heave of the vehicle can become resonant.
It follows from the heave equation (7.43) that nondimensional frequency (the
Strouhal number) of free oscillations k f is given by
204 7. Aerodynamics of a Lifting System Near Curved Ground

g I, ; Far from Resonance g -----1-

h = 0.1
Nea r ResonarlQe ;

0.3 - \ - - - - - - ' - - - , - - - - - - - - i
I
h = 0.15

3 ---~H+~-~-----~

0.0 ~~....:....--LI---.J..
o 23 4 5678
! ---::---:---=9:-"":""0 0.0. 0.08 0.12 0.16 020
211 211
k= - k= -
Lw Lw
Fig. 7.18. Wave-induced heave accelerations of a rectangular wing versus the
Strouhal number k for different magnitudes of the relative ground clearance (>' = 2,
J.L = 75, CYO = 0.65, a w = 0.03).

kf ~ j_!. 8G,
J.L ßh '

or by using the relationship between ßCy/ßh and the lift coefficient in cruise
CYO '

kf = j~~. (7.51)

When the nondimensional frequency of the forced oscillations approaches k f ,

the heave motion due to waves becomes resonant. For large first-generation
ekranoplans the magnitudes k f are smaIl. If, as an example, we take J.L =
75, h = 0.1, and C yO = 0.65, the resulting critical Strouhal nu mb er can be
found to be 0.294 which corresponds to the ratio of the wavelength to the
chord equal to 21.3. It means that for a wing chord of the order of 18 meters
the resonant wavelength would be about 385 meters. Further consideration of
calculated results demonstrates that with an increase in the relative density
and relative ground cIearance the wave-induced heave motions diminish. At
the same time, increases in the wave height and the cruise lift coefficient lead
to an increase in wave-induced motions.
The structure of the formula (7.49) reflects the character of the depen-
dence of wave-induced heave motions on the most important design parame-
ters, such as the relative density of the vehicle, the cruise lift coefficient, the
aspect ratio of the main wing, the relative ground clearance, as weIl as upon
the parameters of the waviness of the underlying surface, Le., wavelength and
amplitude.
 7.3 The Aerodynamic Response of a Lifting System 205

It should be noted that heave accelerations, exemplified above, are associ-

ated with fiight above asolid wavy wall. When considering the corresponding
effects of sea waves, we should take into account that with a decrease of the
wavelength (an increase in the Strouhal number) the height of the wave dimin-
ishes. 8 Therefore, the expected amplutude ofthe wave would be considerably
smaller. It is also worth mentioning that the estimations discussed in this
paragraph are very approximate and provide some qualitative information
on the infiuence of different factors on accelerations due to the waviness of
the ground.

7.3 The Aerodynamic Response of a Lifting System

to Aperiodic Perturbations

The aperiodie behavior of aerodynamie characteristies versus time can result

from aperiodie motions of the lifting surface (takeoff, landing, acceleration-
decceleration, the variation of cruise parameters due to the action of control
surfaces) and also due to aperiodic perturbations, caused by unevenness ofthe
ground or gusts. If the problem is linearized with respect to the perturbations,
then the Fourier integral transform method can be used for the analysis of the
response of the lifting system to the aperiodie perturbation or motion. In what
follows, the analysis will be restrieted to the aerodynamies of a lifting surface
in the ground effect under the action of aperiodic perturbations (irregularities
or steps on the ground or gusts). The solutions of the problems of motion
of the wing in elose proximity to a wavy sinusoidal wall or under the action
of a vertieal sinusoidal gust for arbitrary Strouhal numbers can be used as
transfer functions for the application of the Fourier integral.
It is weH known that if a stahle system that has behavior described by a
linear differential equation with linear boundary conditions is subject to the
prolonged action of harmonie perturbation, then the response of the system
will change in time harmonieally and with the same frequency. Consequently,
within the linear theory, the aerodynamic characteristics due to harmonie
perturbation, say the lift coefficient Gy< (t), can be written as

Gy< (t) = cGy(k) exp(ikt),

where Gy(k) is a complex transfer function of the lift coefficient and c is the
amplitude of the kinematic parameter, characterizing the harmonie pertur-
bation. For example, when a wing is subject to the action of regular waves on
the ground surface, the parameter c can be set equal to the wave amplitude
aw , whereas for a vertical harmonie gust, c can be equated to the maximum
velo city in the gust vg , k is the Strouhal number, corresponding to the per-
turbation under consideration, and i = A is an imaginary unit. Transfer
8 For large waves the height-to-wave ratio constitutes approximately 1/20.
206 7. Aerodynamics of a Lifting System Near Curved Ground

functions of moment coefficients can be introduced in a similar manner. We

consider a lifting surface, moving at a constant speed and without slipping,
in elose proximity to the ground and subject to action of aperiodic pertur-
bations such as an arbitrary irregularity on the ground or a vertical gust
of arbitrary form. For simplification, assume that motion of the wing takes
place in a direction perpendicular to the front of the perturbation. Assume
also that the speed of the perturbation is negligiably small compared to the
speed of the lifting surface, 9 so that the waves can be viewed as still with
respect to the wing. Introduce a fixed coordinate system Xl, Yl, and Zl in
which the directions of all axes coincide with those of the moving coordinate
system X, Y, z, introduced earlier. Planes XIOZI and XIOYI coincide with un-
perturbed position of the ground plane and the plane of symmetry of the
wing, respectively. Let the perturbation have the following form in a fixed
coordinate system:

where again c is a parameter that characterizes the magnitude of the per-

turbatiOon, for example, the relative (in terms of the root chord) height of
an irregularity on the ground surface, and f(XI) is a function of the order of
0(1) describing the form of the perturbation. Let the form function of the
perturbation be absolutely integrable on the interval -00 < Xl < 00, that is,

Then, in a wide dass of piecewise continuous functions, f(Xl) is representable

by the Fourier integral, see, for example, Smirnov [150]:

f(Xl) = -1
27r
1 00

-00
f2(k) exp(ikxt) dk, (7.52)

where
(7.53)

The function f2(k) is a Fourier trans form of the function f(xt) and conversely.
Counting time from the moment to, when the leading edge of the root chord
is located just above the frontline of the perturbation and passing over to the
moving coordinate system,

P(x, t) = 27r
c 1 00
-00 f2(k) exp[ik(x + t)] dk. (7.54)

9 For a wing-in-ground-effect vehicle fiying over the sea surface, this assumption
is reasonable. It follows from the statistics of sea waves that the velocity of
progressive waves does not exceed 1 knot, whereas the speed of existing and
future wing-in-ground-effect vehicles varies in the range from 50 to 300 knots.
 7.3 The Aerodynamic Response of a Lifting System 207

It follows from consideration of expression (7.54) that in the coordinate sys-

tem, attached to the wing, the function P can be viewed as the result of the
superposition of harmonie perturbations of the traveling wave type and the
quantity dl(k)/21f represents the amplitude of a harmonie of relative fre-
quency k. Knowing the transfer function, i.e., the aerodynamie response of a
lifting system to the action of a traveling wave of perturbation of arbitrary
frequency and unit amplitude, it is easy to derive the following expressions
for the coefficients of the unsteady lift and the moment upon a wing which
is subject to the action of aperturbation of arbitrary form f(XI):

Cy(t) = -c 1
21f
00

-00
GyD(k) exp(ikt) dk, (7.55)

mz(t) = 21f
C 1 00
-00 GmzD(k) exp(ikt) dk, (7.56)

where Gy and Gm. are complex transfer functions of the lift and the mo-
ment coefficients. As seen from (7.55) and (7.56), to determine the transient
aerodynamic characteristics of a wing subject to the action of an irregular
perturbation of a given form, we have to calculate the Fourier amplitudes
D(k).
We find the form of the function D(k) for perturbations of simple shape.
According to (7.53), the amplitude of the harmonie of relative frequency k

I:
can be written in a complex form

D(k) = f(XI) exp(-ikxI) dXI = DI(k) +W2(k). (7.57)

For an "even" perturbation, f(-XI) = f(XI),XI E [-Lp,L p], and

{OO (~~
D(k) = DI(k) = 2 10 f(XI) COSkXI dXI = 2 10 f(XI)coskxldxI,
(7.58)
where L p is the length of the perturbation. For an "odd" perturbation,
f(xI} = -f(-XI}, XI E [-L p, Lp], and
(L p /2
D(k) = W 2 (k) = -2i 10 f(xI) sin kXI dXI. (7.59)

As examples of "even" perturbation consider

• eosine perturbation, defined as

f( ) - {cos(1fxdLp) for lXII ~ L p/2,

Xl - 0 for lXII;:: Lp /2.
Integration gives the Fourier amplitude in the form
{OO 1f 21fL p ( Lp )
D(k) = 2 10 cos L pXl COSkXI dXI = 1f2 _ k2L~ cos k2" . (7.60)
208 7. Aerodynamics of a Lifting System Near Curved Ground

• triangular perturbation, given by the function

In this case, the following is the expression for the Fourier amplitudes:

(7.61)

As an example of an "odd" perturbation, consider a sine-type perturbation of

the form
f(xI) = {sin(21l' x d L p ) for lXII::; L p /2,
o for lXII 2: L p /2.
Calculation of the Fourier amplitudes in this case results in

(7.62)

Besides the "even" and "odd" forms considered above, it is interesting to

consider a step-type perturbation. The form of such perturbation is described
by a Heaviside step function

f(XI) = { 1 for Xl > 0,

o for Xl < O.
Representation of the step function by the Fourier integral is known to have
the form
f( ) =
Xl 2'
_1_100

1l'Z -00
exp(ikx l ) dk
k '
wherefrom
(7.63)

Before one performs calculations for concrete cases of irregularities on the

ground or vertieal gusts, the origin of the fixed coordinate system should be
shifted to the front of the perturbation and the transform {}(k) rewritten
accordingly:
{}*(k) = {}(k)exP(ik L;),

where {}* (k) is the amplitude of a harmonie in the new coordinate system.
Separating the real part of the expression for the transient lift coefficient, we
obtain the expression
c (OO
Cy(t) = 21l' Loo lR{Gy(k){}(k) exp[ik(t + L p /2))} dk. (7.64)
 7.3 The Aerodynamic Response of a Lifting System 209

Taking into account the preceding results for the unsteady lift coefficient of
a wing in motion over wavy ground obtained in paragraph 7.1, we can write
the lift coefficient transfer function in the form
iJ .
Gy(k) = h[A(k) - IB(k)], (7.65)

where for a rectangular wing, the parameters A(k) and B(k) depend on
the Strouhal number and the aspect ratio and are determined by using the
previous analysis of the aerodynamics of a wing in motion above the waves.
For example, for a wing of infinite aspect ratio advancing in close proximity
to sinusoidal wavy ground, the formulas for A(k) and B(k) are
A = Al cosk - BI sink, B = Al sink + BI cosk,
where

AI =2 (
1 - cos
k2
k) -1+k
1.
[2ksmk+(1-k)cosk-1];
2
2

2 (Sin k ) 1 2 .
BI='k -k-- 1 +1+k 2 [(1-k )smk+k(1-2cosk)].
Denoting the height of the irregularity on the ground as Uh and employing
the wave amplitude aw as €, we can rewrite (7.37) as

ah iJ
Cy(t) = 27rh 100
-00
kL p + B(k) sm(kt
nl(k) A(k) cos(kt + -2-)
{ [ •
+ kL] T)
-n2 (k) [A(k) sin(kt + kL T) - B(k) cos(kt + T)
kL ]} dk, (7.66)
so that in concrete cases
• for an "even" irregularity n(k) = n 1 (k), n 2 (k) = 0,

• für an "üdd" irregularity n(k) = W 2 (k), niCk) = 0,

ah iJ
Cy(t) = - 27rh 100
-00 n 2 (k)[A(k) sin (ktk+L ) T - B(k) cos (ktk+L
T )] dk;
(7.68)
• for a step-type irregularity,

Cy(t) = a~iJ {l(t)A(O) + ;7r i: [A(k) ~ A(O) sinkt - Bkk) cüskt] dk},

where let) is a Heaviside step functiün, defined as

1 (t)
°
= {1 for t > 0
for t < O.
210 7. Aerodynamies of a Lifting System Near Curved Ground

0.8,----,------,----r---,---,

0.4 H--+I\--\+--i-----+---i

0.2 1t-f--++--l-~___::j.L---+T--1

0.01---+--\--+-+-+--+---1

~.2L-_~ _ _-L_ _J -_ _ L--~

0.0 0.2To 0.4To 0.6To 0.8 To 1.0 To

t
Fig. 7.19. A history of the additional lift eoeffieient of a flat plate flying over a
triangular irregularity, >. = 00.

0.8

Cy(t>
"\

I~
0.6
ahCyo
1,.=1
0.4 \

\
I/
~ A=2
--- I

>< I
0.2

~
~

0.0
~~
j
2~
JI
·0.2 /

-0.4
0.0 0.2 0.4 0.6 0.8 1.0 1.2 t 1.4 1.6

Fig. 7.20. A history of the additional lift eoeffieient of a flat plate of finite aspeet
ratio flying above a eosine-type irregularity, L p = 0.5.

Some results of the calculations of the transient lift coefficient of a rectangu-

lar wing in motion above different types of uneven ground are presented in
Figs. 7.19 and 7.20.
The vertical coordinate of each graph stands for Gy (t) jäh G yO ' where G yO is
a steady-state (cruise) lift coefficient of a Hat rectangular wing in the ground
effect and äh = ahj h is characteristic vertical dimension of the irregularity
related to the ground clearance.
 7.3 The Aerodynamic Response of a Lifting System 211

1.0 I--r--
Cy
~Yl~
ahCyo
0.8
1..= />.3 / Vv "
/
! l/~ Vi I..Fl

0.6
// / 1..= 0.5

,// !

0.4 I /j
1/
Ij I

i
0.2
ff II

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
t
Fig. 7.21. A history of the additional lift coefficient of a rectangular wing of finite
aspect ratio, flying over a "step" in the extreme ground effect.

An analysis of calculated data for an isolated irregularity shows that

• for wings of small-aspect ratio at initial moments, the additional lift first
increases, then drops and tends to zero. In so me cases the influence of
the perturbation vanishes before the trailing edge of the wing finds itself
beyond the irregularity;
• with an increase of the aspect ratio, the character of the curves changes
somewhat. After the first maximum of additional lift, the second one ap-
pears, which is smaller in size. The influence of the perturbation may be
observed longer than it takes for the wing to overfly the irregularity com-
pletely. For short convex local deformations of the ground and for certain
time periods, negative magnitudes of additional lift due to the perturbation
mayappear;
• the magnitude of the maximal additional force due to the local irregularity
increases with an increase in the amplitude of the irregularity and the
aspect ratio and with a decrease of the relative ground clearance measured
between the trailing edge and the unperturbed ground plane.
The data presented in Fig. 7.21 characterize the variation in time of the
additional lift force for a step type irregularity.
It follows from Fig. 7.21 that with a decrease in the aspect ratio, the
transition to the steady state takes place faster.
Conclusions formulated for flight above an irregularity on the ground gen-
erally remain valid when the wing is subject to the action of vertical aperiodie
gust. Presented here are only the results associated with the influence of the
so-called step gust. In the case 01 a step gust the function that character-
izes the distribution of the vertical velocities in the gust has the form of a
212 7. Aerodynamics of a Lifting System Near Curved Ground

Heaviside step function

v (x t) = {vgo for x > t,

g, 0 for x < t.

For this case, as shown earlier, the Fourier amplitude function takes the
form described by equation (7.63). For an infinite aspect ratio .A = 00, the
expression for the transient lift coefficient of a wing under the action of a
vertical step gust can be derived in an analytical form. It can be shown that
for .A = 00, the corresponding complex transfer functions of the lift and the
moment coefficients are given by

G _ ~ (1-
ik) 2 (1 -
ik)
Gm. = 3h 1 + ik . (7.69)
y - h 1 + ik '

Taking into account (7.69) the expressions for the coefficients of transient
lift and the moment coefficients for a wing of infinite aspect ratio under the
action of a vertical step gust will be

C = vgo [ ( )
y h It
_ ~
1f
1 00

-00
(k sin kt + cos kt) dk]
1 + k2

= vgo [1 - exp( -t)] for t > 0; (7.70)

h

mz = 2;~o [1 - exp( -t)] for t > 0, (7.71)

where l(t) is a Heaviside function.
It is remarkable that for the case und er consideration, the point of the
application of the additional lift force due to the gust remains at the same
time-independent position, namely, at a distance of one-third of the chord
length from the leading edge. Some results of the calculations, carried out
with help of formula (7.70) for .A = 00 and by numerical integration for
rectangular wings of a finite aspect ratio, are presented in Fig. 7.22.
This figure features the dependence on time of the ratio Cy (t) / CYoo ' where
C Yoo is a steady-state magnitude of the lift coefficient. In the same figure, the
dashed lines designate similar results for an unbounded fluid obtained ana-
lytically by Küssner [152] for a wing of infinite aspect ratio (see Bisplinghoff
et al. [151]) and numerically by the method of discrete vortices; see Belot-
serkovsky [139]. A comparison of the results for the two limiting cases of the
extreme ground effect (h « 1) and an unbounded fluid (h = 00) shows that
for wings with small and medium aspect ratios, corresponding variations of
the transient lift are steeper near the ground compared to the out-of-ground
effect.
At the same time, comparison with calculations of Küssner demonstrates
that for wings of very large aspect ratios transition to a steady state also
occurs faster, although at the initial period of time, the curve of the transient
lift coefficient goes upward more steeply for an infinite fluid.
 7.3 The Aerodynamic Response of a Lifting System 213

Fig. 7.22. A time history of the lift coefficient of a rectangular wing of finite aspect
ratio under the action of a "step" -type vertical gust (solid lines: present theory for
h « 1; 2: >. = 0.3; 4: >. = 1; 5: >. = 00; dashed lines: calculated results for h = 00;
1: >. = 0.25; and 3: >. = 1 (from Belotserkovsky [139]); and 6: >. = 00 (from Küssner
[152]).
214 7. Aerodynamics of a Lifting System Near Curved Ground

Appendix: The Motion of a Wing Over a Wavy Wall

at an Arbitrary Course Angle: Derivation
of the U nsteady Lift and the Moment Coefficients
Assume that the instantaneous gap between a lifting surface and the ground
is described by the form

h*(x, z, t) = h + (}x - a w cos k(xcosß + zsinß + t cosß) (Al),

where (} is the vehicle's course angle with respect to the wave front between
the normal to the latter and the direction of motion in the unperturbed
ground plane, aw is the amplitude of the wave, k = 211"/ Lw is the Strouhal
number, Lw is the length of the wave related to the root chord of the lifting
surface. Relating all quantities in (Al) to hand introducing new notations
K, = kcosß and p = ksinß, we can rewrite (Al) in complex form as

h*(x, z, t) = 1 + iJx - ä w exp[i(K,x + pz + K,t)]. (A2)

Representing the channel flow potential in the form of an expansion

(A3)
and using a general scheme of solution, we obtain the following problem for
"the wave-induced" component: CPiia w

ß2CPiiaw ß2CPiia w ß [. ßcpö]

ßx2 + ßz2 = ßx exp(lI~x + pz + r;,t) OX

+ :z [exp(iK,X + pz + K,t) O;:ö] , (A4)

where
~ [COsh(qn x ) ]
CPii(X, z) = L.J Qn . h( ) - 1 cos(qn z ),
n=O SIn qn

Q _ 4(-1)n 11"
n- >.qn3 ' qn = ~(2n + 1),

ßCPii ~ sinh(qnx )
7)
x
= n=O
L.J Qnqn h()
COS qn
Cos(qn z ),

ßCPii ~ [COsh(qn X) ].
(A5)
-ß = - L.J Qnqn h() -1 sm(qn z ).
z n=O COS qn

Writing CPÖa w = rpexp(ir;,t) and differentiating, we obtain the following equa-

tion for the complex amplitude of the potential:

Llrp = exp(iK,x) [exp(ipz) + iK,exp(ipz) ~: + iPexP(iPz/;:ii] = P. (A6)

 Appendix to Chapter 7 215

We represent the right-hand side of equation (A6) as the sum of even and
odd terms
(A7)
where

(A8)

. ~ [COSh(qn X ) ]. }
(Ag)
-lpCOS(pz) ~ Qnqn cosh(qn) - 1 sm(qn z ) .

Alternatively, we decompose P in the following fashion:

00

P = Pe + Po = L[Xe(x) COS(qlZ) + Xo(x) sin(2qlz)], (AIO)

1=0

where

(All)

where
h = 1~, 12 = i",Qnqn1:;', 13 = -pQnqnI;,
J1 = iJ~, J 2 = -",QnqnJ:;', J 3 = ipQnqnJ;.
We take into account the following identities:

2j>-.f 2 4ql(-1)lcos(pAj2) *
"\ cos(pZ)COS(qlZ) dz= \( 2_ 2) =11 ,
A -),,/2 A ql P
 216 7. Aerodynamics of a Lifting System Near Curved Ground

qz - qn ] I*
+( qz- qn )2 - P2 = 3'

21)../2 . , 8ql sin(p>'/2) *

'\ sm(pz) sm(2qlZ) dz = >'( 2 _ 4 2) = J 1,
1\ -)../2 P qn
21)../2. . 16(-I)npqlqnCOS(p>'/2) *
'\ sm(pz) cos(qn Z) sm(2qlZ) dz = >'[(2 )2 2)[(2 )2 2] = J2,
1\ -),,/2 qz - qn - P qz - qn - P
21)../2 . . 2( _l)n cos(p>'/2) [ 2ql + qn
'\ sm(pz) cos(qn Z) sm(2qzz) dz = >. (2 )2 2
1\ -)../2 qz + qn - P
+ 2qz - qn ] _ J*
(2qz - qn)2 _ p2 - 3'
We seek cp in the form
00

cp = L[Xe(x) COS(qlZ) + Xo(x) sin(2qlz)] , (AI3)

1=0

where Xe and Xo have to satisfy the equations

X~' - 4qf X o = Xo • (A14)

To construct particular solutions corresponding to the right-hand sides of

(A14), we represent Xe and Xo as

Xe(X) = lt exp(i~x) + ~ (2 Co:~(qn) {exp[(qn + i~)x]- exp[(i~ - qn)X]}

+I3 exp(i~x) - 2 CO:~(qn) {exp[(qn + i~)x] + exp[(i~ - qn)]}), (AI5)

00 J
Xo(x) = J1exp(i~x) + ]; (2 COS~(qn) {exp[(qn + i~)x]- exp[(i~ - qn)X]}

+J3 exp(i~x) - 2 CO:~(qn) {exp[(qn + i~)x] + exp[(i~ - qn)]}) (AI6)

or, regrouping,

Xe(x) = exp(i~x) (h + ~ I 3) + ~ {2(~~s~(~l) exp[(i~ + qn)x]

(I2 + I 3 ) . }
(A17)
- 2 cosh(qn) exp[(l~ - qn)X] .

Xo(x) = exp(i~x)(J1 + ~ J 3) + ~ {~~~~(~~) exp[(i~ + qn)x]

 Appendix to Chapter 7 217

(A18)

Particular solutions corresponding to the right-hand side of equation (A13),

can be sought in the form
1
L h() [ße exp[(i~ + qn)xJ
00

Xep(x) = O:e exp(i~x) +

n=O cos qn

+1'eexp[(i~ - qn)XJ], (A19)

Xop(x) = 0: 0 f:
exp(i~x) + n=O cos~()
qn
[ßo exp[(i~ + qn)xJ

+1'0 exp[(i~ - qn)XJ], (A20)

where the coefficients have been found in the form

I3 -h
ße = 2[(i~ + qn)2 + q?J '

1'e = 2[('l~ - qn )2 + ql2J; (A2l)

J 1 + L:~-o J3 J3 - J2
~2 + q? ' ßo = 2[(i~ + qn)2 + q?J '
h+J3 (A22)

At the leading and trailing edges of the lifting surface these particular solu-
tions yield the following magnitudes:
00 00 ß
X ep (0) '"'
= O:e + n=O
~
+ 1'e X op (0) = O:e + '"'
ße h() ~
e h()
+ 1'e , (A23)
cos qn n=O COS qn

The derivatives of the particular solutions at the edges are given by

218 7. Aerodynamics of a Lifting System Near Curved Ground

X~p(O) = aoi~ + n=O eosf= :() [ßo(i~ +

qn
qn) + 'Yo(i~ - qn)]. (A27)

We find the eoefficients ae , be and a o , bo of the homogeneous solution from

the boundary eonditions at the edges. It follows from
ocp-
o~w - ;t =
ocp-
0, x = 0, Z E [-.\/2, .\/2], (A28)

and
cp(l, z, t) = 0 (A29)
that

The eoefficients have been found in the form

a __ Xep(l)(qz + i~) + exp( -qz)[X~p(O) - i~Xep(O)] (A31)
e- exp(qz)(ql + i~) + exp( -qz)(qZ - i~) ,

a = _ X op (1)(2ql + i~) + exp( -2qz)[X~p(0) - i~Xop(O)] (A32)

° exp(2qz)(2qz+i~)+exp(-2qz)(2ql-i~)'
b _ exp(qz)[X~p(O) - i~Xep(O)]- Xep(l)(ql - i~)
e- exp(ql)(ql + i~) + exp( -ql)(qZ _ i~) , (A33)
exp(2qz)[X~p(0) - i~Xep(O)l- Xep (1)(2qz - i~) (A34)
bo = exp(2qz)(2qz + i~) + exp( -2qz)(2qz - i~) .
We ealculate the lift coefficient in the form Cy(t) = Gy exp(i~t), where the
eomplex amplitude of the lift eoefficient is given by

or
(A36)

where

lo
1X ( )d -
e x x - ae
exp(qz)-l_be exp(-qz)-l
qz qz
+ l
0
1x ( )d
ep X x, (A37)

(A38)
 Appendix to Chapter 7 219

The complex amplitude of the longitudinal moment coefficient m z is

4(-1)1 1 1
or

mz =
A

L00

n=O
-,,\-
ql
[- 1
Xe(x) dx -
0
xXep(X) dX],
0
1
(A39)

1
where
1 XX ep (X) d X -"
_ Ge [ exp (")
1K _ exp(iK)
" - 1]
o 1K 1K

+~
~
1 { ( "1K + qn ) -
ße [exp eX P(ik+ qn )-l]
---''--'---=---'---
n=O cosh(qn) iK + qn iK + qn

" 'Ye
+ 1K [(" P(iK-qn)-l]} "
exp 1K - qn ) -eX"
- qn 1K - qn
The complex amplitude of the lateral moment coefficient m x due to the asym-
metry of wave-induced loading can be determined with help of the formulas

or, finally,

(A40)

where

1
o
1X ( )d -
o X X - ao
exp(ql) -1 _ b exp(-ql) -1
ql
0
ql
+ l0
1x ( )d
op x x, (A41)

(A42)
8. Schematized Flow Models
for a Power-Augmented Lifting System

One of the problems that developers of wing-in-ground-effect vehieles have

to solve is related to necessity to reduce the power required for detaching the
craft from water. An efficient way to facilitate takeoff consists of blowing air
under the main wing of the craft from special engines. This mode of vehiele
operation is often called power augmentation or, briefly, PAR. Power aug-
mentation provides additional dynamic head to support the vehiele at small
speed and alleviates hydrodynamic loads due to the impact of waves upon
the structure of the craft. From the viewpoint of aerodynamics and hydrody-
namics, the problem of power-augmented takeoff is extremely complicated.
It features the interaction of turbulent jets with the vehiele and water sur-
face, the resulting spray effects, and the transient motion of the vehiele. In
what follows, only very simplified models of power-augmented flows will be
considered for a lifting system moving very elose to the underlying surface.
When modelling the aerodynamics of wing-in-ground-effect vehieles in
regimes of takeoff and transition to cruise by using power augmentation of
the main lifting system, different schemes of the flow of air blown from the
upstream PAR engines onto a wing may be adopted. Some of these schemes
are depicted in Figs. 8.1-8.3.
The first scheme of power-augmented ftow suggests that the suction force
at the leading edge is not realized and that inter action of the exhaust from
the engines with this edge manifests itself in the generation of a reentrant jet,
oriented at a certain angle ßj with respect to the downstream direction. It can
be assumed that the far-field direction of the jet is subject to the requirement
of the conservation of total moment um. The scheme with the reentrant jet
was used by Gallington et al. [154, 156] in the analysis of the efficiency of
power-augmentation regimes. These authors considered simplified cases of
flow around a flat plate at zero incidence and deflected flap, assuming that
there is no leakage from the gaps under the endplates.
The second scheme of PAR ftow can be introduced on an assumption
that the suction force is completely realized, and, moreover, that part of
air jets from the upstream engines flows along the suction surface of the
wing and continues to propagate downstream without separating from the
upper surface of the flap. The argument behind the second scheme, shown in

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
222 8. Schematized PAR Flow Models

L x

Fig. 8.1. Scheme (1) of power-augmented fiow with a reentrant jet, [153, 154, 155J .

L x

Fig. 8.2. Scheme (2) of power-augmented fiow with the realization of a suction
force and unseparated streamlines along the upper surface of the wing and the fiap.

Fig. 8.2, is certain experimental evidence that the jets envelop the rounded
leading edge due to the Coanda effect (see Krause [156]) .1
This scheme does not seem to be completely plausible for takeoff regimes
at small relative ground clearances, when sud den decceleration of the turbu-
lent jets generated by the upstream engines and rat her slow motion of air in
the channel under the wing is observed. These circumstances together with
the fact that the flap deflection angles at takeoff are quite considerable (of
the order of 20-30°) does not give a basis for assuming nonseparated flow
past the suction surface of the wing.
At the same time, a well-known property of the Coanda effect, especially
for relatively thin upper part of a bifurcating jet, to delay separation on the
upper surface of the wing, does not allow us to reject consideration of flow
models with the realization of the suction force.
In this connection the third scheme 0] power-augmented flow shown in
Fig. 8.3 may be considered. This scheme may be based on a suggestion that
the Coanda effect forces part of the bifurcating jet to envelop the leading edge
of the wing, but for a certain combination of system parameters separation

1 Another argument exists in favor of this scheme based on some data showing
that the magnitudes of drag, obtained in theory, are somewhat larger than the
experimental ones.
 8.1 Local Flow Problems for Analysis of PAR Mode 223

L x

Fig. 8.3. Scheme (3) of power-augmented flow with the partial realization of the
suction force and the jet leaving the rounded leading edge (Coanda PAR flow
model) .

of the jet from suction surface of the wing may occur at some angle ßsep. In
the third scheme it is assumed that the flow past the flap is separated.
To develop a description of power-augmented air flow past a vehicle in
the extreme ground effect, it is appropriate to use the method of matched
asymptotic expansions, taking the relative ground clearance h as abasie small
parameter, Le. , treat the PAR problem in a fashion, which has been adopted
throughout this book. Then the channel flow under the wing with small gaps
under the endplates can be assumed to be described with a certain degree of
adequacy by equation (4.53) or (in the steady case) byequation (4.65). The
solution of these equations enables us to determine both the span-averaged
velocity and pressure and, eventually, the lift, moment and induced drag. To
calculate the flow parameters near leading and trailing edges and associated
suction force and ideal pressure drag on the flap, it is necessary to consider
the corresponding local flow problems. Local flow formulations can be used,
e.g., for determining the appropriate deflection of the leading edge flap for
shock-free entry of the flow and analysis of other possibilities of controlling
the efficiency of cruise and PAR modes of performance of wing-in-ground-
effect vehicles. A set of relevant local leading edge and trailing edge flow
problems for power-augmented regimes is summarized in Figs. 8.4-8.10. In
what follows, solutions are presented of selected local flow problems.

8.1 Local Flow Problems for Analysis of PAR Mode

Notwithstanding a certain degree of idealization, the scheme with the reen-

trant jet (scheme 1) gives the possibility of obtaining the results which are
sufficiently plausible and compatible with experimental data on PAR effi-
ciency. We consider some relevant local problems of flows around the leading
and trailing edges. Solutions of these problems were obtained by Fridman
[157] by the method of S.A. Chaplygin of singular points within the frame of
the ideal jet theory.
224 8. Schematized PAR Flow Models

y @ Tl c
©
0 if F

B id 0
V·
~
B Sj A 0 B C
-1 0 b ;
a. b.
Fig. 8.4. The fiow region past a leading edge with a reentrant jet: physical and
auxiliary planes.

