Wilbeck, J.S.

, Impact behavior of low strength projectiles, in Report, AFML-

TR-77-134 1978, Air Force Materials Lab Wright-Patterson AFB OH.

J EVELVh

NPACIT *EHAVIOR OF LOW STRENGTH PROJECTILES

Metals Behavior Branch

Metals and Ceramics Division

ILaJ
Cww~s, TECHNICAL REPORT AFML-TR-77-134

S 7 FinalFRe,11 I JulAB7ORATOR

pproved for public release; distribution unlimited].

AIR FORCE MATERIALS LABORATORY

AIR FORCE WRIGHT AERONAUTICAL LABORATORIES
2,
AIR FORCE SYSTEMS COMMIAND3
WRIGHT-PATTERSON AIR FORCE JE OHIC 454MJ 1 2
 NOTICS
When Government drawings,, specifications. or other data are used for enty pur-
pose other than In connection with a definitely- related Government procurement
operation. the United states Government thereby Incurs no responuibility n~r any
obligation whatsoever; and the fact that the government may have formulated,
furnished. or In any other way supplied the said drawings, specifications, or other
data, is not to be regarded by Implication or otherwise as In any manner licen-
sing the holder or any other person or corporation, or conveying any rights or
permission to manufacture, use, or sell any patented inventlon that may In any
way be related thereto.
This report has been reviewed by the Information Office (01) and is releasable
to the National Technical Information Service (NTIS). At NTIS, it will be avail-
able to the general public, including foreign nations.
This technical report has 'been reviewed and ic approved for publication.

7rTngiRgneer

FOR THE1 COWANDIR

WALTh1U H. RhIMA0i, Ar~t!,.Che

Metals Behavior Bronch
Metals and Ceramics Division
Air Force Materials Laboratory

"If yo~~r address has changed, If you wish to be removed fromi our mailing iiet,
or If the addreasie is no longer employed by your organinatioc. please notify
A~~/!1kL-0 W-PAPP, OR 45433 to help us maintain a current smailing list".

Copies of this report should not be returned tunless return is required by se-
curity considerations, contractual obligations, or notice on aspecific dociment.
AIR FORCE"I6,gO s~ - 300w
NOWb 107
 UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS PAGE (When Data 81119e0 ,

REPORT DOCUMENTATION PAGE REU CO INGORM

B UF
T IMPORT u.
NH OOVT ACCESSION NO, I. RECIPIENT'S CATALOG NUMIER

AFML-TR-77-134
S7 TITL.2 (and •,beftie) S. TYPE OF REPORT G PERIOD COVERED
IMPACT BEHAVIOR OF LOW STRENGTH PROJECTILES Final Report,
1 July 1974 - 31 May 1977
6. PERFORMING ORG. REPORT NUMUER

T, AUTNOR(a) 4. CONTRACT OR GRANT NUMPER(a)

James S. Wilbeck
S. PERFORMING ORGANIZATION NAME AND ADORESS/ WQO(RAM ELEMENT.
ARE A & WORK '•NITN PROJECT.
UNUME~qS TASK

Air Force Materials Laboratory

Air Force Wright Aeronautical Laboratories 2418030
24180302
Wright-Patterson Air Force Base, Ohio 45133
11. CONTTOLLIN0 OFFICE NAME AND ADDRESS It. REPORT DATE

Air Force Materials Laboratory July 1978

Air Force Wright Aeronautical Laboratories is. NUMUER orPAGES
Wright-Patterson Air Force Base, Ohio 45433 129
17t MONI71ORIN4 AGENCY NAME & ADORESS(If dliferent from Cnltrolling Olfio) IS. SECURITY CLASS. (of thie eport)

Unclassified
IS&, SICATION DOWN OR AO'NO
IA
5•N LOULE

S. STRIGUTION STATEMENT (of AI#e Ropoffl

Approved for public release; distribution unlimited.

I?. DISTRIBUYION STATEMENT (*I tIh' *e , u,,t entered in 310. 20. It dilffeent from Report)

IS. SUPPLEMENTARY NOTES

lIS. K1%* WORDS (Coetlhww an ,revo. aide i neoessary and identify by block number)

Impact, Foreign Object Damage, Soft Bcdy Impact, Bird Impact, Fluid Impact,
Hydrodynamic Theory, Fluid Jet

0.ISTRACT (Coltnue an revese eldelt nr eessary and ln'ini~ty by bleck number)

An unsteady, hydrodynamic model of the impact of a low strength projectile on

a rigid or semi-rigid target has been developed. Impact tests of various mate-
rials including birds, gelatin, and RTV rubber have been conducted to validate
this theory. At velocities above 100 m/s, these materials were seen to flow
as a fluid upon impact. The resulting impact pressure histories were very simi-
lar to those predicted using the model. The impact process was dominated by an
initial shock phase which was quickly followed by a release to a steady flow
DD om 1473 EDITION OF IR oV 65 1 OBSOLETE UNCLASSIFIED
SECURITY CLASIiFICATION 3F THIS PAGE (fien Doesa srota)
 UNCLASSIFIED
JCtI CLASSIFICATION OF THIS PAOEIR(Whe Del. bolomd)

,I O. Cont'd)
condition. This model should be applicable to the impact of any material
for whi ýh the shock pressures generated during impact are much greater than
the strength of the projectile but are less than the strength of the target.

T1

UNCLASSI FIED
SECURITY CLASSIFICATION OF THIS PAOIIbefn Des. Untore)
AFML-TR-77-134

FOREWORD

This report summarizes some of the inhouse research conducted by

Captain James Wilbeck, Metais Behavior Branch, Metals and Ceramics Division,
Air Force Materials Laboratory. The work was conducted under Prodect 7351,
Task 735106, "Work Unit 735106B3," Dynamic Behavior of Engine Materials."
The experimental data was obtained by the University of Dayton Research
Institute under Contract F336i"5.-73-C- 5027 which was monitored by Capt.
Wilbeck and Dr. Alan K. Hopkins.

The research was conducted during the period July 1974 to May 1977.
This report was initially submitted to the Graduate College of Texas A&M
University as a dissertation in partial fulfillment of the requirements
for the degree of Doctor of Fiilosophy in May 1977. It was subsequently
submitted for publication as an AFML report in June 1977.

ACCESSION for

NTIS )I section%

DDC IIIIff section 0

ju~IC~T~
I
...........
i.....
AFML-TR-77-134

TABLE OF CONTENTS
SECTION PAGE
I INTRODUCTION 1
II LITERATURE SURVEY 4
III IMPACT THEORY 16
1. Hydrodynamic Theory 16
2. Special Considerations 34
3. Equations of State 45
4. Limitations of Theory 55
IV EXPEkIMENTAL TECHNIQUE 57
1. Gun Systems 57
2. Hopkinson Bar 59
3. Pressure Plate 64
V RESULTS AND DISCUSSION 67
1. Hopkinson Bar Tests 67
2. Pressure Plate Tests 70
VI CONCLUSIONS 96
REFERENCES 97
APPENDIX A RADIAL PRESSURE DISTRIBUTION 103
APPENDIX B GENERAL PRESSURE-DENSITY RELATIONSHIPS 105
APPENDIX C MATERIAL PROPERTIES 112

78 40
AFML-TR-77-134

LIST OF ILLUSTRATIONS
FIGURE PAGE
1. One Dimensional Shock Flow. a) Shock Propagating into
a Fluid at Rest, b) Flow Brought to a Rest Across
the Shock, c) Standing Shock. 18

2. Effect of Compressi ility on the Hugoniot Pressure

for Water. (1 MN/m - 10 Bars) 20

2. Shock and Release Waves in Fluid Impact. 21

4. Comparison of the Shock Velocity, us, and the Sound
Speed in the Shocked Region, cr. 23
5. Variation of the Duration of the Hugoniot Pressure with
Impact Velocity for a Cylindrical Projectile of Water. 23
6. Variation of Critical Length with Impact Velocity
for Water. 25
7. Radial Distribution of Steady Flow Pressure.
8. Compressibility Effects in Steady Flow. a) Stagnation
Pressure, b) Pressure Distribution. 31
9. Standing Shock in Steady Flow. 32
10. The Impact of a Yawed Projectile. 35
11. The Impact of a Curved Surface on a Flat Target. 37
12. Oblique Impact of a Cylinder on a Rigid Plate. 37
13. Variation of Hugonlet Shock Pressure with Angle of
Impact for Water. !I MN/m4 - 10 Bars) 39
14. The Steady Flow Phase of an Oblique impact. 39
15. Steady Flow Pressure Distribution for a 300 Impact of
a Plane Jet of Water (Taylor, Reference 59). 40
16. Relationships Between the Hugoniot Shock Pressure and
Particle Velocity for Water with Various Porosities.
(1 MN/mn2 = 10 Bars) 42

17. Relationships Between the Stagnation Pressure and the

Impact Velocity in Steady Flow for dater with Various
Porosities. (1 MN/m' - 10 Bars) 42

vi
AFML-TR-77-134

LIST OF ILLUSTRATIONS (CONTINUED)

FIGURE PAGE

18. The Variation in Steady Flow Pressure Distribution

for Water Due to the Addition of Porosity. 43

19. Impact on an Elastic Target During the Early

Shock Phase. 43

20. Hugoniot Pressure-Volume Rel tionships for Water with

Various Porosities. (1 MN/mr " 10 Bars) 52

21. Relationships between the Shock Velocity and the

Particle Velocity for Water with Various Porosities. 53

22. Isentropic Pressure-Volume Relationships for Water

with Various Porosities. (MN/mr - 10 Bars) 54
23. Overall View of the Gun Range Facility Used for Bird
Testing. 58

24. Sabot Stopper for Use with Powder Gun. 59

25. Cylindrical Projectile with Balsa Wood Sabot. 60

26. Sabot Stripper for Use with Compressed Air Gun. 60

27. Schematic of Hopkinson Pressure Bar. 62

28. Overall View of the Hopkinson Bar Apparatus. 63

29. Pressure Plate Apparatus. 65

30. Typical Hopkinson Bar Data, Normal Impact. 68

31. Relationship Between the Normal Component of

Momentum and the Impulse Imparted to the Target
for the Impact of Porous RTV. 69

32. Relationship Between the Normal Component of Momentum

and the Impulse Imparted to the Target for the
Impact of Porous Gelatin. 69

33. Relationship Between the Normal Component of

Momentum and the Impulse Imparted to the Target for
the Impact of Birds. 70

34. Typical Pressure Plate Data, Normal Impact. 72

34. Typical Pressure Plate Data, Normal Impact (Continued) 73

vii
AF.L-TR-77-134

LIST OF ILLUSTRATIONS (CONTINUED)

FIGURE PAGE

35. Typical Pressure Data at Various Impact Velocities

for Porous RTV, Normal Impact. 76

36. Typical Pressure Data at Various Impact Velocities

for Porous Gelatin, Normal Impact. 77

37. Typical Pressure Oata at Various Impact Velocities

fcr Birds, Normal Impact. 78

38. Typical Distribution of Pressure Plate Data for the

Impact of Porous RTV, Normal Impact. 80

39. Typical Distribution of Pressure Plate Data for the

Impact of Porous RTV, Oblique Impact (cx - 45O), 81

40. Huqoniot Shock Pressures Measured During Normal

Im act of Karous RTV. I(MN/m 2 - 10 Bars,
3 ). 83
1us kg/mr - .036 lb/in
41. Hugoniot Shock Pressures Measured During Normal
O kg/mJ
0Imact Porous
oj - Wbin ). (I MN/m2 - 10 Bars,
.036 Gelatin. 84

42. 4ugoniot Silock Pressure• Measureu Durinig Normol

Impact of girds. (1 MN/m 2 - 10 Bars, 103 kg/mJ =
.036 lb/in3 ). 85

43. Steady Flow Pressures Measured at the Center of

Impact •urinq Normal Impact 3f Porous RTV,
(0 MN/ryn - 10 Bars, 103 kg/m3 .036 lb/in3 ). 87
44. Steady Flow Pressures Measured at the Center of
Impact
(0 MN/m4Quring
- 10 Normal
Bars, ImRact of Porous
10J k9/m3 . .036 Gelatin.
Ib/in3). 88

45. Steady Flow Pressures Measured at the Center of

Impact uring Normal I(pact Qf Birds.
(0 MN/NI - 10 Bars, 10- kg/mna .036 lb/in3 ). 89

46. Padial Distribution of Steady Flow Pressures tor

the Normal Impact of Porous RTV. 90

47. Radial Distribution of Steady Flow Pressures for

the Normal Impact of Pure Gelatin. 91

48. Radial Distribution of Steady Flow Pressures fc"

the Normal Impact of Porous Gelatin. 91

viii
AFNL-TR-77-134

LIST OF ' LUSTRATIONS (CONCLUDED)

FIGURE PAGE
49 Radial Distribution of Stoady Flow Pressures for the
Normal Imact of Birds. 92
50 Hugoniot Shock Pressures Measured at Various Points
Along the Target Surface During t~e Oblique Impact
(a - 450) of Porous RTV. III MN/:n*. 10 Bars), 94
51 Hu~oniet Shock Pressires Measured at the Initial Point
of Impact During the Oblique I act (Q a 250) of
Porous RTV. (1 MN/n 2 - 10 Bars). "4

