Design and analysis of an open circuit

subsonic wind tunnel

Cite as: AIP Conference Proceedings 2080, 040005 (2019); https://doi.org/10.1063/1.5092923
Published Online: 08 March 2019

Duane Clive Gonsalves, Ashley Glen D’Souza, Gavin Mark Vas, Karthik Madhyastha, Ramya Moodakare,
and Alister Gleason D’Souza

ARTICLES YOU MAY BE INTERESTED IN

Experimental investigation on influence of molybdenum content on tribological properties of

hybrid PTFE composite
AIP Conference Proceedings 2080, 020017 (2019); https://doi.org/10.1063/1.5092900

Simplified method to predict fatigue crack growth life by combining finite element method
and Willenborg’s model
AIP Conference Proceedings 2080, 040006 (2019); https://doi.org/10.1063/1.5092924

Comparative study of journal bearing performance with ferrofluid and MR fluid as lubricant
AIP Conference Proceedings 2080, 040008 (2019); https://doi.org/10.1063/1.5092926

AIP Conference Proceedings 2080, 040005 (2019); https://doi.org/10.1063/1.5092923 2080, 040005

© 2019 Author(s).
 Design and Analysis of an Open Circuit Subsonic Wind
Tunnel
Duane Clive Gonsalves1, a), Ashley Glen D’Souza1, Gavin Mark Vas1, Karthik
Madhyastha2, b), Ramya Moodakare2, c) and Alister Gleason D’Souza2, d)
1
Student, Department of Mechanical Engineering, St Joseph Engineering College, Vamanjoor, Mangaluru - 575028
2
Assistant Professor, Department of Mechanical Engineering, St Joseph Engineering College, Vamanjoor,
Mangaluru – 575028
a)
Corresponding author: [email protected]
b)
[email protected]
c)
[email protected]
d)
[email protected]

Abstract. Wind tunnel is an instrument having wide application in aerodynamics field. It is used to study the effect of
airflow around test objects. The present work deals with the design of an academic open circuit subsonic wind tunnel
using J. B. Barlow’s method. The diffuser and settling chamber geometry are obtained using standard formulae. The
contraction cone profile which is very crucial in ensuring uniform flow with negligible losses and boundary layer growth
is obtained using MATLAB. The modeling is done using SOLIDWORKS and the geometry is analyzed for flow velocity
and static pressure using ANSYS FLUENT. Analysis is performed for three contraction cone profiles and second order
polynomial profile is found to be the most appropriate.

Keywords – Wind tunnel, Computational fluid dynamics (CFD), ANSYS, Aerodynamics

INTRODUCTION
Wind engineering is a field that has been evolving over centuries. A vast part of wind engineering depends
specifically or in a roundabout way on wind tunnels. Wind tunnels are utilized for a wide range of reasons such as
their scope in testing prototypes in early design cycles be it automobiles or airplanes [1]. Computational fluid
dynamics (CFD) is a part of fluid mechanics that utilizes data structures and numerical techniques to solve problems
involving fluid flows. Computers are utilized to perform laborious calculations to simulate the interaction of liquids
and gases with surfaces characterized by boundary conditions. CFD can be used to simulate full scale models in
order to decrease the cost of prototypes and increase the efficiency.
In this paper, the goal is to design and analyze an open circuit subsonic wind tunnel according to academic
standards in order to better understand basic concepts of aerodynamics.

MODEL DEVELOPMENT

Wind Tunnel Design

The very first step in developing the design for the wind tunnel is to find the test section size and shape based on
the requirements. The wind tunnel being designed has primary use in vehicle and aerofoil testing. The design of the

Emerging Trends in Mechanical Engineering 2018

AIP Conf. Proc. 2080, 040005-1–040005-7; https://doi.org/10.1063/1.5092923
Published by AIP Publishing. 978-0-7354-1810-3/$30.00

040005-1
facility must be done efficiently since the test section’s cross-sectional area determines the size of the facility. The
power required by the facility is determined by the test section size, speed and design. Obtaining the largest size of
the test section and the highest speed for the available funds is almost always desired [2]. The design procedure
followed is as follows:

Test section
Test section is the component from where the design procedure starts. The test section is the part of the wind
tunnel in which the model which is to be analyzed is placed. In the operation of a low speed wind tunnel, the test
section has the smallest cross-section while having the highest velocity within the tunnel [3].
The cost and power is dependent on the cross sectional area and the length of the test section. The length of the
test section is usually twice its width. As the air proceeds along the test section, the boundary layer increases which
reduces the effective area of the airflow. This causes an increase in the flow speed outside the wall boundary layers.
Contraction cone outlet usually does not have a constant velocity distribution. To counter this, a settling chamber is
placed before the test section. The test section design procedure starts from determining the Reynolds number (Re).