8.1.1 Reentrant Jet Flow Near the Leading Edge

We consider scheme 1 of the power-augmented flow past a wing in the extreme

ground effect. We introduce stretched local coordinates near the leading edge
as
- x-I - Y
X=~+Ls, Y=--l,
h1
where h1 = h*(l) is a local ground clearance at the leading edge and L s
is the stretched distance of the stagnation point from the tip of the leading
edge. The plane of physical fiow is depicted in Fig. 8.4.
Assurne as known the width of the incoming jet 8j and the velo city on its
boundary [Ij = 1. To determine the relative velo city potential of the flow, we
apply the method of singular points of S.A. Chaplygin, which employs the
idea of determining a holomorphic function in the complex plane, knowing
the function's zeros and poles and applying Liouville's theorem; see Gurevich
[138].
We map the domain of the flow in the physical plane z = x + iy onto
the first quadrant of an auxiliary complex plane ( = ~ + i1], so that the
solid boundary AOBC of the flow in the Z plane is transformed into a pos-
itive semiaxis ~ 2': 0 and the free surface part ADFC of the boundary in
the physical plane be transformed into a positive semi-axis 1] 2': O. Taking
into consideration that the complex conjugate velo city in the physical plane
dw/dZ has a zero ofthe first order at the stagnation point (( = 1) and per-
forming an analytic continuation of this function onto an entire plane (, we
obtain
dw (-1
dZ = - (+ l'
(8.1)

The complex conjugate velo city of the flow in the auxiliary plane ( is
found as
 8.1 Local Flow Problems for Analysis of PAR Mode 225

dw (((2 - 1)
d( = N ((2 + d2)((2 _ b2)' (8.2)
Expression (8.2) can be commented on in the following way. The complex
potential w should have a zero of the second order at the point ( = 1, that
is, w = 0[(( - 1)2]. Because there is no violation of conformity at this point,
the complex conjugate velo city may be assumed to be dw/d( = 0(( - 1),
i.e., having a simple zero at ( = 1. At the point ( = id, which is the im-
age of infinitely distant point D of a free jet, the function dw/d( must have
a pole of the first order (a stream with a finite flow rate). Physically, such
behavior corresponds to that of a sink in an auxiliary plane and models the
disappearance of the jet on the second leaf of aRiemann surface. In addi-
tion, the function dw/d( has another simple pole at the point ( = b, which
corresponds to a stream of a finite flow rate inside the channel. The resulting
expression (8.2) was found by analytic continuation of singular points of the
function dw /d( with the subsequent application of the Liouville theorem.
Dividing left and right sides of formula (8.2) by corresponding sides of
formula (8.1) and integrating the resulting expression taking into account
the requirement Z(l) = 0, we obtain the relationship between the planes
Z = Z(() in the form

- r< (((+1)2
(8.3)
Z(() = -N Jl ((2 + d2)((2 _ b2) d(.
We find the constants
- = Z(O)
ZA
- = Ls = N Jor 1 t(t+1)2
(t2 + d2)(t2 _ b2) dt. (8.4)

The condition of mass conservation is given by 2

(8.5)

(8.6)

(8.7)

wherefrom
(8.8)

The requirement, accounting for angle ßj of orientation of the reentrant jet,

lS
dw -. 1- id
dZ (id) = exp[-(7r - ßj)l] = 1 + id' (8.9)

2 All velocities are rendered nondimensional with respect to the velo city Uj on the
jet's external boundary.
226 8. Schematized PAR Flow Models

The eonjugate eomplex veloeity of the flow far downstream in the ehannel is

d~(b)=_b-l (8.10)
dZ b+ l'
It follows from (8.9) and (8.10) that

d = eot d'
ß·
(8.11)

* b-l (8.12)
va = b + l'
The magnitude of v; is determined by matehing the loeal flow to the flow in
the ehannel under the wing. Within the theory of the extreme ground effect
for a reet angular wing with endplates, deseribed by equation (4.65), it turns
out in particular that v;= v(l), where v is a spanwise-averaged velo city in
the ehannel under the wing. Other unknowns ean be ealeulated by using the
formulas
b = 1 +v~ (8.13)
1- v~'
8rj = 8j - v~, (8.14)
ßj = -2areeot d, (8.15)
- 2 2
N = _ 28rj b + d = _ ~ 8. (8.16)
7r l+d 2 7r J'

The flow velo city on AB (the ''wetted'' surfaee of the wing), ( = e, 'TJ = 0,
eE[O,b),is
dw l-e
dZ(e) = 1 +e'
where eis determined from
(8.17)

(8.18)

The solution derived herein gives the possibility of ealeulating the loeal
pressure distribution on the wing and the ground near the leading edge, as
weIl as the inelination angle ßj and the flow rate of the reentrant jet as
funetions of the flow rate of the ineoming jet and the elevation of the leading
edge above the ground.
 8.1 Local Flow Problems for Analysis of PAR Mode 227

8.1.2 Reentrant Jet Flow Near the Leading Edge

with a Deflected Tip

In order to analyze the possibilities of controlling the lifting capacity of a

winged vehicle in the ground eifect and of enhancing the efficiency of power-
augmented modes of operation, the local problem of practical interest is that
of fiow past aleading edge with a defiected tip. Herein, this problem is treated
within the reentrant jet formulation. The anticipated eifect consists of gen-
erating of press ure thrust on a defiected part of the leading edge with the
goal of reducing the drag and, consequently, the installed thrust required to
provide power-augmented lift of a vehicle of a given weight. As in the pre-
vious analysis, the problem is solved by the method of singular points. The
fiow patterns in the physical plane ( and the auxiliary plane ( are shown in
Fig.8.5.
Stretching of coordinates is carried out in the following way:

The complex conjugate velo city in the auxiliary plane ( is

dw _ (-a ((-l)(he/ 7r
dZ - - ( + a (+ 1 .
(8.19)

The exponent accounts for the violation of conformity at point 0 with

( = 1, where the angle between the adjoining parts of the solid surface
changes when passing from Z to the auxiliary plane (. The derivative of the
complex velocity potential in auxiliary plane ( is

(8.20)

Therefore,

y c
® D
11
©
if F

B
V· E id D
~
B 8j A 0 E B C
0 a b ;
a. b.
Fig. 8.5. Flow region past a leading edge with a deflected tip within a reentrant
jet scheme: physical and auxiliary planes.
228 8. Schematized PAR Flow Models

The unknown constants are determined from the following conditions:

1 J dw 7rN a 2 + d 2
QD = 2i hd d( d( = - 2 b2 + d2 = Jrj , (8.22)

1
QB = 2i
idW
b d( d( = -
7rNb 2 -a 2
2 b2 + d2 = v~. (8.23)

It follows from the condition of mass conservation that

(8.23a)
Therefrom
(8.24)

and, in addition,
(8.25)

because
~~(id) = exp[-(7r - ßj)iJ = -exp(ißj), (8.25a)

Oie d Oie + 7r - ßj
- arctan d + arctan - = 2 ' (8.26)
7r a
and
(8.27)

The velo city v~ is determined by matching with the channel flow, for
example, with a solution of equation (4.95). Because ZA = blef exp( -Wie)
and aN = -28j/7r, there is one more equation

2Jj 1r t(t+a)2
-:;;:- Jo (t 2 +d2)(b2 -t 2)
(1+t)(he/
I-t
7f

dt=b lef ,
-
(8.28)

where blef is the chord of the deflected part of the leading edge (as a fraction
of the distance of the hinge point ahove the ground). Thus, the system of
three equations (8.25),(8.27), and (8.28) was obtained for determining the
three unknowns a, b, and d. The magnitude of ßj is determined from (8.26).
The coefficient of the normal force acting on the deflected part is given by

C - 1-
n-
2J.
_J_
7rblef
10
1

(t 2
t(t-a)2
+d2 )(b 2 -t 2 )
(1-t)(he/
--
l+t
7f
dt
.
(8.29)

These results facilitate the analysis of the influence of parameters of the

deflection of the leading edge tip upon the efficiency of power-augmented
regimes.
 8.1 Local Flow Problems for Analysis of PAR Mode 229

8.1.3 Loeal Separated Flow Past a Flap (Endplate)

When modelling power-augmented lifting flows by using flow sehemes 1 or 3

(see Figs. 8.1 and 8.3), it is necessary to have a solution of the flow problem for
the jet escaping from under the flap or the endplate. Earlier, some results of
the solution of this problem by Gurevich [138] were used for determining the
contraction of the escaping jet and of the press ure drag component due to the
rear flap. Here we give a complete solution of this local problem by a method
different from [138]; see Fridman [157]. Note that knowledge of characteristics
of this type of loeal flow is quite practical beeause both available experimental
data and theoretical results show the significant influence of effective gaps
under endplates and flaps upon the efficiency of cruise and power-augmented
modes of motion. In addition, the structure of equation (4.53) and expressions
for determining the aerodynamic centers and static stability criteria show
that it is possible to control both the lifting capacity and the stability of
longitudinal motion by devices that provide variation (chordwise and in time)
of the gaps under the endplates. 3
The flow problem under consideration and the corresponding auxiliary
plane ( are depicted in Fig. 8.6 with designations : X = xjh1, Y = yjh 1;
the stretched length of the device is bf = bdh for the flap and hep(x) =
hep(x)jhl(x) for the endplate. In the latter case, hep and h1 are the local
length of the endplate and the local ground clearance of the wing at the cross
section x = const.
The complex conjugate velo city in the auxiliary plane is

d~ = _(( - i)(h/'Tr
(8.30)
dZ (+i .
The derivative of the complex velo city potential in the auxiliary plane is
dw (
d( = N (2 +c2· (8.31 )

3 This can be realized by deflectable or retractable plates.

y
® 11 ©
ie C

0 v** c X 0
A ~
C A B
-1 0 ;
6. b.
Fig. 8.6. Local separated flow past the flap: physical and auxiliary planes.
230 8. Schematized PAR Flow Models

The determination of constants is given by

1 J dw 7fN
Qe = 2i hc d(d( = 2' (8.32)

N = 2Qe = _ v;* . (8.33)

7f 7f
It is not difficult to find the velo city potential by integrating (8.31):

re; dw N (2
w(() = Jo d(d(+w A =2 ln (l+ c2)' (8.34)

- re; ( (( + i)Or/ 7r - (8.35)

Z(()=-N Jo (2+ c 2 ( - i d(+ZA,

where ZA = -bf exp(ilh). As Z(i) = 0, then

2v**
_0_
7f
l i
__"'_
1'2
0'" C
2
I' (I' + i)Or/7r dl' = b_.e
+
_"'_
1"
",-I
'" r
lOe
,

wherefrom
(1+'fJ)Or/ 7r
d'fJ=b-r
1
-2v;*1 "I -- (8.36)
7f 0 c2 - "1 2 1 - "I
with
dw . ** (C - 1) or/7r
dZ(lC) = -vo = - C+ 1 (8.37)
or
v** = (C-l){}r/7r. (8.38)
o C+ 1
The magnitude of constant c can be found from the equation

7f C + 1 J0 c2 - "1 2 1 - "I
r
~(C-l){}r/7r 1 "I (l+'fJ)Or/ 7r d 'fJ=br. (8.39)

Taking into account the equation QB = Qe,

-
c5f = V o** = (C-l)Or/7r
-
c+-1 . (8.40)

Expression (8.40) determines the width of the effective gap and the flow
contraction factor.
We evaluate the behavior of the free surface far from the flap (endplate).
On part of the boundary in the auxiliary plane, corresponding to the free
surface (lAB!), "I = O. Therefore,

- (~( ~ + i) Or/7r _
Zrs(~) = -N + c2 ~ _ i
Jo ~2 d~ - br exp(iBr/7f), (8.41)
 8.2 Unseparated Flows Near Leading and Trailing Edges 231

where ~ E [0, (0). Taking into account the identity

~+i)(h/7f
( --. 2Bf 1 2Bf
= exp(i- arctan t) = exp(iBr) exp( - i - arctan~), (8.42)
~-l 1f <, 1f

-
Xfs(~) = -N
je ~ cos ( ~ arctan
~2 2
t) d~ - b_cos Bf ,
f (8.43)
a +C
-
Yfs(~) = -N
je ~sin(~e arctan
2
t) d~ - b_sin B
f f. (8.44)
a +c
For ~ ---+ +00,
(8.45)

Therefore,
- - 2B N -
Yfs rv -1 + 6[ + -1ff- exp(XfsjN). (8.46)

The latter result shows that for the escaping fiow under a fiap or an end-
plate, .Kfs ---+ -00, the jet boundary resolves into a horizontal line exponen-
tially. From the physical viewpoint, this means that perturbed zero pressure
conditions apply immediately outside of the fiap or the endplate.

8.2 Unseparated Flows Near Leading and Trailing Edges

The local fiows near the leading and trailing edges for the eonventional (with-
out power augmentation) eruise of a wing-in-ground-effect vehicle were eon-
sidered in seetions 2 and 3. Here attention is paid to some loeal problems
that describe the unseparated fiows near the leading and trailing edges of the
wing in the power-augmented mode, when the ineoming and (or) eseaping
stream has a finite width.

8.2.1 Unseparated Coanda Flow Past a Leading Edge

This regime of fiow in the power-augmented mode was observed experimen-

tally in [153, 155, 156] and represents a praetieal interest from the viewpoint
of enhancing the efficiency of takeoff. The eorresponding fiow pattern and
auxiliary plane ( are shown in Figs. 8.7, with
- 1- x - y
X=T' y = h1 '
where h 1 is a the loeal elevation of the leading edge above the ground (the
leading edge relative to the ground clearanee).
The approach for deriving the solution is similar to that applied in previ-
ous sections. In what follows, the main results of this solution are presented.
232 8. Schematized PAR Flow Models

® TI D
©
D
D
C
E

E~____~+- ____________
C ~ E C B A D
o X 0 c 1 a ;
a. b.
Fig. 8.7. Unseparated fiow of a finite width past a leading edge(the Coanda effect):
physical and auxiliary planes.

The complex conjugate velo city and the potential derivative in the auxiliary
plane ( are
dw (-1 (+a
= -- --,
---=- (8.47)
dZ (+ 1 ( - a
dw (2-1
d( = N (((2 _ c2)' (8.48)

Therefore,

(8.49)

The unknown parameters can be found from the following conditions

v* = d~(c) = 1-c a+c, (8.50)

o dZ 1 +c a-c

1 1 dw 7rN-
QE = 4i !o d( d( = 2c2 = 8 j, (8.51)

Q = _~ 1 dw d'" = _ 7rN(c2 -1) = *

c 2i lc d( .. 2c2 V0 •
(8.52)

The flow rate of air should remain the same far downstream, Le.,

(8.53)

where
(8.54)
Therefore,

(8.55)
 8.2 Unseparated Flows Near Leading and Trailing Edges 233

Equations (8.50), (8.51), (8.52), and (8.53) enable us to determine four

constants out of five. The fifth unknown, v~, is determined from the procedure
of matching. Due to the circulatory fiow at the leading edge a suction force
occurs (the image of the ti p of the leading edge in the ( plane is located at a
point ( = a, 'Tl = 0). The suction force coefficient can be derived by singling
out the intensity of the square root singularity at the leading edge.
For (-+ a,
d~ r-.J2a a-1_1_, (8.55a)
dZ a+ 1 ~ - a
- (a + 1)2 2
X(~) r-.J N 4 2( 2 2) (~- a) . (8.55b)
a a -e
Therefore,
dw (a - 1)..;N
d~ r-.J va2 _ e2
1
VX =
A
VX' x -+ 0, (8.55e)

where
(8.56)

is the intensity of the square root singularity. The coefficient of the suction
force in 2-D fiow is given by

(8.57)

The parameters contained in the formula for Aare found in the form

(8.58)

a=
v~(1 + e) + (1 - e)
e (8.59)
v~(1+e)-(1-e) .
Substituting formulas (8.56), (8.58), and (8.59) in the formula (8.57) for the
suction force coefficient leads to the following expression:

(8.60)

It is easy to see that this result coincides with formula (4.62) with U(t) = 1
(steady fiow). Thus, it turns out that within the assumption 8uj = 0(1), the
suction force acting on the leading edge in a fiow of finite width is identical
to that for unlimited fiow.
We write the asymptotic expansion of the function dw /dZ (the complex
conjugate velo city) in more detail. From (8.47) for ( -+ a,

dw a- 1 1 a 2 + 4a - 1 2
d Z- -+ 2a -a+1.,,-a
-c- + (a+l )2 + O(~ - 1) . (8.61)
234 8. Schematized PAR Flow Models

Based on this expression, we ean single out a constant eomponent of the

horizontal veloeity near the leading edge,

u _ a2 +4a-1 (8.62)
1- (a+1)2 .

Taking into aceount (8.47)-(8.49), the velo city and pressure eoefficient in the
loeal flow problem ean be determined by the following formulas valid on AB:
( = e; eE [a,oo), 1J = 0,

e e
..-
dw -1 +a
v = dZ- = (: + 1 Z--a' (8.63)
.
p = 1 _ v = 1 _ (e - 1 e+ a)
2 2 (8.64)
e+1e-a '
where

8.2.2 Coanda Flow of Finite Width Past a Deflected Leading Edge

This loeal problem is of interest in eonneetion with the possibility of in-

ereasing the reserve of efficieney of the power-augmented modes by providing
shoek-free entry ofthe tip of the leading edge. The seheme of the flow and the
auxiliary plane are shown in Fig. 8.8, where x = (x -l)jh!, yjh!, h 1 = h(l),
and bier is the length of the defleeted part of the leading edge as a fraction of
the leading edge ground clearanee h 1 • Here are the main steps and the results
of the solution obtained by the method of singular points of S.A. Chaplygin.
The eomplex eonjugate velo city in the auxiliary plane ean be derived in the
form
dw _ (- b (+ a (( - 1 (+ d)fhe /7r (8.65)
dZ - (+ b (- a ( + 1 ( - d .

Cf) TI F
F
F
C
E
E~ ____~~__________
C ~ EC B D A DJF
-1 X 0 c

a.
Fig. 8.8. Flow of finite width around the leading edge with a deflected tip (the
Coanda effect): physical and auxiliary planes.
 8.2 Unseparated Flows Near Leading and Trailing Edges 235

The derivative of the complex velo city potential in the auxiliary plane is

(8.66)

Mapping of the physical plane onto the auxiliary plane is realized by the
function

(8.67)

Unknown quantities are determined from the following conditions:

(8.68)

(8.69)

From mass conservation,

(8.70)
where QD = 6uj .
It follows from the preceding relationships that
- 11"
8uj = QE - Qe = 2. N. (8.71)

Prom (8.69), (8.70) we can find that

~ = V1 - ~;. (8.72)

In addition

v*=dw c =b-ca+C(l-cd+C)(}le/ tr • (8.73)

o dZ () b+c a - c 1+ c d - c

Prom the conditions

ZD = ZD 1 = i, (8.74)
ZA = blef e-i(tr-(}le) +i (8.75)
follow the formulas

l1
d (t+b)2 t-a (t+l d-t)(}le/ tr
~::--~ -- -- --
t(t 2 - c2) t + a t - 1 d + t
dt = 0
'
(8.76)
236 8. Schematized PAR Flow Models

N J
1
a(t+b)2 a-t (t+l d-t)()le/ 7r
-- -- --
t(t 2 - c2) a + t t - 1 d + t
-
dt = blef. (8.77)

The magnitude of v~ is determined by matching. For the remaining six

unknowns, six equations are obtained, namely, (8.68), (8.71), (8.72), (8.73),
(8.76), and (8.77). Accounting for the relationships

where

_ b-c (l-C d+C)()le/ 7r

M- b + c 1+c~ ,
b= C/ VI - v~/ 8 j ,

we can reduce this system to one containing three unknowns a, c, and d and
including equations (8.76), (8.77), and the additional condition

a+c M -v* = o. (8.78)

a-c 0

The suction force at point A can be determined by the formulas of S.A.

Chaplygin

i 1 dw dw
Cs = 2" !a dZ d' cl"
_ (a - b)2 (a -1 d + a)()le/7r
Cs - 2rrN
2 2 -- -- • (8.79)
a -b a+l d-a
Consideration of the last formula shows explicitly that it is possible to
choose an angle of deflection of the leading eclge tip that corresponds to
shock-free entry, that is, zero suction force. This takes place at a = b. For a
given length of the deflected part of the edge (forward flap) blef, this angle
can be determined from the solution of the system

C C)
( 1 - d + ()le/
7r _ v*
(8.80a)
1 +c d - C - 0'

l d +1-
_t:-:2:--_b-:2:- (t-
t(t - c ) t - 1 d + t
2 2
d - t)()le/ 7r dt-O
- ,
(8.80b)

l d _,--b2:::---_t-;:2.,.. (t-+-
t(t 2 - c2 ) t - 1 d + t
1 d - t)()'le/ 7r
-- dt=b- f (8.80c)

with
 8.2 Unseparated Flows Near Leading and Trailing Edges 237

The normal force coefficient for the defiected part of the leading edge
normalized with respect to blef is

Cn =!I r d (t-b)2 t-a (t-l d+t)(he/ 7r dt (8.81)

blef J1 t(t 2 -c2 ) t+a t+l d-t '
and for zero suction force,

t2 - b2 (t---1 -d +-t)lIfe/ 7r
dt (8.82)
t(t 2 - c2) t +1 d - t .

8.2.3 The Model of Flow Past a Leading Edge with

a Winglet

Following we consider aseparate small wing (winglet) in front of the leading

edge. This winglet may be used to control the efficiency of power-augmented
mode. In a corresponding local fiow, this device is approximately represented
by a point vortex of a given circulation r. The fiow pattern and auxiliary
plane ( are presented in Fig. 8.9.
The images of the vortex and the stagnation point, appearing in the fiow
domain due to it, have (in the auxiliary plane) coordinates (0 = 9 + if and
(1 = d + ic. In the physical plane z, the vortex of circulation (-r) is located
at a point Zc = Xc + iYc . The method of singular points gives the following
results. The complex conjugate velocity in physical plane can be obtained in
the form

(8.83)

where ( designates the conjugate complex coordinates. The complex potential

and its derivatives in the ( plane are represented by the following expressions:

® Tl 0
©
0
o l; 0=g+if
0
l; = d+ic
C o 1
E
C ~
E~____~____________ E C B A 0
o X 0 c 1 a l;

a. b.
Fig. 8.9. The flow region past a leading edge with a forward winglet: physical and
auxiliary planes.
238 8. Schematized PAR Flow Models

. 2 2
+ Ir 1
A

(I') = 2QE 1 I' _ Qe 1 (1'2 _ 2) ( - (1 (8.84)

W "
7r n" 7r n " c 27r n ,1'2, - "I
1'2 '

or
dw (2_1 ((2-(6)((2_(6)
(8.85)
d( = N (((2 - c2 ) ((2 - (r)((2 - (f)'
It follows from (8.84) that

dw 2(QD (2 - QE c2 ) 4r de(
d( = 7r(((2- c2) +-;- ((2_(f)((2_(f)' (8.86)

The associated mapping function is

The unknown constants can be found from the following conditions:

* dw c - 1 c + a (C 2 - 2gc + g2 + j2) (C 2 + 2dc + d2 + e2 )

va = dZ(c) = c+1 c-a c2 +2gc+g 2 +f2 c2 -2dc+d2 +e 2 '
(8.88)

(8.89)

1 1. dw * [C 4 - 2(g2 - j2)C 2 + (g2 + j2)2]

Qe =- 2i Je d( d( = va = 7rN c4 _ 2(d2 _ e2)c2 + (d2 + e2)2 . (8.90)
Because the rate of flow should remain the same, Le.,

we obtain
(8.91 )
The coordinates of the vortex point can be found by means of the expres-
si on

Therefore,

Xc = ~ [N 1'1 (( + 1)2 (- a (( + (0)2(( + ~0)2 d(], (8.92)

a (((2 - c2) (+ a ( + (1)2(( + (1)2

-Vc = ~ [N 1'1 ((+ 1)2 (- a (( + (0)2(( + ~0)2 d(] + 1. (8.93)

a (((2_ c2) (+a ((+(1)2((+(1)2
 8.2 Unseparated Flows Near Leading and Trailing Edges 239

The circulation should be given by

- r = iR i (1
dw
d( d(, (8.94)

0= CJ i1 ~~d(. (8.95)

Using expression (8.84), we obtain

Thus, condition (8.95) is equivalent to

0<[ ((f -1) 12(1'2

:es ((f _ c2) '>1 '>1 -
1'2)(1'2
'>0
12)]_
'>1 - '>0 - 0

or

Eventually,

(8.96)

8.2.4 Unseparated Flow Near a Trailing Edge with a Flap

We start by considering a flow of finite width past a trailing edge with a

flap. This scheme corresponds to the second power-augmented flow model,
discussed at the beginning of this section. The flow picture and the aux-
iliary plane ( are shown in Fig. 8.10, where X = -x/h, Y = y/h - 1,
bf = brJ h, bf sin ef ::; 1. In the latter relationship the equality sign cor-
responds to the case, when the flap touches the ground.
Application of the method of singular points gives the results, presented
below. The complex conjugate velo city in the ( plane can be derived in the

,+
form

dw _ (' - 1 C)lh/ 1f (8.97)

dZ - (+ 1 ( - C .

The derivative of the complex velo city potential with respect to the aux-
iliary variable ( is obtained as

(8.98)
240 8. Schematized PAR Flow Models

y
® Tl C ©
D UJ=1

D SUJ c:::=::!:>
8 E
V" 0
c:::=::!:>
8 Sj D 0J A 0 8 C
----------~--------~
-1 o c a 1 b ;
a. b.
Fig. 8.10. The flow region past a flap of the wing in the extreme ground effect:
physical and auxiliary planes.

The resulting mapping function is

(8.99)

It follows from mass conservation that

Qn + QB = Qe, Qn = 8uj , QB = v~*, Qe = 8j ,

1 1 dw 7rNa 2 _
Qn = 4i 10 d( d( = 2b2 = 8 uj , (8.100)

Q =2-1dwd(=7rN(b2-a2). (8.101)
B 2i h d( 2b 2
Prom (8.99), (8.100), and (8.101), we obtain
- 7rN
Qe =8j = T· (8.102)

As IOAI = IOlAI = bf ,ZA = bfexp(-iIJr), then

Z(a) = bfexp(-it'h), Z(c) = 0,
or for ( = e, TJ = 0,
N 1 a
1 t2 - a 2 (1 + t t - c) Or/7r
-- --
t(b 2 - t 2) 1- t t +C
-
dt = bf , (8.103)

1 c
1 t2_a2 (1+tt-c)or/7r
--
t(b 2 - t 2) 1 - t t + c
dt-O
- .
(8.104)

In addition, for ( = b, v~* = dw / dZ, or

Vo
** = (b-1 b+c)Or/ 7r
b+ 1 b - c .
(8.105)
 8.2 Unseparated Flows Near Leading and Trailing Edges 241

Note that in this problem the channel flow velo city near the flap is cal-
culated as one of the results of the solution. In a two-dimensional problem
the magnitude of Jj is assumed to be known because the flow rate and the
velo city on the jet boundary are the same as those in front of the wing. Then
from conditions (8.101)-(8.104), we can obtain the following system of three
equations for the determination of constants a, b, and c:
-
-28j
7r
1a
1
-
f(t, a, b, c)dt - bf = 0, (8.106a)

1 1
f(t, a, b, c)dt = 0, (8.106b)

( 1 - -a2)_8· - (b-1 b+c)th/1f_

-- -- - 0 (8.106c)
b2 J b+1b-c '
where

t2 - a 2 (1 +t t - c) or/1f
f(,tabc
, ,) = t(b2 - t2) -
1 - -t -
t +-c . (8.107)

Knowing a, b, and c, we can determine the remaining unknowns using the

formulas
_ a2 _
8uj = b28j , (8.108)

2-
N= -8j , (8.109)
7r

(8.110)

8.2.5 Flow of Infinite Width Past a Trailing Edge with a Flap

This problem is connected with the aerodynamics of the cruise regime when
(in stretched coordinates) the flow around the trailing edge is semi-infinite.
The corresponding pattern of the flow and the auxiliary plane ( are depicted
in Fig. 8.11, where Y = yjh, X = xjh, (h is the angle of the deflection of
the flap (B f = 7rn, 0 < n < 1), and bf is the chord of the flap, normalized
by the ground clearance.
Using the method of singular points, we find the derivatives of the complex
potential in form of the expressions

~~ =C, (8.111)

dw (+b
(8.112)
d( =N ((-c)((-1)2'
242 8. Schematized PAR Flow Models

y
@ 1) C ©
VI
A 01
C v**
~ °
L==:==f::==~~=~A
C
-1
01 B
-b~
0 C
c1l;
A
01

a. b.
Fig. 8.11. Flow of infinite width past a fiap of the wing in the extreme ground
effect: physical and auxiliary planes.

Taking into account that

Qc = Vo** = 2i
1
1'1c dw d/'
d( .. = 'Ir
N c+b
(1 _ c)2' (8.113)

we obtain the following expression for N:

(1 - c)2 cn
N= -. (8.114)
c+ b 'Ir
The mapping function can be constructed from (8.111), and (8.112) in
the form
(8.115)

From the condition IOBI = IBOll = bf , we obtain the following system

of equations:

Ja
r- b dZ -
d( d( = bf exp( -iOr), (8.116a)

r-
Ja
oe dZ
d( d( = 0, (8. 116b)

wherefrom upon introducing a new variable ( = -t we can transform (8.116a)

and (8.116b) into

-N r b
t - b
Ja tn(t + c)(t + 1)2
dt = b
f,
(8.117a)

1 a
00 t - b
-----dt=O.
tn(t + c)(t + 1)2
(8.117b)

From the second equation of the system (8.117a,b), we can find that
 8.2 Unseparated Flows Near Leading and Trailing Edges 243

b= 1o
00 tl-n
(t+c)(t+1)2
dt/
1 0
00 t- n
(t+c)(t+1)2
dt.

Performing integration by using the residue theorem, after some calcula-

tion,
b = -;nc-:-n.. :. (l_---,--:.c)_-_c--.:.(-;-c1_-_cn-,'-) (8.118)
(l-c n ) -ncn (l-c)'
Using this expression, we can combine (8.114) and the first of the two
equations of system (8.117) into one transcendental nonlinear equation for
determining the constant c:

(8.119)

where n = (h/7r and b is determined from (8.118). It should be kept in mind

that 0 < c ~ 1. The magnitude of velocity v~* is determined by the formula

Vo
dw ()
** = dZ c = cn . (8.120)

The pressure coefficient is calculated as

p = 1_ (:~) 2 = 1 _ (2n. (8.121)

For X ~ -00, Y = 0-, we obtain ( ~ c-, TJ = 0+. Therefore, far

downstream in the channel,

p=l- ..c2n ""l-c2n [1--c-~,

2n ( )] (8.122)
c
It follows from (8.115) that
- c+b 1
X"" N In(c-~) = -ln(c-~),
(1 - c)2cn 7r
Then
p"" 1 - c2n + 2nc2n - 1 exp(7rX), X~ -00, Y = 0-. (8.123)
Estimate (8.123) is valid for X ~ -00, Y = 0- and ~ ~ 1+, TJ = 0+.
Thus, finally,
e
p = 1- n "" -2n(~ -1).
It follows from (8.115) that

X"" -N (11-c
+ b) _1_
~-1'

Eventually, one obtains the asymptotics of the pressure on the upper

surface of the wing far from the fiap in the form

X~ -00, Y = 0- .
244 8. Schematized PAR Flow Models

8.3 Envelopes of the Efficiency of Power Augmentation

for a Scheme with a Reentrant Jet
Suggestions to utilize blowing under the main lifting wing of the wing-in-
ground-effect vehicle for improved performance date back to early projects
by Warner; see Belavin [IJ. The technical development and implementation
of the concept for large takeoff weight ground-effect vehieles is due to the
effort of Russian engineers R. Bartini and R. Alexeyev. It is known that
power augmentation by turbulent jets from upstream engines, directed under
the wing, can be used either as a temporary, transition al mode of motion at
takeoff and landing or permanently for cruise.
The efficiency of power augmentation (PAR) is determined by two main
factors. The first of these factors is the magnitude of the required thrust of
PAR engines for a given weight of the vehiele (the inverse quantity equal
to the ratio of the lift R y to the required thrust T can be characterized as
the efficiency of power augmentation, later on designated as KpAR = Ry/T).
The second factor is the thrust recovery, representing the excess thrust which
can be used for forward motion. Gallington et al. [153, 154, 156J proposed a
rather convenient form of diagrams which were called PAR envelopes. The
vertical axis of these envelopes represents the PAR efficiency K pAR mul-
tiplied by a characteristic relative ground elearance h. Such scaling of the
vertical axis of the PAR envelope implies that the order of the lift coefficient
is Gy = 0(1) and that of the drag coefficient Gx = O(h) (or the coefficient
of required thrust) assumed and realized within a one-dimensional model of
the fiow. Hence, the efficiency factor K pAR should be of the order of O(l/h)
and KpARh = 0(1). A simple explanation of the correctness of such a scaling
consists of the fact that in strongly deccelerated fiow under the wing with
a defiected fiap, the pressure is elose to the dynamic head. In this situa-
tion, the lift is directly proportional to the chord length, whereas the drag
coefficient should be of the order of the ground elearance. The horizontal
axis of the PAR efficiency envelopes represents the thrust recovery fraction,
that is, the ratio of excess of thrust over the drag T - R x to the thrust
Tr = (T - Rx)/T = 1 - Rx/T. Having used some of the PAR energy to
generate lift, it is important to have a certain reserve of thrust for furt her
acceleration to cruise speed. Using the scheme of PAR fiow with a reentrant
jet, Gallington and Chaplin [154J analyzed the KpARh - Tr diagram for the
simple case of a fiat plate with a defiected fiap and zero incidence, assum-
ing that there is no leakage from the channel fiow region. In the analysis,
they employed two parameters, the width of the incoming jet in front of
the wing Oj and the width of the jet escaping the channel under the wing
oe, both parameters expressed as fraction of the characteristic ground elear-
ance h (Le., 6j , 6e). As a result, on the diagram were plotted two families
of curves, corresponding to the constant magnitudes of each of these pa-
rameters when varying another one, namely, a (8j = const.,6e = var.) and
(6e = const.,8j = var.). On the PAR diagram presented by Gallington et
 8.3 Envelopes of Efficiency of Power Augmentation 245

al., the domain of points with coordinates K pAR - Tr is bounded below by a

straight line, corresponding to Jr = 0 (zero gap under the flap). The upper
boundary of this domain is obtained by using the momentum equation to
a surface, including flow cross sections on the incoming jet, outcoming jet
and the reentrant jet. This implies a realization of the momentum equation
which corresponds to a minimal possible (within the PAR flow model with
a reentrant jet) width of the incoming jet Jj = Jjmin achieved for a jet orien-
tation angle ßj = 7r. Evidently, the latter case corresponds to maximum of
PAR efficiency (KPAR h)max. Practical tasks that can be set forth to extend
the approach of Gallington et al. are as follows:
• Using PAR efficiency envelopes based on a scheme of a reentrant jet and
applying a mathematical model of the aerodynamics of the wing with small
endplate tip clearance in a strong ground effect, analyze the influence of
different factors (such as the angle of pitch, the curvature of the wing's
lower surface, the aspect ratio, gaps under the endplates, waves on the
underlying surface, etc.) upon the efficiency of power augmentation;
• Consider approaches of modelling the PAR flow scheme with a realization
of the Coanda effect with subsequent determination of power-augmentation
efficiency;
• Analyze the reserves of enhancing PAR efficiency through an appropriate
choice of the geometry and the kinematic parameters of the lifting surface
and use of devices, such as rigid and jet flaps, forward flaps, etc., to control
the efficiency of power-augmented regimes (e.g., by providing shock-free
entry and fuller realization of the suction force);
• Develop a scheme of using PAR efficiency diagrams based on a potential
theory for prediction and maximization of the efficiency of blowing the ex-
haust from the upstream PAR engines. Therewith, account should be taken
of the most significant factors, such as momentum losses due to impact of
the jet upon the ground, the mixing of turbulent jets, the entrainment of
the surrounding air by the turbulent jet ejected by the engine on its way
to the leading edge of the wing, the COalescence of a system of initially
circular jets into an almost two-dimensional jet, and the space orientation
and reciprocal position of engines in the PAR power system in the lateral,
horizontal, and vertical directions, as wen as their locations with respect
to the wing.