52 Steady Flow Pressures Measured at or Near Stagnation

Point for Normal ano Oblique (a - 450) Impact of
Porous RTV. (1 MNN/• 10 Bars) 95
53 Shock Hugoniot for Air. 109
54 Experimental Values of Shock Velocity vs. Particle
Velocity. 109

ix
AFML-TR-77-134

LIST OF SYMBOLS
A, B Material constants in Cogolev's pressure-density
relctionship
A Cross-sectional area of projectile
a Initial radius of projectile and circular jet
b Initial width of planar jet
c Wave speed
C• Longitudinal, elastic wave speed
co Isentropic wave speed
Cr Velocity of initial release wave (speed of sound in the
shocked region)
0 Initial diameter of projectile
E Modulus of elasticity
F Force
fm Mass fraction
fv Volume fraction
I Impulse
K Bernoulli's constant
k Constant in linear Hugoniot equation
L Initial length of projectile
Lc Critical length of projectile, defined by Equation 12
M Initial mass of projectile
P Pressure
PH Hugoniot shock pressure, defined by Equation 4
P Free-stream pressure, atmospheric pressure
Ps Steady Flow pressure at the stagnation point
q Defined by Equation 49

x
AFML-TR-77-134

LIST OF SYMBOLS (CONTINUED)

R Radius of jet at which steady flow approaches atmospheric
pressure

r Radt,'l distance from center of impact

T Duration of an elastic impact

t Time

tB Duration of the full shock pressure at the center of impact

tc Time needed for release wave to reach center of shock wave

to Duration of impact for a fluid projectile

u Velocity of fluid

uo0Impact velocity

Up Particle velocity

us Velocity of shock propagating into fluid at rest

xr Distance traveled by release wave

xs Distance traveled by shock wave

y Distance from center of impact

z Amount of porosity, volume fraction of air

a Angle of impact obliquity (relative to target surface)

3 Material compressibility

y Ratio of specific heats of air

Strain

Constants in Equations 24 and 25

0 Empirical constant, Engel (Reference 41)

p Density

Pf Density of pure (nonporous) material

Pz Average density of porois material

0 Angle of projectile yaw
Ocr Critical angle of projectile yaw

xi
AFML-TR-77-134

SUBSCRIPTS

B Center point of impact

c, cr Critical values

H Hugoniot

m Soft body material

n Air

p Particle

p Projectile

r Release wave

s Shock

t Target
0 Initial (free-stream, atmospheric) conditions

I Region in front of shock (unshoL~ed region) in projectile

2 Region in rear of shock (shocked region) in projectile

3 Region in rear of shock (shocked region) in target

4 Region in front of shock (unshocked region) in target

xii
AFML-TR-77-134

LIST OF UNITS

Mass g (gram) = 0.0022046 lb ,,(pounds-mass)

kg (kilogram) - 10' g

Length m (meter) = 3.2808 ft (feet)

= 39.37 in (inches)

cm (centimeter) = 0.01 m

= 2.54 in

Time s (second)

Force N (Newton) = 0.2248 lbf (pounds-force)

MN (Meganewton) = 1O6 N

Density kg/mr3 = 0.0624 lb m/ft3

3
= 3.613 x 10-5 lbm/in

Pressure MN/mi2 = 10 Bars

= 145.04 lbf/inL

xlii

.~- - - - - - -- - - - - - - - - - - - - - - ----- Q---

AFML-TR-77-l134

SECTION I
INTRODUCTION

One of the major hdzards to flight safety today is the inflight

impact of birds. Aircraft windshields and engine blading are especially
vulnerable to damage. Because of the importance of this problem. both
NASA and the Air Force have extensive programs aimed at improving the
impact tolerance of these critical components by developing better
materials and better structural designs. However, this work is currently
being hindered by a lack of understanding of the actual impact event.
Fundamental research is needed to define both the loads generated
during impact and the response of various structures tc these loads.

Work is currently being conducted in the Air Force Materials

Laboratory to define bird impact loads. This is a very complicated
problem since the loads are actually coupled to the response of the
structure. In order to simplify this, researchers have decided to
decouple the loads from the target response. This is accomplished by
studying the impact of birds against a rigid surface. The actual coupling
mechanisms will be studies in a later program.

Another program being conducted in the laboratory in parallel with

the bird loading program concerns the development of a "bird substitute"
material. Thi.i synthetic bird would replace real birds in actual impact
tests of aircraft components. Several materials are presently being
considered, inrluding gelatin and RTV (room ter.:,"%rature vulcanized)
rubber.

In order to fulfill the requirements of both programs, it is obvious

that a complete understanding of the fundamental principles of the impact
process is required. It seems reasonable to expect that the basic
mechanisms of impact of gelatin or RTV rubber should be similar to those
of bird impact. This is based on the belief that during impact each of
these materials tends to behave as a fluid. This is true for most mate-
rials at sufficiently high velocities. The requirements for a material
to "flow" is that the stresses generated during impact substantially
AFML-TR-77-134

exceed the strongth of the material. For many prujectile materials such
as aluminum or steel, the stresses required to cause the material to fail
arE sufficient to cause local failure in the target. For example, the
result of a hypervelocity impact is a crater.

However, there would seem to be a large group of materials, often

considered solids, whose strength are very low relative to the strength
of many target materials. For the purposes of this study, this group of
materials will be referred to as soft body materials. Soft body impact
will be defined as impact in which the stresses generated substantially
exceed the strength of the projectile but are well beiow the strength
of the target material. This implies that soft body projectiles wil1
flow upon impact while the target may see little or no plastic deformation.
Examples of soft body impact are birds and tire fragments striking engine
blading, hail striking aircraft windshields, rain drops falling on a leaf,
lead bullets striking a steel plate, or even bugs hitting an automobile
windshield.

With this in mind it was felt that a general survey of the state-
of-the-art in soft body impact would aid greatly in an understanding of
the basic mechanisms of bird impact as well as the impact of various
bird substituti' materials. However, as seen in the Literature Survey
of this report, the results of the literature search were limited. With
the exception of water droplet and water jet impacts, there appears to
have been no concerted effort to study the basic processes involved in
soft body impact.

For the past 20 years, researchers have studied the impact of

water droplets on a solid surface in order to develop an understanding
of the mechanisms involved in the erosion of steam turbine blades and
the erosion of aircraft and missile structures in rain. Much progress
has been made in this area, although there is still controversy over
the exact amplitude and distribution of loads, as well as the response
of both the liquid droplet and the impacted surface to these loads.

2
AFML-TR-77-134

Based on the insights gained from the literature search, iL appears

possible to directly apily much of the kvowledge gained from fluid i-npact
studies to the more general area of soft body inpact. Wt'h this in mind,
the author decided to undertake the present research program. lhe
ourpose of this research was to understand the basic processes involved
in soft body impact. Both theoretical and experimental studies have
been conducted on the impact of various soft body materials, including
bird, gelatin, and RTV rubber.

A study of this magnitude should be of great benefit since the

results could be used in many areas. As discussed above, the most
important application to the Air Force would be in aiding in the under-
standing of bird impact as well as in the development of a bird
substitute material. However, there are many other practical applications.
For example, knowledge fr.,m this program would aid in the study of hail
damage, rain erosion, and damage due to debris impact during a hurricane
or tornado.

3
AFML-TR-77-1 34

SECTION II

LITERATURE SURVEY
Th3 impact of low strength materials has only been studied exten-
sively in the last few years. Thus, there has been very little research
in the area of soft body impact and almost none concentrated on the
actual mechanisms involved in the impact process. Early researchers
such as Hodgkinson (Reference 1), Tait (Reference 2), Vincent (Reference
3), and Raman (Reference 4) studied the coefficient of restitution
(ratio of the rebound velocity to the initial velocity) of various
soft objects. They found that the coefficient of restitution decreased
with increasing velocity, demonstrating that the percentage of kinetic
energy dissipated during impact increased with increasing velocity.
Very little work has been done in this area except for recent studies of
birds and bird substitute materials.

In the 1960's, several English engine people, including Ailcock

and Collin (Reference 5), studied the deflection of beams due to the
impact of various soft objects including birds. Based on a comparison

,
of the results, they determined that gelatin projectiles and birds
produced comparable loadi during impact. Tudor (Reference 6), studying
the impact of gelatin projectiles on a cantilevered beam, developed a
functional relationship between the beam deflection and the initial
momentum of the projectile. In America work was carried out by Tsai,
et al. (Reference 7) who found a similar relationship for the impact of
RTV rubber balls. The only theoretical study of bird loading was
performed by MacCauley (Reference 8) and Mitchell (Reference 9) in
Canada. The work of both men were only crude approximations of the real
impact case. MacCauley assumed the bird to behave as a perfect fluid
and Mitchell assumed the bird to behave as a semi-rigid projectile.
With other simplifying assumptions, both men derived expressions for tile
pressures generated during impact. No experimental work was performed
to check their results.

4
AFNL -TR-77-1 34

Since so little has been done in the specific area of soft body
impact, itwas felt that a survey of the general area of impact might
be of benefit. Specific interest was paid to studies conlcentrating on
the response of the projectile to the impact, as opposed to the response
of the target. It was found that numerous studies of the mechanics of
the impact event and the related impact strength of materials have been
conducted over a large range of impact velocities. Much of this work is
summnarized in textbooks by Goldsmith (Reference 10)., Kolsky (Reference 11),
and Johnson (Reference 12).

For the impact of solids, the mechanisms dominating the impact

process vary with impact velocity. Realizing this, several individuals
have attempted to divide the field of impact of solids into various
physical regimes. Hopkins and Kolsky (Reference 13) defined five regimes;
(a)elastic impact, (b)plastic impact, (c)hydrodynamic impact, (d)impact
at sonic velocities, and (e)explosive impact. In elastic impact, the
stresses generated do not exceed the yield strength of the material.
Thus, the nature and duration of the impact only depend on the elastic
moduli and the elastic wave velocities of the material. With increased
velocity, the stresses generated cause plastic deformation. The material
strength is still a dominating factor. As velocities increase still
further, the stresses generated by deceleration of the projectile greatly
exceed the yield stress. In this hydrodtynamic regime, the projectile and
target can be treated as fluids, and it is the material density which
dominates instead of strength. In each of these regimes, stress waves
propagate into the material, thus dissipating energy throughout the
projectile and target. As the velocity of impact approaches or exceeds
the wave velocity, more energy must be dissipated in the local region of
impact. Wave motion plays an increasingly important role in determining
the local stress distribution. Shock waves are also generated. As the
velocity continues to increase, all of the energy is deposi-%'ed in the
local area. The heat produced is concentrated in a very small region, and
is thus sufficient to melt and eventually to vaporize the material.
This process is analogous to a small explosion taking place on the target
surface.

5
AFML,-TR-77-134

Although these regines have been defined for the impact of high
strength projectiles, there is no reason to believe that these same
regimes would not be applicable to the impact of low strength materials.
Barber (Reference 14) found that cylindrical projectiles of RTV rubber
and gelatin bounced at low velocities and "flowed" at high velocities.
Similar results were also found for real birds. Thus, a further look
into these various regimes should lead to a better understanding of soft
body impact.

Johnson (Reference 12) described initial studies of elastic impact

which applied elementary, one-dimensional, elastic wave theory to the
impact of cylinders. Using this simple theory to study the problem of a
cylinder impacting a rigid plate, an interfacial pressure P = pc uo and a
duration T - 2L/u 0 can be derived, where c. is the longitudinal, elastic
wave speed in the projectile. However, this simple theory neglects the
effects of transverse strain and inertia as well as all dissipative
forces in the projectile material.

Theories for elastic impact which take into account the effects of
radial inertia were first proposed by Pochhamer (Reference 15) and
Chree (Reference 16) independently. Later, theories were proposed by
Love (Reference 17) and Rayleigh (Reference 18) and more recently by
Redwood (Reference 19). Using Love's theory, Conway and Jakubowski
(Reference 20) analyzed the coaxial impact of nearly perfect, square-
ended, finite length bars. Their results showed a slight change in pulse
shape due to the transverse effects.

Donnell (Reference 21) in 1930 was apparently the first to investi-

gate longitudinal plastic wave propagation. However, major contributions
to the theory of plastic wave propagation were not made known until
after World War II when Taylor (Reference 22), Rakhmatulin (Reference 23),
and Karman and Duwez (Reference 24) published their theories which had
been developed indepetciently during the war. In a later paper, Taylor
(Reference 25), in an attempt to develop a simple method for determining
the dynamic yield stress of materials, considered the propagation of both

6
AFML-TR-77-134

elastic and plastic waves during impact of a cylind2r on a rigid wail.

His simple, theoretical model led at least to qualitative, if not
quantitative, understanding of projectile mushrooming upon impact.
Experiments by Whiffen (Reference 26) showed that Tay~or's method was
inaccurate except over a limited range of velocities. Hawkyard (Ref-
erence 27) examined the consequences of establishing an energy balance
across the discontinuity at the plastic wavefront. His predictions of
the f~nal shape of the mushroomed projectile corresponded much more
nearly to the shapes found experimentally by Whiffen than did the pre-
dictions of Taylor.

Although the theories of Taylor and Hawkyard aided in underrtanding

the mechanisms of deformation during impact, they were too idealized to
give good quantitative results. Both effectively consIdered the pro-
Jectile to be a rigid, perfectly plastic material. In an attempt to
overcome this limitation, Barenblatt and Ishlinskii (Reference 28)
considered the impact of a linear, viscoplastic rod on a rigid wall.
Ting (Reference 29) extended their work to study the impact of a non-
linear, viscoplastic projectile and contrasted his results with those of
Barenblatt and Ishlinskii. All of this work considered one-dimensional
wave propagation.

When the stresses generated during impact greatly exceed the yield
stress of the material, the problem can be approached hydrodynamically.
In the hydrodynamic approach used to study high velocity impacts, the
material strength and viscosity are neglected and a simple pressure-
density-energy equatioi of state is used to describe the material. For
most materials this approach is applicable over a wide stress range.
Hopkinson (Reference 30), in studying the impact of a lead bullet
on a steel plate, found experimentally that the stress in the target was
approximately that to be expected from a fluid jet whose density was the
same as for lead. Birchoff, et al. (Reference 31) in their study of
explosives with lined cavities, may have been the first to incorporate
into a mathematical theory the idea that a high strength material such
as steel could be treated as though it were a perfect fluid when sub-
jected to very high stress levels.