Reynold’s number

ρvd
Re  (1)
μ
where,
v = velocity of the fluid
d = the characteristic diameter
ρ = the density of the fluid
µ = the dynamic viscosity of the fluid

Friction factor

We have Prandlt universal law of friction relating Reynolds number and friction factor (f).

f = [2log (Re √f)-0.8] -2 (2)

Flow Loss

For constant section, we have Darcy – Weisbach equation. For loss

ΔP flv 2
 (3)
ρ 2D h

Diffuser
The high velocity air at the test section outlet decelerates in the diffuser, thereby achieving static pressure
recovery and reducing the load of the drive system. To reduce the power losses, it is of critical importance that the
diffuser design must be optimum. Since there are chances of flow separation in the diffuser, it is of utmost
importance to minimize the design errors. After a brief study on flow behavior for different angles of diffuser it is
found that an optimum diffuser angle of 5° can be taken [4]. The inlet side of the diffuser is same as the test section.
Area ratio can be calculated from dimension R1 and R2 which are shown in the Fig 1. The length can be calculated
from the angle and area ratio using trigonometric relations.

040005-2
 R2
Area ratio=
R1
(4)

(a) (b)

FIGURE 1. a) CAD model of diffuser b) Diffuser diagram depicting the dimensions

 (0.5R 2  0.5R1) 
θ  tan  1   (5)
L
 

Diffuser Loss

The theoretical calculation of expansion loss is more difficult and is obtained presently only by experimental
correlation. The expansion loss coefficient is represented as the product of two factors, one a function of equivalent
conical angle and the other a function of diffuser area ratio. The loss coefficient can be decomposed as a sum of a
friction loss coefficient and an expansion loss coefficient as indicated by

K K K (6)
where,
friction loss coefficient,
 1  f
K f   1  2
 (7)
 AR  8sinθ
expansion loss coefficient,

  A r  1
2

K ex  K e θ  (8)
Ar

Contraction Cone
The design of contraction cone must be done to produce a uniform and steady stream at its outlet, and also
requires the avoidance of flow separation [5]. Also, the contour should give low adverse pressure gradient at the
ends of the contraction cone. For a good quality flow, the contraction length must be short, and the contraction ratio
must be high [6]. Taking all these into consideration the contraction cone is designed using the method developed by
Jonathan H Watmuff according to which a variety of contraction shapes can be specified by two simple power law

040005-3
arcs which are joined together at an inflection point [7]. The shape is plotted in MATLAB using the two equations
specified below:

A = R1 - A1xn (9)
B = R0 + A2 (Lc-x)n (10)

‘A1’ and ‘A2’ are matching constants and exponent ‘n’ controls the steepness of the contraction shape. ‘x’ is the
increment. As the exponent is reduced, the wall curvature near the entry and the exit decreases so that the pressure
extreme and the adverse pressure gradients also maybe expected to decrease. The exponent n in equation 9 and 10 is
varied to get the required profile for different order curves. Contraction cone profiles for order 2, 3, 4 are obtained
from MATLAB. The resulting profile for order 2 is shown in Fig 2.

(a) (b)

FIGURE 2. a) CAD model of contraction cone b) Contraction cone profile obtained from MATLAB software

Losses in contraction cone

The loss in the contraction cone is considered to be from friction only. A reasonable approximation for contraction
cone loss coefficient (K) is given by
K 0.32 f (11)
Where f - Average of entry and outlet friction factor
L - Length of Contraction Cone,
D - Height of test section.

Dynamic pressure (q) is

q =0.5 x ρ x V2 (12)
Therefore, the contraction cone loss is
Kxq (13)

To reduce the turbulence of air entering the contraction cone, a settling chamber is inserted as shown in Fig 3.
The settling chamber diameter matches the contraction cone diameter. A settling chamber length of 0.5 times the
inlet diameter is often used. Settling chambers having constant area sections are inserted at the inlet and exit of
contraction cone.
The purpose of installing a flow straightener or honeycomb is to remove the effect of swirl and lateral mean
velocity variations, provided the flow yaw angles are not greater than about 10o. The honeycomb cells tend to stall if
the yaw angles are large and consequently reduce the effectiveness whilst increasing the power loss. The typical
length-to-cell diameter ratios are in the range of 6-8 and porosity is typically in the vicinity of 0.8. Porosity is taken
as the ratio of void volume to total volume.