8.3.1 Estimate of the Efficiency and the Thrust Recovery

of PAR Based on a Reentrant Jet Scheme

At the beginning, it is worthwhile to clarify the procedure for determin-

ing PAR efficiency envelopes by using simple examples. First, we adopt the
scheme of PAR flow with a reentrant jet (Fig. 8.1). We assume the potential
character of the flow which enables us to use, in particular, the theory of
jets in an ideal fluid [138]. We suppose additionally that the pressure on the
246 8. Schematized PAR Flow Models

boundaries of free jets is equal to the atmospheric pressure (zero perturbed

pressure) and that the velocity is equal to [lj. We can write the PAR efficiency
K pAR as

KpAR = -R C,
y y (
= - 8.124 )
T Ct
where C y = 2Ry / pUl S is the lift coefficient, S is the reference area of the
wing, and Ct = 2T/pUlS is the thrust coefficient. Note that both coefficients
are based upon the dynamic press ure head of the jet and the reference area
of the wing. We introduce thrust recovery jraction Tr

T, - T - R x _ 1 _ R x _ 1 _ C x (8.125)
r - T - T - Ct '

where C x = 2Rx / p~2S is the drag coefficient. Applying horizontal projection

of the momentum theorem to the surface bounded by fiow cross sections at
the nozzle of the PAR engine, the incoming jet (I-I), the reentrant jet (II-
II) and the escaping jet (III-III) , we deduce the relationships between
the thrust (drag) forces and the width of the participating jets. For example,
applying the above mentioned theorem to the fiow cross sections at the nozzle
and (I-I) accounting for the direction of the outward normal, we obtain

T cosBj = p~28jCo, (8.126)

where Co is the wing's root chord. Hence, by using the definition for the
thrust coefficient,

Ct cos Bj = 28j. (8.127)

It follows from (8.127) that, within the assumption of the potential fiow
theory, the horizontal projection of the thrust coefficient is equal to the dou-
bled magnitude of the relative (Le., expressed as a fraction of chord) width
of the jet in front of the wing 8j . 4 Later on, for simplification of presentation,
cos Bj is omitted, so that Ct will be understood as the horizontal component
of the same coefficient. Generally speaking, coefficients Cy and Ct depend on
quite a number of factors (e.g., the fiap defiection angle, the angle of pitch,
the geometry of the wing, particularly on the form of its lower surface, the
magnitude and distribution of the effective gaps under endplates in direction
ofthe chord, etc.). However, for the moment, to illustrate the method oftrac-
ing PAR efficiency envelopes, an example of a wing with no lateralleakage at
zero incidence B = 0 and with a fiap set at an effective gap Jf is considered. 5
In this case, it follows from the previously derived formulas that
4 For the practical situation of blowing turbulent jets this result does not hold
because due to certain moment um losses related to jet impingement upon the
ground, the decrease in the averaged jet velocity and the entrainment of the
surrounding air causes an increase in the jet width in the downstream direction.
5 This case was considered by Gallington et al.
 8.3 Envelopes of Efficiency of Power Augmentation 247

-2
Gy = 1 - 8r , (8.128)
- 2
Gx = h(1 - 8r) , (8.129)
where;Sr = 8rJh, and h is the ground clearance at the trailing edge. Note that
for zero incidence, the lift and the ideal pressure drag acting upon the wing
are due only to the deflected flap. The suction force is not accounted for in
the reentrant jet model.
Taking into account (8.124), (8.125), (8.127), and (8.128), we write the
PAR efficiency and the thrust recovery fraction as

(8.130)

(8.131 )

where 8j = 8j /h.
Excluding the parameter 28j from (8.130) and (8.131), we obtain the
relationship of the PAR efficiency to the thrust recovery fr action for given
magnitudes of 8r:
-2 -
K h = (1 - 8r ) (1 _ T, ) = 1 + ~f (1 - T, ). (8.132)
PAR (1 _ 8r)2 r 1 _ 8f r

On the other hand, excluding the parameter 8f from the same equations
results in the expression of KpARh as a function of Tr for a given 8j . In fact,
it follows from (8.131) that

8f = 1- V28j(1- Tr ). (8.133)
Substituting (8.133) in (8.130),

KpARh -
_1- [1- j25 (1-T,)]' _ P(l- T,) +Tr-l.
28j
j
- 8j (8.134)

Formulas (8.132) and (8.134) define sets of curves of constant values 8f =

const. and 8j = const. Both the formulas and the corresponding diagrams
show how to meet the required PAR efficiency and thrust recovery fraction
magnitudes by an appropriate setting of the trailing edge gap for a given
thrust coefficient or by securing the required thrust coefficient for a given
trailing edge gap. Note, that for a thrust recovery fraction Tr ranging from
zero to unity, the domain ofvariation ofthe PAR efficiency KpARh is bounded
both from below and from above. The lower bound can be determined from
formula (8.132) for Jj -* 0 as
248 8. Schematized PAR Flow Models

(8.135)
As stated earlier, the equation of the upper bound of the PAR efficiency
diagram can be obtained by applying the momentum theorem for a control
surface, including cross sections I-I, II-II and III-III. As a result,

(8.136)

or, in nondimensional form,

(8.137)
Accounting for the mass conservation condition

(8.138)

and substituting a concrete expression for Cx in the case under consideration

by (8.129), we obtain the following expression:

Cx h(l - 8f)2
Ct = 28j = 28f + 1 -cos ßj = 28f + 1
-cos ßj .
(8.139)

A certain magnitude of the reentrant jet orientation angle ßj corresponds

to each pair of parameters (Sj, Sf). Therefore, it is possible to find a magnitude
of this angle, for which the coefficient of required thrust (or the required width
of the incoming jet) would be minimal. This takes place at ßj = 7r, resulting in
the following equation for the upper boundary of the PAR efficiency diagrams:

K h _ Cy (8r) . h _ C y (8r)
( PAR' )max - C . (8) - 28· (8 )' (8.140)
t mm f Jmin f

where
- 2
Ctmin .. = 28 + (1 - 8r) = ~(1 _ S )2
= 28Jmm (8.141)
h f 2 2 f·

As can be shown from (8.128) and (8.141), for the optimal blowing case,
parameter Sr is related to the thrust recovery fr action Tr in the following way:

Tr -_ Ct
- Cx _
-1-2 - - - .
Sf) 2 (1 -
(8.142)
Ct 1 + 8f
It follows from (8.142) that in the case under consideration, nonnegative
values of Tr are reached if
(8.143)

The maximum PAR efficiency, corresponding to optimal blowing, can be

expressed by the formula
 8.3 Envelopes of Efficiency of Power Augmentation 249

1- 8r)
(KpAR ·h)max=2 ( 1+8r . (8.144)

Combining (8.142) and (8.144), we can derive a simple relationship between

the efficiency of power augmentation and the thrust recovery fr action for
optimal organization of blowing:

(K pAR ' h)max = V2(1 - Tr). (8.145)

Note that for a fixed magnitude of the thrust coefficient C t and a varying flap
setting 8r , the efficiency of power augmentation varies between its minimal
and maximal values, and the thrust recovery fraction ranges between its lower
and upper bounds

1- ~ < T, < 1 -
C -
t r -
2(
Vf2
C; _ 1) 2. (8.146)

On the other hand, for a fixed trailing edge setting 8r and varying thrust
coefficient Ct , the thrust recovery fraction changes in the range

1 - 8r
max { 0, 1 - 2 ( 1 _ 8r
)2} ::; T r ::; 1. (8.147)

Based on the preceding analysis, we can conclude that the realization of

the optimal blowing regime associated with maximum PAR efficiency im-
plies "tuning" of parameters 8j (or Cd and 8r in accordance with equation
(8.141). If, for a given setting of the trailing edge gap, the incoming jet is
wider than the optimal width, it spills over the leading edge. In this case, as
indicated in [155], this redundant part of the jet does not participate in the
PAR mechanism of lift. On the other hand, if the jet coming to the wing is
thinner than required by the upper bound, the blowing may become com-
pletely ineffective owing to insufficient pressure recovery at the entrance of
the channel flow. It is remarkable that theoretical analysis of the local flow
around the leading edge, based on a reentrant jet scheme (see paragraph
1.1.1), shows that the corresponding solution exists if
- - 1- 2
20fj 2: 28r + 2'h 1 (1 + v~) ,

the upper bound of the domain of existence of the solution, can be shown
to coincide with the equation for the curve of maximum PAR efficiency. For
the simple case under consideration, v~ = -8r, h1 = 1, wherefrom it can
be seen that the equality

- - 1 - 2
. = 20r
20'Jmm + -(1
2 - or)

coincides with requirement (8.141). The PAR efficiency envelope for the sim-
ple case considered above (i.e., zero gaps under tips of the endplates, a flat
250 8. Schematized PAR Flow Models

1.6,-----,------;----,----,----,

KpAR h

0.2 0.4 0.6 0.8 1.0

Thrust Recovery Fraction
Ct-Cx
Tr= - - -
Ct
Fig. 8.12. Envelopes of PAR efficiency based on a reentrant jet scheme.

lower foi! surface and a zero pitch angle) is presented in Fig. 8.12. In some
cases, for reasons of practical use of the PAR efficiency diagrams, it is con-
venient to trace these diagrams, using a set of constant thrust coefficients
Ct = const., Jr = var. rather than a set of constant incoming jet thicknesses
(8j = const., 8r = var.).

8.3.2 Margins of Efficiency of PAR Based

on a Reentrant Jet Scheme

The simple nonlinear one-dimensional theory of a rectangular wing with end-

plates in the extreme ground effect, introduced in section 4, can be used
together with PAR efficiency envelopes for qualitative prediction of the ef-
ficiency of power augmentation in a range of design parameters, including
the adjusted pitch angle, the height of the endplates, the aspect ratio, and,
the geometry of the lower surface of the wing. Simultaneously, such a math-
ematical model can be used as a tool for development of technical measures
to enhance the performance of the vehicle both for power-augmented takeoff
and cruise.
Taking into account these results, we can write the PAR efficiency and
the thrust recovery fraction as

KpAR h = Cy(Jf , other parameters) (8.148)

Ct
 8.3 Envelopes of Efficiency of Power Augmentation 251

Tr = 1 _ Cx (8f, othe~:arameters), (8.149)

where the lift coefficient Cy and the drag coefficient C x depend on the param-
eters G = 20ep / >'h = 28ep / >., 6 (j = () / h (() the angle of pitch), ;Sf = od h, Of the
effective gap under the fiap, €n = E:n/h, E: the parameters of the form of the
lower surface of the wing, Ct = Ct/h, and Cx = Cx/h. In the potential fiow
theory, the thrust coefficient and the width of the incoming jet are connected
with each other as Ct = 20j (or Ct = 2;Sj). The coefficients C y and C x can be
determined by using the formulas of section 4, the latter without accounting
for the suction force. Sets of curves corresponding to constant magnitudes of
the thrust coefficient Ct and the effective gap of the fiap ;Sf can be obtained
from (8.148)-(8.149) by exclusion of Ct or ;Sf respectively. The lower bound
of the diagram is a straight line with the equation

(KpARh)min = (CC y )_
x cif=O
(l-Tr )· (8.150)

As seen from (8.150), the lower bound can be lifted if one manages to in-
crease the lift-to-drag ratio of the wing in conditions of complete blockage of
the fiow under the fiap (;Sf = 0). In the general case of a wing of finite aspect
ratio, the upper bound of PAR efficiency in the scheme with a reentrant jet
can be found through application of the momentum theorem and accounting
for the mass conservation of incoming and escaping jets in the system. We
adopt a general, though somewhat simplified, scheme of the interaction of
jets due to power augmentation with the wing in the three-dimensional case,
based on the PAR fiow model with a reentrant jet. One should take into ac-
count that outgoing jets escape from under the rear fiap and through the gaps
under the tips of the endplates. Applying the momentum theorem in a pro-
jection onto the horizontal direction and rendering the resulting relationship
nondimensional, we obtain

(8.151)

where Ouj is the width of the reentrant jet,7 and I ep is the lost momentum due
to leakage through the gaps under the endplates. The momentum loss due to
leakage can be determined by chordwise integration of the local momentum
loss at cross sections x = const.:

f
1

I ep = 2 v;p(X)oep(x)dx (8.152)
o

6 As designated earlier, >. is the aspect ratio ofthe wing and Oep is the characteristic
relative gap under the endplates.
7 Note that this and other quantities that may vary spanwise are assumed to be
averaged in the lateral direction.
252 8. Schematized PAR Flow Models

or taking into account the relationship (4.51) between the velo city of the
leakage vep(x) and the span-averaged velo city ofthe channel fiow v(x) in the
concrete case of a steady fiow,

(8.153)

We introduce the distribution of an effective gap under the endplates oep(x) =

o~pLl(x), where, as earlier, C5ep is an effective gap under the endplates at the
trailing edge and Ll(x) is a function ofthe order ofunity that characterizes the
form of this distribution. Taking into account the designation of parameter
G = 2oep />"h, we rewrite (8.152) as

/[1-
1

I ep = h>..G v 2 (x)]Ll(x)dx
o
1

= h>..G / ß(x)Ll(x)dx = h>..GI:p, (8.154)

o
where h is the relative ground clearance, ß(x) = I-v 2 (x) is the span-averaged
I:
pressure under the wing, and factor p is determined by integration:

(8.155)

In the particular case when the gap under the endplates is uniform chordwise,
that is, Ll(x) = 1, the magnitude of I ep can be expressed in terms of the lift
coefficient in the following way:

(8.156)

As a matter of fact, I ep represents the momentum drag and should be added

to the drag coefficient Cx of the wing. Note, that in the PAR fiow model,
employing the scheme of a reentrant jet, the suction force is not realized.
To exclude the width of the jet escaping downstream Ouj from (8.151), we
apply the requirement of mass conservation in the form

(8.157)

where Qep is a nondimensional expression for the rate of air fiow leaking
through gaps under the endplates. This expression can be determined in the
following way:
1

Qep = / VI - v (x) oep(x)dx

2

o
 8.3 Envelopes of Efficiency of Power Augmentation 253

J
1

= ~h G >.. v'ß(x)Ll(x)dx = ~ h >..G Q:p' (8.158)

o
where
(8.159)

Substituting (8.147) and (8.149) in (8.144), accounting for (8.150),

C - J. - 215 >.. Cx + 2Iep - 2Qep cos ßj

t - 2J- [ + >.. (1 _ cos ßj) . (8.160).

Setting ßj = 7l', we obtain the equation of the upper boundary of the PAR
efficiency diagram

(8.161 )

or
(8.162)

- - - C x + 2G(I:p + Q:p)
Ctmin = 2Jjmin = 2J[ + 2 '

C x + 2GI:p
Tr =l- , (8.163)
C tmin
where Ctmin = Ctminlh, bjmin = Jjmin/h, and Cx = Cx/h, and I:p and Q:p are
calculated by using formulas (8.147) and (8.150).
It important to note that for fixed magnitudes of iJ and Sn characterizing
the angle of pitch and the geometry of the channel between the wing and
the ground, as weH as for a given form of the endplate gap distribution, both
the upper and lower bounds of the PAR efficiency envelopes depend on a
similarity criterion G = 2Jep / >"h. Both bounds of the diagram can be shown
to move down with an increase in G. In its turn G increases with an increase
in the relative gap under the endplates Jep , a decrease in the aspect ratio >..,
and a decrease in the ground clearance h.
Figures 8.13-8.16 show some calculated results that illustrate the influence
of different factors on PAR efficiency diagrams.
In particular, Fig. 8.13 presents the influence of parameter iJ (where e is
measured in radians) for no leakage (J~p = 0). It can be seen that an increase
of {j leads to shifting of the PAR efficiency domain downward, so that for
254 8. Schematized PAR Flow Models

2.0,-----,-----r---,---...,----,

KpAR h
Upper boundary
1.6 f - - - - + - - - + - - - 1
Lower boundary I

~
c: 1.2 I---+..=".-..:--I---:::-+----!----I
.~
=
w
a:
(1=0

~ 0.8

0.4 f----=--t...-::--"-+::--=-~-""'"_<:+----->,r---I

9=1

0.0 '--_----L_ _- ' -_ _--'-_ _..I-._....=.

0.0 0.2 0.4 0.6 0.8 1.0
Thrust Recovery Fraction
Ct-C x
Tr= - - -
Ct
Fig. 8.13. The influence of the angle of pitch upon the PAR efficiency (zero gap
under the endplates).

any given pair of magnitudes Gt , 8r , the efficiency of power augmentation

diminishes.
We turn to consideration of the influence of parameter C, which plays the
role of an equivalent (generalized) gap under the endplates. 8 Just for sim-
plicity of the analysis in these calculations, it was assumed that the effective
gap under the endplates is constant chordwise ..1(x) = l.
It follows from Figs. 8.14, plotted for a flat plate at zero incidence and
a deflected flap, that augmentation of parameter G leads to a significant
reduction of the PAR efficiency (the whole diagram .drops down).
It is interesting to evaluate the influence of the curvature of the lower
surface of the wing for jj = 0 and a different G. It is seen from the PAR
efficiency envelope (see Fig. 8.15) for a curved wing that an increase in the
relative curvature Je = 8e /h results in an increase in KpARh, especially in
the range of larger magnitudes of the thrust recovery fraction. Similarly to a
wing without curvature a nonzero gap under the endplates leads to reduction
of the efficiency of blowing for a wing with endplates.
It is not difficult to explain a decrease in the PAR efficiency for a nonzero
angle of pitch jj > 0 and an improvement in efficiency for a wing curved
upward. As the matter of fact, within the scheme of the reentrant jet an
increase in the angle of pitch results in an increase in the pressure drag. At

8 In fact, the same rate of leakage from under the endplates allow for different
combinations of the width of the effective gap under the endplates, the wing's
aspect ratio and the relative ground clearance.
 8.3 Envelopes of Efficiency of Power Augmentation 255

I
U
i
I I 1
I
1

KpAR h I (
1.2 r - i_ _ -+-G~=:-O--+_-I' u~~e: ~~u_n~~ry •
Lowe, bou....~! ;

I ä= 0 I

04 1----
• I

i 0.5

0.0 1'------'--_-'--_---..l_ _ L-___=_~

0.0 0.2 0.4 0.6 0.8 1.0
Thrust Recovery Fraction
Ct-C x
Tr= - - -
Ct
Fig. 8.14. The influence of the endplate gap similarity parameter G = O~p/ >"h
upon the efficiency of the PAR regime.

0.0 L I_ _- - - ' -_ _ --1-_ _-'--_ _--'---_--'-'

0.0 0.2 0.4 0.6 0.8 1.0

Thrust Recovery Fraction
C t- C x
Tr= - - -
Ct
Fig. 8.15. PAR efficiency envelopes for a wing with a curved lower surface; Oe is
the relative curvature.

the same time (at least for 5f not elose to unity) the relative increment of Cy
due to the angle of pitch is insignificant because the fiow under the wing is
already deccelerated due to the defiection of the fiap.
256 8. Schematized PAR Flow Models

Curving of the lower surface at a zero pitch angle (0 = 0), we can gain
the same increment of lift as that of a flat wing with a nonzero pitch angle.
At the same time, the corresponding press ure drag becomes much smaller
the because horizontal component of the pressure force acting upon rear
part of the wing is canceled by the pressure force (thrust) acting upon the
forward part of the wing. Practically, because the vehicle normally has a
certain (design) angle of pitch adjusted for cruise conditions, it is desirable,
when using power augmentation at takeoff, to generate effects analogous to
the action of curvature. This can be done by forward flaps.
Figure 8.16 illustrates the influence of a forward flap upon the PAR effi-
ciency of the wing in the extreme ground effect. In this example, the forward
flap has a chord, constituting 20% of the wing's root chord Co' Deflection of
the flap was chosen to locate its leading edge tip at the level of or lower than
the hinge of the rear flap. In the example under discussion, the former posi-

°
tion is achieved for a flap deflection angle Oie = (1 - ble)O jble . For example,
when = 2°, Oie should be not less than 8°.
It follows from consideration of Fig. 8.16 that for nonzero angle of pitch
and forward flap deflected downward, the efficiency of power augmentation
increases for a given 0. If Oie> (1 - ble)O jb le , the efficiency domain shifts
upward, so that the power augmentation efficiency becomes higher than that
of a wing at zero incidence with a deflected rear flap. Essentially, this effect
is due to the fact that, when the leading edge of the forward flap is lower

2.011-T(~=~=)l
KpAR h Upper boundary
1.6 ,,=:::---+~;:--+---I

~
c 1.2 ' ,

~
W
, 0:5, ' ,

~ 0.8 f---'+~'"7'+,-,-+-""""'----'k-~--1
9= 0' , "
" '
0.4
........... i
, ''l,'
I ~"'~ ....

0.2 0.4 0.6 0.8 1.0

Thrust Recovery Fraction
Ct- C x
Tr= - - -
Ct
Fig. 8.16. PAR efficiency envelopes for a wing at an angle of pitch with a deflected
forward flap (zero gap under the endplates).
 8.4 Discussion of a Mathematical Model of PAR with the Coanda Effect 257

than the hinge of the rear fiap, the resulting pressure thrust re duces the total
pressure drag of the wing.

8.4 A Discussion of a Mathematical Model of PAR

with the Coanda Effect

The PAR model based on a reentrant jet scheme does not account for the
suction force at the leading edge. At the same time experiments show that
the incoming jet has a tendency of enveloping a rounded leading edge. For a
relatively thin jet spilling over the edge, this phenomenon can be identified
with the Coanda effect. Gallington [155] and Krause et al. [156] remark that
PAR theory,9 on average, underestimates the efficiency of power augmenta-
tion predicted by experiments, because it completely neglects the Coanda
effect. In this connection, an attempt can be made to develop a PAR fiow
scheme which would incorporate a manifestation of the Coanda effect. One
model of this type is proposed herein for further discussion and evaluation.
The main idea of this model is that, whereas the angle ßj of the overspilling
jet with respect to the x axis is governed by the moment um law, the detach-
ment of the jet takes place tangentially to a rounded leading edge at a certain
point with an abscissa X sep (see Fig. 8.17). In this case, the suction force is
realized only on the "wetted" part of the surface of the leading edge.
To calculate the suction force acting upon the "wetted" part of the
rounded leading edge, we consider a local fiow in the immediate vicinity of
the leading edge with a rounded nose. It can be shown that for a wing with

Fig. 8.17. The definition of a PAR fiow model with a partial realization of suction
force (the Coanda effect).

9 Based on a reentrant jet model.

258 8. Schematized PAR Flow Models

analytical cross sections at z = const,10 the rounded leading edge at distances

of the order of its radius of curvature PIe is approximated by a semi-infinite
parabolic cylinder.
As derived in Rozhdestvensky [44], the press ure coefficient pi on an oscu-
lating parabola has the following quasi-steady form: l l

(8.164)

where parameters U1 and U2 characterize the velocities of the local incoming

and the circulatory flows around the rounded leading edge, thus determining
the position of the stagnation point. These parameters have to be determined
by matching with an appropriate outer flow field; X = Xle/ PIe is the stretched
local abscissa and Xle is the unstretched local abscissa.
It should be noted that for an analytical wing section, PIe = O(b;), where
bt is relative thickness of the wing section. On the other hand, within the
assumptions of the theory of the extreme ground effect adopted in this book,
the relative thickness bt of a wing section is of the order of the relative
ground clearance O(h); see (2.1). Therefore, for analytical foils, the radius of
curvature is assumed to be of the order of 0(h 2 ).
We calculate first the suction force coefficient when the flow past
the osculating parabola is not separated. Representing the press ure
coefficient on the surface of the wing section near the rounded leading edge
as

i XU'f 2U1 U2 v'X Ui

p = p = 1- X + 1/2 =f X + 1/2 - X + 1/2' (8.165)

we can observe that the two first terms of (8.165) are associated with sym-
metrical flow around the parabola and therefore make no contribution to the
suction force. The third and the fourth terms are due to the circulatory flow
around the edge.
Hence these terms have to be taken into account when calculating the
suction force. It is easy to see that the third term of (8.165) changes sign,
when passing from a point on the upper branch of the parabola to a mirror
reflection of this point on the lower branch of the parabola. This means that
the sum of the longitudinal press ure forces due to the third term, acting upon
parabola, should be equal to zero. In connection with these conclusions, the
calculation of the suction force can be reduced to the integration of just the
fourth term of (8.165), Le.,

es = J U? dy
X + 1/2 ds ds on the parabola, (8.166)

10 Most of the practical foil families have analytical descriptions.

11 This is valid in a wide practical range of Strouhal numbers.
 8.4 Discussion of a Mathematical Model of PAR with the Coanda Effect 259

where ds is an arc element. Taking into account the equation of a parabola

Y = ±PleV'2X, we obtain

Cs = Pie 1
o
00
X
U2
2
+1
/2 dY = PleUi, h 1 0
00
1
/VdX
(X + 1/2)v X

= 4Ui,Ple arctan v'2X1~ = 27rUi,Ple, (8.167)

where Y = Y/ Pie. To find U2 we match the asymptotic expansions for the

velo city on the foil in three characteristic regions: D h 2 the immediate vicinity
of the leading edge of the order of Pie = O(h 2 ), the larger vicinity D e of the
order of O(h) of the leading edge, and the channel flow D I under the wing.
In D h 2 expression for the velocity on a parabola, corresponding to (8.165), is

(8.168)

In D e by using the formula (2.39) of Section 2, the flow velocity of the relative
motion in points on the leading edge can be written as

Ocj;le = -U(t)
OX
+h 0 al ox + O(h 0 ) ,
o<pae 2 (8.169)

where <Pae is determined by formula (2.43). Matching (8.169) with the channel
flow velo city v(x, t) and accounting for the asymptotics of (2.47), leads to the
following expression for al:

al = h*(l, t)[U(t) + v(l, t)]. (8.170)

On the other hand, recalling that the asymptotics of o<pae/ox in the imme-
diate vicinity of the edge is described by the expression (2.45), we can match
the solution (8.169), obtained in D e with description of the flow velocity
(8.168) in D h 2. The latter matching results in the following expression for
parameter U2 :
U2 = al [U(t) + v(l, t)]. (8.171)
J27rh*(1, t)Ple
Now, we can write the final expression for the suction force coefficient, taking
into account (8.167) and (8.171), as

Cs = 27rUi, Pie = 27r h*(l, t) [U~~ + v(l, t)]2 = hh 1[U(t) + v(l, t)]2, (8.172)

where h 1 = h1(0, z, t)/h. It can be seen that the expression for the suction
force coefficient, obtained by direct integration of press ure forces, acting on a
parabolic leading edge, coincides with that determined by using the strength
of the square root singularity of the perturbed velo city at the leading edge
of an infinitely thin wing section.
260 8. Schematized PAR Flow Models

Later on, it is assumed that the suction force is realized only on

the "wetted" part of the osculating parabola. Then the suction force
fraction "'s
(that is, the degree of realization of the suction force) can be
calculated by direct integration of the local pressure coefficient pi, given by
formula (8.165), along the corresponding "wetted" part ofthe parabolic edge.
This integration leads to the following result:

"'s = ~ (1 + ~ arctan y'2Xsep), (8.173)

where "'s = crrtialjCs is the suction force fraction, ranging from zero to
unity. At the same time, the angle ß;e p is calculated with use of the equation
of a parabola, i.e, Y = y'2Xsep as

ßjsep = arctan ~.
1 (8.174 )
Y 2Xsep
Comparing the last two equations, we can see that the angle of detachment
of the jet from the contour of the parabola can be expressed by the suction
force fraction by using the following expression:

(8.175)

Thus, a PAR flow scheme can be composed of one model parameter (the
suction force fraction "'s)12 and accounting for Coanda effect. To determine
the PAR efficiency diagrams for this case, we can write the corresponding
momentum equation as
- Ct - - C x - crrtial
Ct = -h = 20j = 20f + 1 ßsep . (8.176)
- cos j
Employing the suction force fraction "'s, we can rewrite the previous equation
in the following way:

(8.177)

Supplementing this equation with expressions for the PAR efficiency

K pAR hand the thrust recovery fraction Tr, in the simplest case of zero
gap under the endplates, we obtain

(8.178)

- - 2
Tr = 1 _ Cx - Cs = 1 _ (1 - "'s)Cx = 1 _ (1 - "'s)(1 - or) (8.179)
Ct 20j Ct '

where Ct can be calculated with help of (8.169).

12 In the reentrant jet scheme the role of such a model parameter is played by the
angle of inclination of the jet with respect to the downstream direction.
 8.4 Discussion of a Mathematical Model of PAR with the Coanda Effect 261

We can use the relationships (8.169-8.171) that feature the parameter

1 ~ ks ~ 0 to plot PAR efficiency diagrams accounting for the Coanda effectj
see Fig. 8.18.
As seen from this figure where both (reentrant jet fiow and Coanda fiow)
types of PAR efficiency envelopes are shown simultaneously, the overall pre-
diction accounting for the Coanda effect gives higher magnitudes of K pAR
and somewhat higher magnitudes of the thrust recovery fraction for the same
combination of 8j and 8f . We can also see from the observation of Fig. 8.18
that for certain combinations of the thrust coefficient and the rear fiap set-
ting, the efficiency of power augmentation with the Coanda effect reaches
its maximum value. As follows frorn (8.179), at a fixed fiap setting 8f , the
minimal magnitude of the thrust coefficient is attained at ks = 0.258 and can
be found from the expression
- - - - 2
Ctmin (ör) = 2Öf + 0.439 (1 - ör) . (8.180)
The corresponding thrust recovery fraction and the (maximal) PAR efficiency
depend On 8f and can be calculated by the formulas

PAR model
KPAR h
based on
1.6 Coandaflow
scheme

~
c:: 1.2
.~
IE
w
I%:
0.8
~
PAR model
0.4 based on
re-entrant
jetscheme
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Thrust Recovery Fraction
Ct-C x
Tr= - - -
Ct
Fig. 8.18. Comparison of PAR efficiency envelopes based on two fiow models:
a reentrant jet model and the Coanda fiow model with partial realization of the
suction force.
262 8. Schematized PAR Flow Models

Table 8.1 PAR efficiency and thrust recovery for optimal settings of a rear flap
8r/h 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Ctmin 0.617 0.681 0.747 0.815 0.886 0.958 1.033 1.11
(KPARh)max 1.583 1.409 1.255 1.116 0.991 0.877 0.772 0.676
Trop • 0.132 0.303 0.441 0.554 0.646 0.721 0.783 0.833

Table 8.1 presents the adjustment magnitudes of the rear fiap settings and
the (minimal) thrust coefficents, as weH as the corresponding magnitudes of
the PAR efficiency and the thrust recovery fraction.
9. The Aerodynamic Efficiency of a Wing
in the Extreme Ground Effect

9.1 Optimal Wing-in-Ground Effect

The theory of a wing in the extreme ground effect, as developed herein,

enables us to formulate a number of extremal problems, which are of both
theoretical and practical interest.
We consider the conditions of the minimality of the induced drag of a
lifting system in the extreme ground effect in Munk's sense [159]. We as-
sume that a thin vortex wake extending behind the trailing edge represents
a cylindrical surface with a generatrix parallel to an unperturbed underlying
boundary. This assumption permits us to assurne that, at a sufficient distance
downstream from the lifting system, the flow is elose to two-dimensional in
the plane normal to the direction of motion (a 'frefftz plane). It is known
that the induced drag coefficient can be expressed by the integral over the
whole 'frefftz plane ST of the kinetic energy of a fluid per unit length in the
direction of motion, see Ashley and Landahl [161] and Barrows and Widnall
[160]:
(9.1)

In formula (9.1), all quantities were rendered nondimensional by using

some characteristic (cruise) speed Uo and the length of the root chord Co. The
differential dST represents a differentially small area element in the 'frefftz
plane, related to the square of the root chord. Quantities >. and l represent,
respectively, the aspect ratio and the relative span of the lifting surface, <p is
the velocity potential of the perturbed motion of the fluid.
Using Green's theorem, we can transform integral (9.1) into a contour
integral:
>. 1 a<p
CXi = l2 l.cE <p an dl, (9.2)

where CE is the contour in the 'frefftz plane that incorporates a contour

Cl that eneloses the wake and a contour C2 that coincides with the line of
intersection of the ground and the 'frefftz plane. Taking into account the
tangency condition on the ground, we can transform (9.2) into an alternative
form

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
264 9, Aerodynamic Efficiency

(9.3)

where 1:, is the contour of the wake passed in one direction. Within the as-
sumptions of linear theory, the expression for the lift coefficient for a wing of
small aspect ratio with a curvilinear lateral axis can be written in the form

,X
Cy = l2 J'Js(p)cos(n,y)dS=7,2
r 2'x J'Js
r (aCP_
ax - acp+)
ax cos(n,y)dS

2,X r 2,X
=7,2 Jc(cp+-cp_)cos(n,y)dl=7,2 Je r(l)cos(n,y) dS,
r (9.4)

where n is anormal to a curvilinear cut that represents the wake in the

Trefftz plane. To determine the conditions of the minimum of the induced
drag for a given magnitude of the lift coefficient, we find a minimum of the
function
(9.5)
where ,x* is the variational Lagrange multiplier. Then, in accordance with
variational calculus, a variation of the function V must be equal to zero, that
is,
(9.6)

Turning to contour integration with the help of Green's formula, we obtain

(9.7)

Taking into account (9.7), we obtain

2'x
8V = 7,2 J [aCP
an -,x * cos(n,y) ] 8rdl = O. (9.8)
c
For an arbitrary variation of 8r, the equality (9.8) is possible only if

~~ = ,x* cos(n, y). (9.9)

It can be seen from (9.9) that ,x* represents the vertical downwash Wo in the
middle of the wake. Thus, the expression (9.9) can be rewritten as
acp
an = Vn = Wo cos(n, y), (9.10)

which corresponds to the following theorem (Munk [159]): The induced

drag of a wing is minimum if the normal component of the induced
 9.1 Optimal Wing-in-Ground Effect 265

downwash at each point is proportional to the angle of inclination

of the lifting element at this point. Taking into account (9.10), the
expressions for the coefficients of the lift and the induced drag of an optimal
wing take the form

Gx ,
J
= Ti
)..2 8cp
T(l) 8n dl =
)..
[2W o
/l/2
-l/2 T(z) dz, (9.11)
t:.