7
AFML-TR-77-134

Thus, it is seen that even high-sLrength materials "flow" when

subjected to very high pressures. The major drawback to this area of
rtsearch is that the stresses generated at the projectile-target inter-
face are so great that the target material also behaves hydrodynanmically,
resulting in large deformatione (craters) in the target. Thus, it is
not possible to uncouple the response of the projectile from that of
the target. A review of the literature, as performed by Chakrapani and
Rand (Reference 32), for example, showed that most of the research in
this regime is concerned with predicting or measuring crater shape and
depth. Therefore, it was decided that further study of this regime or
the higher velocity regimes (sonic and explosive regimes) would be
fruitless.

A much more promising area of research appears to be the study of

the impact of fluids, in which the ?luid projectile is always assumed
to behave hydrodynawically. An extensive amount of theoretical and
experimental work has been conducted on the impact of fluids on solid
targets, in v'iich the target is assumed to have little or no plastic
defot•m .1 on.

The majority of work ii this area has been focused on the impact of
water jets and water drop.ets, both in the study of rein and steam
erosion as well as water jet cutti;,9, In these investiqtions, there are
several features of thi inmpact process of greatest importance. Heymann
(Reference 33) stated that these werec (a) the initial impact pressure,
(b) the area over which the pressure acts, (c) the vplocity of the lateral
outflow of liquid after impact, and (d) the app'roximate decay time -f
high impact forces.

Some of the early work wis conducted by Cook (Reference 34) in 1928,
in which he compait.d the impact of a column of water to the waterhamner
problem. He stated that in both cases, a moving column of water is
suddenly arivsted by a fixed surface, thereby ca'ising an instantaneous
pressure in the front layer of the column given by the expression
P - Uo0,4T.,where u0 is the impact velocity, p is the initial density

8
AFFL-TTr-77-134

of the water, and8 is its compressibility. Noting the magnitude of this

initial "watprhammer" pressu.'.e, he stated that "the oressir" n r-,-4'P
on an ,Oemo'nt of :u. U it. first encounter with water moving at a
finite velocity is dffferent from that produced by the steady impact
of a moving stream of water at the same velocity". Heymann (Reference 35)
phrased it Jifferently, noting that in this type of impact, compressibility
effects rather than flow phenomena initially predominate in the liquid-
Thus, Cook and Heymann identified the two basic regimes of wdter jet
impact, the initial shock regime and the late-time steady flow regime.

These two regimes were later identified by Bowden and Brunton

(Reference 3f) in their more detailed description of the impact of a
cylinder against a rigid plate. In their paper, they stated that upon
impact of an ideal, flat-ended cylinder on a rigid plate. stress waves
propagate into the liquid and the pressure at the interface is the
waterhammer pressure given by Cook (Reference 34), P - pcou 0 , described
in terms of the wave speed, co, instead of the compressibility, S.
According to their description, as release waves propagate inward to the
center of the cylinder, liquid is turned outward and flows radially with
the same velocity u0 and the pressure at the interface is released and
decays to the steady flow pressure. The pressure reduces to zero whert
all of the liquid has been turned. The duration of the high pressure
was determined by the time taken for the release waves from the sides of
the liquid cylinder, traveling at velocity c0 , to reach the center of the
cylinder.

For the impact of a sphere on a rigid plate, Bowdei; and Field

(Reference 37) described the event to be somewhat the same. Upon initial
impact, the pressure at the interface again reaches a maximum given by
the waterhammer equation. Release waves propagate inward sooner than
for a cylinder of equal radius so that the duration of the peak pressure
is less. They also stated that the waterhammer equation was only correct
for low velocity impacts. For high values of impact velocity, the pres-
sures are such that co is more correctly replaced by the shockwave
velocity, us. Thus, they stAted that the initial pressure generated
upon impact of a cylinder or sphere is given by the relationship

9
AFML-TR-77-131

Ou -us0 . (This pressure will ýe referred to as the Hugoniot pressure,

since it can be derived from the Hugoniot relationships across a shock.)

Although these two models of cylinder and sphere impact are

generally accepted, there is much disagreement in the literature over
certain features of the impact event. Most of this disagreement centers
around the peak pressures generated initially during impact and the
initial radial velocity of the released liquid.

It is generally agreed that Bowden and Field (Reference 37) were

correct in stdting that the instantaneous pressure qenerated during
impact of a cylinder is the Hugoniot pressure. However, nreat difficulty
has arisen in verif•ing this experimentally due to the inability to
generate a flat-ended cylinder of water which can be impacted on a rigid
surface. Bowden and Brunton (Reference 36) described a technique for
accelerattng a jet of water agaii-st a plate. However, these jets suffer
from an inherently rounded front surface. Both Brunton (Referencc 38)
and Kinslow, et al. (Reference 39) were unable to obtain instantaneous
peak pressures upon impact, probably a result of the rounded front.
Thus, Kinslow recorded a peal pressure of only half that predicted.
Field, et al. (Reference 40), also recorded peak pressures less than
the Hugoniot pressure. Thus, work continues to develop better methods
of testing for cylindricl41 impact.

Unlike the case of cylinder impacts, thert 's much less agreement
concerning the magnitude and distribution of the initial pressure in
spherical droplet impacts. Engel (Reference 41), an early pioneer in the
field of erosion, stated that the average pressure in the region of the
droplet which had been traversed by the shock could be expressed by
P •pcoUo where 0, a constant, approaches unity for high velocity
-

impact and the 1/2 is a consequence of the spherical shape of the droplet.
However, she made no attempt to specify instantaneous distribution of
pressure. Jenkins and Booker (Reference 42) agreed with Bowden and
Field (Reference 37) that the maximum impact pressure for a curved surface
is the Hugonlot pressure, and that this pressure is uniform over the
liquid-solid interface until lateral jetting begins. In contrast to

10
AFML-TR-77-134

this, Skalak and Felt (Reference 43) concluded that the "average" pressure
over the liquid-solid interface was the Hugoniot pressure but also con-
cluded that the pressure whs nW. uniform, with maximum pressure (greater
than pu su ) near the edge of contact. Heymann (Reference 44) developed
a theory based on the assumptioii that oblique shocks exist at the droplet
edge and used this theory to predict that the pressure increases during
early impact. This increased pressure exists at the droplet edge and
continues to increase until lateral jetting starts. Pressures exceeding
twice the Hugoniot pressure were predicted for an impact of Mach a 0.3.

Even greater disagreement resulted frum contradicting experiments.

As stated earlier, the jet launching technique used by Brunton (Ref-
erence 38) produced a cylinder with d rounded impact face. Thus, this
jet was used by several researchers to simulate droplet impact. However,
Brunton's experiment permitted the determination of average pressure
only, instead of the actual distribution. Using this same technique,
Kinslow, et al. (Reference 39), found the maximum pressure, about half
the predicted Hugoniot pressure, to occur at the center of the impact.
Johnson and Vickers (Reference 45) found approximately the same pressure
at the center but, contrary to the work of Kinslow, et al., measured
maximum pressures at the edge greater than the Hugoniot pressure. Using
a two-dimensional droplet impact test, Rochester and Brunton (Reference
46) verified the work of Johnson and Vickers. Using high speed "shadow-
graph" photography which allowed them to observe the stress waves
generated in the droplet during impact, Rochester and Brunton observed
a region of high pressure located at the edge of drop prior to lateral
jetting. Field, et al. (Reference 40), used an experimental technique
of launching a target projectile at a suspended droplet. Using a
similar high speed photographic system to that of Rochester and Brunton,
they also observed a region of high pressures near the edge of the
droplet.

Not only is there disagreement in the literature over the amplitude

and distribution of peak pressures generatec Juring initial stages of
impact, but there is also disagreement over the lateral flow velocities

11
 AM-R71341
resulting during these initial stages. Using the various experimental
techniques, Engel (Reference 41), Jenkins and Booker (Reference 42),
Fyall (Reference 47), Camus (Reference 48), and Rochester and Brunton
(Reference 46) observed radial velocities from' three to seven times the
initial impact velocity of the jet or droplet. Bowden and Brunton
(Reference 36) postulated that the high radial velocity resulted from
jetting action of the water between the confines of the drop and the
impacted surface. The results of Rochester and Brunton (Reference 46)
correlated well with this theory. Contrary to this, Jenkins and Booker
(Reference 42) found that their results compared favorably with thE
theory that the radial wash originates as steady, incompressible
Bernoulli flow, where the maximum stagnation pressure ir the initial
peak impact pressure. The experimental work of Fyall (Reference 49)
agreed with this theory. Even studies of ice impact by McNauqhton,
et al. (Reference 50) showed this theory to accurately predict initial
flow of the ice particles upon impact.

Besides development and refinement of experimental techniques, a

large effort has been applied to the use of numerical techniques to
develop a better understanding of the impact event. Huangi, et al.
(Reference 51) spent a considerable amount of time in the study of
cylindrical and spherical impact at s~ubsonic velocities. For both of
these cases, they found the maximum pressure to occur at the center of
impact and to have a magnitude less than the Hugoniot pressure. However,
these results are in doubt since the numerical technique was unable to
properly handle the strong shocks generated. The apparent introduction
of artificial viscosity probably prevented the program from responding
rapidly enough to measure the actual peak pressures. Also, the numerical
solution predicted precursor waves upon impact which propagated at
velocities much greater than the shock velocity. However, there is no
apparent physical reason for the existence of these waves.

Glenn (Reference 52) did similar studies in order to point up some

of the discrepancies of the work by Huang, et al., and to extend the
scope to include initially supersonic impact. All of his work concentrated
on the impact of a flat cylinder on a rigid plate. For both subsonic and

12
AFML-TR-77-134

supersonic impact, the peak pressure at the surface was the Hugoniot
pressure, thereby throwing doubt on the work by Huang, et al. Glenn
also found the maximum radial velocities to be greater than the initial
impact velocity. Thus, Glenn's results compared well with experimental
results. Numerical studies by Lesser and Field (Reference 53) and
Prichett and Riney (Reference 54) also predicted the initial peak pressure
at the center of a cylinder to be the Hugoniot pressure.

As stated earlier, once radial release waves have propagated into

the center of the cylinder (or sphere) the pressure begins to decrease.
After several wave reflections, the flow of a cylinder of fluid will
approach a steady state condition. At the center, the pressure will be
equal to the stagnation pressure of a jet. From Bernoulli's Equation,

the stagnation pressure is seen to be P = l/2puo2 for an incompressible

fluid, For a low velocity impact, this pressure is an order of magni-
tude less than the maximum shock pressure occurring initially at the
center.
The major area of study in the steady flow regime is the distribution

of pressure and velocity over the target surface. The pressure and
velocity are related by Bernoulli's Equation. Thus, knowledge of one
implies knowledge of the other. However, it is not possible to calcu-
late the distribution of pressure (or velocity) in the impact region for
an axisymmetric jet. An exact solution can be obtained for a two-
dimensional jet. Milne-Thompson (Reference 55) derived the following
expression for the potential flow velocity distribution along the
surface of a plane jet impinging on a flat, rigid surface:
y/b = 2/rr [arctan (u/u ) + arctanh (u )] where y is the distance from
0 0
the center along the surface, b is the initial width of the jet, u is
the velocity along the surface, and uo is the initial impact velocity.
This solution assumes a free-streamline jet, thereby implying that at
large distances from the center of impact, u = u0 . This expression is
substituted into "Ornoulli's Equation of the form P - Po = 1/2 pu 0 2
[1 - (u/uo) Using elementary momentum considerations, Banks and

13
AFML-TR-77-134

Chandrasekhara (Reference 56) developed a normal curve approximation to

the exact pressure distrib-ition of the form P - Po0 1/2 Puo 2 exp [()

This distribution closely resembles that of Milne-Thompson. Using

a similar approach, Banks and Chandrasekhara developed an error curve
approximation for2 an axisynuietric jet impact of the form
P " Po = 1/2 Puo exp (•)2] where r is the radial distance froi

the center and a is the initial radius of the jet. This expression was
a fair approximation to the data obtained by Gibson (Reference 57).

Leach and Walker (Reference 58) found that their experimental data
a 1/2 Pu0
2
was fit well by a polynomial expression of the fopm P - Po
1l-3(r)2 + 2(r)3] where R was the value of r where P P0. They

found from momentum considerations that R/a = 2.58. From this it was
concluded that the region where the pressure is significantly greater
than the ambient pressure is confined to about 2.5 jet radii.

Taylor (Reference 59), studying oblique impact, stated that the

total downward force on the target plane decreases as the angle of
incidence decreases. However, the maximum (stagnation) pressure is the
same for all angles of impact, namely 1/2 puo0, although the area over
which this maximum pressure acts is much reduced as the angle of
incidence becomes small.

Based on the above information, several points are clear. There

is general agreement about the pressures generated during the impact
of a cylinder of water on a rigid plate. The pressure at the center is
initially pusu 0 , the Hugoniot pressure, and remains at the "evw until
lateral release waves permit radial flow to begin. After a sufficient
time, steady flow is set up with the pressure at the center given by
Bernoulli's Equation.

There is less agreement about the impact of a sphere (or droplet)

of water. Physically, one expects the -initial pressure at the center
to be the same as for a cylinder, pusu0 , and to remain at this level
until radial release. In addition there is extensive experimental

14
AFML-TR-77-1 34

evidence that as the droplet impinges on the target, the peak pressure
moves outward with the point of contact and increases in magnitude until
lateral jetting (radial flow) begins. The duration of the droplet impact
is probably too short to allow sufficient time for steady flow to be
completely established.