040005-4
 Settling Chamber

(a) (b) (c)

FIGURE 3. Flow Straighteners (a) Circular (b) Square (c) Hexagonal

Drive System
Since the wind tunnel is of suction type, the fan is located at the downstream of the diffuser section. The thrust
of the fan and the drag of the various tunnel components vary with the square of the fan rpm. Therefore, to maintain
uniform velocity in the test section, speed adjustments should be made by varying fan rpm rather than fan pitch.
Therefore, discharge (Q) is given as the product of area and velocity at outlet.
Q = av (14)

RESULTS AND DISCUSSIONS

The geometry is modeled in SOLIDWORKS and imported to ANSYS FLUENT. K-omega SST model is used to
analyze the model since it is widely used in aerospace applications. The geometry is analysed for its flow velocity
and static pressure. Analysis is carried out on the three geometries and the variation of velocity along contraction
cone and test section in axial direction is displayed in Fig 4.

FIGURE 4. Graph depicting the velocity from the start of the contraction cone to the end of the test section in axial direction

040005-5
 An inlet velocity of 3.737m/s is required for a velocity of 30m/s in the test section which is found out by using
continuity equation. The air is made to settle after contraction. After this the velocity steadily increases along the test
section as found out through analysis and this is verified by literature survey. This can be corrected by adding a
positive angle throughout the length of the test section. The velocity increases more sharply from 0.3m to 1.3m as
seen in Fig. 4. Steeper the curve of the geometry, higher are the chances of flow seperation.

(a) (b)

FIGURE 5. Order 2 contours a) Velocity magnitude b) Static pressure

As seen in Fig 5a the velocity in the axial direction is close to 30m/s which is indicated by the legend. The
velocity increases greatly at the cone outlet. Boundary layer is formed at the wall due to the friction between the
airflow and the wall. The boundary layer increases along the test section. Due to this phenomenon there will be a
slight increase in velocity than expected at the outlet. The boundary layer must be as small as possible in the test
section for efficient use. Therefore, slight positive angles are usually given to the test section to avoid this.
Atmospheric pressure is assumed at the inlet, since the settling chamber was not considered in the simulated
cases. Due to area decrement and friction losses, the static pressure drops downstream. Most of local losses are due
to turbulence dissipation and flow separation. The region where the probability of flow separation occurring the
most is in the region where there is a dramatic decrease in static pressure [8]. The maximum static pressure drop was
detected close to the contraction cone outlet as seen in Fig 5b.
Order 2 geometry produced the least drop in pressure whereas order 4 geometry produced the highest drop in
pressure at the contraction cone outlet. Since order 2 geometry has the least chance of flow separation it is employed
in the wind tunnel.

CONCLUSION
Subsonic wind tunnel can be designed using the above-mentioned method. Losses can be calculated, and a
suitable drive system can be selected. For the selected dimensions order 2 contraction cone profile is best suited
among the other profiles since the drop in static pressure is minimum compared to the other profiles. To counter the
increase in velocity in the test section, a slight positive diffuse angle can be provided throughout the test section.

REFERENCES
1. J. D. Anderson Jr., Fundamentals of Aerodynamics, (Mc-Graw-Hill Inc., U.S.A., 1984), pp. 1-140.
2. J. B. Barlow, H. R. Jr. William, and A. Pope, Low speed Wind Tunnel testing (John Wiley & Sons, U.S.A.,
2010), pp. 1-100.
3. L. Cattafesta, J. Mathew and C. J. Bahr, “Fundamentals of Wind Tunnel design”in Encyclopedia of Aerospace
Engineering, edited by Richard Blockley and Wei Shyy (John Wiley & Sons, Ltd, USA, 2010), pp. 1-9.
4. R. D. Blevins, Applied Fluid Dynamics Handbook, (Van Nostrand Reinhold Co., New York, 1984), pp. 144-
146.
5. R. D. Mehta and P. Bradshaw, Aeronautical Journal-New Series 83,443-449(1979).
6. T. Morel., J. Fluids Eng 97, 225-233(1975).

040005-6
7. Jonathan H. Watmuff, “Wind Tunnel Contraction Design” (Proc. of 9th Australasian Fluid Mechanics
Conference, Auckland, 1986), pp. 472-475.
8. D. E. Ahmed and E. M. Eljack, “Optimization of model wind-tunnel contraction using CFD” (Proc. of 10th
International Conference on Heat transfer, Fluid Mechanics and Thermodynamics, Florida, 2014), pp. 87-92.

040005-7