Gy = Ti J2)..
T(l) cos(n, y) dl
2)../l/2
= Ti T(z) dz. (9.12)
t:. -l/2

Using Prandtl's representation of the relationship between the lift and the
induced drag coefficients, we obtain

(9.13)

where
4 /l/2
J-l = -1-
2 - T(z) dz. (9.14)
7r Wo -l/2

The product )..J-l = )..e is called the effective aspect ratio of the wing and can
be used as a measure of the aerodynamic efficiency of the lifting system.
Now, we can find the form of the optimality condition for a wing in dose
proximity to the ground. Recalling that in the asymptotic theory, discussed
herein, both the gap and its chordwise derivatives are assumed small
8h* 8h*
h*(x, z), 8z «1, (9.15)
8x'
and by using the formula for the determination of induced downwash at points
of the wake in steady motion, we obtain the following form of the optimality
condition for a laterally curvilinear wing in the extreme ground effect:

8cp
8n = V n :::' O!Wl
8 [ *
= 8z h (0, z) 8z = Wo = const.,
8T] (9.16)

where h* (0, z) is the distribution of the distances of points of the trailing edge
from the ground and T = T(z) is the distribution of the circulation along
the trailing edge of the wing. Integrating (9.16) taking into account that the
loading must vanish at the tips of the wing,
z z-G
Topt = Wo / -l/2 h*() dz, (9.17)
0, Z

where
G = /l/2 zdz /l/2 dz .
(9.18)
-1/2 h*(O, z) / -1/2 h*(O, z)
266 9. Aerodynamic Efficiency

The constant C was determined by using the requirement that the circulation
should vanish at the extremities of the wake (wing).
If the lateral curvature of the wing's surface is negligible, Munk's theorem
is reduced to the requirement that the downwash should be constant span-
wise. In the limit of vanishing clearances hand taking into account (9.10),
the requirement of optimality of a wing in the extreme ground effect can be
reduced to

x = O. (9.19)

Integrating (9.19) and taking into account that the tip loading should be
equal to zero, i.e., <Pll (0, ±>"/2) = 0, we obtain

(9.20)

Thus, the optimal wing in the extreme ground effect has a parabolic
spanwise distribution of circulation. This conclusion reveals a distinction
of the aerodynamics of the extreme ground effect from that in an unbounded
fluid, where the optimal wing has an elliptic loading distribution spanwise. It
is also compatible with the results of de Haller [134], who obtained an exact
solution for an optimal wing in the Trefftz plane in terms of an infinite series of
elliptic functions and demonstrated by calculations that, when the clearances
diminish from infinity to zero, the optimalloading distribution changes from
elliptic to parabolic. Returning to the solutions derived in paragraph 3.5, we
can see that a semielliptic wing is optimal for any aspect ratio and
that a flat wing of small aspect ratio in the extreme ground effect is
optimal independently of its planform. Similar conclusions follow from
the theory of a lifting line(s) in the extreme ground effect, set forth in section
10, where parabolic spanwise loading also furnishes the minimal induced drag
for a given lift.
Using the results of linear theory, stated in paragraph 3.4, we can study
the requirements for the optimality of a rectangular wing of an arbi-
trary aspect ratio. In a sufficiently general case, the spanwise distribution
of the circulation at the trailing edge of a rectangular wing for h -t 0 can be
written as follows:
CXJ
7r
<pll(O,Z) = LCYnCosqnz, qn = ~(2n + 1). (9.21 )
n=ü

Comparing expressions (9.20) and (9.21), we can determine the coefficients

subject to the optimality condition:

(9.22)
 9.1 Optimal Wing-in-Ground Effect 267

Due to the fact that the coefficients an refiect the specifics of a concrete
problem, condition (9.22) enables us to find the optimal spanwise distribution
of the different parameters, such as aerodynamic twisting, jet fiap momentum
distribution, etc.
For example, we can find such a distribution of the angle of pitch for which
a rectangular wing of arbitrary aspect ratio has a minimum induced drag.
From the solution of the corresponding fiow problem for an arbitrary spanwise
distribution of the angle of pitch B(z) = B0 8(z), presented in paragraph 3.4,
it follows that at a point on the trailing edge

(9.23)

where
Bn = >:2/>./2 8(z) cos qnz dz.
->'/2
The coefficients Bn in accordance with (9.22) can be derived from the equation
BoB n qn 4w o( _l)n
Bn = -h2'
qn
tanhqn tanh - = -
2
A 3h
qn
Consequently,

B _ 4w o( -l)n f(_l)n cosqn z

(9.24)
n- ABo n=Ü qn tanhqn tanh(qn/2)'

Therefrom, the optimal spanwise distribution of B(z) for a rectangular wing

(aerodynamic twist) of arbitrary aspect ratio A is described by the following
equation:

For a wing of small aspect ratio A --+ 0 (qn --+ (0), this expression yields

8 () ~ 4w o ~ (-l)ncosqnz = Wo
opt Z - AB L..t B
o n=Ü qn 0

and therefore, no aerodynamic twisting is required for the optimiza-

tion of a reet angular wing of small aspeet ratio.
For a wing of large aspect ratio A --+ 00 (qn --+ 0),

8 opt (z) = 8w o
ABo
f (-l)n~OSqnZ = WO (A 2 _ z2).
qn Bo 4
(9.25)
n=Ü
It follows from (9.25) that an optimal rectangular wing of large aspect
ratio in the extreme ground effeet should have a parabolie distri-
bution of the angle of pitch. It can be shown that a noticeable gain in the
268 9. Aerodynamic Efficiency

lift-to-drag ratio ean be aehieved only for wings of suffieiently large aspeet
ratios.
In another example, we find the optimal distribution of the jet mo-
mentum along the trailing edge of a reet angular wing with a jet fiap. Based
on the results obtained in paragraph 6.2 of section 6, at points of the trailing
edge of a jet-fiapped reet angular wing-in-ground effeet,

(9.26)

where
v'2j)../2
an =T JCJ-L(z) eosqn z dz.
-),,/2

Aeeording to (9.26) and (9.22),

from whieh
4w ( _1)n
o
an = --;:-"'---::'---
Arq; tanh qn '
so that the optimal distribution of the jet velo city distribution along the
trailing edge is given by the following expression:

lCJ-L(Z)
2
= 4w o
Ar
t
n=ü
(_1)n eosqn Z .
q; tanh qn

Setting the aspect ratio to infinity, we derive a parabolie distribution of the

jet velocity in the form

fl = ;~ (~2 _z2).
For a wing of small aspect ratio,

(9.27)

Bearing in mind that the derivative of S(z) ean be summed up in a closed

form (see Gradshtein and Ryzhik[147], p. 52),

dS
dz
=_ t
n=ü
(_1)n sinqn z
qn
= _~ In tan(~ +
21r 4
1rZ)
2A'
(9.28)

and integrating (9.28), we ean express the right-hand side of (9.27) by the
Lobaehevsky function L(u) (see Gradshtein and Ryzhik [147]),
 9.2 The Lift-to-Drag Ratio of a Wing in the Ground Effect 269

A -+ 0,

This formula describes an optimal law of ejection of air along the trailing
edge of a rectangular wing with a jet Hap.

9.2 The Lift-to-Drag Ratio of a Wing

in the Ground Effect

The economic efficiency of wing-in-ground-effect vehides is directly related

to the lift-to-drag ratio, which can be viewed as one of the principal design
parameters. In what follows, some estimates of the lift-to-drag ratio of a
lifting system in dose proximity to the ground will be discussed on the basis
of the simple Prandtl's formula and the relevant results of the asymptotic
theory set forth herein.
First of all, we write the induced drag coefficient of a lifting surface in the
conventional form, introduced by Prandtl,

(9.29)

where the function J.L(h, A) characterizes the influence of the ground and the
wing aspect ratio upon the induced drag coefficient of a wing for a fixed
magnitude of the lift coefficient. The quantity Ae = AJ.L can be interpreted
as an effective aspect ratio. As follows from the results of paragraph 3.4, the
coefficient J.L of a rectangular wing with a Hat lower surface in the extreme
ground effect can be calculated from the formula

71'
qn = -;x(2n + 1). (9.30)

It follows from (9.29) and (9.30) that for h -+ 0 and Cy = const., the mag-
nitude of the induced drag coefficient diminishes proportionally to h as the
wing approaches the ground. Hence, it is dear that for a lifting system op-
erating near the ground, its effective aspect ratio depends, at least,
upon one specific new parameter, the relative ground clearance.
From a practical viewpoint it is convenient to have at hand some sim-
ple formulas for evaluating the lift-to-drag ratio K = Cy/Cx as a function
of Cy for the analysis of the existing margins for enhancing the maximum
magnitude of K max or K for a given Cy .
Taking into account (9.29), we can write the expression for the aerody-
namic fineness (lift-to-drag ratio) in the form
270 9. Aerodynamic Efficiency

(9.31)

where Cxo is the viscous drag coefficient.

Similarly to the unbounded flow case, we can augment the lift-to-drag
ratio of the lifting system by
• increasing the aspect ratio
• decreasing the induced drag by securing optimal spanwise distribution of
the circulation by choosing an appropriate planform of the wing, as weH
as the distribution of the angle of attack (pitch) in the lateral direction;
• realizing the leading edge suction force; and
• reducing the viscous drag of the lifting system.
In addition, specific features of ground effect aerodynamics indicate that the
magnitude of the lift-to-drag ratio can be increased by bringing the wing
closer to the underlying surface (decreasing the relative ground clearance h)
andjor mounting endplates at the tips of a wing.
The maximum of the function K takes place at a certain optimal magni-
tude of the lift coefficient
(9.32)
and the corresponding magnitude of the maximum lift-to-drag ratio can be
found by substituting (9.32) in (9.31):

K max = ~j~)..Jj. (9.33)

Xo

Whereas the lift-to-drag ratio K specifies the aerodynamic efficiency of

the lifting system, the product K Uo (where Uo is the design cruise speed) is
closely related to its range of flight. Because the speed is inversely propor-
tional to the square root of the lift coefficient, to find the maximum K Uo,
we can consider the foHowing function:

The maximum of this function takes place at a certain magnitude of the lift
coefficient, which differs from (9.32), namely,

(9.34)

The magnitude ofthe lift-to-drag ratio, corresponding to the maximum range,

is
(9.35)
 9.2 The Lift-to-Drag Ratio of a Wing in the Ground Effect 271

We can draw some practical conclusions from the preceding results. First of
all, the ratio of speeds, corresponding to the maximum lift-to-drag ratios for
the ground proximity and an unbounded fluid is given by the expression
(Uo)K max 1
(9.36)
(Uo)Kmax ~'
which shows that, the more one gains in the lift-to-drag ratio by flying closer
to the ground, the less the cruise speed of the vehicle. Second, going for
a larger range entails a certain loss in the lift-to-drag ratio, compared to
its maximum possible magnitude. This loss can be determined (in relative
terms) by dividing expression (9.35) by expression (9.33):

KICKUolmax =
K max
v'3
2
~ 0.866. (9.37)

It is interesting to evaluate the reserves of enhancement of the maximum

lift-to-drag ratio near the ground h --t 0 in comparison with the same property
in unbounded flow, h --t 00. Suppose, that we compare optimal flat wings of
the same relative span l. Keeping in mind that for an optimal wing operating
out of the ground effect, h = 00, the loading distribution is elliptic, i.e.,

(9.38)

we can derive from (9.14) the following expression for the coefficient J.l = J.l~t:

J.l~t =
7rl
4;m
Wo
1 J~ -
1 2
/
-1/2 4
z 2 dz = Fm
7rWo
1 ~d( 1

-1
= Fm.
2wo
(9.39)

On the other hand, we can calculate the downwash Wo, induced in the Tr-
efftz plane by an optimal wing in unbounded fluid by using the following
relationship:

Wo
1
= 27r V.p.
/1
-1
dF d(
d( z _ (
Fm
= - 27r V.p.
/1 ( -1 ~Z
d(
_(
Fm
= 2' (9.40)

Combining the two preceding expressions, we obtain the well-known Prandtl

result: for an optimal wing in an unbounded fluid,
C2
Cx I
--*.
= 7r /I (9.41 )

The magnitudes of the maximum lift-to-drag ratio and the corresponding

(optimal) lift coefficient for this case are

7r AJ.l~t _ 1 r;>:
----c::- - 2Yc::'
272 9. Aerodynamic Efficiency

As shown in paragraph 9.2, for an optimal wing in the extreme ground effect
(h -+ 0), the distribution of the loading is parabolic. Therefore, the corre-
sponding distribution of the circulation can be written as

(9.43)

Substituting this formula in (9.4), we obtain the following expression for the
factor f..L = f..Lopt when h -+ 0:

f..Lopt = 4;m
1fl Wo
1 (~
12
/
-1/2 4
_ z2) dz = rml
21f Wo
1 -1
1
(1 _ (2) d( = 2rml.
31fWo
(9.44)

Now, we can derive a relationship between the downwash Wo and the param-
eter rm , using formula (9.20) for the downwash behind the (straight) trailing
edge of a wing in the extreme ground effect:

(9.45)

Finally, using the two preceding equations, we find that

l 1
f..Lopt = 31fh = 31fh 1 ' (9.46)

where hl = hll represents the distance from the trailing edge to the ground,
related to the span of the wing.
Eventually, the capacity of the lifting surface without endplates to take
advantage of the closeness to the ground can be evaluated with the help of
the following formula:

(9.47)

Note that this formula was obtained by assuming that the loading is optimal
both for h -+ 0 and h -+ 00.
The corresponding ratio of (optimal) lift coefficients in the extreme ground
effect and an unbounded fluid has a similar form:

(9.48)

Employing formulas (9.32) and (9.34), we can see that the cruise speed cor-
responding to the maximum range always exceeds that corresponding to the
maximum lift-to-drag ratio. The ratio of these speeds is constant, and

(9.49)
 9.2 The Lift-to-Drag Ratio of a Wing in the Ground Effect 273

Therefore, the expected gain in range, when flying in the extreme ground
effect as compared to flying far from the ground (without accounting for the
variation of the density of air) can be estimated by using the ratio

(9.50)

To evaluate the margins connected with the realization of the suction force,
it is beneficial to know the ratio of magnitudes of the maximum lift-to-drag
ratio with fuHy realized suction force K max to that with no suction force
Kmax . We take example of a rectangular wing for h --+ O. If the flow near
the leading edge is not separated, suction is realized, and the factor J.l can
be determined by equation (9.30). If there is no suction force, which may
happen due to improper profiling of the leading edge, the factor J.l for a wing
of rectangular planform can be found in the form

__ ~ ~ tanhqn tanh(qn/2) _ Cz (9.51 )

J.l - J.l 7f h)...3 L... q4 - 7f A
n=O n

Taking into account relationships (9.30) and (9.51), the loss in the lift-to-drag
ratio when the suction force is not realized, can be assessed by the formula

~ tanh qn tanh( qn/2) / ~ tanh 2 qn tanh 2 (qn/ 2)

2L... 4 L... 4 .
n=O qn n=O qn
(9.52)
For moderate aspect ratios, expression (9.52) can be approximately
rewritten as
(9.52)
yltanh(7f/A)tanh(7f/2A)
For rectangular wings of smaH aspect ratios, it follows from expression (9.52)
that
~max = y12. (9.53)
K max
This is exactly the same result as that for unbounded flow. In Fig. 9.1, the
calculated fraction K max / K max is plotted versus the aspect ratio A for h « l.
In the same figure, the dashed line represents calculated data, corresponding
to the motion of a rectangular wing in an unbounded fluid; see Belotserkovsky
and Skripach [130].
We can deduce from the above analysis that for wings of moderate and
large aspect ratios, realization of the suction force results in a larger increment
of the maximum lift-to-drag ratio than in an unbounded fluid. Therefore,
considerable attention should be paid to profiling the leading edge of the
wing in the extreme ground effect.
274 9. Aerodynamic Efficiency

5.0
i
Kmax
I I
Kmax 4.5

4.0
/
3.5

3.0
I .'
/
,/
/
.}' /
h ----7 0

2.5
V
/'
----1---
2.0 - -
V h = co
/ ---
1.5 ~

1.0
I
o 2 3 4 5 A 6

Fig. 9.1. The ratio of the maximum lift-to-drag ratio of a rectangular fiat wing
with suction force to that without suction force versus the aspect ratio (solid line:
extreme ground effect; dashed line: out-of-ground effect [130]).

As already discussed, it is possible to improve the aerodynamic quality

of lifting systems near the ground by mounting endplates on the tips of
the wings. In this case, the efficiency factor J1. should be replaced by J1.ep.
Restricting consideration to the case of lower endplates, we can find the
latter coefficient as

(9.54)

where hep/h is the ratio of height of the endplate to the ground clearance and
the factor K: ep = Cyep/Cy characterizes the relative augmentation of the lift
coefficient due to the influence of the endplates and can be calculated by using
(6.30). The maximum relative gain in the lift-to-drag ratio, resulting from the
installation of lower endplates on a rectangular wing, can be evaluated by the
formula 1
K;;fax
(9.55)
K max Jl + (2hep + ())/)..'
where the denominator takes account of the augmentation of the wetted area
of the wing due to mounting of the (lower) endplates and () is the adjusted
pitch angle in radians.
Returning to Fig. 6.4, we can see that for a wing of a small aspect ratio
in the presence of endplates, the gain in aerodynamic quality can be quite
noticeable. For example, if the height of the lower endplates at the trailing
edge constitutes 60% of the ground clearance, then, for a reet angular wing of

1 This formula assumes that the wing is fiat and the tips of the endplates are
parallel to the ground.
 9.2 The Lift-to-Drag Ratio of a Wing in the Ground Effeet 275

1.4
1

opt
Kmax
Kmax 1.3

1.2 f---j---+--f--+-+-+-

I
1.0 L '_ _-=::::::=:..--'-_--'_ _-L-_--'-_
o 2 3 4 5 6

A.
Fig. 9.2. The relative inerease of the lift-to-drag ratio of a reet angular wing in the
extreme ground effeet due to minimization of the indueed drag: a. optimal spanwise
distribution of the piteh angle; b. optimal jet momentum distribution when using
a jet Rap.

aspect ratio A = 0.7, flying at relative distance from the ground h = 0.07 at
incidence () = 0.05, we can expect augmentation of the order of 25% of the
maximum lift-to-drag ratio compared to that with no endplates.
For the optimal wing,2 the maximum lift-to-drag ratio can be determined
by the formula

(9.56)

where J.lopt is the optimal (maximal) magnitude of the coefficient J.l.

To decide whether it is worthwhile to attain an optimal spanwise loading
distribution, it is practical to evaluate the following ratio:

K~~ _
K max -
JJ.lopt _
---;:- -
Cx;
Cx; m;n •
(9.57)

Figure 9.2a presents the fraction (9.57) versus the aspect ratio A for an
optimal distribution of the angle of pitch for a rectangular wing in the extreme
ground effect. Figure 9.2b shows the relative increment of the maximum lift-
to-drag ratio for a wing with optimal organization of jet ejection along the
trailing edge versus the aspect ratio in comparison with uniform blowing.
It is interesting to be able to evaluate the influence of the dynamic com-
pressibility of air upon the lift-to-drag ratio. It was shown in section 5 that for

2 In the sense of minimal indueed drag.

276 9. Aerodynamic Efficiency

1.00 i ........~~=--I-II-I

0.95 f---+---~~;:-">+--_t

0.90 f----\----l----~~~_t

0.851----\----l----+----''IT-"'rl

0.80 I----\----l----+--~

0.75 L-._ _-L._ _ _L-._ _- L ._ _- - - '

0.0 0.2 0.4 0.6 0.8
M
Fig. 9.3. The relative influence of compressibility upon the maximum lift-to-drag
ratio of a rectangular wing in the extreme ground effect versus the aspect ratio for
different magnitudes of the Mach number.

a given Mach number, incompressible flow results can be used in compress-

ible flow but for a smaller aspect ratio>..' = >-Jl - M;, smaller clearance
h' = hJl - M;, and larger angle ofpitch ()' = ()Jl- M;. Using the expres-
sion for the maximum lift-to-drag ratio given by (9.33) and assuming that
the viscous contribution remains the same, we can roughly estimate the rel-
ative effect of compressibility on the maximum lift-to-drag ratio for different
aspect ratios and Mach numbers; see Fig. 9.3.
We can use formula (9.31) to analyze the influence of different factors
upon the lift-to-drag ratio of a fiying wing of rectangular planform with end-
plates on the basis of the one-dimensional nonlinear flow model developed
in section 4. Following Rozhdestvensky [63] and Kubo [162], some results
are represented herein of calculations illustrating the behavior of the lift-to-
drag ratio of the previously mentioned simple configuration versus the design
parameters. To evaluate the aerodynamic efficiency of a wing, a simplified
calculation of the viscous contribution to drag is used, based on the con-
ce pt of an equivalent flat plate, see Voitkunsky et al. [163] or Raymer [164].
Hence, it is assumed that a lifting system has viscous (friction) drag identical
to that of a flat plate that has the same wetted area, length,3 and speed.
Correspondingly, the friction drag coefficient for fully developed turbulent
flow is determined as
Cf _ 0.455
(9.58)
- (logRe)2.58'
where Re = UoCo/v is the Reynolds number based on the root chord and
cruise speed. The area of the wetted surface of the flying wing configuration,
3 The length of the plate is equal to that of the wing's chord.
 9.2 The Lift-to-Drag Ratio of a Wing in the Ground Effect 277

related to the square of the root chord Sw, is determined by the formula

(9.59)

where Swep is the wetted surface of (two) endplates of given configuration,

related to the square of the chord length.
The latter quantity can be defined as

(9.60)

where hep(x) is a given chordwise distribution of the height of the endplate

as a fraction of the chord of the wing.
In the particular case of a fiat wing and uniform distribution of the gap
between the endplates and the ground, formula (9.60) yields the following
result:
Sw = 2('\ + 2hep + (}), (9.61 )
where hep represents the relative height of the endplate at the trailing edge.
Accounting for the fact that in the calculation of the lift and the induced
drag coefficients, the reference area was that of the wing's pi anform , we can
obtain the resulting expression for determination of the viscous drag coeffi-
cient 4 based on the wing reference area:

SW
C xo = Cf ( 2+ T ep )
. (9.62)

Figures 9.4-9.7 are graphs of the lift-to-drag ratio versus the lift coef-
ficient. Figure 9.4 illustrates the infiuence of the relative ground clearance
and the design lift coefficient upon the aerodynamic efficiency of a fiat thin
wing of reet angular planform with endplates. For eaeh magnitude the design
lift coefficient, the gap between the tips of the endplates and the ground is
uniform chordwise.
It can be seen from the graph that, when a wing operates in the extreme
ground effect, a decrease in ground clearance results in a considerable in-
crease in the lift-to-drag ratio. At the same time, the optimal lift coefficient 5
increases. Figure 9.5 demonstrates the same tendencies for the infiuence of
the gap under the endplates upon the aerodynamic efficieney of the configu-
ration. The dashed lines correspond to the case when the leakage of the fiow
from under the endplates occurs with contraction; see the considerations on
the effective gap in paragraph 4.3 and the solution of the local problem for
contracted leaking fiow under the endplate (or fiap) in paragraph 8.1.3 of
section 8.
4 Note that this formula does not include viscous pressure (form) drag and, con-
sequently, gives a very approximate estimate of the viscous drag.
5 That is, the lift coefficient, corresponding to the maximum aerodynamic effi-
ciency.
278 9. Aerodynamic Efficiency

iJ
40

K=
C
....L I
Cx
I/Ih=r'°t'
30 ~
/ /-t--- I

"'" '" '\

I

,( --+--.
Ih = p.07 ~
20
/
i
i
~
I. . .
I ~ "~
I Ih = .08 ~

10
//
It--I---_. B--- '"
h 1°.0
I II
0.0 0.2 0.4 0.6 0.8 C 1.0
Y

Fig. 9.4. The lift-to-drag ratio of a rectangular wing with endplates in the extreme
ground effect versus the design lift coefficient for different magnitudes of the relative
ground clearance (hep = 0.06, Re = 6 x 10 8 ,'\ = 0.8).

45
K , -- - --
40 -,
'~ÖepO= 0.015 ,

35 ,

30
,
,
L-
,'/ -~~
'/;-' K ' , ). ,Öep = 0.025
25
CI

f / ~ ----..:.
20

15

10
/
r öepc = 0.035
---
5 I
11
o I
0.0 0.2 0.4 0.6 0.8
Cy
Fig. 9.5. The lift-to-drag ratio of a rectangular wing with endplates in the extreme
ground effect versus the design lift coefficient far different gaps under the tips of the
endplates (flat plate, gap uniform spanwise; solid lines: no leaking flow contraction;
dashed lines: with contraction; h = 0.1, Re = 1.34 x 109 ,'\ = 1).

It is easy to conclude from Fig. 9.5 that a decrease in the gap under the
endplates may result in considerable augmentation of the lift-to-drag ratio.
Simultaneously, the optimum lift coefficient increases. In addition, realization
 9.2 The Lift-to-Drag Ratio of a Wing in the Ground Effect 279

50
Cy
K=-
Cx 1.,=2
40
/ I'-- ~

30 L/ ' ~

"'-< ~= .5
--......

1/ 1.,= 1.0 ~
i'----
20 ~ - ---,... :------
-.... I'---
ß A.l~R I'--- :--..... I'---
:f I
10
!
11
I !
o
0.0 0.2 0.4 0.6 0.8 1.0
Cy
Fig. 9.6. The lift-to-drag ratio of a rectangular wing with endplates in the extreme
ground effect versus the design lift coefficient for different aspect ratios (fiat plate,
gap uniform spanwise, h = 0.1, Re = 1.34 x 109 ,.A = 1).

of contraction of the fiow escaping from underneath the vehicle may lead to
significant gains in the aerodynamic efficiency. 6 The infiuence of the aspect
ratio upon the lift-to drag ratio for a fixed gap under the tips of the endplates
is shown in Fig. 9.6. This figure shows that the augmentation of the aspect
ratio leads to an increase in the aerodynamic efficiency and increases optimun
lift coefficient.
The dependence on the Reynolds number of both the lift-to-drag ratio
and the magnitude of the optimum lift coefficient is illustrated in Fig. 9.7
and gives rise to an obvious conclusion: the larger the Reynolds number
the larger the aerodynamic efficiency, and the smaller the optimum lift co-
efficient. The latter circumstance together with observations related to the
infiuence of other design parameters, as discussed above, shows that the way
to increase the cruise speed of the vehicle, 7 when fiying close to the ground,
consists of increasing the design Reynolds number. Other factors, leading to
the enhancement of the lift-to-drag ratio, such as increase in the aspect ratio,
decrease in the ground clearance and/or height of the endplates, result in a
diminution of cruise speed for a given magnitude of wing loading.
Figure 9.8, plot ted on the basis of the one-dimensional nonlinear theory of
section 4, confirms the result discussed previously in this paragraph, namely,
if suction force is not realized (e.g., due to stall) the efficiency may drop
considerably.
6 This can be achieved by using endplates with sharp tips or keels with sharp tips
fixed on fioats.
7 That is to reduce the magnitude of the lift coefficient, corresponding to the
maximum lift-to-drag ratio.
 280 9. Aerodynamic Efficiency

25

K=-
Cy 2.. 105 Re 1= 3.~·10!
,
Cx
20 11~ r---..
~ ~

W/ P\ ~~
15
r7 1.1 109 '" ~
S·I 08
V
~
~
~e~ 108 ~
10

5
0.0 0.2 0.4 0.6 0.8 1.0
Cy
Fig. 9.1. The lift-to-drag ratio of a rectangular wing with endplates in the
extreme ground effect versus the design !ift coefficient for different magnitudes
of the Reynolds number (gap uniform spanwise, Hat plate, h = 0.08, A = 0.8,
8ep jh = 0.25).

25
C
K= ~
Cx
20 / ~ with suc ion force

Ir~ ~
~

7 '\
15
""-
withou suction !orce

I
10

5
I~
~
o
0.0 0.2 0.4 0.6 0.8 1.0
Cy
Fig. 9.8. The !ift-to-drag ratio of a rectangular wing with endplates in the extreme
ground effect versus the design lift coefficient for realization of the suction force and
no suction force (Hat plate, gap uniform spanwise, h = 0.08, Re = 6 x 108 , A = 0.8,
8~pjh = 0.25).
10. Integral Formulations for Lifting Surfaces
in the Extreme Ground Effect

As mentioned in the Introduction, asymptotic analysis can be applied directly

to the integral equation of the lifting surface in the ground effect. Essen-
tially, in the extreme ground effect, we study the limiting process, when the
distance between two double (vortex) layers, representing the wing and its
mirror image, becomes vanishingly small. In the limit, the two double layers
merge into a quadruple layer so that the procedure can be characterised as
quadruplication. 1 The main result of quadruplication is the confluence (for
h -+ 0) of the integral equation of the wing-in-ground effect into a differen-
tial equation (ordinary for two-dimensional flow, and in partial derivatives
for three-dimensional flow). The resulting differential equation can be shown
to be identical to that, obtained in the course of solving the corresponding
boundary problem by the method of matched asymptotic expansions. The
quadruplication approach in the aerodynamics of wings in the ground effect
was first introduced by Panchenkov [64]. In what follows, all derivations will
be based on a different scheme of quadruplication proposed in [66].

10.1 A Slightly Curved Foil

in a Two-Dimensional Ground Effect

For the two-dimensional steady motion of a slightly curved foil near a flat
ground plane, the corresponding singular integral equation has the form

- 1 V.p.
211'
1
0
1
'Y(~) JC(x -~, h) d~ dys,
= --d
X
(10.1)

where 'Y(x) is the strength of vorticity that replaces the foil in the mathe-
matical model, and dys/dx is a function that represents the slope of the foil
camber line with respect to the horizontal axis;

(10.2)
(x - ~)[(x - ~)2 + 4h 2 ]

1 The term was introduced by A.N. Panchenkov [64].

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
282 10. Integral Formulations

is the kernel of the integral equation. The first (singular) term of the kernel
represents the contribution to the downwash from the foil-bound vorticity,
whereas the second (regular) term accounts for the contribution of the down-
wash from the vorticity of the foil mirror image. Abbreviation ''v.p.'' implies
that the integral is understood in the sense of the Cauchy principal value.
We set the ground clearance h to zero and expand the kernel of the equa-
tion in h. It can be seen that in the course of a straight forward expansion, the
kernel would vanish to the lowest order. The second- and higher-order terms
would contain divergent integrals. The principal difficulty in constructing this
expansion is that parameter his not always small compared to (x - e). To
single out the subdomain x - e = O(h), we split the integral on the left-hand
side of equation (10.1) into three integrals:

= 1 1 ')' K de = 1X-1'J ')' K de + 1x+1'J ')' K de + 1

1
I ')' K de = 11 + h + h,
o 0 X-1'J X+1'J
(10.3)
where a small parameter TJ is chosen so that h « TJ « 1. It can be shown
that contributions 11 and h are of the order 0(h 2 ), whereas the integral
12 = O(h). In what follows, only the middle integral 12 will be retained, and
11 and 13 will be neglected.
For 12 , the variable of integration varies in the range of x - TJ < e < x +TJ.
We introduce a new stretched variable (= (e - x)/h, which has the order of
0(1). Then,
(10.4)

Outside of the small vicinities of the edges (x = 0 and x = 1), the vortex
density ')' can be expanded into a Taylor series

(10.5)

Substituting this expansion in (10.5) gives

(10.6)

Recalling that both hand TJ should be small quantities with h = o(TJ), we

can derive the following lowest order representation of the integral equation
(10.1) for h -+ 0:
h '( ) = h d')' = dys (10.7)
')' x dx dx'
Thus, it has been shown that for very small relative ground clearances
the integral equation for a foil-in-ground effect degenerates into a
simple ordinary differential equation of the first order (10.7).
Recalling the relationship between the vortex density and the perturba-
tion velo city potential <p,
 10.2 A Foil with a Jet Flap in Proximity to the Ground 283

'Y - dcp __ dcp+ (10.8)

- dx dx'
and assuming domination of the channel flow velo city potential, Le., cp+ =
O(hcp-), it is easy to identify equation (10.7) with the equation (3.32), ob-
tained by the matched asymptotics technique within the boundary problem
formulation. The equivalence (to the lowest order) of boundary conditions
for the aforementioned equations (10.7) and (3.33-3.34) follows from simple
physical reasoning, namely,
• the circulation of the velo city r(x) = f;
'Y(e) de should be zero at the
leading edge, Le., the potential at the leading edge should be continuous:

r(l) = cp-(1) - cp+(l) ~ cp-(1) = 0; (10.9)

• the flow velo city (pressure) should be continuous at the trailing edge, Le.,
at x = 0,
dcp_ dcp+ dcp_
'Y=---~-=O. (10.10)
dx dx dx

10.2 A Foil with a Jet Flap in Proximity to the Ground

Note that the quadruplication procedure can be applied in all cases when the
kernel of the integral equation has a form simHar to (10.2). For a foH with a
jet flap near the ground, the corresponding integral equation can be written
as (see Menshikov [165])

1
271' v.p.
1
1
-00
[1 x-
'Y(e) x _ € - (x _ €)2
e 4h 2 ]de =
+ f(x), (10.11)

where
dy·
f(x) = 0 for 0< x ~ 1; f(x) = d~ for x ~ 0, (10.12)

'Y is the strength ofthe vorticity that replaces the foH and the jet; dYj/dx is the
unknown distribution of slope of the jet. The vorticity (pressure difference)
on the jet is assumed proportional to the jet momentum coefficient, and the
curvature of the jet

x ~ o. (10.13)

In these relationships, yj, yj' represent the local slope and the curvature of
the jet and q is the jet momentum coeffcicient. Quadruplicating the inte-
gral operator, we obtain the following differential equations for the vorticity
distributions on the foH and the jet, as weH as for the jet camber line:
284 10. Integral Fürmulatiüns

h&'f =0 0:::; x:::; 1, (10.14)

dx '

h &-y= !::. q d 3 Yj = d Yj , x:::; O. (10.15)

dx 2 dx 3 dx
Because the jet is blown from a slot at the trailing edge at a small angle T with
respect to the chord and, eventually, be comes horizontal far downstream, we
can apply the following conditions:

~~ (0) = T, (10.16)

Accounting for (10.16), it is easy to derive the solution in the form

l'(x) = T/fi, 0< x:::; 1; (10.17)

~(x) ~ r/fi exo(xJ h~). x'; 0 (10.18)

The corresponding lowest order lift coefficient is given by

(2C;
Cy = TV-';' (10.19)

This expression is identical to that obtained within the boundary problem

formulation für a jet-Happed wing in the ground effect; see formula (6.123).

10.3 A Wing of Small Aspect Ratio

The technique of quadruplication demonstrated above can also be applied

to a wing of finite aspect ratio. For example, for a Hat wing of small aspect
ratio, the integral formulation leads to the equation

1 1 1
ar [ 1 z- ( ]
271" v.p. -1 a( z _ ( - (z _ ()2 + 16h~ d( = -0, (10.20)

where h).. = hj>.., >. is the aspect ratio (h« >. « 1), and r(z) is the circulation
of the velo city. The kernel of equation (10.20) has the same structure as that
of (10.2). Therefore, the approach considered earlier is applicable here, too.
Replacing for h).. ---+ 0 the integral operator by a corresponding differential
one

(10.21 )

we obtain the equation

 10.4 Lifting Line(s) in Close Proximity to the Ground 285

(10.22)

Integrating (10.22) and imposing the condition of zero loading at the tips of
the wing, finally, we obtain the following expression for the lift coeflicient of
a small-aspect-ratio wing:

(10.23)

which is identical to formula (3.69), obtained from the boundary problem

formulation.

10.4 Lifting Line(s) in Close Proximity to the Ground

10.4.1 A Single Lifting Line in the Extreme Ground Effect

The following example illustrates the application of the same technique to the
integrodifferential equation of the lifting line in the ground effect. Assuming
for simplicity that the longitudinal curvature of the wing's sections is zero,
we can write the lifting line equation in the presence of the ground in the
form 2

r(z) =
2C(z) { 1 11
dr [ 1 z- ( ]}
-71" - l - O(z)- 471" -1 d( z _ ( - (z _ ()2 + 16h? d( , (10.24)

where r(z) is the distribution of the circulation of the lifting line spanwise,
C(z) and O(z) are spanwise distributions of the local chord and angle of
pitch,3 l is the relative span (span to chord ratio), and hl = hll is the height
to span ratio. The planform function C(z) is normalized so that

1 1
C(z) dz = 1. (10.25)

Exploring the limit hl -t 0 for the flow around a lifting line near the
ground, note that using the concept of a lifting line implies that the chord of
the wing is much less than the ground clearance and the latter is much less
than the wing span.
Thus, the limiting result will be different from that, obtained in the large
aspect ratio limit from our previous analysis in Section 3. In the latter case,
the distance from the ground is much smaller than the chord, and the chord
is much smaller than the span.