15
AFML-TR-77-1 341

SECTION III

IMPACT THEORYI
In the following section, hydrodynamic theory is applied to the
impact of a soft body against a flat target. Equations are derived for
the pressures generated during impact. Much of this theory is based on
the ideas presented in the Literature Survey. Concepts from various
disciplines have been combined to develop a comprehensive impact theory.
Initially, the normal impact of a right circular cylinder onto a rigid
plate is considered. Next, several special cases are considered,
including oblique impact, impact of a yawed projectile, impact of a
projectile with a curved leading edge, addition of porosity to the
projectile material, and the impact onto a non-rigid target. A subsection
is included which explains the development of the equations of state
used to represent soft body materials. A significant portion of this
work deals with altering the equations of state to account for the
presence of porosity. Finally, the limitations of this impact theory are
presented.I

When a projectile of any material impacts a target plate, the

particles on the front surface of the projectile are instantaneously
brought to rest relative~ to the target face and a shock is formed. The
purpose of this shock wave is to bring each succeeding layer of particles
to rest. The shock compression of a layer of particles is so rapid that
the particles away from the edge of the projectile do not have time to
"commnunicate" with the free surface. This implies that these particles
behave as if they are in a semi-infinite medium which can undergo only
plane strain compression. Thus, shock compression in a bounded medium
is usually consi~dered to be a plane strain process.

The pressure in the shock compressed region is very high initially

and is constant throughout the region at early times. As the shock
propagates up the projectile, the particles alung the projectile's edge
are subjeui...,. to a very high pressure gradient due to the shock loading

16
4 on one side and the free curface on the other. This pressure gradient
A causes the particles to be accelerated radially outward and a release
wave is formed. The function of this release wave is to relieve the
radial pressures in the projectile.

A very complicated state of stress begins to develop in the pro-

jectile. The radial pressure release causes shear streo~ses to appear.
The radial acceleration of the particles also results in tensile
Ii stresses being developed. If,at any time, the state of stress is such
that the strength of the material is exceeded, the material will "flow".
For soft body impact, it will be assumed that the stresses throughout
the impact event greatly exceed the material strength, so that the flow
will continue indefinitely. For these materials, to a first approximation
the material strength can be neglected so that they can be considered to
behave as fluids.

After several reflections of the release waves, a condition of

steady flow is established. A const-ant pressure and velocity field is
set up in the projectile, and the particles flow along paths which are
fixed in space, called streamlines.

In the following section, a hydrodynamic impact theory will be

used to study the mechanisms involved in soft body impact. In order to
* simplify the analysis, the case of a homogeneous, right-circular
cylinder impacting normally on a rigid plate will be considered initially.
By assuming the target to be rigid, the response of the target material
to -the impact pressures can be neglected. This simplifies the problem
and allows the researcher to concentrate on the response of the projectile.
Several assumptions are implicit in this theory. The projectile material
will be considered homogeneous, even when high amounts of porosity are
considered. As stated previously, the strength of the projectile material
will be neglected. Also, in order to simplify the analysis, the material
viscosity and the frictional forces along the impact surface will be
neglected.

17
AFML-TR-77-134

a. Shock Regime
For the normal impact of a cylinder on a rigid plcte, the flow
across a shock can be considered one-dimensional and adiabatic, irrevers-
ible. Figure 1-b illustrates a shock wave propagating into a fluid at
rest, where us is defined as the velocity of the shock propagating fi,'to
the fluid at rest and up is the velocity of the particles behind the shock
in this reference system. From this figure, it can be seen that the
particle velocity is actually the change in velocity across the shock.
Figure 1-b illustrates the case for which the velocities are all
measured relative to the fluid in the shocked state. This case is iyn-
onymous with the impact of a cylinder on a rigid plate. The projectile's
initial velocity is seen to be u0 and it is brought to rest. behind the shock.
In order to write the conservation laws across the shock, the steady
state shock condition must be considered. This case is shown in Figure 1-c.
l--b- us

Ug =up Ul = 0

b)
-ua 0 U u

-44
a Up J , U

C)

Figure 1. One Dimensional Shock Flow. a) Shock Propagating into

a Fluid at Rest, b) Flow Brought to a Rest Across
the Shock, c) Standing Shock.
18
 Iv
AFML-TR-77-134

For this case, the equations of conservation of mass (continuity)

and momentum may be written

P us P2(U-) (1)

PI + P1 Us P2 + 0 2 (Us-U p)2 (2)

Combining these two equations, the pressure behind the shock is found
to be
P
2sp1UU (3)

The pressure in the shocked region, given by Equation 3, is often

referred to as the Hugoniot pressure. Throughout the remainder of
this report, this pressure will be represented by the notation, PH'
For the impact of a cylinder on a rigid plate, it can be seen that
up a uo. Thus, for this case, Equation 3 becomes

PH M P1UsUo (4)
For very low impact velocities, the shock velocity, Us, can be
approximated by the isentropic wave speed in the material, co. Thus,
for low impact velocities, Equation 4 may be approximated by the relation

PH = Pleuoo (5)

As stated in the Literature Survey, many researchers believed

that the shock pressure generated during the impact of water on a rigid
surface could be given by Equation 5. Kinslow (Reference 39) apparently
used this relation in expressing shock pressures for a water jet impact
at 633 m/s. Although this relation may be adequate for very low velocity
impact, it deviates markedly from Equation 4 for higher velocities.
Figure 2 demonstrates the differences in pressures found from these two
relationships for water. At Mach - 1.0, Equation 5 gives a shock pressure
of only 1/3 that of the actual value given by Equation 4. Kinslow's
results, for a velocity of 633 m/s, are seen to be in error by more
than 40%.

19
AFML-TR-77-134
7000

6000

z 5000

4000
W

2000 8U

0
0 300 600 900 1200 1500
IMPACT VELOCITY, u0 (m/a)

Figure 2. Effect of Compressibility on the Hugoniot Pressure

for Water. (1 MN/m 2 . 10 Bars)

b. Release Regime
As soon as the shock is formed. it begins to propagate up the
projectile and radial release waves propagate in towards the center axis
of the projectile. The problem can no longer be considered one-dimensional
in nature. For the normal impact of a cylinder, the problem becomes two-
dimensional, axisymmetric.
Figure 3 shows the release regime of impact for a cylinder with
an original L/D of 2. Figure 3-b illustrates the shocked region in the

20
 AFML-TR-77-134

U'

a) b) J
B A B A

a A B A

Figure 3. Shock and Release Waves In Fluid Impact.

ProJectile Just after impact. The pressure in the shocked region is

given by Equation 4. It should be noted that the velocities of the shock
and release waves are much greater than the initial velocity of the
projecti'e. Figure 3-c shows the moment when the release waves have
converged on Pt. 8, the axis of the cylinder. Figure 3-d shows the
roment when the release waves have just caught the center of the shock,
Pt. C. The region of fully shocked material no longer exists. The
curvature of the shock is due to the release process, which has progres-
sively weakened the shock, thereby decreasing the shock velocity.

21
AFNL-fR-77-134

Several reflections of the release wave are probably necessary

to establish steady flow. The existence of steady flow, of course, is
dependent on the length (or L/D) of the cylinder. For very short
cylinders, the Impact event will be over before the release regime has
ended. Although there is no simple, analytical method for studying
the release regime, several critical Items can be identified.

The duration of the shock pressure at the center of impact can

be approximated by the amount of time that it takes the initial release
wave to reach the center, Pt. B of Figure 3-c. The release wave can be
considered to be a fan of wehk expansion waves, which implies that the
release process can be considered isentropic. The velocity of the
initial release wave is just the speed of sound in the shocked material,
cr, which is given by the expression

2 = (dP/dp)p, (6)

which is the slope of the isentropic pressure-density curve at the

IHugoniot state. Thus, the expression for the time necessary for the
release wave to initially reach the center is
tB : a/cr (7)

where a is the initial radius of the cylinder. Figure 4 shows the speed
of sound in the shocked region, cr' plotted versus impact velocity for
water. Using this information, the duration of the Hugoniot pressure,
tB, can be obtained from Equation 7. Figure 5 shows the relationship
between tB and the impact velocity, uo, for cylindrical projectiles of
water of various radii. (The equation of state of water is given in a
later section).

Another important item is the time that it takes the front of the
release wave to just capture all of the shock wave. As stated previously,
immediately upon impact the release wave begins to interact with the
shock at the edge of the projectile. Since the wave speed in the
fully shocked medium is always greater than the shock speed, the
release wave will interact with progressively more of the shock as the

22
AFI4L-TR-77-134
2500

ZgsO0 s
*.2000

W 1000

500

01
0 100 200 300
IMPACT VELOCITY, % (m/*)

Figure 4. Comparison of the Shock Velocity, us, and the Sound

Speed in the Shocked Region, cr

25

20

0
0 0 100 150 200 250 300
IMgPACT VELOCITY, ta. (rn/A)

Figure 5. Variatic'a of the Duration of the IHugoniot Pressure with

Impact Velocity for a Cylindrical Projectile of Water.

23
AFML-TR-77-134

impact continues. Figure 3-d shows the condition in which the release
waves have just converged on Pt. C. After this time, the pressure in
the region behind the shock will rapidly decay. The time of this
occu~rn%.ce, tc, may be derived from geometric considerations. In order
for the release wave starting at Pt. A to interact with the shock at
Pt. C, it must travel a radial distance, a, and an axial distance equal
to the axial distance traveled by the shock. At the time of intersection,
the shock has propagated a distance
X8 = (uS-Uo)tc

where (us-uo) is the velocity of the shock relative to the target as

seen in Figure 1-b. From this it can be seen that the release wave will
have traveled a distance
X = (X2 + d2)½ (9)
r "
in the time, tc. given by the expression
te Xr/cr (10)
Thus, it can be seen that

((Cr-_US_Uo0)2}½ ( 1

From Equation 11 an expression can be derived for the critical

projectile length, Lc, which is the length for which the radial release
wave will just intersect the shock on axis, Pt. C, as the shock reaches
the end of the projectile. Again, from geometry, it can be seen that
L ut
s (12)
Combining Equations 11 and 12 and nondimensionalizing, it is seen that
US
(L/D)c (1)
2{C2-(U )2}
r s 0

For a projectile with an L/D < (L/D) , the shock will reflect off the
c
projectile rear surface before it has all been captured by the radial
relebse waves. The shock will be reflected in the form of a rare-
faction wave. This rarefaction wave could complicate the impact event

$ 24
AFlqL-TR-77-134

by causing a tensile failure in the projectile, decreasing the incoming

velocity of the end of the projectile, and disturbing the radial release
process.

Hemver, for a projectile with an L/D > (L/D)c, the shock will
be substantially weakened by the release waves prior to reaching the pro-
jectile rear surface and its effects will be reduced or effectively
cancelled. Figure 4 showed both us end cr as a function of impact
velocity S'or water. Using these values, (L/D)c for a cylinder of water
can be obtained from Equation 13. Figure 6 shows a plot of (L/D)c versus
impct wvlocity for water.

2.0

.1J

-J
o 1,05

S 1.0
-J

E
o, 0 50 100 1 200 250 300 350
I

IMPACT VELOCITY, u0 (m/r )

Figure 6. Variation of Critical Length with Impact Velocity

for Water.

2S
AFM4L-TR-77-1 34

For a projectile of sufficient length, steady flow should be set up

up after several reflections of the radial release waves. It seems likely
that for a projectile with a length somewhat greater than Lca condition
of steady flow should exist for some time.

c. Steady Flow Regime

During the release phase, the shock is constantly being weakened
by the release waves, with an accompanying decrease in shock velocity.
For a low velocity impact (subsonic), the shock wave will be weakened to
the point that itwill disappear. By this time, for steady flow, stream-
lines will be established. However, for a supersonic impact, the shock
wave will not disappear. Instead, its velocity will decrease until it
becomes a standing shock. Behind this standing shock, the flow will be
subsonic and will follow steady state streamlines. Because of the
standing shock, the velocity and pressure fields set up in the fluid will
be different for the two cases.

(1) Subsonic Flow

If it is assume~d that the shearing forces and body forces
are negligible throughout the flow field and that all stream properties
vary continuously in all directions, then along each streamline,
Bernoulli's Equation can be written

P x+ (u du =K (14)

whiere K is constant along the streamline and may vary from one stream-
line to another. For the case of a cylinder impacting a rigid plate,
the flow field is essentially uniform at some distance away from the
impact surface, so that the K must be the same for each of the stream-
lines in this region. This implies that K has the same value throughout
the entire flow field. For this case, the pressa.re P, at any point in
the flow field can be related to the velocity, u, at that point by the
relation
f P
+fJ
U
udu =0 (1s)
0 0

26
AFML-TR-77-134

where P0 and u are the pressure and velocity of the uniform flow field
some dismance away from the impact surface and are approximated by the
atmospheric pressure and the initial impact velocity.

In order to obtain the pressure at any point along the impact

surface fr'om Equation 15, the velocity at that point must be known and
the equation of state of the material, p = p(P), must be known. The
equations of state for soft body materials will be discussed in a later
section. The expression for the velocity at a point is not found so
easily. For a two-dimensional fluid jet, the velocity field in the
fluid was obtained by Milne-Thompson (Reference 55) from potential flow
theory. However, no closed form solution of potential flow theory has
been developed for a circuldr jet. The normal approach is to assume an
expression for the velocity field based on empirical data.

In this approach, the pressure at the center and the pressure at

the edge are found using Equation 15. Then a general expression for the
pressure distribution is assumed based on empirical data. This expression
is forced to satisfy the pressure boundary conditions and conservation
of momentum.

For the normal impact of a uniform, cylindrical projectile on a

rigid plate, axial symmetry dictates that the point at the center of
the plFte be the stagnation point. The pressure at the center of the
plate ib the stagnation pressure, Ps (gage pressure), and the velocity
at the center is zero. Therefore, at the center of the plate, Equation
15 takes the form
f (Ps + P o)dP = 'o
)P p 2 (16)

If the fluid is assumed incompressible, Equation 16 gives

Ps= PUo 2 (17)

For most materials, the density tends to increase the applied pressure, so
that Equation 16 implies
P8 • ½ PU0 2 (18)

27
AFML-TR-77-134

The second pressure boundary condition is that the pressure must go to

zero at some radial distance from the center.
Simple momentum considerations require that during steady flow,
the impulse imparted to the target by a unit of fluid must be equal to
the axial momentum lost during impact. This may be written

fl
J
'F dt)
I
(M du) (19)
U0

where the force, F, is assumed constant during steady flow and the
duration of impact is represented by tD' For a unit of fluid with
initial values of mass M, density p, length L, and cross-sectional
area A, this expression becomes
F tD = (u-uo) (0)

For most soft body materials, the rebound velocity after impact, u, is
so small that it may be ignored. Also, during steady flow, the duration
of impact for a unit of fluid of length, L, is simply the time required
for the fluid to flow through its length. That is,
tD = L/u (21)
Thus, the force generated in the steady flow regime is seen to take the
form
F = pAu 2 (22)
0

Since the force is simply the integral of the pressure over the impact
surface, this expression may be rewritten

21rj Prdr = pAu2 (23)

From the above arguments, it is evident that any assumed expression for
pressure must predict the pressure to be maximum at the center, given
by the stagnation pressure Ps, decreasing with radius to a value of
zero at some point, and that the expression for pressure must satisfy
Equation 23.

As stated in the Literature Survey, both Banks and Chandrasekhara

(Reference 56) and Leach and Walker (Reference 58) developed expressions

28
AFML-TR-77-1 34

for the pressure distribution due to the nomal impact of a water jet.
The expression of Banks and Chandrasekhava took the form

P = PU0 exp C (24)

and that of Leach and Walker took the form

p = 0 {1-3(, + )3) (25)

"2
where r is the radial distance from the center, a is the initial radius
of the jet, and ýl and r.2 are constants used to make Equations 24 and 25
conform to Equation 23. For this case, cl - 0.5 and ý2 - 2.58.
Figure 7 shows a plot of Equations 24 ard 25.
The main drawback to these two expressions is that they define a
maximum pressure at the center, Ps. of 1/2 pu . As was shown earlier,
this is only true exactly for an incompressible fluid. For a compres
sible material, the pressure will be greater. The effect of compres-
sibility on the stagnation pressure can be demonstrated by

1.0

0.8

Q2
Eq. (25)
GL2

0
0 0,8 1.6 2.4 &2
r/ci

Figure 7. Radial Distribution of Steady Flow Pressure.

29
.
 "* AFML-TR-77-134

considering the impact of a cylinder of water. Using the equation of

state for water which is given in a later section, Bernoulli's compressible
equation in the form of Equation 15 can be solved for Ps. Figure 8-a
shows the results. Although the stagnation pressure is not dramatically
different from 1/2 Puo 2 for low velocities, the difference is evident
at higher velocities. This pressure difference would be even greater
for very compressible materials.

For so1't body impacts, Equations 24 and 25 may be generalized to

allow for the increased stagnation pressure due to compressibility. Thus,
they may be written

P P X
-xp (-C(E)2} (26)

and
P = PS (1- 3 ( 1 )2 + 2(r (27)
2a S

respectively, where the constants c, and 42, derived in Appendix A, are

given by Equations A-7 and A-9
P/PUo
P1
.1 S 0

C2 (3.33 Puo2/PS)½

The effect of compressibility on the pressure distribution should

be evident. Figure 8-a shows that for an impact velocity of uo Z 1400 m/s,
Ps M 1.14 (1/2o0%2). Figure 8-b shows a plot of Equation 27 using this
value of Ps.

(2) :upersnnic Flow

For the supersonic impingement of a fluid jet normally
onto a rigid plate, a standing shock is set up in the flow. The change
in particle velocity across the shock may be small. However, the flow
behind the shock will be subsonic because of the greatly increased
local sound speed.

30

i - - * -
AF4L-TR-77-134

1.0

1. i,- - - -

0.5

0
0 500 1000 1500
a . IMPACT VELOCITY, u m/*)

u,., 1400 m/s

1.0 uO. 100 rn/s

1-of

S0. 00 10 20 30

Figure 8. Compressibi1ity Effects in Steady Flow. a) Stagn.ation

Pressure, b) Pressure Distribution.

31
 AFML-7'R-77-l 34

I In the previous section on subsonic flow, Bernoulli's Equation

was used to relate the pressures along the impact surface to the pressure,
P , density po, and velocity, u0, in the uniform flow region. However,

I for supersonic flow this approach cannot be used since the standing shock
a discontinuity surface and Bernoulli's Equation only holds ina
is
continuous flow field. Instead, the change in stream properties across
the~ shock must be evaluated using shock relations. Bernoulli's Equation
can then be used throughout the flow field behind the shock. Figure 9
illustrates this case. Region I in front of the shock is auniform flow
field and the stream properties in this region are the initial properties,
namely P0, P0, and uo. In Region 2 just behind the shock, the properties
may not be constant over the cross section since the shock may not be a
perfectly plane shock. However, at the center axis, the shock can be
assumed plane and one-dimensional shock equations can be used. The
properties along the axis Just behind the shock Wiill be designated
'21 P2 * and u2.

STAGNATION
STREAMLINE

t \ ,STAGNATION POINT

Figure 9. Standing Shock in Steady Flow.

32
AFML-TR-77-134

Since no uniform flow field exists behind the shock, Bernoulli's

Equation will take the form

-- + u du 0 (28)
f p dP U
2 f2

and will hold only along the streamline which passes through the center
of the projectile and the stagnation point in the center of the impact
region.

The case of a standinq shock was shown in Fiqure 1-c. For this
case, Equations 4 and 1 take the form
p 2 ." poU.Up (2g)

where

P0 u P2( Up) (30)

u =u (31)

u = Uo-U 2 (32)

One more equation is needed in order to solve this set of equations.

Such an equation, relating the shock velocity, us, with the particle
velocity, up, will be given for various soft body materials in a later
PI
section. Thus, for a given set of initial conditions and a given material,
Equations 29, 30, 31, and 32 can be solved for P2, p2, and u2 .

Using these values, the pressure at the stagnation point on the

surface of the plate can be found using Bernoulli's Equation of the form

d U(33)
fP 2 P 2'

In this relationship, the equation of state of material must be the

pressure-density relationship for an isentropic compression passing through
the points P2 ` P2.

33
AFML-TR-77-1 34

As in the subsonic flow case, once P5 is determined, Equations 26

and A-7 or 27 and A-9 can be used to develop an expression for the
pressure distribution along the surface of the plate.

d. Termination of Impact
As stated previously, the particles of fluid flow along stream-
lines which are set in space. These streamlines are curved near the
plate surface so that the impacting material is "turned" during impact.
For subsonic flow, as the fluid nears the target surface there is a
gradual decrease -invelocity with a corresponding increase in local pres-
sure. Thus, during steady flow there is a pressure field set up in the
fluid in which the local pressure is maximum at the stagnation point and
decreases to the atmospheric pressure, Po, at a substantial distance
from this point. As the end of the projectile nears the impact surface,
it enters this field of increasing local pressure and disrupts the field
due to the inmmediate drop in pressure behind the projectile rear surface.
For supersonic impact, this happens when the projectile end passes
through the standing shock. Ineither case, release wdves emanate from
the re-r surface and propagate to the impact surface, thereby causing a
sliqht decrease in the impact velocity of the rear surface as well as a
:'-~eIn t. pressure along the impact surface of the target. This
d.
p. --.-
ss continues until the end of the projectile reaches the surface of
the plate and the impact event isended.I
As stated earlier in Equation 21, the total duration of the
impact can be approximated by the time needed for the projectile to
"flow through" its lenqth, or tD 'L/u 0 Deviation from this could be
caused by the decrease in projectile velocity due to release waves from
the back i.f K' ojectil1e.

2. SPECIAL CONSIDERATIONS
In the last sectior' a theory was presented concerning the normal
impact of a right-circil cylindrical projectile of fluid against a
rigid target. However, there are several other cases of special interest.
Included among these is the oblique impact of a right-circular cylinder,
the normal impact of a yawed projectile, and the normal impact of a

34
AFML-TR-77-1 34

projectile with a curved leading edge. Next, the porosity of the

projectile material is seen to have a great effect on the impact~ pres-
sure. Finally, the impact of a fluid projectile against a nonrigid
target is considered.
a. Projectile Yaw

Figure 10-a presents the normal impact of a right-circular

cylinder in which the axis of the cylinder isyawed by an angle * to the
velocity vector. The amplitude of the shock Hugoniot pressure will be the
same as that for a projectile impacting without yaw. However, the dura-
tion of this pressure at various points along the target surface will
differ if the yaw angle, *,is greater than the critical angle, Ocr*

a) bb

Figure 10. The Impact of a Yawed Projectile.

In Figure 10-b, the projectile is seen during initial impact.

If the shock originating at point a arrives at point b prior to point b
reaching the impact surface, then a release wave will develop and propa-
gate back into the shocked region. The critical angle, 'cr' is the
angle of yaw for which the shock will just reach point b at the instantJ
point b reaches the surface. From the figure, it can be seen that for
* ~.,the
= ratio of the distance traveled by point b, EF, to the

35
AFML-TR-77-1 34

distance traveled by the shock, ab, must be equal to the ratio of the
velocity of point b, uo to the shock velocity, us. this may be written

0
for~~ cr E- = (34)

From geometry, it is evident that

Cr (35)

Thus, the relationship for the critical angle becomes

er Sin- (36)

The critical angle changes with velocity for most materials since the
shock velocity is normally a function of the impact velocity.

For an impact inwhich the yaw is less than the critical angle,
the amplitude and duration of the shock pressure at the various points
over the surface will be essentially the same as that for the impact of
a projectile without yaw. However, for an impact inwhich the yaw is
greater than the critical ancle, the full shock Hugoniot pressure willI
exist only at the initial point of impact. At each of the other points
on the impacting surface of the projectile, the material will have been
shocked and partially released before impacting the surface, thereby
,educing the shock pressure generated upon impact. Also, the duration
of the shock pressures at the center will be decreased due to the
inmmediate formation of release waves.

b. Projectile Leading Edge Curvature

For the impact of a curved surface on a flat target, the angle.
tp,which the impacting surface makes with the target surface increases
as the impact progresses. This condition is demonstrated in~ Figure 11.

36
AFIML-TR-77-134

Figure 11. The Impact of a Curved Surface on a Flat Target.

Once the angle 0 exceeds ocr as defined by Equation 36, a release wave
will form and propagate into the shocked region. This implies that the
duration of the Hugoniot shock at the center will be less for the impact
of a sphere than for a cylinder (without yaw) of like diameters,

c. Oblique Impact
Figure 12-a demonstrates the oblique impact of a right-circular
cylinder against a rigid plate in which the velocity vector is directed
along the axis of the cylinder and intersects the target at an angle a
relative to the target surface.

(1) Shock Regime

For the oblique impact of a projectile on a rigid plate,
a coordinate transformation aids in the understanding of the shock
process. From Figure 12-a it is evident that the component of projectile
velocity normal to the plate is uo sin a and the component tangential j
to the plate is uo cos a. Figure 12-b shows the impact in the transformed

a) b) 'N

Figure 12. Oblique Impact of a Cylinder on a Rigid Plate.

37
 AFM4L-TR-77-134 /

coordinate system in which the projectile is seen to impact normally onto

the moving plate. A-ssuming the frictional forces along the surface are
,negligible, this case is identical to the normal impact of a yawed
projectils with an initial'velocity of u0 sin a. Correspondingly, the
Hugoniot shock pressure will- be identical to that produced by a normal
impact of a projectile with an initial velocity of u0 sin a. Figure 13
demonstrates the decrease in shock pressure with impact angle of obliquity
for the impact of a cylinder of water.

(2) Release Regime

The release process is similar to that of the yawed
projectile. For this case, the angle that the impacting projectile
surface makes with the target surface is seen to be governed by the
relation *- 900-m. For the impact in which * > ýcrl the full Hugoniot
pressure will occur only at the initial point of impact.

(3) Steady Flow

The steady flow phase of impact for a fluid cylinder is

shown in Figure 14. From momentum considerations it can be seen that
the majority of fluid will flow "downstream". This causes the stagnation
point to shift "upstream" of the center of the plate. As long as a
stagnation point exists, the pressure at that point will be given by
Equation 16. Thus, the maximum pressure generated during steady flow
will be independent of the angle of impact. However, the distribution
of pressure along the surface will be greatly dependent on the impact
angle.

Although no expression for pressure distribution in an

oblique impact of a circular Jet has been developed, Taylcr (Reference 59)
developed a distribution for a two-dimensional (plane) jet of water. This
distribution is shown in Figure 15 for an impact angle, a, of 300.
In this figure, the profile of the projectile is superimposed on the
pressure distribution tc demonstrate the distance that the stagnation
point, s, has shifted upstream from the center of impact.

38
AFML-TR-77-134

bOQo
i32900
(NORMAL)

h M aOO450
atO

I1000 0
0 300 600 900 1200 1500
IMPACT VELOCITY, uo (m/i)

Figure 13. Variation of Hugoniot Shock Pressure with Angle of

Impact for Water. ( MN/mni 10 Bars)

U0

//

Figure 14. The Steady Flow Phase of an Oblique Impact.