2 The integrodifferential equation of the lifting line for an unbounded fluid can be
found in Ashley and Landahl [161].
3 C(z) is a ratio of loeal ehord to the root ehord.
286 10. Integral Formulations

Quadruplieating the integral part of equation (10.24), we obtain the fol-

lowing ordinary differential equation of the seeond order:

(10.26)

Equation (10.26) should be solved with boundary eonditions of zero load-

ing at the tips of the wing, Le, r(±l) = O. Suppose that O(z) = 0 is eonstant,
Le., the wing is flat.
Consider first the case of a constant chord G(z) = 1 (rectangular
wing), for whieh the resulting expressions for the distribution of the cireula-
tion along the lifting line and lift eoefficient ean be obtained as

r(z) = 27rO [eosh(pz) _ 1],

l eosh(p)

Gy = 27rO(1 _ tanph P ),
p=V~·
rr- (10.27)

Reealling that the indueed downwash in the extreme ground effeet is propor-
tional to the seeond derivative of the cireulation with respect to the spanwise
eoordinate z, Le.,
(10.28)
we ean derive the indueed drag eoefficient for a reet angular wing of large
aspeet ratio in the extreme ground effeet in the form

Gx . = 7r0 2 sinh 2p - 2P. (10.29)

I P eosh2p + 1
As in Prandtl's classicallifting line theory, the indueed drag eoefficient ean
be shown to be proportional to the square of the lift eoeffieient. We ean write
G2 Aeff 4(eosh2p + l)(p - tanhp)2
G - y p,--- (10.30)
Xi - 7rAp,' - A - Ap(sinh2p-2p) .
Examining equation (10.28), we ean eonclude that in the extreme
ground effect, the optimal4 spanwise distribution of loading for a
wing of a large aspect ratio is parabolic rather than elliptic, as in the
unbounded fluid ease. As follows from (10.26), for a flat untwisted wing, the
spanwise ehord distribution seeuring a parabolie loading distribution is also
parabolic. Substituting G(z) = k (1 - z2) (where from normalization eondi-
tion (10.25), k = 3/2, A = 3lj2), and r(z) = r o(z2 - 1) the equation (10.26)
we obtain the following formula for r o:
27rO
(10.31)
r o = l(l + 47rhJ/l) .
4 In Munk's sense, Le., ensuring minimal induced drag for a given lift.
 10.4 Lifting Line(s) in Close Proximity to the Ground 287

ing formula: _
Cy - rol
/1 2 _
To obtain the lift coefficient based on the chord, we have to apply the follow-

(1 - z )dz - (
81['0
) . (10.32)
-1 3 1 + 47rhJ/l
The downwash corresponding to parabolic spanwise loading is constant
along the lifting line. Simple calculation shows that
d2 r 8h l 7r(}
Qj(z) = -h\ dz 2 (z) = -2h\ro = l(l + 47rhJ/l)' (10.33)

The induced drag coefficient of the optimal lifting line is calculated as the
lift coefficient times the induced drag, i.e.,
C - C . _ 8hllro2 _ 3h\C2 (10.34)
Xi - Y Ql - 3 - 2l y
or, rewriting (10.34) in the Prandtl's format,5

C _ C; Aeff 4
Xi - 7rAJ.L' J.L= T = 97rh\' (10.35)

These results show that in the limiting flow problem of a lifting line
in the extreme ground effect, the effective aspect ratio is inversely
proportional to the ground clearance related to the span. Figure 10.1
presents the inverse efficiency factor 1/J.L versus the relative ground clearance
(based on the span) for a single wing with rectangular and parabolic plan-
forms of the same relative span l = 5, operating in the extreme ground effect.
Figure 10.1 was obtained by using formulas (10.30) and (10.35).

3,--------,--------,-------,

Rectangular
Planform
2
formula ( 10.30 )

Parabolic
Planform
hl = h/l
0
0.0 0.1 0.2 0.3

Fig.l0.1. The inverse efficiency factor 1/p, for wings of a large aspect ratio and
different planform geometries in the extreme ground effect (h, ~ 0, l = 5).
5 Note that for a parabolic planform, >. = 3l/2.
288 10. Integral Formulations

10.4.2 A Comment on "Span-Dominated"

and "Chord-Dominated"Extreme Ground Effects

In most of this book, the modelling of the extreme ground effeet implies that
the ground clearanee is mueh less than the ehord and the span. This type of
ground effeet, associated with pronouneed stagnation under the wing, will be
designated as chord dominated ground effect (CDGE).6 On the other hand,
when eonsidering the aerodynamics of a wing of large aspeet ratio in the ex-
treme ground effeet on the basis of lifting line theory, it was assumed that the
ehord is mueh less than the ground clearanee, the latter being mueh smaller
than the span. In this ease, we ean introduee the notion of a span dominated
ground effect (SDGE). Although both of these effeets imply an inerease in
the lift-to drag ratio at smaller ground clearanees, it ean be shown that they
have somewhat different natures. To distinguish this differenee, eompare the
behavior of the lift and the indueed drag eoeffieients of a wing of large aspeet
ratio versus the ground clearanee within the previously mentioned models.
In the ealeulated examples, the relative span and angle of piteh of the wing
were I = 8 and () = 0.05, respeetively. In the CDGE, ealculations of of lift
and the indueed drag eoeffieients were made for eases of rectangular and
(optimal) semielliptie Hat wings by using formulas (3.65), (3.67), (3.78), and
(3.80). The aerodynamic eoefficients, eorresponding to the SDGE, were de-
termined by formulas (10.27), (10.29), (10.32), and (10.35), eorresponding to
reet angular and (optimal) parabolie planforms. 7
Figure 10.2 shows that for a fixed piteh angle with a deerease in the
relative ground clearance (based on the chord), the CDGE model responds
by an inerease of both the lift and the indueed drag eoefficients. As seen
from formulas (3.65), (3.67), (3.78), and (3.80), both eoefficients within the
CDGE model are inversely proportional to the ground clearanee. Note that
Standingford and Thek[101] eame to the same eonclusion in their accurate
numerical analysis of the aerodynamies of lifting surfaees for small ratios of
the ground clearance to the ehord.
As seen from Fig. 10.3, the behavior of lift and indueed drag eoefficients
versus the relative ground clearanee (based on span) within the SDGE model
is different. In this ease for a fixed piteh angle we ean observe a deerease
in the indueed drag eoefficient as the wing flies closer to the ground. The
lift eoefficient inereases with a deerease in the ground clearanee although
somewhat more slowly than in the CDGE model. s

6 In the literature one may often encounter the alternative term ram in association
with the ground effect.
7 Note that for span-dominated and chord-dominated ground effects, the optimal
planforms of the wings are different, although both give rise to a parabolic dis-
tribution of loading.
8 If the lift coefficient is kept constant, while the wing approaches the ground, the
induced drag coefficient decreases in both models of extreme ground effect.
 10.4 Lifting Line(s) in Close Proximity to the Ground 289

1.0

,
0.8

\~Y
.!!
I:
.~
==0CD 0.6
(.)
.S:!
'\
E
111
I: 0.4 ~-
>.
~
eCD
"C

<C
_ ~OO Cx; ~
0.2 - -- - -
-- - - - - - - - - - - -
0.0
0.04 0.08 0.12 0.16
hl = h I I

Fig. 10.2. The behavior of the lift and the induced drag coefficients versus the
relative ground clearance based on the chord (CDGE model) in the extreme ground
effect (l = 8, B = 0.05; solid lines correspond to a semielliptic planform; dashed lines
correspond to a rectangular planform).

10.4.3 A Tandem of Lifting Lines in the Extreme Ground Effect

In what follows the tandem, comprising two wings of large aspect ratio and in
steady motion near a solid plane wall, is modelled by a system of two lifting
lines and their "mirror images." The Cartesian coordinate system adopted
herein is attached to the front wing and has a z axis directed to the starboard
side of the tandem and the x and y axes directed downstream and upward
respectively.
To avoid tedious algebra and simplify the solution, the analysis is re-
stricted to the case when both wings of the tandem have identical ground
clearances, planforms, areas, and aspect ratios. Thus, the rear wing operates
in the wake of the front one. Ground clearances are measured from the mid-
chord of corresponding wings. As earlier, all quantities and functions will be
rendered nondimensional by using the semispan of the wing and the velocity
of the incoming stream.
For hl ~ 0, using the same technique as for a single lifting line in the ex-
treme ground effect, we can reduce the system of Prandtl's integrodifferential
equations describing the aerodynamics of a tandem near the ground to a set
of two ordinary differential equations of the second order; see Rozhdestvensky
[67):
(10.36)

(10.37)
290 10. Integral Formulations

0.5

,/""'
0.4
I-- r--
k
J!l
c:
Cy 1-1""00 ClCj
CD
'0
IE 0.3 /" r---..
--
--/V.-
CD
0
() Cy _ -- ........... 100 ClCj
.!:.1
E - ........ - ...
- ---- -
V'
t1I
c: 0.2
>-
~
~
0.1
,-j
/

0.0
0.00 0.05 0.10 0.15 0.20 0.25
h 1= h I I

Fig. 10.3. The behavior of the lift and the induced drag coefficients versus the
relative ground clearance based on the span (SDGE model) in the extreme ground
effect (l = 8,9 = 0.05, solid linescorrespond to a parabolic planform, dashed lines
correspond to a rectangular planform).

In these equations r 1,2(Z) stand for the distributions of the loading in the
direction of the span of the front and rear wings of the tandem, C(z) repre-
sents the form of the chord distributions, and fh,2(Z) are the distributions of
the pitch angle along the span of the wings.
It can be seen from observation of the right-hand side of equation (10.36)
that in the extreme ground effect, the downwash induced by the rear wing
upon the front wing is negligible. At the same time, the front wing affects
the aerodynamics of the rear wing, see equation (10.37). The system admits
closed form solutions.
Suppose that both wings are rectangular, C(z) = 1, and flat (h(O) =
01 , O2 (0) = O2 . Then, the first equation of the system (10.36)-(10.37) can be
integrated to yield

r 1 (z) = 21[-(}1 (coshpz

A coshp
-1), p= V~ = V~·
(T fl (10.38)

Substituting solution for Tt(z) in the second equation ofthe system (10.36)-
(10.37), we obtain the following nonhomogeneous ordinary differential equa-
tion for function n (z ):
d 2 r 2 (z)
dz 2 - p
2
r2 (z) = p + qcoshpz, (10.39)
where p = 02/hl, q = 2(h/h 1 coshp. Integrating (10.39) and using the require-
ment that the loading should vanish at the wing tips r 2 (±I) = 0, we obtain
the following solution for r 2 (z):
 10.4 Lifting Line(s) in Close Proximity to the Ground 291

r2 (z) = (A - 4!2 cosh2pz) coshpz

+-q (Sinh 2pz
2p 2p
+ z )smhpz
.
-
82
--,
h 1p 2
(10.40)

where
2P
A = - q - -q (Sinh
-- + 1) tanhp + 82
.
4p2 2p 2p hlp2 cosh p
Using express ions (10.38) and (10.40) for the loading along the front and rear
wings, we can readily obtain both the lift and the induced drag coefficients
for the above case of a rectangular planform of lifting elements of the tandem.
However, in what follows, the accent will be on a parabolic loading dis-
tribution for which each of the wings and the tandem as a whole have min-
imal induced drag for a given lift. Writing the circulations and the planform
r r
equations of both lifting lines as 1 ,2(Z) = O, ,2(z2 -1),G1 ,2(z) = G(z) =
k(l - z2), k = 3/2 and substituting these expressions into equations (10.36)
and (10.37), we obtain the following simple system of algebraic equations
with respect to the amplitudes of the loading distributions:

(10.41 )

wherefrom
r _ 21['8 1 r. _ 21T(82 - 4h1r lO )
(10.42)
10 - l(l + 41Tht/l) , 20 - l(l + 41Tht/l) .

The lift coefficients of each wing and the overall lift coefficient of the tandem
Gy, are obtained in the form

c _ 4l r lO _ 81T8 1
(10.43)
Yl - 3 - 3(1 + 41Tht/ l ) '
G _ 4lr20 81T (82 - 3h 1Gy, /l)
3 (10.44)
Y2 -
3 1 + 41Tht/l
(10.45)
In the latter coefficient, the reference area used was half that of the tandem.
For optimal wing loading, the downwash on both lifting lines is uniform along
the span:
(10.46)
Consequently, the induced drag coefficients for the front and rear wings can
be found in the form
292 10. Integral Formulations

The induced drag of the tandem as a whole 9 will be

3hl 2 3hl
Cx = C X1 + CX2 = 2i:(CY1 + 2CY1 CY2 + Cy22 ) = 2i:(C 2
Y1 + C y2 ) . (10.48)

It is worthwhile to remember here that an optimal tandem in an unbounded

fluid (hl = 00) has the following relationship between the lift and the induced
drag coefficients of its elements:

(10.49)

10.5 Quadruplication of the Integral Equation

for a Wing of Finite Span

The above approach can be extended to the case of an arbitrary aspect ratio
A :» h. The relevant integral equation can be written in the form

(10.50)

where Ys = Ys(x, z),

1 1
G(x -~, z - () = - - " r = V(x -~)2 + (z - ()2, (10.51)
r r
r' = V(x - ~)2 + (z - ()2 + 4 h 2 • (10.52)
We introduce 1 :» 'fJ :» hand single out the integral contribution over a small
square S." = 2 'fJ x 2'fJ in the vicinity of a point (x, z):

I= J"fJs = J.JS'lr + J.Js-S'l

r = 1 + h. 1 (10.53)

As in two-dimensional flow, with an asymptotic error of O(h 2 ), it is sufficient

to consider the first of the two integrals. It is convenient to introduce a

l
function
r(x, z) =
x
'Y(~, z) d~, (10.54)
Xl.

where Xle = Xle (z) is the equation of the leading edge. We turn to considera-
tion of the following integral

111 = J.Jsr 'Y {Ja dS

OX
= t+."
Jz-."
l +'" ox ox
X

x-."
{Jr {Ja de d(

9 Again, with the reference area equal to half that of the tandem.
 10.5 Quadruplication of the Integral Equation for a Wing of Finite Span 293

= jz-'1Z+'11X-'X+'1 1 -(3' - (3)(~-x)d~dx.

8r
8x r
1
r
1
(10.55)

We introduce the stretched variables ~ = (~- x)/h, ( =( - z)/h into In:

j _ii'j'j ji'j_i'j 8r - - ( 1 1) - -
In = 8x (x + h~, z + h() R3 - R'3 d~ d(, (10.56)

where fj = Tl/h, R = r/h, R' = r'/h. The Taylor series expansion holds
outside the vicinities of the edges:
8r - - 8r _8 2 r - 82 r 2
8x (x + h~, z + h() = 8x (x,z) + h~ 8x 2 (x, z) + h( 8x8z (x,z) + O(h ).
(10.57)
Integrating in (10.56) and taking into account (10.57), we obtain

For h ~ 0 and fj = Tl/h ~ 00, we obtain to the lowest order,

I n -'" 471' h 82r8x(x,z)

2 • (10.59)

Now, we turn to the integral

112 = 118",
')'(~,()
00
82G
z
8 IX z+'11 x+'1
I x-82dxdS=82
z Z-'1 X-'1
2 l
')'(~,()Gd~d(dx
00

(10.60)

Taking into account the Taylor expansion of ')' in the vicinity of the point
(x, z), we obtain

(10.61)

Therefrom, for h ~ 0, Tl ~ 0, and fj ~ 00, it is easy to find the lowest order

representation of 112 :

82
112 ~ 471' h 8z 2
IX ')'(x, z) dx = 471' h 8 r(x,
00
2z)
8z 2 . (10.62)

Summing up expressions (10.59) and (10.62), we find that the limiting asymp-
totic form of the integral equation for a wing of finite aspect ratio is identical
to the Poisson differential equation
294 10. Integral Formulations

r + 88zr) __ dys
h (8 2
8x 2
2
2 dx'
(x, z) E S. (10.63)
-

Equation (10.63) has to be solved in the interior of the two-dimensional do-

main S, bounded by the wing's planform contour. The boundary conditions
for (10.63) can be either derived by the matching procedure or adopted on
the basis of appropriate physical requirements (e.g., continuity of the circu-
lation at the leading and side edges and continuity of the pressure at the
trailing edge). As expected, equation (10.63) is identical to that obtained
earlier within the boundary problem formulation; see formula (3.59).
11. Equations and the Stability
of Motion of a Lifting System
in the Extreme Ground Effect

The analysis of the dynamics of wing-in-ground-effect vehicles provides data

for assessing the stability of motion, controllability, and ride comfort of the
craft under development.
A study of the linearized equations of the motion of wing-in-ground-effect
vehicles was carried out by Irodov [166], Kumar [167]-[170], Zhukov [171]-
[175], and Staufenbiel et al. [176]-[178]. This research revealed a significant
distinction between the dynamics and the stability criteria of these vehicles
and the aircraft that normally operates out of the ground effect. It was also
found that one of the typical problems of the design of ground-effect machines
is due to the strong coupling between their aerodynamic configuration and
the fiight dynamics.
Irodov[166] considered the linearized equations of perturbed longitudinal
motion in terms of the variation of the angle of attack and the ground clear-
ance. Assuming that the speed of the vehicle remains constant, he derived a
( quartic) characteristic polynomial equation of the fourth order. Applying the
Gurvitz-Ruth criteria of stability, he came to the conclusion that aperiodic
static stability is ensured when the aerodynamic center of height is
located upstream of that of the angle of attack. This important prac-
tical conclusion signifies that if the aerodynamic configuration is not selected
properly, it is impossible to secure static longitudinal stability to the motion
of a wing-in-ground-effect vehicle by choosing the position of the center of
gravity.
According to Irodov [166], to secure the oscillatory stability of the vehicle
one has to provide an appropriate location of the center of gravity upstream
of the center of the angle of attack. In the same work, Irodov indicated
that account of the variation in cruise speed practically does not modify the
previously mentioned condition of static aperiodic stability.
Kumar [166] studied the dynamics of a wing-in-ground-effect vehicle in
both longitudinal and lateral motion, incorporating the effects connected with
perturbation of the speed of forward motion. His stability analysis was based
on a quintic characteristic equation.
A thorough study of the dynamics of wing-in-ground-effect vehicles ac-
counting for the perturbation of speed and incorporating stability analy-

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
296 11. Equations and Stability of Motion

sis with special reference to controllabilityl and design, was carried out by
Zhukov, starting in the 1970s and finalized in [175J. He revealed several dis-
tinct parameters, defining static stability and dynamic behavior of wing-in-
ground-effect craft. In particular, he introduced the notion of binding to the
ground, as a capability of a vehicle in cruising fiight to stay in ground effect
after the action of controls or gusts of wind.
Staufenbiel also studied stability criteria, used the quintic characteristic
equation for the analysis of the dynamics, and discussed nonlinear effects.
This section covers some linear formulations related to the longitudinal
dynamics of wing-in-ground-effect vehicles. First of all, a derivation is given
in terms of the perturbations of the relative ground clearance hand the
pitch angle () of the linearized equations of motion without (after Irodov)
and with (after Zhukov) account of perturbation of forward speed. Then, we
consider an approximate derivation of an asymptotic form of the linearized
equations of the longitudinal motion of a wing-in-ground effect vehicle in
the extreme ground effect, Le., for vanishing relative clearances between the
lifting surface and the ground. The orders of magnitude of the terms are
evaluated formallyon the basis of a simplified nonlinear unsteady theory of
the extreme ground effect, discussed in section 4. Eventually, an asymptotic
form of the equations of motion is derived for h --t 0 and sm all periods of
time from the moment of the action of the perturbation. It is shown that on
(nondimensional) time scale t = 0(1), which corresponds to distances of the
order of the chord from the moment of perturbation, the equations of motion
correspond to the quartic formulation of Irodov [166], Le., the speed of the
vehicle remains almost constant. From a praetieal viewpoint, this signifies
that Irodov's eriterion of static stability is valid, although it was derived
on the basis of the somewhat restrictive assumption of no perturbation of
speed. Differing from Irodov, the asymptotic form of the equation, valid for
a vanishing h, does not depend explicitly on the relative ground clearance,
but rat her on the reduced density fl = J.th and the ratios of the design pitch
and the curvature of the lower surface to h, Le., the number of parameters
is fewer by one compared to the initial formulation. On larger time sc ales
of the order of l/h and 1/h 2 the variation of speed is first driven by height
and pitch perturbations and later is determined by the speed perturbation
proper. The latter conclusion confirms the results derived by Zhukov [175J.

11.1 Linear Equations of Longitudinal Dynamics

Here, the linearized versions of the equations of longitudinal motion will
be deduced, corresponding to what is known as the quartic and the quintic
descriptions ofthe dynamics ofwing-in-ground-effect vehicles. In other words,
linearized equations of motion and corresponding characteristic equations
1 Matters of automatie eontrol are eovered in Diomidov [179J.
 11.1 Linear Equations of Longitudinal Dynamics 297

are written with and without accounting for perturbation of the forward
speed. Instead of the representative kinematic parameters a (angle of attack)
and h (relative ground clearance) utilized by Kumar and Staufenbiel, the
parameters () (pitch angle) and h (relative ground clearance) will be employed.
In what follows, all unsteady aerodynamic derivatives will be incorporated.
The pair of parameters (), h is more practical in the aerodynamics of
ground-effect vehicles than a, h. The use of (), h was first proposed by
Treshkov [180] and since has been adopted in most of the Russian devel-
opments on wing-in-ground-effect vehicles.

11.1.1 The Quartic Characteristic Equation

We consider first the case when the cruise speed remains constant
in otherwise perturbed motion. In what follows, all quantities and functions
will be rendered nondimensional with respect to the cruise speed Uo and the
root chord Co of the main wing of the vehicle. In this case, the equations of
perturbed uncontrolled motion of the WIG can be written as

(11.1)

d 2Ö
ILi zdt2 =m z, (11.2)

where IL is relative density of the vehicle defined by the formula

2M
IL = pSCo' (11.3)

In (11.1) and (11.2), Gy and mz are the perturbed lift and the moment
coefficients, hand Ö represent the perturbed relative ground clearance and
the pitch angle, M is the vehicle's mass in cruise, S is the wing's reference
area, pis the specific density of air, and i z is a coefficient of the longitudinal
moment of inertia of mass, determined by the relationship

. Iz
Zz = MC2' (11.4)
o

where I z is the longitudinal moment of inertia of mass with respect to the

center of gravity. To track the relationship between the dynamic stability of
the vehicle and its cruise speed, it is sometimes convenient to use the following
obvious equation:
Uo
Fr = r::7'l' (11.5)
ygvo

where Fr and CyO are correspondingly the cruise Froude number and the
cruise lift coefficient. Accounting for (11.1), (11.2), and (11.5), the (two)
298 11. Equations and Stability of Motion

equations of motion with respect to the perturbations of pitch and ground

clearance can be written in an alternative form as

(11.6)

Using a representation of the perturbed lift and the moment coefficients in

terms of the derivatives with respect to the principal kinematic parameters
() and h, we can write

(11. 7)

(11.8)
Now we can rewrite the equations (11.7), (11.8) in the following form:

d2 h h dh h - iJ dB () -
J-l dt 2 - C y dt - C y h = C y dt + C y (), (11.9)

d2 B iJ dB () - h dh h -
J-l Zz dt 2 - m z dt - m z () = m z dt + m z h. (11.10)

Excluding h or Bfrom the equation set (11.9), (11.10), we obtain the following
fourth-order (quartic) characteristic equation of the system:

(11.11)

where the coefficients Ai(i = 1,2,3,4) are

A 1 = - -l . (iJ . ch)
mz+z z y ' (11.12)
J-lZz

(11.13)

(11.14)

(11.15)

The last coefficient can be rewritten to be expressed in terms of the abscissas

of the aerodynamic centers of height Xh and of pitch ():

(11.16)

The stability of motion for the fourth-order equation will be ensured if

(11.17)
 11.1 Linear Equations of Longitudinal Dynamics 299

It can be shown from consideration of formulas (11.12)-(11.17) that con-

ditions Al > 0, A 2 > 0 and A 3 > 0 always hold. Hence, the requirements of
stability in proximity to the ground can be reduced to the two inequalities
A 4 > 0 and A l A 2 A 3 - A~A4 - A~ > O. Satisfaction of the former condition
provides aperiodic stability to the craft, Le., the absence of positive real roots
of the characteristic equation. Meeting the latter requirement ensures oscil-
latory stability, Le., the absence of positive real parts of the complex roots of
the characteristic equation. It follows immediately from these considerations
and observation of (11.16) that to secure the aperiodic stability of wing-in-
ground-effect vehicle, one should select an aerodynamic configuration of the
craft for which the center of height is located upstream of the center of pitch,
Le., Xh - XI} > 0 (note that here the X axis is directed upstream).

11.1.2 The Quintic Characteristic Equation

If we account for the perturbation of the speed of the forward mo-

tion, the corresponding equations can be written as

(11.18)

d 2 ji , I} (}:. h- h~
J.L dt 2 = 2U CyO + CyB + CyB + Cyh + Cyh, (11.19)

. d2 0 (), I} - h- h~ (} :.
J.L2 z dt 2 = Yt(2C xo - C t )U + mß + mzh + mzh + mß· (11.20)

In these equations, U' represents the relative perturbation of the cruise speed,
cf is a derivative of the thrust coefficient with respect to the relative speed of
forward motion, C x is a static drag coefficient,2 and Yt is the vertical distance
of the thrust line from the vehicle's center of gravity. As earlier, all quantities
are rendered nondimensinal using the cruise speed Uo and the root chord Co.
The drag coefficient is related to the instantaneous cruise speed Uo (1 + U ').
Excluding two of the three unknown parameters, we obtain the following
fifth-order (quintic) characteristic equation of the perturbed system:

ChD
x + Chx C(}D
x
+ CI}x
(J.LD2 - C;D - C;) -(C~D + C~)
-(m~D+m~) (J.L i z D2 - m~D - m~)
(11.21)
We write the quintic characteristic equation (11.21) as

(11.22)

2 Incorporating both viscous and induced drag.

300 11. Equations and Stability of Motion

The corresponding necessary and sufficient requirements for the stability of

the system will be

Bi > 0 (i = 1, ... ,5), (11.23)

(B 1 B 2 - B 3 )(B3 B 4 - B 2B 5 ) - (B 1 B 4 - B 5 )2 > O. (11.24)

Coefficients Bi can be found in the Appendix to Chapter 11.

11.2 The Equations of Motion

in the Extreme Ground Effect
11.2.1 Order Estimates and Assumptions

We turn to the evaluation of the form of the quintic equations of motion for
h --+ 0; see Rozhdestvensky [181J. In section 4, order estimates were obtained
for the major aerodynamic coefficients on the basis of a mathematical model
of a simple flying wing configuration in immediate proximity to the ground.
In particular, for an adjusted angle of pitch (in radians) and a curvature of
the wing sections of the order of O(h),

(11.25)

As per the previous analysis, the derivatives of the aerodynamic coefficients

have the following order of magnitude:
- - 1
(c y, m Z
~
)h,h,8,8
~
= 0(-)
h ' (11.26)

c~;h,jj,jj = 0(1). (11.27)

Additionally, we assurne that the viscous drag of the configuration does not
vary with a small variation in the ground clearance and the pitch angle. To
evaluate the order of magnitude of the coefficient which represents the cF,
derivative of the thrust coefficient with respect to the relative speed, it is
assumed that the drop of the thrust versus the cruise speed is linear, so that
the (current) thrust T of the engines can be expressed in terms of the relative
speed of motion, instalIed thrust Tm and cruise thrust To as

- u (11.28)
u= Uo'

where Uo is the design cruise speed. Introducing the thrust coefficient as

 11.2 The Equations of Motion in the Extreme Ground Effect 301

wherefrom the derivative of the thrust coefficient with respect to the relative
speed is given by

(11.29)

where CT m = Tm/W is the instalIed thrust-to-weight ratio that characterizes

the relative power capacity of the vehicle.
Recalling previous order estimates and assuming additionally that the
instalIed thrust-to-weight ratio CTm = O(h),

(11.30)

It seems rational to consider the magnitude of Yt as that of the order of 0 (h).

In other words, the ordinate of the thrust line is assumed comparable with
the ground clearance.
Another convention to be adopted is related to the density factor Jl that
enters the equations of motion. Based on the statistics for existing and pro-
jected wing-in-ground-effect craft (see Rozhdestvensky [182]) we can assume
that the product of the vehicle's density and relative ground clearance is of
the order of 0(1). In this case, it is appropriate to introduce, instead of Jl, a
new quantity Jlh = Jl h = 0(1), which can be called the reduced density.

11.2.2 Asymptotic Form of the Equations of Motion for h 4- 0

Employing these estimates and conventions about the orders of magnitude

(in terms of h) and neglecting terms of the order O(h) and higher, we can
reduce the equations (11.18)-(11.20) to
dU'
Jlh dt = 0, (11.31)

d2 h - . :. _ :.
Jlh dt 2 = al h + a2 h + a3 () + a4 (), (11.32)

d 2Ö _ . :. - :..
Jlh i z dt 2 = b1 h + b2 h + b3 () + b5 (), (11.33)

where al = h C;, a2 = h cb,

a3 = h a4 = hci, ce, and Jlh = Jlh. Coefficients
bi (i = 1,2,3,4) are given by the formulas

b1 = hm~, b2 = hm~, b3 = hm~, b4 = hm~.

Note that coefficients ai and bi are of the order of unity, because for each of
the above derivatives,
302 11. Equations and Stability of Motion

. iih09
coefficlent ' " = 0 (1)
h . (11.34)

The system of equations (11.32) and (11.33) has a structure similar to that
of equations (11.9) and (11.10), derived On the basis ofIrodov's assumption
of nO perturbation in speed. It gives birth to a quartic characteristic equa-
tion (11.11) whose coefficients are identical to Ai, i = 1 ... 4, though written
somewhat differently:
Al = _ _ 1_. (b 4 + i z a2), (11.35)
J.Lh ZZ

A2 = -~[a4 b2 - a2 b4 + J.Lh(b3 + i z at}], (11.36)

J.Lh Zz

(11.37)

1
A4 = - 2 - ' (al b3 - a3 bl ). (11.38)
J.Lh Zz
The advantage of the formulation presented above consists of the
reduction in the number of parameters on which Ai depend. In par-
ticular, for h -+ 0, the relative clearance h does not enter the coefficients of
the quartic equation explicitly. Thus, the coefficients of the quartic depend
(nonlinearly) only on the reduced density J.Lh and ratios e/h that characterize
the design geometrical and kinematic parameters of the vehicle. The parame-
ter e = O(h) can be the adjusted angle of pitch () or the maximum curvature
C of the lower surface of the wing related to h, etc.
The variation of speed can be analyzed by introducing "Iarge time" t =
O(l/h) and "very large time" t = O(1/h 2 ). It can be shown that On the scale
of a "Iarge time," the variation of the speed of the vehicle is mostly driven by
perturbations in height and pitch, whereas on the scale of "very large time"
the variation of speed is determined by the perturbation of speed proper.
In the latter case, the perturbed equation for speed is completely uncoupled
from those for height and pitch and has the form

dU' -0 - ,
d = (Ct
J.Lh- - 2Cx ) U, (11.39)
72

where 72 = h2 t is a "squeezed" time variable and cf = cf /h and Cx = Cx/h

are quantities of the order of unity.

11.3 Static Stability and "Binding "Near the Ground

One of the difficult points in the design of lifting systems in the ground effect
is to provide a sufficient margin for the static stability of longitudinal motion.
 11.3 Statie Stability and "Binding "Near the Ground 303

As shown by Irodov [166], Kumar[167], Staufenbiel[l77], and Zhukov [171],

the longitudinal static stability of the motion of a wing-in-ground-effect ve-
hicle depends on the reciprocallocation of the aerodynamic centers of height
and of pitch. The reserve of static stability also depends on location of the
center of gravity. We define positions of the aerodynamic centers of height
and pitch, respectively, as

(11.40)

where the superseripts h and B are ascribed, respectively, to the derivatives

of the lift and the moment coefficients with respect to the ground clearance
and the angle of pitch. Through analysis of the linearized equations for the
longitudinal motion of wing-in-ground-effect vehicles, Irodov [166] showed
that static stability is ensured if the aerodynamic center of height is located
upstream of the aerodynamic center of pitch, so that for x axis directed
upstream, the corresponding static stability criterion can be written as

Xh - xe> O. (11.41)

For the coordinate system adopted in this book, with x axis directed upstream,
the formulation of the condition of the static stability of longitudinal motion,
used by Zhukov and Staufenbiel, implies that the fuH derivative of the lift
coefficient with respect to the ground clearance (for a fixed zero magnitude
of longitudinal moment around the center of gravity) should be positive, Le.,

dCyI = acy _ acyamz/am z > O. (11.42)

dh m.=O ah aB ah aB
Essentially, the latter inequality shows that for a statically stable vehicle,
the stabilizing effect of aCy/ah should exceed the destabilizing influence of
nose-down moment. Note that Zhukov designated the fuH derivative ofthe lift
coefficient with respect to the relative ground clearance as a force stability pa-
mmeter and pointed out that this factor has an effect upon the controHability
of the vehicle and its response to the action of wind. For a properly designed
vehicle, the derivative aCy/ah should be negative, and we can rewrite (11.42)
as
acyamz/am zacy -1 = Xh -1 = Xh - Xe > o. (11.43)
aB ah aB ah Xe Xe
It can be seen from (11.43) that Irodov's criterion deals with pitch stability,
implicitly assuming that height stability is provided, Le., C; < 0, whereas
the criteria of Zhukov and Staufenbiel represent the combined effect of pitch
and height stability.
We designate the above difference in the location of the aerodynamic
centers as SSM = xh - Xe and, when SSM is positive, we refer to it as
to the static stability margin. Suppose that both derivatives (of height and
304 11. Equations and Stability of Motion

of pitch) and the positions of corresponding aerodynamic centers are defined

with respect to a certain reference point, say, to the trailing edge. In practice,
they have to be defined with respect to the center of gravity which can be
viewed as a pivot al point. Introducing the abscissa x cg of the latter and noting
that changing the reference point does not affect differentiation with respect
to h, whereas
(11.44)

we obtain the following formulas for the new positions of the centers of height
and of pitch (when the reference point coincides with the center of gravity)
expressed by corresponding parameters referred to the trailing edge:
o mz0 - xcgm hZ J(
( mz) _ Xo - XcgXh
XOCg = Coy cg
= Coy - x cg Chy - J(
- x cg
(11.45)

where J( = C!/C;. For a foil that possesses height stability and C!/C; < 0,
factor J( is negative. It may be practical to evaluate the variation of the
static stability margin as a function of the position of the center of gravity
(the pivot al point). Simple calculations lead to the following equation:

J( J(
= J( - Xcg (Xh-XO) -+ SSMcg = JC
- X cg
SSM. (11.46)

It can be seen from (11.45)-(11.46) that if a wing is found to be statically

stable with reference to the trailing edge, its static stability is ensured for any
other upstream position 01 the relerence point (center 01 gravity), see Irodov
[166]. A wing that is statically unstable with reference to the trailing edge
remains statically unstable for any other position of the reference point. Si-
multaneously, equation (11.45) shows when the center of gravity is shifted
upstream from the trailing edge, the center in pitch moves in the same di-
rection, whereas the center in height retains its position. Consequently, the
static stability margin diminishes.
At present several optional aerodynamic configurations ofwing-in-ground-
effect vehieles are known that enable us to secure the static stability of
longitudinal motion. In a wing-tail combination, employed in Russian first-
generation ekranoplans, the main wing operating in elose proximity to the
underlying surface was stabilized by a highly mounted tailplane, taken out
of the ground effect. This measure shifted the center of pitch downstream,
thus increasing the static stability margin for a practical range of design pitch
angles and ground elearances. The negative effect, associated with the use of
a large nonlifting tail unit consists in an increase in structural weight and a
noticeable reduction of the lift-to-drag ratio.
 11.3 Static Stability and "Binding "Near the Ground 305

Another possibility is connected with the use of a tandem aerodynamic

scheme with both lifting elements located close to the ground. When de-
veloping his first 3-ton piloted SM-l prototype, Alekseev borrowed a tandem
configuration from his own designs of hydrofoil ships; see Rozhdestvensky and
Sinitsyn [19]. Jörg [7, 8] applied a tandem configuration in the design of his
"Aerofoil Flairboats." Note that to provide static stability to a configuration
comprising two low flying wings, one has to adjust the design parameters
of these wings (pitch angles, relative ground clearances, and curvatures of
pressure surfaces of the wings) in a certain way. The shortcoming of a tan-
dem as an option in providing static stability of longitudinal motion to a
ground-effect vehicle, consists of a somewhat narrow range of pitch angles
and ground clearances for which the flight is stable [22].
A way to reduce the area of the tail stabilizer (or to get rid of it) is
related to appropriate profiling of the lower surface of the main wing. It
means that instead of a wing section with an almost flat lower surface, known
to provide a considerable increase in lift in proximity to the ground, one
has to give preference to wing sections with curved lower surfaces which
secure static stability, although are less efficient aerodynamically. Staufenbiel
and Kleineidam [177] proposed a simple way to augment the static stability
of the Clark-Y foil with a flat lower surface, which consists of providing
this foil with a trailing edge flap, deflected to an upward position. Later
on, the same authors found that if unloading of the rear part of the foil
is combined with decambering of its fore part, the range of static stability
can be enhanced noticeably. A family of foils with an S-shaped mean line
may be shown to possess such a property. In their stability prediction for
an S-shaped foil, Staufenbiel and Kleineidam [177] used an approximation
of the foil's mean line with a cubic spline function. The parameters of this
curve fitting function were selected to provide the maximum range of lift
coefficient in which the foil would be stable. An experimental investigation of
the influence of the form of the airfoil upon its static stability was carried out
by Gadetski [183]. Based on his experimental data, the author concluded that
it is possible to control the positions of the aerodynamic centers by proper
design of the foil. He demonstrated experimentally that an upward deflection
of the rear part of the foil moves the center of height upstream and the center
of pitch downstream. A similar investigation using the method of conformal
mapping and an experimental technique of fixed ground board was done by
Arkhangelski and Konovalov [184].
In what follows, a qualitative analysis will be carried out of the static
longitudinal stability of schematic aerodynamic configurations by using the
mathematical models of the extreme ground effect, see Rozhdestvensky [185].
The simplest case involves a wing of infinite aspect ratio moving in immedi-
ate proximity to the ground. In this case, within the assumptions of extreme
ground effect aerodynamics, it is possible to determine the characteristic
centers, whose reciprocal positions define both the static stability and the
306 11. Equations and Stability of Motion

controllabity of the configuration, in analytical form. The effect of the finite

aspect ratio is estimated by applying the nonlinear one-dimensional theory of
a rectangular wing with endplates in motion elose to the ground. A study of
the static stability of a tandem configuration of infinite aspect ratio is made,
assuming that for h -+ 0, the foHs constituting the tandem work indepen-
dently.