39
AFML-TR-77-1 34

Figure 15. Steady Flow Pressure Distribution for a 300 Impact of

a Plane Jet of Water (Taylor,Referenice 59).

From simple momentum considerations, Equation 22 stated that

the total force exerted on the target at any instant of time for a normal
impact could be given by the relation F -pAu 2. However. the momentum
transferred to the target during an oblique impact is only the normal
component of the total momentum. Thus, for an oblique impact, Equation 22
can be rewritten
F = Au2 sin QL (37)

For normal impact, a = 9('* and Equation 37 reduces to Equation 22. Thus,
the pressure distribution for an oblique impact is seen to be different
from that for a normal impact, due both to the decrease in total force
and the loss of axial -symmnetry.

d. Material Porosity
In the previous section on Hlydrodynamic Theory, the Hugoniot
shock pressure was presented as a function of impact velocity for water
in Figure 2. Later, in Figure 7, the distribution of steady flow pres-
sure was given for the impact of a jet of water. In both of these cases,
the water was assumed to have no porosity. As will be seen in the next
section on Equations of State of soft ýndy materials, the presence of

40
AFML-TR-77-134

porosity has a great effect on both the shock velocity and compressibility
of a material during impact. Thus, the equation of state for a porous
material differs from that of a nonporous material.

The decrease in shock velocity due to proosity results in a

marked decrease in the shock pressures. Figure 16 presents the variation
in Hugoniot pressure with porosity, z, for the normal impact of a cylinder
of water. (As defined in the next section, z is simply the volume
fraction of air in the water). Likewise, the decrease in density due
to the addition of porosity causes a decrease in the stagnation pressure
during steady flow. However, the increase in compressibility effects
tend to counteract this, so that the resulting decrease is relatively
small. Figure 17 presents the variation in the steady flow stagnation
pressure for water with porosity.

In Figure 8, the alteration of the steady flow pressure

distribution due to comnpressibility effects was presented. In that
case, the increased material compressibility resulted from increased
impact velocity. Compressibility effects are even more pronounced
for materials which have porosity. Figure 18 presents tha pres~ure
distribution for both porous and nonporous water using Equation 21.
in this figure, the pressure is nondimoeisionalized with respect to
the incompressible flow pressure 1/2p u20 2, where p;, defined in the next
section by Equation i8, is the Average density of the porous ',ater for
this case. Also, the radial distance from center is nondimensionalized
relative to the initial projectile radius, a.

e. Non-Rigid Target

Up to this point the theory has been concerned with impact on a

rigid plate. However, all real target materials are non-rigid to some
extent. Thus, the response of the target during impact c,nnot be
overlooked. If the target materiol can be assumed to behve elastically,
the initial tarqet rcsponse is straightforward. At very early times
during the impact, only the local area of the target directly under

41
AFML-TR-77-134

800 Z0
" Bo =.lO
Zzz,

""LZ600
; Z z.20

U) 400 z=.50
( =z.40
0.

o 200

01
0 45
0 ., ý75 150 225 300 375 450
PARTICLE VELOCITY, Up(r/S)

Figure 16. Relationships Between the Hugoniot Shock Pressure and

Particle Velocity for Water with Various Porosities.
(I MN/m 2 110 Bars)

100 000

~i,
0

S300 .. 400

IMPACT VELOCITY, uo (m/u)

Figure 17. Relationships Betweert the Stagnation Pressure and the

Impact Velocity in Steady Flow for Wiater with Various
'orosities. (1 MNA•' 2 - 10 Bars)

42
 AFML-TR-77-134

1.4
1.2

00 ,8 1.6 20 30
NONDIMENSIONALIZED RADIUS, rle

Figure 18. The Variation in Steady Flow Pressure Distribution

for Water Due to the Addition of Porosity,

the impacting projectile is affected by the impact. Figure 19 illustrates

:•' this early shockuW
regime. From this figure it is evident that to a f~rst
-i!: approximation the shock waves may S00
be considered initially planar, and
one-dimensional shock equations can be used for the target as well as
t;he projectile,

' PROJECTILE

0 1.2ARGET

* Figure 19.Tempact on ainElastic Target During the Early

ShS Pae•ec.tile

43
•*1 AFML-TR-77-134

In Figure 19, Region 2 is the shocked region in the projectile

and Region 3 is the shocked region in the target.
For the impact on a non-rigid target, Equation 4 is seen to
take the form
P PU (38)
H s p

where up , the particle velocity, is no longer equal to the initial impact

velocity. In Regions 2 and 3, Equation 38 may be written

P = Pp U (39)

P3 Pt= u
Ut UPt (40)

At the projectile-target interface, equiiibrium requires that

2 3

(41)
U2 = u3

From the definition of particle velocity, it can be seen that

U : U (42)

0 2
:u-43p 3 (3

Thus, the'particle velocities can be related by the expression

UP U-U (44)

Subst:tuting Equation 44 ioto Equatior 40 and eouating the pressures F2

and P3 gihes

pUspU P PtUSt (Uc-_1 ) (4i)

44
 AFML-TR-77-134

From this, it can be seen that

Y~ I.

u Uo• -t- - (46)

Thus, the pressures generated at the projectile-target interface for

the impact of an elastic target material are seen to be

P ~Puu

To demonstrate the difference between the pressures predicted

by Equation 4, which assumes a rigid target, and Equation 47, several
calculations were made for the normal impact of water on various targets
at a velocity of 200 m/s. For the impact on steel, the shock pressure
found using Equation 47 is approximately 4% less than that found using
Equation 4. For titanium, the difference is approximately 8%; for
atuminum, approximate'iy 11%; and for polycarbonatc, approximately 35%.

For thin targets, the initial shock wave in the target reflects
off the rear surface as a tensile wave of similar strength. This wave
propagates back to the impact surface, causing a decrease in the shock
pressures at the surfhce. Continuing wave reflections cause an increase
in the local pirticle velocity of the target Pnd a decrease in pressure.
As the impect proceeds, the entire target structure develops a velocity
along the initial axis of impact. The net result of this target
deformation is that the relative velocity between the projectile and
target decreases with a resulting decrease in the flow pressures.

3. EQUATIONS OF STATE
In order, to determine the pressures generated during impact, the
material properties of the projectile and target must be known. These
are applied to the problem through the equation of state of the materials.

43

* - ---
AFML-TR-77-1 34

a. Shock Compression
The expression for the initial shock pressure is given by
Equation 4, PH = pousuo. In this expression, p0, the initial density,
and u0, the initial velocity, are known for any given impact. Thus, in
order to solve for the shock pressure, a second expression is needed
whicl i .ýs the response of the projectile material under shock loading.

It has been found that for most solids and fluids (including
w,;er and air) the relationship between the shock velocity and particle
velocity can be expressed as

us C +Ik U (48)

where k is a constant for the material and c0 is the sound speed

(velocity of propagation of an infinitesimal disturbance) in the
material. This relationship, often called the "linear Hugoniot", will
be used to represent all soft body materials in this research. It can
be substituted into Equation 4 directly (by recalling that for normal
Impact. of a rigid target, the particle velocity, up,, is simply the
initial velocity of the projectile). Thus, Equations 4 and 48 are
sufficient for determining initial shock pressure.

However, it is also desirable to obtain a relationship between

the pressure and density on both sides of the shock by combining
Equation 48 with Equations 1 and 4. Equation 1, the continuity equation
across the shock, may be written in the form

U
1 P (49)

where the subscripts 1 and 2 refer to Region 1 in front of the shock

and Reginn 2 behind the shock.

46
AFML-TR-77-134

Substituting this relationship into Equation 48 gives

u = 1/(l-kq) (50)

and Equation 4 may be written in the form

P2 =PlUýUp (51)

Thus, substituting Equations 49 and 50 into Equation 4 gives

P = PlC2/1-kq)2 (52)

Another approximate expression for the pressure-density relation-

ship across a one-dimensional shock was given by Cogolev, et al.
(Reference 60). and takes the fornt
•P.)B l
P2= A (53)

where A and B are material constants. Ruoff (Reference 61) has demon-
strated that for a material which exhibits a linear Hugoniot, Equation 48,
the constants may be approximated by the expressions

A PCo2/4k-)

B 4k-i

where k is the constant in Equation 48.

Thus, the pressure and density behind the shock can be obtained
from either Equations 52 or 53 if the relationship between the shock
velocity and particle velocity can be approximated by Equation 48.
For water, Heymann (Reference 35) showed that Equation 48 provides a
qiite accurate fit to the data up to Mach 1.2 (approximately 1800 m/s).

47
AFML-TR-77-134

b. Isentropic Compression

The expression for the steady flow pressure -isBernoulli's com-

pressible equation, Equation 15,

f PdP f udu = 0
P f
0 0

In order to solve this equa-1.i,a • given impact velocity, the

pressure-density relationship .;;s•troptc compression must be given
for the projectile materir'S.

For very low pressure levels, two simple expressions are often
used to approximate the isentropic pressure-density relationship for
solids and liquids. The first is simply

p = constant (55)
Although never exact, the assumption of incompressibility can often be
used for low velocity impacts without causing considerable error.
This, of course, will not be true for porous materials.

The second relationship is obtained from the definition of the

bulk modulus
"• ~dP
-P (56)

For very low pressures, s, the bulk modulus, is often assumed constant.
For more accuracy, a relationship of the form

B = + 81 P + B2p 2 +.. (57)

may be assumed.

However, for the pressures normally generated during soft body

impacts, the assumptions of incompressibility and constant bulk modulus
are inadequate. Instead, equations of state accurate to pressures of
several hundred MN/m2 (several kilobars) are needed. In this pressure

48

~ - -
AFML-TR-77-134

range, the isentropic pressure-density relatiotiships for most solids

and fluids, including water, can be approximated by the Hugonlot shock
relationship (Reference 62) given by Equations 52 or 53.

c. Effect of Porosity

It has been stated that for most soft body materials, Equations
52 or 53 can be used to describe the pressures and densities resulting
from both isentropic compression and shock compression. Both of these
equations are based on a linear Hugoniot relationship, Equation 48.
It has been observed that departure from linearity is usually due to
porosity or phase change (Reference 63).

Although phase changes do not occur in the soft body impacts of

interest, many of the soft body materials are porous. The cellular
structure of animal and vegetable matter normally contains a small
amount of porosity. Rubber, in its many uses, is often used with large
amounts of porosity. Thus, for these materials, Equations 52 and 53 are
not representative.
In order to develop isentropic and shock relationships which
will hold for porous materials, a simple theory by Torvik (Reference 64)
will be considered. Torvik's theory predicts the pressure-density
relationship across a shock for a homogeneous mixture based on the
pressure-density relationships of each constituent. Although Torvik
used this theory to develop a pressure-density relationship under shock
I
compression, there is no reason why it should not be used also to
develop the isentropic relationship. Thus, in this report, Torvik's
theory has been used to develop both shock and isentropic relationships
for mixtures of which one component s air. The derivation of both the
shock and isentropic relationships using this theory are presented in
Appendix B. The main assumptions made are that the porous material is
macroscopically homogeneous and isotropic and that the density of each
constituent of the mixture is the same as that for a homogeneous sample
of that constituent of the same pressure.

49
AFML-TR-77-134

In the context of this report, a porous material is defined

simply as a mixture in which one of the constituents is air. The
amount of porosity of the material, z, will be defined as the volume
fraction of air in the mixture. If the full density of the soft body
material (density with no porosity present) is represented by pf
and the average density of the soft body material with porosity is
represented by pz, then from Equation B-1 it can be seen that the
average density of the material can be defined by the expression

Oz Pair + (1-Z) Pf (58)

where the volume fraction of air, fvn, has been replaced by z and
the volume fraction of the soft body material by the term (l-z).
For the materials of interest pf>> Pair' so that Equation 58 can be
approximated by the relationship
P = (1-z) Pf

For a porous soft body material, the isentropic pressure-density

relationship can be obtained from Equation B-9. Substituting the terms
just defined, this equation becomes

= -z)(!2 + P2z (60)

where the subscript 1 refers to the initial (unstressed) state and the
subscript 2 refers to the final (stressed) state, A ind B are empirical
constants given by Equation 54, and a is the ratio of specific heats
of air. Similarly, the shock pressure-density relationship carn be
obtained from Equation B-18 and takes the form

- /B
Z (1-) + 1 +z (1-q) (61)
Pz2

50
AFML-TR-77-134

where. q is defined in Equation B-17. The difference in the isentropic

and shock relations is due to the fact that the pressure-density
relationships for air are different for these two cases, whereas these
relationships have been assumed the same for the soft body material.

The effect of porosity can best be demonstrated by applying

Equations 60 and 61 to the case of water with varying amounts of porosity
added. The properties of water are given in Appendix C. Figure 20 shows
two pluts of Equation 61 for this case in which porosity is varied from
0 to 40%. Figure 20-a shows the variation in specific volume (density )
with pressure up to 500 MN/m 2
(5 kbar) and Figure 20-b looks at the lower
portion of these curves for pressures up to 20 MN/m 2 . The relationship
between the particle velocity and the shock velocity for porous materials
is obtained by simultaneous solution of Equations 1, 3, and 61. This
relationship is shown in Figure 21 for water with porosity added.
From the solution of these three equations, the relationship between the
shock pressure and particle velocity is also cbtained. This relationship
was used in an earlier section to determine the effect of porosity on the
shock pressures generated during impact. Figure 16 shows the Hugoniot
shock pressure as a function cf impact velocity and porosity for the
impact of water onto a rigid plate.

The isentropic pressure-density relationship, Equation 60, for

water with porosity is shown in Figure 22 (for water, volume = (densityl).
In isentropic compression, the volume of air in the mixture is seen to
vanish under substantial pressure. Simultaneous solution of Equations 15
and 60 result in the relationship between steady flow pressure and impact
velocity. This relationship was used in a previous section to derive
the variation in the steady flow pressure at the stagnation point to
the amount of porosity in a projectile of water. This i~iformation was
presented in Figure 17.