11.3.1 A Single Wing in the Extreme Ground Effect

We consider first a simple example of a single foH at a full rear fiap opening.
As follows from (4.92), for Je = 1 the Hft and the moment coefficients of a
single foH in the extreme ground effect are given by the formulas

r dx 1 1 r xdx
1
(11.47)
Gy = 1 - Jo h*2(X) , mz = 2- Jo h*2(X) ,

where h is the relative ground elearance defined at the trailing edge. Writing
h*(x) as h*(x) = 1 + Öx + tf(x) (where Ö= O/h, t = c/h, and c = O(h) is a
small parameter that characterizes the curvature of the lower surface of the
foH ), we can differentiate (11.47) with respect to hand 0 to obtain

hG~ = 2
r xdx
1
Jo h*3(X) , hm~ = 2
r x dx
1 2
Jo h*3(X) , (11.48)

h f 1 1-h*(x) h h _ f 1 x [1 - -
h*(x)] d
h Gy = 2 Jo h*3(x) dx, mz - J
2 o h*3(x) x. (11.49)

It follows from (11.48) and (11.49) that the quantities hGZ, hG~,hm~, and
h m~ depend upon Ö and trather than upon 0, h, and c. In accordance with
the assumed order relationships of the small parameters, this means that the
above quantities are of 0(1). Most important is that the number of
defining parameters is fewer by one (in this case 2 instead of 3).
Now we can determine the position of the centers of height and pitch in the
following fashion:

(11.50)

Note that according to (11.47), both the Hft and the moment coefficients in
the two-dimensional extreme ground effect depend on the simHarity param-
eters Ö,t, Le., Gy = Gy(Ö,t) and m z = mz(Ö,t). Hence, the derivatives with
respect to height and to pitch can be obtained in an alternative form:

(11.51)
 11.3 Statie Stability and "Binding "Near the Ground 307

Using (11.50)-(11.52), we can write the following alternative expressions for

the abscissas of the centers of height and of pitch:

(11.53)

where
CO = -2 [1 ~dx (11.54)
y Ja h*3'
ce = _ (f(x)dx (11.55)
y 2 Ja h*3 '
We can be conclude from (11.53) that for a flat plate (g = 0), the abscissas
of the centers of height and of pitch coincide. Hence, a Hat foH in the ex-
treme ground effect is neutrally stable. However, (positive) static stabH-
ity of a single foH can be achieved by introducing a nonplanar (curved or/and
polygonal) lower surface to the foH. It follows from the extreme ground-effect
theory that the aerodynamic response of the flow depends on the local dis-
tribution of the width of the channel under the wing

h*(x) = 1 + Öx + gf(x),
where h*(x) = h*(x)/h, Ö = O/h, and g = e/h, and € is a parameter of the
curvature of the lower surface of the foH. This means, in particular, that the
stability margin of a curved foH depends upon the ratio of the curvature to the
ground clearance rather than upon the curvature proper. It follows therefrom
that to ensure the same reserve of stability for smaller ground clearances, one
has to turn to proportionally smaller curvatures.
As discussed earlier in this section, one of the known recipes for im-
proving stability of a single foH is S-shaping; see Staufenbiel and Kleinei-
dam [177], Gadetski [183], etc. A simple representative of such a famHy
is a foH with a sinusoidal lower surface described by the form function
f(x) = -sin(27l"x), x E [0,1]. It turns out that other forms of the lower
surface can be proposed which also lead to the enhanced static stability of
the longitudinal motion, see Rozhdestvensky [185].
In Fig. 11.1 a comparison is presented ofthe behavior oft he SSM = Xh -Xe
versus the design lift coefficient for the previously mentioned sine foil, a
special stab foil whose equation for the lower surface is f(x) = -15x(1 - x)5
and adelta foil, whose lower surface is composed of two flat segments joined
in a vertex located at 25% of the foH chord from the trailing edge. For the
latter foH, the form function that characterizes the curvature of the lower
surface can be written as

{ -X/Xd, for x E [0, Xd]; (11.56)

f(x) = x -1/(1 - Xd), for x E [Xd, 1].
308 11. Equations and Stability of Motion

0.5 +----'-_-'---'-_'--.....1---'-_-'-----'._+

0.4

0.3

02

0.1

0.0 +----.--.,.--r'-r---,-~-...,---..-+
0.2 0.3 0.4 0.5 0.6
Cy
Fig. 11.1. Static stability margins for foils (,X = 00, h = 0.1, solid lines) and
rectangular wings with endplates (,X = 0.625, h = 0.1, c5~p = 0.025). The numbers
correspond to 1: "sine" foil section; 2: "stab" foil section; 3: "delta" foil section.

In all calculated cases, the ratio of the curvature parameter to the relative
ground clearance was € = 0.2, which corresponds to maximum curvatures of
the lower surface is equal to c = €h = 0.2h. The tentative geometries of the
sections of the aforementioned three foil types (number 1 corresponds to the
sine Joil, number 2 to the stab Joil, number 3 to the delta Joi~ with thickness
distribution of NACA-0008 on top of the corresponding lower surface and
appropriate rounding of the leading edge are also shown in Fig. 11.1 for the
design ground clearance h = 0.1. To better demonstrate the form of the foils,
the vertical dimension is multiplied by 4.
Plotted in the same figure are some calculated results for rectangular
wings of finite aspect ratio A = 0.625 for the same foil sections, ground
clearances, and relative curvatures. The gap under the endplates was assumed
to be c5~p = 0.025. Figure 11.1 shows that for a wide range of variation
of the lift coefficient, the increase in the degree of three-dimensionality (an
augmentation of the gap under endplates) brings about a deterioration of
the static stability, although qualitatively the behavior of the static stability
margin versus the cruise lift coefficient is similar to that of the 2-D foil.
Note that, based on the results of their theoretical calculations, Staufenbiel
and Kleineidam concluded that the way of shaping the foil for better static
stability has a similar effect on a rectangular wing with a modified airfoil
section.
Figure 11.2 shows the static stability margin of a 2-D foil with a sinusoidal
lower surface versus the cruise lift coefficient and for different ratios of c / h.
Figure 11.3 presents an estimate of the influence of position Xd of the delta
Joil upon the position of the center of pressure x p , the center of pitch xo, and
 11.3 Statie Stability and "Binding "Near the Ground 309

0.6 ,--------,---.-- --=r

Xh - Xe

0.4 f-----f-'\-------J ~

0.2 ,'-.--T-+-----'<;--t-.----+------I

Fig. 11.2. The statie stability margin of a sine loil in the extreme ground effeet
versus the eruise lift coefficient for different ratios of s / h.

1.0
~

g'"
~ ~.

c:
Q)
Ü
u
'E
CI)
0.6
c:
>- I I
e
'0
Q)
xe
<C 0.4
Cy
!
_ _ _I
0.2
0.0 0.2 0.4 0.6
Xd

Fig. 11.3. The infiuence of the position of the vertex of the delta loil upon the
Hft coefficient and the position of the aerodynamic centers: pressure (black circles);
height (empty circles); and pitch (crosses), e/ h = 1.

the center of height Xh. Figure 11.4 illustrates the dependence of the positions
of the aerodynamic centers of the delta foil upon the ratio e/ h of the pitch
angle to the relative ground clearance. In Fig. 11.5 some calculated data
are plotted, showing the effect of a short rear flap upon the static stability
margin. In particular, it follows from Fig. 11.5 that even a small blockage of
the flow near the trailing edge leads to noticeable diminution of the static
stability margin. This is quite natural because once the flow underneath the
foH stagnates, the form of the lower surface has almost no influence upon
310 11. Equations and Stability of Motion

1.2,---r;=======::::;-J

.. 1.0

~
CD
(,) 0.8
t>
'E
1'11
c:
>- 0.6
~
.1!
0.4

0.2
0 2 4 6
e
h

Fig. 11.4. The positions ofthe aerodynamic centers versus the ratio B/h for adelta
= 0.2, Xd = 0.2.
Joil, c/h

0.3 ~-~--r---~---r---,

Xh-Xe

0.2r----n-TiI_~~5, ___

0.1 1------+----'<--+----"~_+--+_----1

0.0 0.4 0.8 1.2 1.6 e 2.0

h

Fig. 11.5. The influence of a short trailing edge flap on the static stability of a
delta Joil in the extreme ground effect, c/h = 0.2, Xd = 0.2.

the aerodynamics. A conclusion that follows from these results is that if the
static stability of longitudinal motion of the vehicle is secured by profiling
the lower surface of the wing, control by a rear flap should be applied with
caution.
 11.3 Statie Stability and "Binding "Near the Ground 311

11.3.2 A Tandem in the Two-Dimensional Extreme Ground Effect

An option for securing the longitudinal stability of a wing-in-ground-effect

craft consists of using the tandem scheme with both wings operating at small
relative ground clearances.
The parameters of the front and aft wings of the tandem should be se-
lected to provide an appropriate stability margin. Owing to the fact that
the interaction effects of small clearances are of the order of O(h), one can
perform a qualitative analysis of the stability of a tandem foil configuration
based on the aerodynamics of a single foil in extreme proximity to the ground.
Here some results are presented of an analysis of the static stability margin
of a tandem foil configuration (>. = 00) for the particular case when the foils
have equal chords and ground clearances h 1 = h2 = h = 0.1. In the calcula-
tions, the center of gravity of the tandem was assumed to coincide with the
trailing edge of the front foil. The distance between the foils constituted 20%
of the chord.
Figure 11.6a depicts the magnitude of the static stability margin versus
the adjusted pitch angles (related to the ground clearance, Le., iJ1,2 = iJ1,2/h)
of the front and rear foils when both foils are flat. We can conclude upon
examination of this figure that, whereas a single foil is not statically stable,
addition of another foil can result in a positive static stability for certain
combinations of adjusted pitch angles. It can also be observed from Fig. 11.7a
that for better stability it is advisable to put more loading upon the front
foil[185].
The static stability of longitudinal motion for a tandem can be further
enhanced by curving the lower surface(s) of one (or both) foil(s). Fig. 11.6b
reflects the increase in the reserve of the static stability of the tandem when
the front foil has a sine type curvature with amplitude s = 0.2h = 0.02.
A simplified analysis of the influence of the design parameters on the
position of the characteristic centers and the static stability of a tandem

a.
0.2

..--82
Fig. 11.6. 3-D charts of the statie stability of a tandem eomprising two foils versus
their adjusted pitch angles related to the ground clearanee h 1 = h2 = h: a. both
foils are flat, b. front foil has a sine eurvature.
312 11. Equations and Stability of Motion

2.0

h2' h1 =0.5
e: 1.8
0
:;:;0)
Gy= 0.4
Oe:
1.6 Ls =0.38 X '
1]'
.t~
=~
... u.
111'" 1.4 Xh

--
.!!C!)
e:.c
C!)- 1.2
Ue - /'

~"E
1110 1.0 -,.-- ___ ---- Xe
.~ .c. ,
.!!U
0_
f!!o 0.8
'"
.c
U 0.6 Cy1 ' Cy
0.4
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

Fig. 11.1. The calculated relative abscissas of the aerodynamic centers of a tandem
craft versus the lift coefficient of the front Wing as a fraction of the total lift
coefficient, h 2 /h 1 = 0.5.

configuration can be conducted within the linear theory of a foH in in extreme

ground effect; see chapter 3. Assuming that the wings of the tandem are Hat,
have identical chord lengths Go, and an infinite aspect ratio, we can derive
the following formulas for the abcissas of the center of pressure (x p ), the
center of height (Xh), and center of pitch (xe):

(11.57)

(11.58)

where
p = xp - Ls-l.
In these formulas the following notations have been introduced: GYl and
Gy are, respectively, the lift coefficients of the front wing (based on its ref-
erence area) and the tandem as a whole (based on the sum of the reference
areas of the front and rear wings); L s is the distance of the trailing edge of
the front wing from the leading edge of the rear wing, related to Go; and
I\:h = h2 /h 1 , Le., the ratio of the relative ground clearance of the rear wing
to that of the front wing. Introduction of the lift coefficient of tandem as a
whole is practical for the analysis because for a selected wing loading and
cruise speed, the vehicle should be designed for a fixed magnitude of the
cruise lift coefficient.
 11.3 Static Stability and "Binding "Near the Ground 313

Observation of the formula allows us to draw some preliminary conclusions

about the position of characteristic centers in the case under consideration.
In particular,
• when design ground clearances of the front and rear wings are indenti-
cal,3 the center of height coincides with the center of pressure Xh = x P '
As per Zhukov[175], the closeness of these centers to each other improves
controllability of the vehicle; see also 11.2.3,
• in the case under consideration, for fixed magnitudes of the lift coefficients
of the front wing and the tandem as a whole, the position of the center of
pressure does not depend upon the ratio h 2 /h 1 .
Based on the analytical results presented above, the simplified analysis
of the stability is reduced to a calculation of the positions of the three char-
acteristic centers and the static stability margin SSM = Xh - X() versus the
parameters C y !> C y , L s , and II:h. The results of some calculations are pre-
sented in Figs. 11.7-11.11.
Figures 11.7-11.9 exemplify the effect of the ratio of the design ground
clearances h2 /h 1 of the tandem wing elements upon the dependence of the
positions of the characteristic centers on CyJCy. One can conclude from
observation of these graphs that by varying the design ratio of the ground
clearances of the front and rear wings of the tandem, we can bring the cen-
ter of height to different positions coincident with the center of pressure,
upstream or downstream of the center of pressure.

2.0

h2' h1 =1
c:
0
:co)
1.8 Gy= 0.4
~.E ls =0.38
.tS: Xp=Xh
rd-g 1.6
UJ'"
... u..
,SICD
c:.c
CD- 1.4 /'
0'0

-
,g"E! .-/ Xe
.!~ 1.2
,SI 0
u_
!!O
1\1
.c 1.0
0
Cy1 ' Cy
0.8
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Fig. 11.8. The calculated relative abscissas of the aerodynarnic centers of a tan-
dem craft versus the lift coefficient of the front wing as a fraction of the total lift
coefficient, h 2 /h 1 = 1. The center of height coincides with the center of pressure.

3 The Jörg TAF vehicles.

314 11. Equations and Stability of Motion

2.0

h2' h1 = 1.5
c: Cy = 0.4
0 1.8 >-
ti~
ls = 0.38 Xh
J:~
III"E: 1.6 ,,' Xp
1110
111 .... ./
.... u.
-SCD
c:J:
CD- 1.4 , ./
Uo . -.
~
~
Xe
~"E
1110
OE:.&:.
1.2
-SU
u_
1110
lii
J:
U 1.0
CY1 ' Cy
0.8
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Fig. 11.9. The calculated relative abscissas of the aerodynamic centers of a tan-
dem craft versus the lift coefficient of the front wing as a fraction of the total lift
coefficient, h2/hl = 1.5.

0.6

0.5 SSM = Xh-Xe

Cy= 0.4
0.4 Gy= 0.5
c: STABLE
.~ 0.3
111
:::!E 0.2
Cy= 0.7
~
:ö 0.1
.l!!
CI)
u 0.0
~
CI) -0.1 h2' h1 =1
-0.2 ls= 0.38
-0.3
Cy1 ' Cy
-0.4
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Fig. 11.10. The calculated static stability margin SSM = Xh -Xe of a tandem craft
versus the lift coefficient of the front wing as a fraction of the total lift coefficient,
h2/ h 1 = 1, L s = 0.38 for different magnitudes of the total lift coefficient.

Figures 11.10 and 11.11 show more explicitly whether the vehicle is stable
and what is the dependence of its static stability margin SSM = Xh - Xo on
different design factors. In all cases the calculations confirm a conclusion of
[185J that for better statie stability, it is desirable to put more load-
ing onto the front wing. Figure 11.10 shows that (for a fixed fraction of the
front wing loading) reducing the lift coefficient of the tandem brings about
the deterioration of static stability. In practical design terms this means,
 11.3 Statie Stability and "Binding "Near the Ground 315

0.8

0.6
SSM =Xh-Xe =
Ls 0.8
=
Ls 0.38-
0.4 1
c:
.~
IV
STABLE Ls= 0.2
~ 0.2
~
:c 0.0
s
CI)
0
h2' h1 =1
-0.2 Gy= 0.4
""S
CI)
-0.4
UNSTABLE
-0.6 [
Cy1 ' Cy
-0.8
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

Fig. 11.11. The ealculated static stability margin SSM = Xh -XIJ of a tandem craft
versus the lift coeffieient of the front wing as a fraction of the total lift coefficient,
h 2 /h 1 = 1, Gy = 0.4 for different magnitudes of the spacing between the wings of
the tandem.

for example, that for a prescribed speed of the vehicle, a reduction of wing
loading may lead to a lower reserve of the static stability.
Figure 11.11 illustrates the effect of the spacing between the wings of
the tandem. The conclusion here is straightforward: enlarging the distance
between the wings entails augmentation of the static stability margin.
A question may arise why the tandem, even with Hat (identical) wing
elements can be designed to be statically stable, whereas an isolated single
wing shows neutral static stability, see 11.2.1. A reasonable answer to this
quest ion is if the wing elements of the tandem sec ure height stability, the
whole system acquires pitch stability.4

11.3.3 The Degree of Binding of the Vehicle to the Ground

One of the parameters relevant to the design of wing-in-ground-effect vehicles

is the so-called binding criterion; see Zhukov [175]. The binding criterion
enables us to foresee what would be the behavior of the vehicle in height and
pitch with variation in the speed of longitudinal motion. In what follows some
simple quasi-static considerations will be used to relate the derivatives of the
ground clearance and pitch with respect to speed to such characteristics as
the static stability margin, the position of the center of mass, the magnitude
of forward speed, the lift coefficient and its derivatives with respect to height
and pitch; see Rozhdestvensky [182].

4 This is valid for small relative distances of the wings of the tandem from the
ground.
316 11. Equations and Stability of Motion

Suppose that the vehicle operates in a cruise mode in astate of static

equilibrium with regard to vertical displacement and rotation around the
lateral axis that passes through the center of mass. In this case the following
equations of equilibrium hold:

Mg=C pU; S (11.60)

y 2 '
where M is the mass of the craft, g is the acceleration of gravity, Gy is the lift
coefficient, Uo is the cruise speed, Mz and m z are the longitudinal moment
around the center of gravity and its coefficient, respectively, p and S are the
density of air and the wing reference area. Differentiating equations (11. 60),
we obtain

(11.61)

Writing out the full derivatives with respect to height, we obtain

(11.62)

dUo 2dmz
2Uo dh m z + Uo dh = 0. (11.63)

Expressing the total derivatives by partial ones,

dGy aGy aGy dB 1 (aGy aGy dB)

dh = ah + {je dh = Xe Xe {jh + Xe {je dh ' (11.64)

dm z am z am z dB aGy aGy dB
dh = ah + aB dh = ah Xh + aB Xe dh . (11.65)

Then, it follows from (11.62) and (11.63), taking into account (11.64) and
(11.65), that

dUo 2 (aG y aGy dB)

2Uo dh XeGy + Uo Xl! ah + Xl! aB dh = 0, (11.66)

dUo 2 (aG y aGy dB)

2Uo dh xcgGy + Uo Xh ah + X() aB dh = 0. (11.67)

Substracting (11.67) from (11.66), we derive

dUo 2 aGy
2Uo dh Gy(Xl! - X cg ) + Uo (Xl! - Xh) ah = 0, (11.68)

wherefrom follows the first binding criterion (for height)

(11.69)
 11.3 Statie Stability and "Binding "Near the Ground 317

In the same fashion writing out the full derivatives with respect to pitch by
using (11.60),
dUo 2dCy
2Uo dB C y + Uo Te = 0, (11.70)

dUo 2dmz
2Uo dB m z + Uo d'e = 0 (11.71)

with
dCy 8Cy 8Cy dh
(11. 72)
dB = 8B + 8h dB'
dm z 8m z 8m z dh
(11. 73)
d'e = 8B + 8h dB'
It can be easily seen that the binding criteriün für pitch can be obtained
from that for height by a simple substitution h ---+ Band B -+ h:

dB = _ 2UoC y (Xh - Xcg ) = 2UoC y (X Cg - Xh). (11.74)

dUo C~ Xh - Xe C~ Xh - Xe

It follows from observation of the binding criteria (11.69) and (11.74) that
for a statically stable vehicle, the location of the center of mass has a direct
effect upon the response of the vehicle to an increment of speed or a horizontal
gust of wind. Suppose that the vehicle is designed so that the abscissa of the
center of mass coincides with the abscissa of the center of height x cg = Xh.
In this case, as seen from (11.69) and (11.74), it will respond to an increment
in speed by an increment in height without a change of pitch. On the other
hand, if the center of mass coincides with the center of pitch x cg = Xe, then
the vehicle is expected to change pitch angle due to the variation in speed
while flying at the same height. This consideration shows that by a thorough
selection of the reciprocal locations of the three important centers (center of
mass, center in pitch and center in height), one can provide static stability
to the vehicle and also render its motion control system more efficient.
318 11. Equations and Stability of Motion

Appendix: Fürmulas für the Cüefficients Bi(i = 1 . .. 5)

1 - ' .
BI = -. [(2Cx -
ILZ z
cf - C;)i z - m~], (Al)

B2 = ~[(Cp - 2Cx ) (izC; + m~ - YtC!) + iz(2CyC~ - /LC;)

/L Zz

(A2)

B3 = ~{(Cp
/L Zz
- 2Cx ) [Yt (C!C; - C~Ci - C!/L)

+/L(m~ + izC;) + Cim~ - C;m~] + 2Cy(C!m~ + C~/Liz

h 0)
- C xmz +/L (Cym h z-
h ()z + C ym ° C()ym hz - COym h)}
z , (A3)

B4 = ~{(Cp
/L ZZ
- 2Cx )[Yt(C!C; - C~C~ - C~Ci + C!C;) - C;m~
+ C y() m hz + COy m hz - C yhm ()]
z + 2Cy(Cox m zh - °
C xh m z + C()x m hz - C xhm ()z

+/L(C;m~ - C~m~)}, (A4)

B5 = ~{(Cp
/L Zz
- 2Cx)[C~m~ - C;m~ + Yt(C!C; - C~C~)]

+2Cy(C!m~ - C~m~)}. (A5)

Coefficients ean be re-arranged to show the influenee of different eriteria (e.g.

statie stability margin as x h - x() somewhat more explieitly. For example,
the eoefficient B 5 ean be re-written in the following form, eoinciding with
aeeount of differenee in nondimensionalization with Zhukov (note that here
x-axis is directed upstream)

(A6)
12. Simple Mathematical Models of Elastic
and Flexible Wings
in the Extreme Ground Effect

The elasticity and flexibility of the lifting surface can playa particular role
in ground-effect aerodynamics due to the expected increase of dimensions of
wing-in-ground-effect vehicles, use of light materials and fabric, etc. Usually,
to account for elastic properties andjor flexibility of the wing, we have to
consider simultaneously the equations of aerodynamics and elasticity. In a
more profound analysis, the formulations should also cover the equations
of the dynamics of the vehicle. As a relationship, linking the deformations
of a lifting surface with aerodynamic loading, we normally use equations of
unsteady bending of an elastic plate accounting for forces, acting in its camber
plane; see Bisplinghoff et al. [151]. In nondimensional form,l this equation
can be written as
Pw 8 2 f - - 82 f - 82 f _ - +
2-c5
p 8t 2
t + DLlLlf + N xx 8 x 2 + N zz 8 z 2 - P - P (12.1)

The following notations are used in (12.1) for functions and parameters: f
are vertical displacements of points of the elastic plate, Pw is the density
of the material of the plate, p is the density of air, c5t is the thickness of
the plate, jj = 2D/pU;C~, Dis the bending (cylindrical) stiffness, N xx and
N zz are factors that characterize the forces acting in the x and z directions,
respectively, Nxx = 2Nxx j pU;Co , and Nzz = 2Nzz j pU;Co •
Equation (12.1) should be solved with boundary conditions, correspond-
ing to the method of fixing the edges of an elastic andjor flexible lifting
surface, as weH as with initial conditions that define the initial magnitudes
of the deflection and the rate of deflection of the surface.
The formulations for an elasticjflexible wing can be used both for eval-
uating the variation in aerodynamic properties due to the deformation of
the lifting surface and for predicting the occurence of static (diveryence) and
dynamic (/lutter) instability. In what follows, so me simplified schemes are
considered for analyzing conditions of the aeroelastic instability of a wing in
the extreme ground effect. The last paragraph contains approximate predic-
tions of the influence of the flexibility of the foH upon its lift coefficient and
some effects of the porosity of a soft foH.
1 All quantities and functions are rendered nondimensional by using the chord
length Co and the speed of forward motion Uo.

K. V. Rozhdestvensky, Aerodynamics of a Lifting System in Extreme Ground Effect

© Springer-Verlag Berlin Heidelberg 2000
320 12. Simple Models of Aeroelasticity

12.1 Evaluation of Speed of Flutter of a Foil

Close to the Ground

Herein, the definition of a fiutterwill be restricted to that of a phenomenon 01

oscillatory instability 01 an elastic wing in a potential fiow without separation
or shock waves. The essential feature of flutter is a certain inter action of the
degrees of freedom of an elastic wing, entailing the influx of flow energy to
this wing.
In a simple mathematical model of the flutter of a wing in the extreme
ground effect, the analysis will focus on the interaction of two degrees of
freedom related to elastic heave and pitch oscillations in a two-dimensional
flow past a flat plate. To evaluate the speed of flutter, we write equations of
the free elastic oscillations of foil in nondimensional form as
;; 2- '':'
p(h + khh) = Gy - pB(x e - X Cg ), (12.2)

(12.3)
where p = 2M/ pSGo = 2M/ pG; . 1 represents the density of the "vehicle",
i e = le/mG;; le is the moment of inertia of mass calculated with respect
to the elastic axis, Le., the point whose abscissa x = Xe; X cg is the abscissa
of the center of gravity, and kh = WhGo/Uo and k e = weGo/Uo are Strouhal
numbers, based on circular frequencies and associated with heave and pitch
springs, which model two elastic degrees of freedom of the foil. Representing
pitch and heave motions as harmonie oscillations,

B(t) = Bo exp(ikt) , h(t) = hoexp(ikt), (12.4)

where k = wGo/Uo is the Strouhal number based on circular frequency w of

oscillations and i = A, we can find the lift and the moment (around the
abscissa Xe of the elastic axis) coefficients Gy and m z by using solutions ofthe
corresponding problems presented in paragraph 3.5. Heave-related coefficients
were determined in the following form:

GYh (t) = ~ A h exp(ikt), A _ k( 4k + ik 2 - 6i)

(12.5)
h - 1 + ik '

B* _ k(5k + ik 2 - 8i) (12.6)

h - 1 + ik '
where h is the relative ground clearance, measured from the trailing edge of
the foil.
The lift and the moment coefficients for pitch oscillations can be found
by the following formulas:

GyO = ~ A e exp(ikt), m zo = ~ Be exp(ikt), (12.7)

 12.1 Evaluation of the Speed of Flutter of a Foil elose to the Ground 321

where

I - 2
Be = 2{ 3 Xe ik[ (1 -
- '8 4 4] C
xe) - Xe + 2
l [(1 - 2 2] .
Xe) - Xe - tk [1(1
'2 4: - 3Xe)

- 30 5 5] + Cl (31 - 2Xe) + 2C2 [ (1 -

ik [(1 - xe) - Xe 2 2] ]}.
Xe) - Xe (12.9)

The constants Cl and C 2 entering the preceeding equations can be found by

the formulas
C _ ikWl + W 2
1 - 1 + ik
where
1 2 3

2'
W.2 -_'kXe
1 -
.2
+ k2Xe
-,
6
Substituting the expressions for B(t) and h(t), as well as the heave and pitch
contributions to the lift and moment coefficients, we can write the following
system of homogeneous equations with respect to the amplitudes Bo and ho
of the oscillations:

[A h + Ilhk2(1 - K:~)]ho + Bo[A e + Ilhk2(xe - x cg )] = 0, (12.10)

[Bh + Ilhk2(xe - xcg)]ho + Bo[Be + Ilhiek2(1 - K:~K:~h)] = O. (12.11)

Note that the system written above contains a combined parameter Ilh = Ilh
that can be called the reduced density. This parameter was initially introduced
in the analysis of dynamic stability; see section 11. The parameters K:h and
K:eh represent ratios of circular frequencies

(12.12)

where w represents that of free oscillations.

The critical condition of flutter is reached for a combination of K:h and
k = wCo/Uo, or for a given wing, for a combination of the circular frequency
wand the speed of forward motion Uo, when the determinant of the ho-
mogeneous system of equations becomes equal to zero. The magnitude of
Uo corresponding to the critical condition of flutter may be defined as the
speed of fiutter and will be designated as Uf. Composing the determinant .::1
of the system and equating it to zero, we obtain the following equation for
determining K:h and k, corresponding to flutter:

(12.13)
322 12. Simple Models of Aeroelasticity

It is worthwhile mentioning that .1 is complex-valued. Therefore, setting the

determinant to zero means equating to zero both its real and imaginary parts,
Le., lR.1 = 0 and ~.1 = O. The latter equations enable us to determine two
real quantities I\;h and k. It can be seen from the structure of the flutter
equation, written above, that it can be rearranged into a quadratic equation
with respect to K = I\;~:

a(k)K 2 + 2b(k)K + c(k) = 0, (12.14)

where the coefficients a(k), b(k), and c(k) are complex-valued functions of
the Strouhal number k, given by the formulas

a(k) = a3a51\;~h' c(k) = (a2 + a3)(a4 + a5) - ab

b(k) = -~[a3(a4 + a5) + l\;~ha5(a2 + a3)],
where ai = ai (k), i = 1 ... 5 are written below:

al(k) = [Ah(k) + JLhk2(xe - xcg)][Bh(k) + JLhk2(xe - X cg )], a2(k) = Ah(k),

a3(k) = JLhk2, a4(k) = B()(k), a5(k) = JLhiek2.
Separating the real and imaginary parts in (12.14), we obtain the following
two equations with respect to K = I\;~:

lRa(k)K 2 + 2lRb(k)K + lRc(k) = 0, 2~b(k)K + ~c(k) = O. (12.15)

The first of these equations with real-valued coefficients is still quadratic,

whereas the second is linear in K. The solution with respect to K is straight-
forward
K = -lRb(k) ± JlRb 2(k) -lRc(k) ~c(k)
and K = - 2~b(k)" (12.16)
lRa(k)
Selecting the appropriate root of the quadratic equation and equating it to
the solution of the linear equation, we can solve the resulting relationship 2
with respect to k = kr and, consequently, find the corresponding magnitude
of K = Kr = I\;~r' Then the speed of flutter Uf can be determined in the
following way:
TT __ WfCo __ WhCo
f
VI kf I\;hrkf' (12.17)

Equation (12.17) enables us to evaluate the speed offlutter for a given circular
frequency Wh of the heave spring and the chord length Co of the foil. For
similarity considerations it is convenient to form a nondimensional speed of
flutter Ur. This can be done by relating Ur to the quantity WhCo which has
the dimension of speed. The resulting expression can be written as
2 In the case when the flutter is possible.
 12.2 An Aeroelastic Wing in the Extreme Ground Effect 323

0.60

0.40
(/
/"
---
---
-
(Oe/(Oh = 3

-
(Oe /(Oh =2

V~ ~ (Oe /(Oh = 1.5

0.20 /
I

0.00
0.0 2.0 4.0 6.0 8.0
J1,=l1 h

Fig. 12.1. The relative speed of flutter of a flat plate in the extreme ground
effect versus the reduced density J-Lh for different magnitudes of WO/Wh, Xe = 3/4,
x cg = 2/3, i e = 1/18.

- 1
Uf = --. (12.18)
"'hfkf

Some ealeulated results for the two-dimensional fiutter of a fiat plate in the
extreme ground effeet are presented in Fig. 12.1 in form of a graph, showing
the behavior of the nondimensional speed of fiutter versus the redueed density
J.Lh = J.Lh for different ratios of the cireular frequencies of the piteh and heave
springs "'Oh = Wo / Wh. These results eorrespond to the following magnitudes
of input quantities: Xe = 3/4, x cg = 2/3, i e = 1/18. One ean observe from
the graph that for a eonsidered "'9h, a deerease in the redueed density J.Lh
leads to a diminution of the speed of fiutter. It is remarlmble that in the
extreme ground effeet, the influenee of the relative ground dearanee hand
the density of the wing 11-, whieh are eombined in just one parameter I1-h, are
somewhat similar. For example, the same relative reduction of either h or 11-
entails an identieal drop in the speed of flutter. If the same wing flies doser
to the ground, its speed of fiutter beeomes smaller, i.e., the proximity of
the ground reinforces flutter-related restrietions upon the speed of
the vehicle as compared to flight in unbounded air.

12.2 An Aeroelastic Wing in the Extreme Ground Effect

Efremov [70]-[72] was one of the first to analyze at length important matters
of the statie and dynamic aeroelasticity of a lifting surfaee in proximity to
the ground. We consider first a wing of infinite aspect ratio. Then, in the
324 12. Simple Models of Aeroelasticity

case of static aeroelasticity, the elastic defiections of the foil contour can be
determined by the simplified equation

-d 4 f -d 2 f _ +
D dx 4 - T dx 2 =P - p (12.19)

with boundary conditions (for a freely supported plate)

d 2f d2f
f(O) = f(l) = dx 2 (0) = dx 2 (1) = 0, (12.20)

where T is the stretching longitudinal force, T = 2T/ pU; Co and p- - p+ is

the nondimensional pressure difference across the foil, causing its deformation
f(x).
We consider an infinitely thin foil in the extreme ground effect. Recalling
the material of section 3, we can formulate the following (linear) problem
for determining the press ure difference (aerodynamic loading) p- - p+, cor-
responding to the fiow past an elastic foil with a distribution of ordinates
Ys (x) = Yo (x) + f (x), where Yo (x) characterizes the basic configuration of the
foil and f(x) gives the perturbed displacements of points on the foil due to
elasticity:
(12.21 )

(12.22)

<Ph (1) = 0, d'P!!

dx
I -0
x=ü - ,
(12.23)

where <Pli is the perturbed velocity potential in the channel fiow.

It is assumed that in the course of elastic deformation, the end points
of the foil (leading and trailing edges) remain in the same positions, i.e.,
Ys(l) = Yo(1) and Ys(O) = Yo(O) or, in other terms, f(O) = f(l) = O.
Representing the perturbation velocity in the channel fiow as a sum

with terms corresponding to the basic foil configuration and its perturbation,
we can integrate (12.22), accounting for the Kutta-Zhukovsky condition, to
obtain

wherefrom
d<p!! = ~f( ) (12.24)
dx h x.
Now, it is possible to replace the right-hand side of equation (12.19), using
expression (12.24). The resulting equation has the form
 12.2 An Aeroelastic Wing in the Extreme Ground Effect 325

Determination of the eondition of the oeeurenee of the statie instability of an

elastic foH ean be redueed to finding eigenvalues of a two-point homogeneous
boundary problem, eomprising equations (12.25) and (12.20). First, we eon-
sider particular eases. For D = 0, we deal with a membrane foH, and the
eorresponding stability equation takes the form

d2 f 2
dx 2 + ATf = 0, (12.26)

where
(12.27)

The eigenvalues and eigen forms ofthe foH that eomply with problem (12.26),
(12.20) ean be easHy found to be
2
Th )* = mr,
ATn = ( PU C0 (n = 1,2, ... ), fn(x) = sin mrx. (12.28)

As seen from this result, the loss of stability of a membrane foH oeeurs at
a eertain eritical speed, whose minimal magnitude ean be associated with
the speed of divergence UOd ' Relating the speed of divergenee to the quantity
y'T / pCo , which also has a dimension of speed, and setting n = 1, we ean
derive the following expression for the nondimensional speed of divergenee
ÜOd :
ÜOd = UOd = rrVh. (12.29)
y'T/pCo
Another eonerete ease ean be obtained for a purely elastic foH when T =
0, D =1= O. In this case, the divergence condition can be deduced from the
equation
d4 f 2_
dx 4 - ADf - 0, (12.30)

and (12.20) in the form

(n=1,2 ... ), (12.31)

wherefrom one obtains the expression for the relative speed of divergenee of
a purely elastic foH
ÜOd = UOd = rr 2 Vh. (12.32)
y'D/pC~
Deriving eigenvalues for the more general equation (12.25), rewritten in the
form

ADT = Jh~ =
326 12. Simple Models of Aeroelasticity

and subject to conditions (12.20), we obtain the following formula for the
relative speed of an divergence of elastic foH under tension:

(jOd = UOd

JD/ pCg
= 7r 2 v'hJ1 + 2~ =
7r
7r 2 v'hJ1 +
7r
;D- = 7r 2 v'h 1+ ~:D~·
"
(12.33)
It is easy to see that in particular cases when D = 0, T :f. 0 and and
T = 0, D :f. 0, expression (12.33) yields the previous results of (12.29) and
(12.32). Comparison of (12.33) and (12.32) indicates that when an elastic foil
is subject to the action of stretching tension (T > 0), the loss of static sta-
bility occurs at higher speeds. 3 It is important to emphasize that for h -+ 0,
the speed of divergence of elastic foH is proportional to the square
root of the relative ground c1earance.
An analysis of the mathematical model of dynamic aeroelasticity with
the pronounced effects of lag both in terms of aerodynamic loads and elastic
deflection in the extreme ground effect leads to the following generalized
equations of free (eigen) oscillations of a flexible foil in the vicinity of solid
ground: 4

(12.34)

(12.35)
Representing the functions of the perturbed velocity potential and elastic
deformations as

'PlI (x, t) = c,Oh (x) exp(ikt), f(x, t) = f(x) exp(ikt), (12.36)

where k = wCo/Uo is the Strouhal number, i = A and eliminating 'PlI from

the previous two equations, we obtain the following homogeneous ordinary
differential equation of the sixth order with respect to j(x):
6' 4' 2' . ' 2
Dd f _t d f _ 8 k2 (Pw ~) d f 41k df ~f' = 0 (12.37)
dx 6 dx 4 2 P t + h dx 2 + h dx + h .

Taking j(x) = jo exp(ipx), p = n7r, we obtain the following characteristic

equation:

p8t k 2 + h1) p2 + h4kp + h2k = O.

(PW
- 6 - 4 2
Dp + Tp - 2 (12.38)

It is convenient to rearrange equation (12.38) as a quadratic equation with

respect to the eigen magnitudes of the Strouhal number k:
3 It is possible to demonstrate that introduction of small porosity augments the
speed of divergence; Efremov [71].
4 Note that structural damping is not accounted for here.
 12.2 An Aeroelastic Wing in the Extreme Ground Effect 327

Pw
2 ( p8t + h1) k 2 - h
4p k - p2 (Dp
- 4
+ Tp
- 4 2) =
- h O. (12.39)

Solving this equation for k = kn , we obtain

kn
_
-
7l'n [ r.
Mt7l'2n2pw/p+l 1±7l'nvh
VI2' ( PW) (-D7l' 2 n 2 +T-) -p8
l+h p8t 7l'2 n 2
PW]t .
(12.40)
The minimal magnitude of the speed Uo, which turns the expression under
the square root into zero, can be associated with speed of flutter; see [71].
The resulting expression for the relative (minimal) speed of flutter ÜOf can
be found from (12.40) for n = l,p = PI = 71' in the form

ü. _! _ UOf -
(12.41)
Of - [) - JD/pG~ -
One can see from equation (12.41) that the speed of flutter diminishes
as the relative ground c1earance tends to zero. The Strouhal number
k f , corresponding to the flutter condition, is given by the expression
k f = wfGo = 71'
(12.42)
UOf M t 7l'2 pw / p +l'
To a certain extent, the asymptotics of the extreme ground effect furnish an
analog of the "piston theory" and permit us to evaluate the magnitudes of
the critical speeds.
In a similar fashion, we can analyze the static stability of a wing of finite
aspect ratio in motion clos'e to the ground. For example, following the lines
of the approach discussed above, the problem of the membrane wing in flow
is reduced to the problem of the stability of a flexible plate subject to the
action of tensile forces. In the latter case, the problem of eigen solutions is
governed by the equation

(12.43)

and corresponding boundary conditions. An analysis of the dynamic stability

of an elastic lifting surface can be carried out on the basis of the equation
2
e L1j _ ~h + 41kh öl
• A

DL1L1L1j - 'i' L1L1j - 2 Pw 8t + ~ j = 0, (12.44)

P öx h
where j = j(x, z) is a complex nmction of two independent variables.
Efremov also examined the influence of compressibility upon the charac-
teristics of flexible foils. It was shown that at supersonic speeds the flexibility
of the foil can give rise to an additional wave drag. In the case of subsonic
flexible foil at small distances from the ground, a differential equation was
obtained, describing the complex interaction of longitudinal elastic waves of
deformation and acoustic waves in the gas.
328 12. Simple Models of Aeroelasticity

12.3 A Flexible Wing - a Simple Theory

of a PARAWIG

The so-called PARAWIG works as a saH wing in proximity to the ground. It

is common knowledge that the main distinction of a saH compared to a rigid
wing is that its surface is subject to deformation under the action oft he flow,
and the difference of the press ures on pressure and suction surfaces of the saH
is proportional to its longitudinal curvature. According to Twaites[186J, the
problem of the aerodynamics of a saH consists of determining the flow for a
given incidence () and excess of length Ö = ls -1, where ls is the length of the
saH. 5 Essentially, this implies predicting the form of the saH wing, the aero-
dynamic loading and the tension on it. To solve for the flow, the conventional
equation of the wing theory has to be supplemented by a condition of static
equilibrium of each element. Naturally, the resulting system of equations is
different from the traditional formulation for a rigid wing. One of the specific
results of saH theory is the theoretical possibility of the existence of more
than one form of saH for prescribed () and ö. However, there is no difficulty
in choosing a real form. For such forms it turns out that the lift coefficient
Gy of the saH exceeds Gy of the rigid wing approximately by the value of
0.636VJ. In what follows, we analyze very simple mathematical models of a
PARAWIG in the concrete case of a flexible foH of infinite aspect ratio in the
extreme ground effect. Fortunately, in this particular case, Twaites's [186J
verdict "... An analytical solution seems out of the question " is not true.
A complete linearized formulation for the perturbed velocity potential of a
steady flow past a thin flexible foi! near the ground can be written as

ä2 rp ä 2 rp
äx 2 + ä y 2 = 0, (12.45)

ärp dy
y = h±O, (12.46)

y = h, (12.47)

(12.48)

"\l rp -+ 0, (12.49)
Equation (12.47) accounts for the interaction of a flexible foH with the flow
field and means that the pressure difference across the wing is proportional
to the longitudinal curvature, and the factor of proportionality is the ten-
sion 1'.

5 Again, all lengths are related to that of the chord.

 12.3 A Flexible Wing - a Simple Theory of a PARAWIG 329

Applying the asymptotic approach advocated in this book for vanishingly

small ground clearances h -+ 0, we can derive the following simplified formu-
lation for a PARAWIG in a steady two-dimensional extreme ground effect
flow:
d 2rp ll 1 dy
(12.50)
dx 2 h dx'
drpll _ 1t d 2 y
dx - -2 dx 2 ' (12.51 )

d:~l (0) = o. (12.52)

Designating drph/dx = 'Y, we can rewrite formulations (12.50)-(12.52) with

respect to 'Y as
1- 11
'Y = --Ty
2
or
I 1 , e I- I
'Y = hYp + h' 'Y = -2Ty~. (12.53)

For a flexible wing, the ordinates y of the foil camber line with respect to the
flat ground can be written as

y(x) = h + ex + 1](x),
where 1](x) represents the deformation of the wing due to the aerodynamic
interaction with the flow, 1](0) = 1](1) = O.
Integrating the first of the equations (12.53) and using the Kutta-
Zhukovsky condition 'Y(O) = 0 for x = 0, y(O) = h, we obtain

1 e 1](x)
'Y=hY=hx+T· (12.54)

Combining (12.54) with the second equation in (12.53), we obtain the follow-
ing equation with respect to 1]( x):

where

a=~.
The solution of this equation was found in the form

1] = Cl cosax + C 2 sinax - ex.

Equations for determining the constants follow from the requirement that
the deformation vanishes at the edges of the foi!. Then,

1](1) = Cl cosa + C 2 sina - e = 0,

330 12. Simple Models of Aeroelasticity

wherefrom
()
C 2 = -.-.
sma
Finally, the deformation function 7](x) is obtained in form of the expression

7](x) = ()(Si~aX - x). (12.55)

sma
The excess length, which should be equal to 5, can be calculated as

Employing the smallness of () and, consequently, y' = O(()),

from where
~ = ~ [~ (1 + sin 2a) _ 2] . (12.56)
()2 4 sin 2 a 2a
For very large tension l' -+ 00, a -+ 0, we find that

- 2()
Tc::: 3hvlIW' (12.57)

Now, we turn to calculation of the lift coefficient of a PARAWIG in the two

dimensions and in very elose proximity to the ground:

C
Y
= 2 11
0
')'(~) d~ = 2
11 [() ()
0
-x + - (SinaX)]
h
- . - - x dx
h sm a ha 2
a (12.58)
= -2() tan-.

For a small relative excess length, we can express the lift coefficient in terms
of 5 -+ 0, the angle of pitch, and the relative ground elearance h:

C = 2()v!e tan ~. (12.59)

Y h~905 2v!e

The relative effect of the flexibility of the foil can be evaluated by relating
expression (12.58) for the lift coefficient of soft foil to that for a rigid foil
(1' = 00, a = 0). Eventually,

CYT 2tan(a/2)
(12.60)
CYT=oo a

Figure 12.2 illustrates the relative variation of the lift coefficient of a flexible
foil versus the combined parameter 5/ ()2 .
 12.3 A Flexible Wing - a Simple Theory of a PARAWIG 331

2.6
CYT
CyT=CXl
/

V
2.2 ,/

/
1.8

,I
1.4

1.0
1
92
o 2 3
Fig. 12.2. The relative variation of the lift coefficient of a flexible foil versus the
available excess length.

Essentially, the approach applied above suggests that the tension in the
foil is sufficiently large Th = 0(1). It can be seen that at certain (eigen)
combinations of the tension parameter and the relative ground clearance
unstable modes occur of flow past a flexible lifting surface. Examining a
homogeneous problem for deformations T'f of the foil,

T'f(0) = 0, T'f(1) = 0,

it is easy to conclude that the eigenvalues of the product Th are 6

- 2
(Th)n = 22' n = 1,2, ...
-rrn
By using the previous relationships, we can calculate the eigenvalues for the
elongation 8 versus the form number n.
The corresponding eigen forms are described by expression

What are the relationships between the amplitude of a certain harmonie and
the elongation parameter 8? Suppose that the amplitudes of the eigen forms
of a foil are designated as T'fon so that accounting for a = -rrn, the equation of
the nth form becomes T'f = T'fon sin -rrnx. Then the elongation parameter 8 can
be expressed as

6 See also paragraph 12.2, dedicated to the approximate prediction of the occur-
rence of instability on an elastic foil under the action of tensile forces.
332 12. Simple Models of Aeroelasticity

1 1 r
8 = ls -1 ~ 2" 10 (1l"n1]0,J 2 cos 2 mrxdx =
1
41l"2 n 21];",

so that for the nth harmonie, the amplitude will be related to the elongation
8 by
2..ß
1]0 = - - .
" 1l"n
The corresponding lift coefficient Gy" will be

For even forms n = 2m(m = 1,2, ... ), the corresponding contribution Gy" to
the lift coefficient is zero.
For odd forms n = 2m -l(m = 1,2, ... ),

41]02"'_1 8..ß
GY2m _ 1 = 1l"h(2m _ 1) = h1l"2(2m - 1)2·

If we include all harmonics in the expression for the form of the fon, then,

dy
L 1]0" sin 1l"nx,
00

L 1]0" 1l"n cos 1l"nx.

00
1](X) = dx =
n=l n=l

accounting for the the fact that the maximum nth form amplitude is inversely
proportional to the index variable n, the above series are divergent.
It is interesting to extend the analysis to a nonlinear two-dimensional
formulation. In this case, within an approximation of the extreme ground
effect, the basie relationships of the mathematical model can be derived in
the form
~ [Y d<Pl1] = d y , (12.61)
dx dx dx
_ + (d<Ph ) 2 _ d2y
(12.62)
P - P ~ 1- 1 - dx = - T dx 2 '

where y = h + (}x + 1] and 1](0) = 1](1) = o. In terms of the vortex strength

"{ and fon ordinates y measured from the Hat ground surface, the set of
governing equations becomes

1 - (1 - "{)2 = -Ty", y = y(x). (12.63)

Integration of the first equation of (12.63) and accounting for the Kutta-
Zhukovsky condition "{(O) = 0 for x = 0, y = h, gives
y-h
"{=--.
y
 12.3 A Flexible Wing - a Simple Theory of a PARAWIG 333

Substituting the previous result in the second equation of (12.63),

2
h
1-2'=-T- y.
"
y
Multipying the latter equation by y',
,
y' - h2 }L =
y2
-Ty'y"
or

y' + h2~
dx y
(!) = _!~(y')2
2dx '
and integrating, we obtain the expression
2 -
Y+-y=-'2Y'2 + C* ,
h T

where C* is a constant to be determined later. Resolving with respect to y',

or

y' = v 2
-= (C*
T
h2
- Y - -)
Y
(12.64)

After integrating (12.64), we obtain

T r dy
(12.65)
x = '2 Jh .j(C* _ Y _ h2 Jy)
To determine the constant C* of integration we impose the conditions of no
deformation at the edges of the wing,

y(O) = h, y(l) = (} + h,
wherefrom follows the equation for determining C*

ff t+ 9 dy
(12.66)
V'2 Jh .jC. _ Y _ h 2 Jy = 1.

Introducing parameter a = .j2JTh, we can write (12.65) and (12.66) in al-

ternative forms:
334 12. Simple Models of Aeroelasticity

r
x=~J1
1 d17
V(G*-17-1j17)'
a- /
- 1
1+9 d17
V(G* - 17 -ljy)'
with 17 = yjh, C* = G* jh, and Ö = ()jh.
The tension parameter t is defined by the condition of a prescribed elon-
gation of the length of the foil:

(12.67)

which together with equation (12.66) enables us to find the relationship be-
tween elongation and tension. Note that the slope of the foil contour y' in the
equation for 8 has to be expressed through G* and y. The lift coefficient of a
PARAWIG in the nonlinear case is determined by straightforward integration
of the press ure jump across the foil:

Gy = 1o
1 - 2
,(x)[2-,(x)] dx = -T[y'(l) -y'(O)] = -2"[17'(1) -17'(0)], (12.68)
a
Le., to the leading order the lift coefficient of a flexible foil is completely deter-
mined by the slopes of the foil contour at the edges. Performing calculations,
we obtain

Gy = hT [J G* - 2h - JG* - h - () - h ~ ()] . (12.69)

The notion of a flexible wing as a sail brings us to the consideration of an

important property of a sail - porosity. Twaites [186] indicates that the lift-
to-drag ratio for a sail is relatively small, partly due to porosity. Porosity is
known to diminish the lift. The through fiow due to porosity is assumed to
be proportional to the pressure jump. For a porous flexible foil moving in
immediate proximity to the ground, the governing equations have the form
d dy
dx (Y/) = dx + a(xh(2 - ,),

_d 2 y
,(2 -,) = -T dx 2 ' (12.70)

where a(x) 2:: 0 is a nondimensional porosity factor.

In the linearized 2-D case, the governing system (12.70) yields

d, dy
h dx = dx - 2a(xh, (12.71)

Assuming a = const. and eliminating , from the previous expressions, we

obtain the following ordinary differential equation with respect to the y or-
dinates of a porous flexible foil:
 12.3 A Flexible Wing - a Simple Theory of a PARAWIG 335

IJ 2
y 111 + 2-y"
h + -=-
Th Y'= 0 . (12.72)

Representing Y = h + Ox + 17(x), where 17(0) = 17(1) = 0, we can write the

following equation with respect to the perturbation of the form of the foH
17(X):

(12.73)

Although implicitly such a formulation implies that IJ = O(h), it will be

assumed that IJ is sufficiently small in comparion with the relative ground
clearance. Investigation of eigen solutions of a homogeneous equation corre-
sponding to (12.72) leads to the following expression for the relative speed of
divergence:
UOd 7rVh
Jl -
(j _ _
(12.74)
°d - JT / pCo - pIJa T /4hCo '

where IJd is a dimensional porosity factor and Co is the chord of the foH. From
observation of (12.74) we can conclude that the porosity of flexible foU
entaUs a delay in the occurence of instability.
Bibliography

1. Belavin, N.I. (1968, 1977) "Ekranoplans," Leningrad, Sudostroenie.

2. Ollila, R.G. (1980) "Historical Review of WIG Vehicles," Journal of Hydro-
nautics, Vol. 14, No. 3, pp. 65-76.
3. Hooker, S. (1982) "Wingships: Prospect forHigh-Speed Oceanic Transport,"
Janes All the World's Surface Skimmers, Janes Information Group Ltd.,
Coulsdon, Surrey.
4. Rozhdestvensky, K.V. (2000) "Dynasty ofthe Caspian Sea Monster," London,
The Institute of Marine Engineers (to appear).
5. Fischer, H. (1988) "Airfoil Boat Technique Explained on X-113 and X-114,"
International High Performance Vehicle Gonference, Shanghai, November,
pp. 1-19.
6. Fischer, H. (1989) "RFB Research and Development in WIG Vehicles," AIAA
89-1495.
7. Jörg, G.W. (1987) "History and Development of the Aerodynamic Ground
Effect Craft (AGEC) with Tandem Wings," Proc. Symposium "Ram Wing
and Ground Effect Graft, " The Royal Aeronautical Society, London, 19 May,
pp. 87-109.
8. Jörg, G.W. (1997) "Tandem Aerofoil Flairboats (TAF)," Proc. International
Gonference on Wing-in-Ground-Effect Graft, the Royal Institution of Naval
Architects, 4-5 December, London, Paper No. 3.
9. Li Shu Ming, Li Kai Xai (1988) "The 902 Single-Seat Ram Wing Surface
Effect Craft," International High-Performance Vehicles Conference, Shanghai,
China.
10. Li Shu Ming, Li Kai Xai, Yuan Chang Hua, Ni Bo, Ye Xiu Nai (1996) "De-
velopment of Wing-in-Ground Effect at CSSRC," Proc. of the International
Workshop "Ekranoplans & Very Fast Graft, " The University of New South
Wales, Sydney, Australia, 5-6 December, pp. 244-257.
11. Kubo, S., Matsuoka, T., Kawamura, T. (1990) "Development of a Wing in
Ground Effect Craft Marine Slider: IL-Sky 1 as a High Speed Marine Boat,"
Fourth Pacific Gongress on Marine Science and Technology, Tokyo, 16-20
July, Vol. 11, pp. 220-227.
12. Kubo, S., Matsubara,T., Higashia, A., Yamaguchi, N., Kawamura, T.,
Matuoka, T., Mizoguchi,Y., Oomura, Y. (1991) "Development ofWing in Sur-
face Effect Craft, Marine Slider: IL-Sky 2 as a High Speed Boat for Sports and
Pleasure," Fourth International Marine Systems Design Gonference, Kobe,
27-30 May, pp. 155-159.
13. Fischer, H., Matjasic, K. (1997) "The Hoverwing Technology-Bridge Between
WIG and ACV," Proc. International Gonference on Wing-in-Ground-Effect
Graft, 4-5 December, London, Paper No. 5.
338 Bibliography

14. Fischer, H., Matjasic, K. (1998) "From Airfish to Hoverwing," Proc. Inter-
national Workshop "WISE up to Ekranoplan GEMs," The University of New
South Wales, Sydney, Australia, 15-16 June, pp. 69-89.
15. Kühmstedt, T. (1996) "Aerodynamic Design Procedure and Results of the
Development of Commercial WIG Craft," Proc. of the International Workshop
on "Ekranoplans €3 Very Fast Graft," The University of New South Wales,
Sydney, Australia, 5-6 December, pp. 20-37.
16. Ebert, J., Meyer, M. (1998) "Hydrowing - A New Efficient Wing-In-
Ground-Effect Craft," Proc. International Workshop "WISE up to Ekranoplan
GEMs," The University of New South Wales, Sydney, Australia, 15-16 June,
pp. 267-273.
17. Chubikov, B.V., Pashin, V.P., Treshchevsky, V.N., Maskalik, A.I. (1991)
"Ekranoplan - a High Speed Marine Vehicle of New Type," Proc. of the In-
ternational Gonference FAST'91, Trondheim.
18. Volkov, L.D., Pashin, V.M., Ponomarev, A.V., Sinitsyn, D.N., Chubikov, B.V.
(1992) "USSR Research and Design Effort in the Development ofMarine Ekra-
noplans. Transportation Capability of These in Different Water Areas," Proc.
Intersociety High Performance Marine Vehicle and Exhibit- HMPV'92, June
24-27, Arlington, VA.
19. Rozhdestvensky, K.V., Sinitsyn, D.N. (1993) "State of the Art and Perspec-
tives of Development of Ekranoplans in Russia," FAST 93, Yokohama, Vol. 2,
pp. 1657-1670.
20. Kirillovykh, V.N. (1995) "Russian Ekranoplans," Proc. of the International
Workshop on "Twenty-First Gentury Flying Ships, The University of New
South Wales, Sydney, Australia, 7-9 November, pp. 71-117.
21. Sinitsyn, D.N. (1993) "Summary of Results of Work on Creation of Rus-
sia's Ekranoplans and Topical Tasks in Development of Advanced Passen ger
Transport Ekranoplans," Proc. 1st International Gonference on Ekranoplans,
S.-Petersburg State Marine Technical University, 3-5 May, S.-Petersburg,
pp. 6-7.
22. Sokolov, V.V. (1993) "On the Contact of Two Elements," Avico Press,
Moscow.
23. Udalov, K.G., Panatov, G.S., Fortinov, L.G. (1994) "Airplane VV-14," Avico-
Press, Moscow.
24. Sinitsyn, D.N., Maskalik, A.I., Litinsky, L.O. (1996) "The Present Day State
and Prospects for the Development of Commercial Ekranoplans," Proc. of the
International Workshop "Ekranoplans €3 Very Fast Graft," The University of
New South Wales, Sydney, Australia, 5-6 December, pp. 163-176.
25. Sinitsyn, D.N. (1996) "Summary of the Construction of the First Commercial
Ekranoplan,' Amphistar'," Proc. of the International Workshop "Ekranoplans
€3 Very Fast Graft," The University of New South Wales, Sydney, Australia,
5-6 December, pp. 146-151.
26. Tomotika, S., Nagamiya, T., Takenouti, Y. (1933) "The Lift of a Flat Plate
Placed near a Wall with Special Reference to the Effect of the Ground upon the
Lift of a Monoplane Aerofoil," Report of the Aeronautical Research Institute,
Tokyo Imperial University, Vol. 97.
27. Tomotika, S., Imai, J. (1937) "Note on a Lift and Moment of a Plane Airfoil
which Touches the Ground with its Trailing Edge," Report of the Aeronautical
Research Institute, Tokyo, Val. 10, No. 154.
28. Dätwyler, G. (1934) "Untersuchungen über das Verhalten von Tragflügelpro-
filen sehr nahe am Boden," Mitteilungen aus dem Institut für Aerodynamik,
E.T.H., Zürich.
 Bibliography 339

29. Sedov, L.I. (1966) "Two-Dimensional Problems of Hydrodynamics and Aero-

dynamics," Izdatelstvo Nauka, Moscow.
30. Wieselsberger, C. (1922) "Wing Resistance near the Ground," National Ad-
visory Committee for Aeronautics TM 77, April.
31. Serebriysky, Ya. M. (1936) "The Ground Effect on Aeroclynamic Characteris-
tics of Aircraft," Trudy TSAGI imeni N.E. Zhukovskogo, Vyp. 267, Moscow.
32. Keldysh, M.V., Lavrientiev, M.A. (1937) "Motion of a Wing under a Free
Surface of Fluid under Gravity," Trudy Konferentsii po Teorii Volnovogo So-
protivlenia, Moscow.
33. Plotkin, A., Kennell, C. (1981) "Thickness Induced Lift on a Thin Airfoil in
Ground Effect," AIAA Journal, Vol. 19, Nov., pp. 1484-1486.
34. Plotkin, A., Dodbele, S.S. (1988) "Slender Wing in Ground Effect," AIAA
Journal, Vol. 26, No. 4, April, pp. 493-494.
35. Plotkin, A., Tan, C.H. (1986) "Lifting-Line Solution for a Symmetrical Thin
Wing in Ground Effect," AIAA Journal, 24, pp. 1193-1194.
36. Panchenkov, A.N. (1965) "Hydromechanics of a Hydrofoil", Naukova Dumka,
Kiev.
37. Nayfeh, A.H. (1973) "Perturbation Methods," Wiley, New York.
38. Van-Dyke, M.D. (1975) "Perturbation Methods in Fluid Mechanics,"
Parabolic Press, Palo Alto, California.
39. Strand, T., Royce, W.W., Fujita, T. (1962) "Cruise Performance of Channel
Flow Ground Effect Machines," Journal of Aeronautical Sciences, Vol. 29,
No. 6, pp. 702-711.
40. Widnall, S.E, Barrows, T.M. (1970) " An Analytic Solution for Two and
Three-Dimensional Wings in Ground Effect," Journal of Fluid Mechanics,
Vol. 41, Part 4, pp. 769-792.
41. Rozhdestvensky, K.V. (1972) "Steady Motion of a Wing at Vanishing Dis-
tances from the Ground," Trudy Leningradskogo Korablestroitelnogo Instituta,
tom 80, pp. 77-85.
42. Rozhdestvensky, K.V. (1972) "Unsteady Motion of a Wing of Finite Aspect
Ratio at Vanishing Distances from the Ground," Trudy Leningradskogo Ko-
rablestroitelnogo Instituta, tom 80, pp. 69-76.
43. Rozhdestvensky, K.V. (1977) "Asymptotic Theory of a Wing, Moving in Close
Proximity to a Solid Wall," lzvestia AN SSSR, Mekhanika Zhidkosti i Gaza,
No. 6, pp. 115-124.
44. Rozhdestvensky, K.V. (1979) "Method of Matched Asymptotic Expansions in
Hydrodynamics of Wings," Sudostroenie, Leningrad.
45. Rozhdestvensky, K.V. (1979) "Influence of Slots upon Hydrodynamic Char-
acteristics of a Thin Foil in Steady Motion near a Wall," Proc. of the Seminar
on Boundary Problems, Kazanskii Gosudarstvennii Universitet, pp. 201-215.
46. Rozhdestvensky, K.V. (1980) "An Estimate of the Influence of Lateral Gaps
upon Hydrodynamic Coefficients of a Wing of Finite Aspect Ratio in Steady
and Unsteady Motion Near a Wall," Izvestia AN SSSR, Mekhanika Zhidkosti
i Gaza, pp. 122-128.
47. Rozhdestvensky, K.V. (1980) "On the Influence of Endplates upon Hydrody-
namic Characteristics of a Wing in Steady and Unsteady Motion Close to a
Solid Wall," Izvestia AN SSSR, Sibirskoye Otdelenie, Asimptoticheskie metody
v Dinamike Sistem, Nauka, Novosibirsk, pp. 142-148.
48. Rozhdestvensky, K.V. (1992) "Matched Asymptotics in Aerodynamics ofWIG
Vehicles," Proc. Intersociety High Performance Marine Vehicles Conference
and Exhibit HMPV 92, VA, pp. WS 17-27, 24-27 lune.
340 Bibliography

49. Kida, T., Miyai, Y. (1973) "A Theoretical Note on the Lift Distribution of
a Non Planar Ground Effect Wing," Aeronautical Quarlerly, Vol. 24, No. 3,
pp. 227-240.
50. Kida, T., Miyai, Y. (1973) "Jet-Flapped Wings in Very Close Proximity to
the Ground," AIAA Journal, Vol. 10, No. 5, pp. 611--616.
51. Read G.M., (1989) "Extreme Ground Effect," Ph. D. Thesis, University of
Adelaide.
52. Thck, E.O. (1980) " A Nonlinear Unsteady One-Dimensional Theory for Wings
in Extreme Ground Effect," Journal 0/ Fluid Mechanics, Vol. 98, pp. 33-47.
53. Thck, E.O. (1983) " Nonlinear Extreme Ground Effect on Thin Wings of
Arbitrary Aspect Ratio," Journal 0/ Fluid Mechanics, Vol. 136, pp. 73-84.
54. Newman, J.N. (1982) " Analysis of Small-Aspect Ratio Lifting Surfaces in
Ground Effect," Journal 0/ Fluid Mechanics, Vol. 117, pp. 305-314.
55. Rozhdestvensky, KV. (1976) "Nonlinear Theory of a Slightly Curved Foil at
Small Distances from a Solid Wall," Trudy Leningmdskogo Korablestroitelnogo
Instituta, tom 104, pp. 88-95.
56. Rozhdestvensky, KV. (1977) "On the Problem of motion of a Rectangular
Wing over a Rigid Wavy Wall," Trudy Leningmdskogo Komblestroitelnogo
Instituta, tom 115, pp. 62-68.
57. Rozhdestvensky, KV. (1979) "Motion of a Rectangular Wing Above Wavy
Wall at Arbitrary Angles to the Wave Front," Trudy Leningradskogo Korable-
stroitelnogo Instituta, Gidromekhanika i Teoriya Komblya, pp. 29-39.
58. Efremov 1.1., Lukaschik E.P. (1986) "Mathematical Model of Motion of a Wing
above the Wavy Solid Wall," Gidromekhanika, Kiev, No. 53, pp. 3-7.
59. Wang, Q.-X. (1991) "Flow Around an Unsteady Thin Wing Close to Curved
Ground," Journal 0/ Fluid Mechanics, Vol. 226, pp. 175-187.
60. Gallington, R.W., Miller, M.K (1970) "The RAM-Wing: A Comparison of
Simple One Dimensional Theory with Wind Tunnel and Free Flight Results,"
AIAA Paper 70-971, AIAA Guidance, Control and Flight Mechanics Con/er-
ence, Santa Barbara, CA, August.
61. Gallington, R.W., Miller, M.K and Smith, W.N. (1972) " The Ram Wing
Surface Effect Vehicle: Comparison of One Dimensional Theory with Free
Flight Results," Hovering Cmft and Hydr%il, Vol. 11, February, pp. 10-19.
62. Ando, S. (1993) "Note on Prediction of Aerodynamic Lift / Drag Ratio of
WIG at Cruise," FAST 93, Yokohama, Vol. 2, pp. 1657-1670.
63. Rozhdestvensky, KV. (1995) "An Effective Mathematical Model of the Flow
past Ekranoplan with Small Endplate Tip Clearances in Extreme Ground
Effect," Proc. o/the International Workshop on "Twenty-First Century Flying
Ships ", The University ofNew South Wales, Sydney, Australia, 7-9 November,
pp. 155-178.
64. Panchenkov, A.N. (1974) "BaSics of Quadrupole Theory of the Wing Near
a Solid Wall," Asimptoticheskie Metody v Teorii Sistem, Irkutsk, Vyp. 7,
pp. 68-97.
65. Kida, T., Miyai, Y. (1976) "An Alternative Analytical Method for Ground
Effect Airfoils," Aeronautical Quarterly, Vol. 27, pp. 292-303.
66. Rozhdestvensky, KV. (1976) "On Asymptotic Solution of the Integral Equa-
tion of the Lifting Surface in Proximity to the Ground," Trudy NTO imeni
akad. A.N. Krylova, Vyp. 34, pp. 36-46.
67. Rozhdestvensky, KV. (1998) "Aerodynamies of a Tandem of High Aspect Ra-
tio in Ground Effect," Proc. International Workshop "WISE up to Ekranoplan
GEMs," The University of New South Wales, Sydney, Australia, 15-16 June,
pp. 166-184.
 Bibliography 341