51

........
......
... .. _
AFML-TR-77-134

z -.00
Z .05
16-0- 16,0O

Z,.20
N.' 12.0 Z =.30
~~~ ----- z=.40
Zz4

S 6.0

4.0

0.0 ifmmuiu
0.0 0.09 0.10 0.11 0.12 0.13 0.14 0.15
VOLUME (10Ar/kg)
a

Z X.00
Z z.0 5
400 Z ".10

"E Z C.20
z 300 Z.30
Z= 40

f 200
w

100

0
0.0 0.08 0.09 0.10 ,i1 0.12 0.1 3 0,14
VOLUME (10mTA/kg)
b

Figure 20. Hugoniot Pressure-Volume Relationships for Water with

Various Porosities. (1 MN/mr2 = 10 Bars)

52
AFM4L-TR-77-1 34

2400 Z20

20001

0 1600 -Z U.20
Z 3

81200

400

0 75 150 225 300 375 450 525

PARTICLE VELOCITY, up (rn/s)

Figure 21. Relationships between the Shock Velocity and the

Particle Velocity for Water with Various Porosities.

53
AFML-TR-77-134

16

i
* 12
12

W
0Z0

w Z81.05

4 ZZ2O
.30
Zv.4O

0 0.10 0.11 0.12 0.13 0.14 0.15 0.16

VOLUME (10 na/kg)

400

IE
1-1 300
z

W 200

100 Z=.00 to 40

0
0 0.06 Q08 QIO 0.12 0.14 0.16 0.18
VOLUME (10 M3/kg)
b

Figure 22. Isentropic Pressure-Volume Pelationshlps for Wter

with Various Porosities. (MN/m 2 = 10 Bars)

54

..........
AFML-TR-77-134

4. LIMITATIONS OF THE THEORY

The basic theory assu,nes that a soft body can be treated as a

homogeneous fluid during impact. The nmaiii case described was the
impact of a right-circular cylinder against a rigid plate, with other
cases being studied briefly. The extent to which the theo;-y is limited
in application will now be considered.

Application of this theory to nonhomogeneous materials may result

in several errors. Variations of density in -he material will result
in variations of the flow pressure. Secondly, mixture theory, which

is the basis of the equations of state for porous materials, requires

that the material be macroscopically homogeneouE. If this is not so,
the propagation of waves in the material will be affected, wi~h an
&ccompanying deviation in the shock velocity and shock pressure.
As stated earlier, a material may be treated as a fluid during

impact if the pressures generated during impact are much greater than
the str'ength of the material. For a projectile with a length-to-diameter
ratio (L/D) of approximately one or less, this probably requires that the
Hugoniot shock pressure be an order of magnitude greater, than the tensile
strength or shear strength of the material. For a projectile with an
L/D goeater than one, there is an added requirement that the stresses
generated in the projectile during steady flow must also exceed the
material strengths. However, as will bc discussed later in the Results
section, it appears that the theory may accurately predict pressure for
impacts in which there are large distortions during impact, even though
the material strength is not !exceeded.

The theory, originally developed for, a cylindrical projectile, is

not limited to this geometry. However, it is probably the most convenient
to study, Doth theoretically and experimentally. Fcr projectiles which
are spherical in shape, there is not sufficient length for steady-flow to
be established. However,the initial shock and release processes as
explained in the theory should still exist. In order for steady-flow to
be established, the projectile must have an L/D greater than one.

55
AFML-TR-77-134

Finally, the theory requires that the ti.rget remain flat, although
l1mitud deflections of the target are allowed. Large distortions of the
surface, stech as "pocketing" of a plite, are not allowed since the paths
of the streamlines will be altered, thereby influencing the flow pres-
sure. Also, the aerial dimensions of the target surface should be
several times the diameter of the projectile, so that full turning of
the projectile results.

56
 AFMIL-TR-77-134

SECTION IV

EXPERIMENTAL TECHNIQUE

The purpose of the experimental program was to measure the forces

and pressures generated during soft~ body Impact on a rigid surface. Two
basic tests were used in this program. A Hopkinson bar with strain gages
was used to measure impulse Imparted to the target during the impact and
a flat plate with pressure tý-ansducers was used to measure the temporal
distribution of pressure at various points on the plate surface.

The Hopkinson bar gave a reliable measure of the impulse (or momentum)
Imparted to the impact surface as well as the relative distribution of
imoilse diuring the impact event. The flat plate gave a reliable measure
of pressure impirted t(, the target during the impact although there
were an insufficient number of pressure transducces to give the distri-
bution of pressures over the entire surfare for any given test. Both of
these tests will be presented ingreater detail along with the gun
system used for launching the projectiles.

1. GUN SYSTEM
All of the impact tests for this program were conducted on the gun
range shown in Figure 23. This gun range is located in the Impact
Mechanics Facility of the Air Force Materials Laboratory. For these tests,
the gun was used in two differen~t c~onfigurations. As shown in Figure 23,
the projectile was explosively driven during early testing. In later
tests, the projectile wa!s driven by compressed air.

I
a. Powder Gun
Figure 23 shows the original powder gun configuration. Dluring
launch, the projectile was housed ina sabot made of high density
polyethylene. This heavy sabot was needed to withstand the high
acceleration forc~es generated by the exploding gas. As the sabot
traveled down the gun barrel, the high pressure gases behind the sabot
were vented in the blast chamber. Complete venting of the high pressure
gases was necessary to establish a constant velocity in the projectile

57
AFML-TR-71-134

TARGET TANK

"13 LAT TAN,,

IBREEC

Figure 23. Overall View of the Gun Range Facility Used for Bird
Testing.

prior to impact. When the sabot and projectile reached the end of the
blast tar'k, the sabot was trapped by the sabot stopper, shown in Figure
24. The stopper was so designed to allow the projectile to continue its
flight without loss of velocity.

The high acceleration forces during early launch and the rapid
stopping of the sabot often damaged the projectiles. Requirements for
a thick sabot wall also limited projectile diameter to about 5 cm.
For these reasons, it was decided to redesign the gun for compressed
air. A more complete description of the powder gun can be found in a
report by Barber, et al. (Reference 65). Also included in that report
is a detailed explanation of the velocity measuring system and the
photographic system.

58
 AFML-TR-77-143
i /SABOT VOTOPER~ PLATE

11 Figure 24. Sabot Stopper for Use with Powder sun.

b. Compressed Air Gun

During the final phases of this project, projectiles were accel-
erated by compressed air vented from a lerge tank. There were many

I advantages to this system. The heavy, polyethylene sabots were

replaced by light, thin-walled balsa wood sabots. This was due to the
Iower acceleration forces inherent in compressed air guns. Figure 25
is a picture of a typical projectile seated in a balsa wood sabot.
In order to separate the projectile from the sabot, • sabot stripper as
shown in Figure 26 was used. In this technique, the last section of the
gun berrel has a decreasing inner diameter. As the sabot enters this
section, it is gently decelerated over a distance of several meters.
This slow deceleration enhances a clear separation of the projiectile
from the sabot. Barber, et al. (Reference 66) presented the compressed
air launching system in greater detail.

2. HOPKINSON BAR
In 1914, Hopkinson (Reference 30) reported the use of a circular
bar to measure the temporal distribution of force generated during the
impact cJ ailead bullet on the end face of the bar. Previous workers,

59

LN TUG
 AFM4L-TR-77-134

IAi

Figure 25. Cylindrical Projectile with Balsa Wood Sabot.

Figure 26. Sabot Stripper for Use with Compressed Air Gun.

60

7- -7
AFML-TR-77-l34

suspending a bar as a pendulum, had measured only the total momentum

transferred to the bar during an impact on an end face. Hopkinson
developed an ingenious method using two bars suspending as pendula. in
which one bar was impaoted by the bullet and the other acted as a
momentumi trap. Upon impact of the first bar, a stress wave, generated at
the impact surface, propagated down the length of the bar. This wave
then passed into the second bar where it was "trapped" and measured by

the height of swing of the second bar. By this method Hopkinson was ableI
to resolve the transferred miome ntum (the impulse) in time, thereby
obtaining the force generated during impact as a function of time.

Although this method worked well, it required many impact tests

using a like projectile and like velocity in order to resolve the force-
time distribution. In the present research programs, the stress waves
were measured using strain gages.

a. Theory
When a soft body projectile impacts a Hopkinson bar, a shock
wave propagates into the projectile as described earlier. At the same
time, a stress wave is generated in the bar. It is essential for a
Hopkinson bar test that the pressures generated at the impact surface
be much less than the yield strength of the bar material. For this case,
the stress wave in the bat, will take the form of an elastic wave. As this
wave propagates Jown the bar, the radial stresses will be released rapidly,
so that the wave will become a plane stress wave, tratveling at the
longitudinal wave speed, c,, of the bar material.

It is essential that the stress wave in the bar be elastic for

several reasons. An elastic wave will propagate essentially undisturbed
into the bar. Also, the strain measured by the strain gages will be
linearly related to the stress at that section of the bar.

The strain gages are normally located at least ten diameters

down the bar from the impact end. This is necessary to give the elastic
wave sufficient time to become planar so that the average strain in the
cross section will be the same as that measured by the gages. For this

61
AFM4L-TR-77-134

case, the strain will be linearly related to the average stress at the
cross section and the strass will be related to the force by the area

of the cr~oss section.

Figure 27 is an illustration of a Hopklnson bar. When the
projectile impacts the end of the bar, the amplitude of the pressures
generated vary over the impact surface. The resultant stress wave in

the bar is nonplanar. However, by the time the wave reaches the strain
gages, it has become planar and the stress measured by the gages is only
the average of the initial stress distribution. Since the exact stress
distribution cannot be measured, only the total force is used. The
relation for force is
r MeA (62)

where E is the modulus of elasticity of the bar, A is the cross-sectional

area of the bar, and c is the strain recorded.

STRAIN GUAGE
II

Figure 27. Schematic of Hopkinson Pressure Bar.

62
AFML-TR-77-1 34

Because of the large diameter of the bar, high frequency com-

ponents of the stress wave are unable to propagate down the bar without
being dispersed (Refer'ence 11). Thus, the measured force profile is
somewhat different from the actual forces generated at the impact surface.
However, even though the force history will be somewhat distorted, the
total impulse imparted to the bar can still be obtained by integrating
the force-timie trace.

b. AFML Facility
Figure 28 is a picture of the aluminum Hopkinson bar mounted in
the AFML range.

Figure 28. Overall View of the Hopkinson Bar Apparatus.

The bar is 7.6 cm in diameter and 370 cm long. The gages are mounted
76 cmifrom the impact end. The bar is mounted in Teflon rings to allow
for free movement and expansion of the bar. This is necessary to prevent
extraneous wave reflections in the bar.

63
AFML-TR-77-l134

The Hopkinson bar was used in three modes. For normal impact,
the end face of the bar was cut at a right angle to the axis of the bar
and the projectile Crajectory was along the axis of the bar. For oblique
impact, the end face of the bar was cut at an angle of 250 and 45* to the
axis of the bar and the projectile trajectory was again along the axis
of the bar.

c. Data Reduction
Equation 62 was used to obtain the force from the strain measure-
ments. The strain data, recorded with oscilloscopes, was converted to'
force, digitized, and integrated with time numerically to obtain the total
impulse delivered during impact.

3. PRESSURE PLATE

As stated e~rlier, the response of the Hopkinson bar is limited by

geometrical wave dispersion. Thus, the force history recorded using this
device is somewhat inaccurate. Also, it is not capable of measuring the
distribution of pressure at the impact surface. The obvious method for
obtaining this-information is to mount pressure transducers on the surface
of a rigid plate. A method similar to this was used by such researchers
as Bowden and Brunton (Reference 67) and Kinslow, et aI. (Reference 39)
to measure the pressures generated during the impact of a water jet.
However, as seen inthe Literature Survey, the results of previousI
researchers using similar methods of pressure measurement for the impact
of water appear to be incorrect. The main source of error was probably
insufficient frequency response of the transducers to measure the short
duration shock pressures (on the order of a microsecond or less). This
problem was compounded by the inability to control the exact size and
shape of the projectile.

Both of these problems were overcome in this study by using much

larger projectiles with sufficient material strength to hold their shape.
The larger diameter and the flat impacting surface of the projectiles
resulted in shock durations of 10 lis or greater at the center of impact.

64
AFML-TR-77-134

a. AFML Pressure Plate Facility

Figure 29 is a picture of the pressure plate used at the AIML

tf aci I i ty.

I PRESSURE
TRANSDUCERS

Figure 29. Pressure Plate Apparatus.

The pressure transducers used for these tests were p'e:oelectric quart:-
pressure transducers manufactured by PCB Corporation. Ihe four trans-
ducers were mounted flush to tke surface and located at the center of
impact and 1.27 cm, 2.54 cnm, and 3.,81 cm from the center. The steee
plate, 15.25 cm in diameter and 5.0,8N cm thick, was mounted as shown in
the fi.'ure. The transducers are described in a report b), Barber, et al.
(Reference 65) to a mLuch .reater extent.

b. Data Reduction

Several exanip les of actual pressure trace. are shown in the ne\t
section. These traces are digitized from the oriqinal oscilloscope
traces . The ampl itudes of the pressures were obtai ned b) multiplyinkl the
output vol taqe of the gage by the kqakqe cal ibration factor.

65
AFML-TR-77-134
The transducers were not originally designed for direct impact
I