68. Efremov, 1.1. (1984) "On the Method of Separation of Solutions in 'Quadrupole
Theory' of the Wing," Dinamika sploshnoi sredy s nestatsionarnimy gmnit-
sami, Cheboksary, pp. 53-58.
69. Efremov, 1.1. (1983) "Problems of Aerodynamics of Flexible Lifting Surfaces,"
Asimptoticheskiye Metody v Teorii Sistem, Irkutsk, pp. 48-79.
70. Efremov, 1.1. (1983) "Problems of Aerodynamics of Flexible Lifting Surfaces,"
Asimptoticheskiye Metody v Teorii Sistem, Irkutsk, pp. 48-79.
71. Efremov, 1.1 (1985) "Unsteady Characteristics and Hydroelasticity of
Thin Wings Near Interfaces," Dissertation of Doctor of Sciences, Institut
Gidromechaniki, Kiev.
72. Efremov, 1.1. (1991) "Oscillations and Stability of an Elastic Plate in Incom-
pressible Flow Near a Rigid Wall," Gidrodinamika, No. 43, Naukova Dumka,
Kiev, pp. 29-34.
73. Lifenko, A.P., Rozhdestvensky, K.V. (1988) "Calculation of Critical Speeds of
Loss of Stability of an Elastic Wing in the Incompressible Flow In Extreme
Proximity to the Ground," Gidrodinamika Ogmnichennykh Potokov, Shuvash-
skiy Universitet, Cheboksary, pp. 76-85.
74. Lifenko, A.P., Rozhdestvensky, K.V. (1989) "Calculation of Critical Speeds
of Static and Dynamic Stability of an Elastic Wing in Very Close Proximity
to the Ground," Trudy Leningmdskogo Komblestroitelnogo Instituta, Matem-
aticheskoe Modelirovanie in Systemakh Avtomatizirovannogo Proektirovania,
pp. 28-37.
75. Kida, T. and Miyai, Y. (1974) "Minimum Induced Drag of Nonplanar Ground
Effect Wings with Small Tip Clearance," Aeronautical Quarterly, Vol. 25, Feb.,
pp. 19-36.
76. Mamada, H., Ando, S. (1973) "Minimum Induced Drag of a Hemicircular
Ground Effect Wing," Journal of Aircmft, Vol. 10, pp. 660-ß63.
77. Mamada, H., Ando, S. (1974) "Minimum Induced Drag of a Semielliptic
Ground Effect Wing," Journal of Aircmft, Vol. 11, pp. 257-258.
78. Ando, S. (1975) "Minimum Induced Drag of Non-Planar Ground Effect Wings
with Small Tip Gaps," Journal of Japanese Society for Aeronautical and Space
Sciences, Vol. 23, No. 256, pp. 304-311.
79. Ando, S. and Yashiro, H. (1975) "Minimum Induced Drag of Ground Effect
Wings," Journal 0/ Hydronautics, Vol. 10, No. 3.
80. Belotserkovsky, S.M. (1965) "Thin Lifting Surface in Subsonic Gas Flow,"
Izdatelstvo Nauka, Moscow.
81. Belotserkovsky, S.M. (1968) "Calculation of the Flow Past a Wing of Arbitrary
Planform in a Wide Range of Angles of Attack," Izvestia AN SSSR, Mekhanika
Zhidkosti i Gaza, No. 4.
82. Belotserkovsky, S.M., Nisht, N.1. (1978) "Separated and Non-Separated Ideal
Fluid Flow Past Thin Wings," Izdatelstvo Nauka, Moscow.
83. Farberov, Va. F., Plissov, N.B. (1967) "Rotational Derivatives of V-Shaped
Wings near the Interface," Trudy NTO Sudproma imeni akad. A.N. Krylova,
Vyp. 88, p. 130.
84. Plissov, N.B. (1968) "Motion of a Heeled Wing above aSolid Ground," Trudy
NTO Sudproma imeni akad. A.N. Krylova, Vyp. 104., pp. 83-86.
85. Konov, E.A. (1970) "Steady Hydroaerodynamic Characteristics of Wings in
Motion with Yaw and Heel near the Interface of two Media," Sbornik NTO
Sudproma imeni akad. A.N. Krylova, Leningrad, Sudostroenie, Vyp. 143,
pp. 89-97.
342 Bibliography

86. Konov, E.A. (1971) "Positional and Rotational Characteristics ofSwept Wings
of Arbitrary Planform and Lateral Cross-Section in Three-Dimensional Mo-
tion near the Air-Water Interface," Sbornik NTO Sudproma imeni akad. A.N.
Krylova, Leningrad, Sudostroenie, Vyp. 168, pp. 275-286.
87. Plissov, N.B., Latypov F.F. (1972) "Research of the Aerodynamic Charac-
teristics of a V-Shaped Wing in Close Proximity to aSolid Surface," Trudy
Leningmdskogo Komblestroitelnogo Instituta, No. 80.
88. Ermolenko, S.D. (1966) "Nonlinear Theory of Small Aspect Ratio Wings,"
Izvestia VUZov, Aviatsionnaya Tekhnika, No. 2.
89. Treschevsky, V.N., Yushin, V.1. (1977) "Calculation of Aerodynamic Char-
acteritsics of Thick Wings of Complex Aerodynamic Configuration, Mov-
ing Above the Boundary of Two Media," Trudy TSNII imeni A.N. Krylova,
Vyp.283.
90. Ivanteeva L.G., Konovalov S.I., Pavlovets G.A. (1980) "Calculation of Aero-
dynamic Characteristics of a Foil Near the Ground for Given Magnitudes of
its Geometrie Parameters," Uchenie Zapiski TSAGI, tom 11, No. 2.
91. Yushin, V.1. (1981) "Use of Continuous Vortex Layers in Treatment of the
Problem for the Incompressible Flow past a Contour," Voprosy Sudostroeniya,
Sero Proektirovanie Sudov, Vyp. 29.
92. Volkov, L.D., Ponomarev, A.V., Treschevsky V.N., Yushin V.1. (1992) "Re-
sults of Aerodynamic Research carried out in Krylov Institute in Support of
Design of Ekranoplans," In Proc. Second International Con/erence on High
Speed Ships.
93. Katz, J. (1985) "Calculation of Aerodynamic Forces on Automotive Lifting
Surfaces," Journal 0/ Fluids Engineering, Vol. 107, Dec., pp. 438-443.
94. Deese, J.E., Agarwal, R.K. (1986) "Euler Calculation for Flow Over a Wing
in Ground Effect," AIAA Paper 86-1765 CP.
95. Kataoka, K., Ando J. and Nakatake K. (1992) "Free Surface Effect on Charac-
teristics of Two-Dimensional Wing," Transactions 0/ the West-Japan Society
of Naval Architects, Vol. 83.
96. Mizutani, N. and Suzuki, K. (1993) "Numerical Analysis of 3-D WIG Ad-
vancing over Still Water Surface," Journal 0/ the Society 0/ Naval Architects
0/ Japan, Vol. 174, December, pp. 35-46 (in Japanese).
97. Volkov, L.D. (1993) "Numerical Investigation of Nonlinear Characteristics of
Foils of Different Shapes in Motion Near the Ground," Proc. 1st International
Con/erence on Ekmnoplans, S. Petersburg State Marine Technical University,
3-5 May, S.-Petersburg, pp. 37-47.
98. Morishita, E. and Tezuka, K. (1994) "Ground Effect Calculation of Two-
Dimensional Airfoil," Transactions Japan Society tor Aeronautical and Space
Sciences, Vol. 36, No. 114, pp. 270-280.
99. Day, A. H., Doctors, L.J. (1995) "A Study of the Efficiency of the Wing-
in-Ground-Effect Concept," Proc. 0/ a Workshop on Twenty-First Century
Flying Ships, The University of New South Wales, Sydney, Australia, 7-8
November, pp. 1-22.
100. Chun, H.H., Park, I.R. (1995) "Analysis ofSteady and Unsteady Performances
for 3-D Airwings in the Vicinity of the Free Surface," Proc. International
Workshop "Ekranoplans - Flying Ships 0/ the 21st Century," The University
of New South Wales, November 7-9, pp. 23-46.
101. Standingford, D.W.F., Tuck, E.O. (1996) "Lifting Surfaces in Ground Ef-
feet," Proc. 0/ the International Workshop "Ekmnoplans fj Very Fast Cmft,"
The University of New South Wales, Sydney, Australia, 5-6 December,
pp. 230-243.
 Bibliography 343

102. Hsiun, C.-M. and Chen, C.-K. (1996) "Improved Procedure for the Inverse
Design of Two-Dimensional Airfoils in Ground Effect," Journal of AircmJt,
Vol. 33, No. 6, November-December.
103. Kühmstedt, T., Milbradt, G. (1995) "Aerodynamic Design ofWing in Ground
Effect Craft," FAST 95, Lübeck-Travemünde, Germany, 25-27 September,
Vol. 1, pp. 597-608.
104. Efremov, 1.1. (1969) "To the Problem of Unsteady Motion of a Thin Foil Near
the Interface," Gidromekhanika, Naukova Dumka, Kiev, Vyp. 15, pp. 29-34.
105. Gur-Milner, S.1. (1975) "The Stable Method of Determination of Aerodynamic
Forces Acting upon a Rectangular Wing near the Ground," Trudy Lenin9md-
skogo Komblestroitelnogo Instituta, Vol. 96, pp. 31-36.
106. Vasil'eva, V.V., Voitkunskiy, Ya. 1., Tkach, A.Ya. (1976) "Unsteady Hydro-
dynamic Characteristics of Bodies Moving Near a Surface," Trudy Leningmd-
skogo Komblestroitelnogo Instituta, No. 104, pp. 17-23.
107. Avvakumov, M.N. (1981) "Unsteady Derivatives of Aerodynamic Coefficients
for Wings of Finite Aspect Ratio, Moving Near Plane and Wavy Sur-
face," Voprosi Sudostroeniya, Seria: Proektirovanie Sudov,TSNII "Rumb" ,
S.-Petersburg, Vyp. 29.
108. Chen, Y.S., Schweikhard, W.G. (1985) "Dynamic Ground Effects on a Two-
Dimensional Flat Plate," Journal of AircmJt, Vol. 22, July, pp. 638-640.
109. Nuhait, A.O., Mook, D.T. (1988) "Numerical Simulation of Wings in Steady
and Unsteady Ground Effect," AIAA Proc. Paper, 88-2546-CP.
110. Nuhait, A.O., Mook, D.T. (1989) "Numerical Simulation of Wings in Steady
and Unsteady Ground Effect," Journal of AircmJt, Vol. 26, pp. 1081-1089.
111. Mook, D.T. and Nuhait, A.O. (1989) "Simulation of the Interaction Between
Aerodynamics and Vehicle Dynamics In General Unsteady Ground Effect,"
Intersociety Advanced Marine Vehicles Conference, Arlington, VA, 5-7 June,
pp. 430-445, also AAA Paper 89-1497-CP.
112. Elzebda, J.M., Mook, D.T., Nayfeh, A.H. (1992) "Numerical Simulation of
Wingships," Proc. of the Intersociety High Performance Marine Vehicle, Con-
ference and Exhibit, HPMV'92, 24-27 June, Arlington, VA, WS28-WS37.
113. Nuhait, A.O. & Zedan, M.F. (1993) "Numerical Simulation of Unsteady Flow
Induced by a Flat Plate Moving Near the Ground," Journal of AircmJt,
Vol. 30, No. 5, pp. 661-617.
114. Ando, S., Ishikawa, M. (1991) "Aerodynamics of a Thin Airfoil Flying over and
in Proximity to a Wavy Wall Surfac-Lifting Surface Theory," Transactions of
Japanese Society for Aeronautical and Space Sciences, 34, No. 103, pp 1-11.
115. Kornev, N.V., Treshkov, V.K. (1992) "Numerical Investigation of Nonlinear
Unsteady Aerodynamies of the WIG Vehicle," Proc. Intersociety High Per-
formance Marine Vehicles Conference and Exhibit HMPV 92, Arlington, VA,
pp. WS 38-48, 24-27 June.
116. Jacob, K. (1987) "Advanced Method for Computing Flow Around Wings with
Rear Separation and Ground Effect," Journal of AircmJt, Vol. 24, No. 2,
February, pp. 126-128.
117. Kawamura, T., Kubo, S. (1988) "Numerical Simulation ofWing in Ground Ef-
feet," Special Publication of National Aerospace Laboratory, SP-9, November,
pp. 223-227 (in Japanese).
118. Akimoto, H., Kubo, S., Yamagata, A. (1998) "Wing Characteristics of Three
Profiles in Surface Effect Using a Simulation at a High Reynolds Number,"
Proc. International Workshop "WISE up to Ekmnoplan GEMs," The Univer-
sity of New South Wales, Sydney, Australia, 15-16 June, pp. 185-190.
344 Bibliography

119. Steinbach D., Jacob K. (1991) "Some Aerodynamic Aspects of Wings Near
Ground," Transactions 0/ Japanese Society /or Aeronautical and Space Sci-
ences, 34, No. 104, pp. 56-70.
120. Hsiun, C.-M. and Chen, C.-K. (1996) "Aerodynamic Characteristics of a Two-
Dimensional Airfoil with Ground Effect," Journal 0/ Aircra/t, Vol. 33, No. 2,
March-April, pp. 386-392.
121. Kim, S. - K., Shin, Y. - K., Kim, J.H. (1995) "Development and Status of
WISES (Wing-In-Surface-Effect Ship)," Proc. Workshop on WIG Ship Tech-
nology, 1-6 June, Korea Research Institute on Ships and Ocean Engineering,
Yusung, Korea, pp. 1-10.
122. Hirata, N., Kodama, Y. (1995) "Flow Computation for Three- Dimensional
Wing in Ground Effect Using Multi-Block Technique," Journal 0/ The So-
ciety 0/ Naval Architects 0/ Japan, Vol. 177, Spring Meeting 17, 18th May,
pp. 49-57.
123. Hirata, N. (1996) "Numerical Study on the Aerodynamic Characteristics of a
Three-Dimensional Power Augmented Ram Wing in Ground Effect," Journal
0/ the Society 0/ Naval Architects 0/ Japan, Vol. 179, pp.31-39.
124. Hirata, N. and Hino, T.(1997) "Investigation of a Three-Dimensional Power-
Augmented Ram Wing in Ground Effect," AIAA Paper 97-0822, 35th
Aerospace Sciences Meeting & Exhibit, Reno, Nevada, 9-10 January, 1997.
125. Barber, T., Leonardi, E., Archer, D. (1998) "Appropriate CFD Technique for
the Prediction of Ground Effect Aerodynamics," Proc. International Work-
shop "WISE up to Ekranoplan GEMs,!! The University of New South Wales,
Sydney, Australia, 15-16 June, pp. 55-68.
126. Barber, T., Leonardi, E., Archer, D. (1998) "Free Surface Deformation Caused
by a Wing in Ground Effect over Water," Proc. International Workshop
"WISE up to Ekranoplan GEMs,!! The University of New South Wales, Syd-
ney, Australia, 15-16 June, pp. 90-101.
127. Patel, V.C. (1998) "Flowat High Reynolds Number and over Rough Surfaces-
Achilles Heel of CFD," 3d Osaka Colloquium, Osaka.
128. Larsson, L., Regnström, B., Broberg, L., Li, D.-Q., Janson, C.-E. (1998) "Fail-
ures, Fantasies, and Feats in the Theoretical/ Numerical Prediction of Ship
Performance," Preprints, Twenty-Second Symposium on Naval Hydrodynam-
ics, Washington, D.C., 9-14 August.
129. Lavrent'ev, M.A. and Shabat, B.V. (1965) "Methods of the Theory of Func-
tions of Complex Variable," Nauka, Moscow.
130. Belotserkovsky, S.M. and Skripach, B.K. (1975) "Aerodynamic Derivatives of
a Flying Vehicle and a Wing at Subsonic Speeds," Nauka, Moscow.
131. Fuks, B.A., Shabat, B.V. (1949) "Functions of Complex Variable and Some
Applications," Gosudarstvennoye Izdatelstvo Tekhniko-Teoreticheskoi Litera-
tury, Moscow-Leningrad.
132. Shadrin, V.P. (1966) "Aerodynamic Characteristics of a Finite-Aspect-Ratio
Wing with a Flap Moving Near a Solid Wall," Gidroaerodinamika Nesuschikh
Poverkhnostei, Naukova Dumka, Kiev, pp. 167-173.
133. Ermolenko, S.D., Rogozin, Yu. A.,Rogachev, G.V. (1972) "Calculation of
Aerodynamic Characteristics of Rectangular Wing with End Plates Moving
with Small Subsonic Speed Near the Ground," Izvestia VUZov, Aviatsionnaya
Tekhnika, No. 3, pp. 105-113.
134. de Haller, P. (1936) "La Portance et la Trainee Induite Minimum d'une Aile au
Voisinage du Sol," Mitteilunger aus dem Institut /ur Aerodynamic der Tech-
nische Hochschule, Zurich, No. 5.
135. Fikhtengoltz, G.M. (1966) "A Course of Differential and Integral Calculus,"
Nauka, Moscow, Vol. 2.
 Bibliography 345

136. Barrows, T.M. & Widnall, S.E. (1970) "The Aerodynamics of Ram-Wing Ve-
hicles for Application to High-Speed Ground Transportation," 8th Aerospaee
Seien ces Meeting, AIAA Paper 70-142, New York.
137. Grebeshov, E.P., Shakarvene, E.P., Tsvetkova, G.1. (1976) "Aerodynamic
Characteristics of a Wing in Proximity of Flat and Wavy Ground," Trudy
TSAGI, Vyp. 1725, pp. 3-28.
138. Gurevich, M.1. (1979) "Theory of Jets of Ideal Fluid," Nauka, Moscow.
139. Belotserkovsky, S.M., Skripach, B.K. and Tabachnikov, V.G. (1975) "A Wing
in Unsteady Gaz Flow," Nauka, Moscow.
140. Söhngen, H., Quick, A.W. (1957) "Schwingungen in Verdichtern," Journal
International de Scienees Aeronautiques, ONERA.
141. Gorelov, D.N., Kurzin, V.B., Saren, V.A. (1975) "Aerodynamies of Cascades
in Unsteady Flow," Nauka, Novosibirsk.
142. Efremov, 1.1. and Unov, S.V. (1984) "Harmonie Oscillations of a Thin Foil
in Subsonic Flow near a Boundary," Trudy OIIMF, Wave Motions in Fluids:
Theory and Experiment, Odessa, pp. 59-64.
143. Spence, D.A. (1956) "The Lift Coefficient of Thin Jet-Flapped Win," Proe.
R. Soe., London, Se ries A, Mathematical and Physieal Sciences, No. 1212,
December, Vol. A238, pp. 46-68.
144. Maskell, E.C., Spence, D.A. (1959) "A Theory of a Jet Flap in Three Dimen-
sions," Proe. R. Soe., London, Series A, June, Vol. 251, pp. 407-425.
145. Bleviss, Z., Struble, R. (1953) "Some Effects of Streamwise Gaps on Aero-
dynamic Characteristics of Low-Aspect-Ratio Lifting Surfaces at Subsonic
Speeds," Report Douglas Aircraft Company, Rpt. SM-14627, April.
146. White, R.B., Landahl, M.T. (1968) "Effeet of Gaps on the Loading Distribu-
tion of Planar Lifting Surfaces," AIAA Journal, Vol. 6, No. 4, pp. 626-631.
147. Gradstein, I.S., Ryzhik, I.M. (1962) "Tables of Integrals, Sums, Series and
Products," GIFML, Moscow.
148. Shumsky, G.M. (1977) "Nonlinear Problem of Motion of a Foil near Wavy
Wall," Izvestia AN SSSR, Sero Tekhieheskikh Nauk, 8, No. 2, pp. 79-84.
149. Belinskiy V.G. et al. (1974) "Maximum and Average Magnitudes of Hydro-
dynamic Characteristics of the Wing, Moving above Nonplanar Ground,"
Gidromekhanika, 11, No. 9, pp. 528-536.
150. Smirnov, V.1. (1957) "Course of Higher Mathematics," Gosudarstvennoye Iz-
datelstvo Tekhiko-Teoreticheskoi Literatury, Moscow, Vol. 2, pp.467-477.
151. Bisplinghoff, R.L., Ashley, H., Halfman, R.L. (1955) "Aeroelasticity,"
Addison-Wesley, Reading, MA.
152. Küssner, H.G. (1936) "Zusammenfassender Bericht über den Instationaren
Auftrieb von Flugeln," Luftfahrforsch., 13, Nr. 12.
153. Gallington, R.W., Chaplin, H.R., Krause, F.H., Miller J.A. and Pemberton,
J.C. (1976) "Recent Advances in Wing-In-Ground-Effect Vehicle Technology,"
AIAA Paper No. 76-874, September.
154. Gallington, R.W, and Chaplin, H.R. (1976) "Theory of Power Augmented
Ram Lift at Zero Forward Speed," DTNSRDC Report ASED 365, February.
155. Gallington, R.W. (1987) "Power Augmentation on Ram Wings," Proe. Sym-
posium 'Ram Wing and Ground Effeet Gruft', The Royal Aeronautical Soeiety,
London, 19 May, pp. 57-86.
156. Krause, F.H., Gallington, R.W. and Rousseau, D.G. (1978) "The Current
Level of Power-Augmented-RAM Wing Technology," AIAA Paper No. 78-
752, April.
346 Bibliography

157. Fridman, G.M. (1996) "Nonlinear Local Solutions to Free Surface Lifting Flow
Problems in Extreme Ground Effect," Proc. of the Second International Gon-
ference "Asymptotics in Mechanics"- AiM 96, S.-Petersburg State Marine
Technical University, S.-Petersburg, pp. 89-96.
158. Gallington, R.W. and Krause, F.H. (1976) "Recent Advances in Wing in
Ground Effect Technology," AIAAjSNAME Advanced Marine Vehicle Gon-
ference, Arlington, VA, September.
159. Munk, M. (1921) "The Minimum Induced Drag of Aerofoils," NACA Report,
No. 121.
160. Barrows, T.M. & Widnall, S.E. (1970) "Optimum Lift-Drag Ratio for a Ram-
Wing Tube Vehicle," AIAA Journal, Vol. 8, No.3, pp. 491-497.
161. Ashley, H. and Landahl, M. (1969) "Aerodynamics of Wings and Bodies,"
Addison-Wesley, Reading, MA.
162. Rozhdestvensky, K.V., Kubo, S. (1996) "A Parametrie Analysis of a Flying
Wing Configuration in Extreme Ground Effect," Proc. of the International
Workshop "Ekranoplans €3 Very Fast Graft," The University of New South
Wales, Sydney, Australia, 5-6 December, pp. 78-96.
163. Voitkunskiy, Ya. I.,Faddeev, Yu.I. and Polischuk, M.A. (1967) "Infiuence of
Viscosity on Profile Lift and Drag Near the Ground", Proc. Leningrad Ship-
building Institute, No. 65.
164. Raymer, D.P. (1989) "Aircraft Design: A Conceptual Approach," AIAA Ed-
ucation Series.
165. Menshikov, V.I. (1974) "Thin Airfoil with High-Lift Devices Near a Screen,"
Samolyotostroeniye i Tekhnika Vozdushnogo Flota, No. 34, Kharkov.
166. Irodov, R.D. (1970) "Criteria of Longitudinal Stability of Ekranoplan,"
Ucheniye Zapiski TSAGI, Moscow, Vol. 1, No. 4, pp. 63-74.
167. Kumar, P.E. (1968) "On the Longitudinal Stability of a Ground Effect Wing,"
College of Aeronautics Report, Aero 202.
168. Kumar, P.E. (1968) " On the Lateral Stability of a Ground Effect Wing,"
College of Aeronautics Report, Aero 207.
169. Kumar, P.E. (1969) "On the Stability of the Ground Effect Wing Vehicle,"
Southhampton University, Ph. D. Thesis, June.
170. Kumar, P.E. (1972) "Some Stability Problems of Ground Effect Vehicles in
Forward Motion," Aeronautical Quarterly, Vol. XVIII, Feb., pp. 41-52.
171. Zhukov, V.I. (1974) "Some Matters of Longitudinal Stability of Ekranoplans,"
Trudy TSAGI, Moscow.
172. Zhukov, V.I. (1977) "Static Longitudinal Controllability of Ekranoplan with
Account of the Thrust Moment," Trudy TSAGI, Moscow.
173. Zhukov, V.I. (1983) "An Investigation of Stability of Ekranoplan with Use of
the Method of Small Increments," Trudy TSAGI, Moscow.
174. Zhukov, V.I. (1983) "On Conditions of Efficient Improvement of Characteris-
tics of Longitudinal Stability with Use of the System of Automatie Control,"
Trudy TSAGI, Moscow.
175. Zhukov, V.I. (1997) "Features of Aerodynamics, Stability and Controllability
of Ekranoplans," Tsentralniy Aerogidrodinamicheskiy Institut im. prof. N.E.
Zhukovskogo, Moscow.
176. Staufenbiel, R. (1976) " Flugeigenschaften von Bodeneffekt-Fluggeraten,"
Zeitschrift fur Flugwissenschaft, Vol. 24, Hefte 1,2, Jan.-Feb., pp. 3-9,
March/ April, pp.65-70; English Trans. "Stability and Control of Ground Ef-
feet Aircraft in Longitudinal Motion", DTNSRDC 77-0048, June 1977.
177. Staufenbiel, R. & Kleineidam, G. (1980) "Longitudinal Motion of Low-Flying
Vehicles in Nonlinear Flowfields," Proc. of the Gongress of the International
Gouncil of the Aeronautical Sciences, Munieh, FRG, pp. 293-308.
 Bibliography 347

178. Staufenbiel, R. W. (1987) "On the Design of Stable Ram Wing Vehicles," Prac.
Symposium "Ram Wing and Graund Effect Graft," The Royal Aeranautical
Society, London, 19 May, pp. 110-136.
179. Diomidov, V.B. (1997) Automatie Control of Motion of Ekranoplan, "Elec-
trapribor," S.-Petersburg, 203 p.
180. Plissov, N.B.,Rozhdestvensky, KV., Treshkov, V.K (1991) "Aerodynamics of
High Speed Ships with Dynamic Support," Sudostroenie, Leningrad.
181. Rozhdestvensky, KV. (1998) "Theoretical Analysis of Dynamies of a WIG
Vehicle in Extreme Ground Effect," Prac. Symposium on "Fluid Dynamics
Problems of Vehicles Operating near or in the Air-Sea Interface, " Research &
Technology Organization, Amsterdam, The Netherlands, 5-8 October.
182. Rozhdestvensky, KV. (1997) "Ekranoplans-The GEMs of Fast Water Trans-
port", Transactions of the Institute of Marine Engineers, London, Vol. 109,
Part 1, pp. 47-74.
183. Gadetski, V.M. (1985) "The Influence ofthe Form ofthe Foil upon its Aerody-
namic Characteristics Near the Ground," Trudy TSAGI, Vyp. 2304, Moscow,
pp. 3-12.
184. Arkhangelski V.N., Konovalov S.1. (1985) "Computational Investigation of
the Influence of Parameters of the Foil upon its Aerodynamic Characteristics
Near the Ground," Trudy TSAGI, Vyp. 2304, M., pp. 12-21.
185. Rozhdestvensky, KV. (1997) "Stability of a Simple Lifting Configuration in
the Extreme Ground Effect," Proc. International Gonference on Wing-in-
Graund-Effect Graft (WIGs), the Royal Institution of Naval Architects, 4 and
5 December, London, Paper No. 16.
186. Twaites, B. (1961) "The Aerodynamic Theory ofSails," Proc. R. Soc. London,
Series A, Vol. A261, No. 402.
Index

acceleration(s) 200,205 - of suction force 45

- of a lifting surface in flight over a composite wing
wavy wall 200, 202, 203 - configuration .2
- wave induced 203 compressibility 121
aerodynamic center - influence of 121,125, 126,323
- of angle of attack 295 compressible flow 121
- of height 295,299,303 - linearised unsteady 9,127
- of pitch 299,303,304 - nonlinear steady 11, 133
aeroelastic wing - past a wing of finite span 124,133
- in extreme ground effect 323 - steady linearized 124
Alexeyev V, 2, 244 cushioning effect 2
- Rostislav V,2
Amphistar 4 definitions 1
aperiodic perturbations 205 derivatives
aspect ratio 58, 63 - of lift and moment coefficients 77,
asymptotic 78,80,82,129
- approach(es) 7 divergence 319
- expansions 7,8 - speed of 16, 325, 326
- solution 25
asymptotics of small clearances 15,16 efficiency 244,2ß3
- aerodynamic 263
Bartini 2,244 - envelopes of the (of power augmenta-
binding 315 tion) 244,245,249,255
- criterion for height 316 endplate(s) 137
- criterion for pitch 317 - of complex configurations 141,145
- degree of 315 - influence of 137, 142
blockage - lower 138,274
- condition of 153, 165 equation( s)
- of longitudinal dynamics 295,296
Caspian Sea Monster 2 - of motion in extreme ground effect
channel flow 8,11,45 301
- in a compressible fluid 121,123 - of motion for h -+ 0 301
Clark-Y airfoil 305 - quartic characteristic 295, 296, 298,
Coanda flow 222,257,261 302
- of finite width past a deflected - quintic characteristic 295,296,299,
leading edge 234 300
- unseparated past a leading edge - Prandtl's integrodifferential 285,
231 289
coefficient (s ) Euler code(s)
- aerodynamic 43 - for steady flows 17
- of induced drag 44 - for unsteady flows 19
- moment 43 extreme ground effect 1
350 Subjeet Index

- ehord dominated 288 - Central 4

- equations of motion in 301
- limiting mathematieal model of 45 indueed drag 44
- span dominated 288 - eoeffieient 44,99, 159
- theory 9 - mlmmum 263, 264
ekranoplan 1,4,5,6 - of the wing 44
elastieity integral
- of the foi! 325 - equation 281,283,284
- equation of the lifting surfaee 292
fiaireraft 1 - formulation 24, 281
fiat plate
- at an angle of pitch 54,64, 104, 110 jet fiap
- with a fiap 55, 70, 103, 110 - a foi! with 111
fiexible wing 328, 329 jet-fiapped wing 10
fiow - of finite span 158
- near a trailing edge with a fiap 239,
241 Kaario 2
- past aleading edge with a winglet KM 2
237 Kutta-Zhukovsky eondition 10, 19,24,
fiow region 37,48,147,166,169
- ehannel 25,27
- edge 25,29
leakage
- upper 25, 28
- lateral 97, 103
fiying wing 2,95,101, 112
fiutter 319,323 lifting line(s)
- speed of 16,320,321,327 - a tandem of 15, 289
foi! - in dose proximity 285
- elastie 324 - single (in dose proximity) 285,287
- membrane 325 - theory 285
- moderately thick 92 lifting surfaee
- slightlyeurved 51 - theory of 23
- thiek 85 lift-to-drag ratio
- with a jet fiap 155 - of a wing in ground effeet 269,304
- with Ioeal suction 109 - maximum 270, 275
frietion drag eoefficient 276 lift
- eoeffieient 43
gap(s) linear theory 47
- effeetive 97,98,111 linearized lifting fiows 47,63
- parameter (generalized) 102,109 - three-dimensional unsteady 79
- under the endplates 96,97,103,112 loeal fiow problems 29,223,231
ground dearanee longitudinal dynamics
- relative 4, 7 - linear equations of 296
ground effeet 1 Loon 2
- extreme 1, 11
- eurved 11,183 Maeh number 9, 122
- maehine(s) 1 margins
gust - of effieieney of PAR 250
- step 212 Marine Passenger Ekranoplans 4
- vertical 205 matehed asymptotic expansions
- method of 6
harmonie oseillations matehing
- of a thin foi! 73 - of fiow deseriptions 36
Hydrofoi! Design and Construetion - stage of 37,38,39,41
Bureau 2 method of
 Subject Index 351

- discrete vortices 17 - near the leading edge

- matched asymptotic expansions 6 with a deflected tip 227
mirror image technique 7,14 rotating flap 166

nonlinear similarity
- flow problems 85 - criterion 102
- unsteady solution for small gaps - parameter 102,255
under endplates 112 slotted wings 171
numerical methods 17 small aspect ratio
numerics 17, 19,20 - wing of 58, 284
- Navier-Stokes solvers 20 speed
- viscous flows 20 - cruise 297,299,300
- of divergence 16,325,326
- of flutter 16,320,321,327
one-dimensional flow problem 11,13
stability 255
- for a flying wing with endplates 13
- height 303
optimal
- to longitudinal motion 302
- jet momentum distribution 268
- pitch 2
- twisting 267
- static 302
- wing in ground effect 16,263,266
static stability 299,302,304
oscillations
- margin 303,304,308,311,313
- of the flap 78
- of a single wing 305,306,307,308
- heave 77, 79
- of a tandem 305,311,315
- pitch 78, 82 Strouhal number 9
osculating parabola 258
suction force 44,81,100,257,360
- coefficient 14,67,70,159
PAR 21,221
- efficiency envelopes 244 tail
- local flow problems 223,231 - surface 2
- mathematical model of (with Coanda - unit 2
effect) 257 takeoff 221
- the thrust recovery of 244 tandem 2,305,311
parabolic arc 54, 90 r-parameter 7
PARAWIG thrust
- a simple theory of 328 - force 82
perturbation(s) - recovery 244,246,260,262
- aperiodic 205 trailing edge
- regular 14 - with a flap 239,241
- singular 8 Trefftz plane 16,44,263
- unsteady height 116
- unsteady pitch 117 uniformly valid flow description 43
power augmentation 4,221 unseparated flow
- envelopes of the efficiency of 244 - Coanda flow past aleading edge
porosity 334 231
- of flexible foil 324,335 - near leading and trailing edges 231
Prandtl-Glauert rule 124 - near a training edge with a flap 239
unsteady
quadruplication 281 - linearised lifting flows 79
- of the integral equation 292 - three-dimensional flow problem 23
quadrupIe theory 14 vortex
- discrete 17
reentrant jet flow 244 - lattice method 17
- efficiency and the thrust recovery of
PAR based on 245 WIG 1
352 Subjeet Index

wing-in-ground effeet 1,4 - flow past aleading edge

- eraft 1 with 237
- vehicle 8 wing(s)
wing-in-surface-effeet ship 1 - aeroelastie 16
wing with a rotating flap 166 - of arbitrary aspeet ratio 63,79
wave(s) - flexible 319
- large-aspect-ratio 66, 84, 285
- indueed forees 186
- optimal 263, 266, 286, 292
- influenee of 13, 183 - reet angular 64,79,96
wavy wall - semielliptie 69
- acceleration of a lifting surface in - slotted 171
flight over 200 - small-aspeet-ratio 58,81,140
- the motion of a lifting system over - with eurvilinear lateral axis 60
- the motion of the wing over (at an - with endplates 95,112,137,274
arbitrary course angle) 191,214 - with a rotating flap 166
winglet WISES 1
 • Search the Springer website catalogue
• Subscribe to our free alerting service for new books
• Look throug h the book series profiles
You want to order? Email to: [email protected]

Journals
• Get abstracts, ToCs free ofcharge to everyone
• Use our powerful search engine LINK Search
• Subscribe to our free alerting service LINK Alert
• Read full-text articles (available only to subscribers
of the paper version of ajoumal)
You want to subscribe? Email to:[email protected]

Electronic Media
• Get more information on our software and CD-ROMs
You have a question on
an electronic product? Email to:[email protected]

: •••••••••••• Bookmark now:

h ."'nger.de/PhyS/
5t>rinto< . Cun_ s.mc.
_ . 7· !H9126_..... - ,
, Springer
Tot +496121145200 · Fu: +49 6121100116
dap . 64J7a1MNTISF · Gh • .