testing. This týact resulted in several problems during their use. The
gages did not have adequate acceleration coinpensatir~n. Thus, plate
vibrations caused spurious Poise to appear in many of the tests results.
Secondly, the resonant frequency of the gages was approximately 300 kHz.
For the strong shocks which displayed high ampli~tudes and short rise times.
this condition caused resonance t~o appear in the output of the gages as
well as apparent overshoot of the initial pressure pulses. This over-
shoot is apparent in several of the actual pressure traces displayed in
the next section. The accuracy of the shock pressure measurements was
also limited by the finite frequency response of the gages (reportedly,
100 k~z) which prevented measurements of rise times of less than 5 u~s.

66
AFML-TR-77-1 34

SECTION V

RESULTS AND DISCUSSION

In order to check the applicability of the soft body theory presented

earlier, a series of tests were conducted using the Hopkinson bar and the
flat plate apparatus. The projectiles for the majority of the tests
consisted of birds (chickens) and the bird substitute materials RTV rubber
and gelatin. The various material properties used in the theoretical
analyses are presented in Appendix C.

All of the projectiles, except the birds, were right circula-.

cylinders. This greatly simplified comparison of theory andi experiment.
Both the RTV and gelatin were tested with and without porosity in order
to see its influence on the pressures generated during impact. Neoprene
and beef steak were tested briefly in order to study the effect of material
strength and homogeneity. The effects of impact obliquity were studied
by conducting tests at three angles, 250, 450, and 900 (normal) to the
surface of the plate.

1. HOPKINSON BAR TESTS

The assumption that the various soft body projectiles behave as
fluids during impact requires that the projectile rebound velocity be
negligible. For normal (900) impacts, this implies that the impulse
imparted to the target must be equal to the initial momentum of the
projectile, assuming the target is rigid. For oblique impact cases,
thi- implies that the impulse must be equal to the normal component of
the initial momentum, that is,
I = Mu0 sin al (63)

Inorder to obtain an accurate measure of the total impulse, various

soft body projectiles were impacted against the Hopkinson bar apparatus
described previously. Projectile mass, impact velocity, and impact angle
were varied for each of the materials tested. The results of these tests
are shown in Figures 30, 31, 32, and 33.

67
AFNL-TR-77-134

S/0

InJ

04

Do
N -- O N• -- 0

I- ,--.,, I
Lu
tm

C~C

kn In• m

S04 'Ui 0.i

N 0 0 N6 0 00

68
 AFML-TR-77-134

16

12

I 00 0 " 5

Figur-90e(NO1AL
0 4 8 12 16 20
NORMAL COMPONENT
0
OF MOMENTUM
m u. %in* (N -s)
20
Figure 31. Relationship Between the Nornral Component of
Momentum and the Impulse Imparted to the Target
for the Impact of Porous RTV.
_o - 900 (NORMAL)
4 0-45*
16 0/

-- 6•-900 (NORMAl)

J ~0°250
4 l 12 16 20

NORMAL COMPONJENT OF MOMENTUM

m uosin•a (N-s)

: Figure 32. Relationshitp Between the Nc-vi.l Component. of Momentum

and the Impulse imparted to zha T~rqet for the
Impact of Por', %•Gelatin,

•. 69
 AFML-TR-77- 34

200

a/0

I . 3
I o 0

0 0 90*0(NORMAL)

0
0 4 S 12 16 20
NORMAL COMPONENT OF MOMENTUM,
MU. uing (N - )
Figure 33. Relationship Between the Normal Component of
Momentum and the Impulse Imparted to the Target for
the Impact of Birds.

Figure 30 presents examples of the measured force histories

from which the impulse measurements were obtained by integration. These
are only approximate force histories since the Hopkinson bar effectively
filters the higher frequency components. The impulse measurements are
compared with the normal component of momentum in Figures 31, 32. and 33
for porous RTV, porous (micro-balloon) gelatin, and birds. From each
of these figut-es, it is apparent that there is no significant deviation
from the relationship given by Equation 63 for each of the materials
tested.

2. PRESSURE PLATE TESTS

a. Normal Impact

The fluid model of soft body impacts against a rigid target,

presented earlier, has several characteristic features. For the impact
of a cylindrical projectile the pressure history at the projectill-target
interfe'ce is~ divided into three distinct phases; the shock, the release,

70
m i ,o-
AFML-TR-77-134

and steady flow. For normal impact, the initial Hugoniot shock pressure
is given by Equation 4, PH = Pousuo' and has a duration at the center of
impact given by Equation 7, tB = a/cV. During the releaso phase, radial
flow is established. Because of the complicated releasL process, no
analytical expressions hav2 been derived for this phase. However, once
flow is completely established, steady flow relations hold. The pressure
at the stagnation point, Ps, located at the center for normal impact, is
derived from Bernoulli's Equation and takes the fom of Equation 16,

yF
f Ps +
P dP
o0
2
O 2o

and the pressure distribution across the surface may be approximated by

Equation 27,
= (1-3 + 2(

The duration of the entire impact process may be approximated by the

time that it takes for the projectile to flow through itself, given by
Equation 21, tD = L/uO-

In order to experimentally determine the pressure histories for

various soft body impacts, the pressure plate apparatus described
previously was used. All of the projectiles, except the birds, were right
circular cylinders, with length-to-diameter ratios of approximately
two. During this testing program, the size, mass, density, and porosity
of the projectile were varied as well as the impact velocity and impact
angle.

Typical pressure histories recorded at the center of impact for

the various soft body materials are presented in Figure 34. Here the
pressure measured with the center transducer (stagnation point) and the
time are plotted in the nondimensionalized form, obtained by dividing the
prassure by P = 1/2 pzuo 2 (the steady-flow stagnation pressure for an
incompressible fluid) and the time by tD = L/uo (the theoretical duration
of impact). Pz, the average density of the porous material was definad

71

. ...
AFML-TR-77-1 34

0 0
40 1 0W 40 1 i

00

C-

eA

C-

0 In-
CM

0 0

72
 AFML-TR-77-134

Inn

zz

a4 o
sD• N 0
'8 t1 N 0
'U

t-4

AV)
inn

inn
CYI

mU

_N • too
go _

3
-7
 AFML-TR-77-1 34

in Equation 58. The impact velocity for each of these tests was
approximately 200 rn/s. The similarities of the pressure histories for
the various materials are evident. Each shows the expected fluid
characteristics; namely, an initial peak pressure followed by a decay to
a lower flow pressure, with a total duration of approximately tD = L/u.
Of course, the exact shape of the pressure histories varies for each
material due to differences in material properties.

The peak pressure is seen to be greatest for the pure gelatin,

due to its high shock velocity and lack of Porisity. Each of the other
materials, with the exception of neoprene, had certain amounts of porosity
which greatly decreased their shock velocities, hence their shock pres-
sures. Neoprene, even with zero porosity, has a very low shock velocity.
On some of the bird shots the pressure did not rise inmmediately to a peak.
This was due to the impact of such things as the bird's feet or feathers
prior to impact by its main body.

For each of these materials, an apparent steady flow region

exists and it is most readily seen in the pressure history of porous RTV.
t;ý Each of the other materials displayed a tendency for the pressure to
slowly decay in this flow regime. This decay from steady flow, most
obvious for the beef and pure gelatin, was described in the theory as
being the unsteady flow effects due to the finite length of the projectile.
As stated in the theoretical section, this effect should be greatest for
subsonic impacts such as those of the beef and gelatin. It should be
least apparent for supersonic impacts such as for the porous RTV a~nd
micro-balloon (porous) gelatin.

Since pure gelatin is almost incompressible at this impact velo-

city, its stagnation pressure is seen to be approximately 1/2 pzu 0 2
Each of the other materials showed varying degrees of compressibility,
evidenced by the fact that in the steady flow regime, their nondimension-
alized pressures had values greater than one.

One of the major features distinguishing the pressure histories

of birds from that of porous RTV and micro-balloon gelatin is the much
larger amount of high-frequency "noise" superimposed on the pressure

74
AFM4L-TR-77-1 34

profile of birds. Initial inquiries into the source of this noise were
fruitless. In order to see if it was caused by the bone structure of the
bird, a beef projectile was constructed and tested. As can be seen,
the noise was even greater for the beef than for the birds. Thus, the
effect of the bone structure was assumed negligible. Another thought
was that the noise might have been caused by the break-up or tearing of
the material (rreation of new surfaces) during impact. This theory
seemed resonable since both the birds and the beef consisted of muscle
fibers which would require more energy to tear than would the RTV or
gelatin. Along this line of thought, neoprene was tested since it bounces
at velocities of 200 m/s and there is no resulting material break-up.
Accordingly, the pressure history of the neoprene impact showed no noise
except that due to acceleration loads on the gages caused by the pressure
plate vibrations. Thus, there appears to be a direct c,.rrelation between
the amount of energy expended in material break-up and the amount of
high-frequency noise recorded. (The extreme amount of noise recorded
during the impact of pure gelatin was apparently due to excitation of the
[gage near its resonant frequency).
point
IiA of interest noted during this testing program was the
fact that the neoprene's pressure history was somewhat similar to that of
the other materials even thou~h it did not "flow" at impact velocities
as high as 200 m/s. This was noted for porous RTV and micro-balloon
gelatin which bounced at impact velocities of 100 m/s. This would seem
to imply that the basic fluid flow theory presented in this report holds
even for materials which do not flow (or shatter) during impact but
which do undergo large amount of deformiation (strain).
A closer comparison of the pressure histories for the impact of
birds and the two bird substitute materials, porous RTV and micro-balloon
gelatin, can be obtained from Figures 35, 36, and 37. In these figures,
typical. pressure histories at the center of impact are shown for various
impact velocities. In each figure a short line representing the pre-
dicted shock pressure amplitude, PHP and duration, tBs is shown as well
as a long line representing the predicted amplitude of the steady flow
pressure at the stagnation point, Ps. The pressure and time are again
presented in nondimensionalized form.

75
 AFNL-TR-77-134

i4

O 22 /

0
W4 -
4
: 4

01
w L

gig: 116 rn/

S0 24 .25_____ __ .___ 1__0_____

4 t 94 rn e
u
2

0 .25 .50 .75 1.00 L25

NONOIMENSIONALIZED TIME, Tut
L
Figure 35. Typical Pressure Data at Various Impact Velocities
for Porous RTV, Normal Impact.

76
i AFML-TR-77-134

6
•'4 uoz 278m/A

8
0A.

-m 4 Uo 171 rn/s

w 4
4
S2u6a
N2 120Om/

2 ___, -_,__, - --

0
0 .25 .50 .75 tOO L•25

S~L NONDIMENSIONALIZED TIME, I.U

Figure 36. Typical Pressure Data at Various Impact Velocities

for Porous Gelatin, Normal Impact.

77
 !i I'
AFML-TR-77-134

2 U

"Q) 4d
0 u. z 225 rn/S
2

W6
N
:3 uO: 197 rn/S
4

4I.

0 .25 .50 .75 h.0 1.I5

NON1DIMENSIONALIZED TIME, "_

L

Figure 37. Typical Pressure Data at Various Impact Velocities

for Birds, Normal Ympact.

78
AFM4L-TR-77-l134

The calculated amplitude and duration of the shock pressures

agree well with the measurements for the two bird substitute materials,
although some transducer resonant excitation and resultant overshoot is
seen on several shots. Discrepancies between theory and measurements
for birds should be expected due to the rounded qeometry and irregular
surface of the bird and the prior impact of su. '-hings as feathers,
wings, and legs.

As was stated earlier, each of the materials seems to display a

regime of steady flow during which the measured pressure agrees with the
calculated stagnation pressure. The deviations in the pressure histories
for the micro-balloon gelatin at high velocity may have been due to
geometric distortion of the projectile during launch or to resonant
excitation.
In order to demonstrate the radial variation of pressure,
Figure 38 presents a typical porous RTV impact in which the pressure
histories at three transducer locations are shown. The peak pressure
appears to be less at the transducer located at 0.6 n/a than at the
center since the duration of the shock pressure at this point is so short
that the transducer could not respond rapidly enough to record it. Of
course, there was no shock at the transducer located outside the initial
impact area of the projectile. This figure also demonstrates the radial
decrease in pressure in the steady flow regime.
Figure 39 shows a typical pressure history for porous RTV
recorded along the major axis of an oblique impact of 450* Lines are
drawn in each figure representing the calculated shock pressure, and
full, steady-flow stagnation pressure. The predicted shock durations are
not shown in these figures. As explained in the theory, the shock
pressure for an impact velocity u0 and an impact angle a should be
identical to that produced by a norm'al impact with an initial velocity
of u0 sin oi. From Equation 36, the 450 impact angle is seen to be
very near the critical angle for RTV. This implies that the full shock
pressure may exist over the entire impact surface and that the duration
of the shock near the center of impact should be sufficient for the
pressure gages to record the full shock pressure. Also, expected from

79
AFML-TR-77-134

G4 CENTER
(&w0.0)

0.

S4 i1.27 cm OFF CENTER

.)
z

4
2.54 cm OFF CENTER

0
0 .25 .50 .J5 1.00
NONDIIENSIONALIZED TIME, T-u

Figure 38. Typical Distribution of Pressure Pltte Data for the

Impact of Porous RTV, Normal Impact.

80
AFML-TR-77-134

42l.. 2.54 cm ABOVE CENTER

4
L24cm ABOVE CENTER

0.

p I

4 1.27CM BELOW CENTER

0 .25 .50 .75 1.00 1.25

NONOIMENSIONALIZED TIME, Tu,

Figure 39. Typical Distribution of Pressure Plate Data for the

Impact of Porous RTV, Oblique Impact (a a 450).

81
AFML-TR-77-134

the theory is the fact that the stagnation point, which is approximately
3 cm above the center transducer, should "see" the full stagnation pressure
during steady-flow, regardless of the angle of impact.

From the figure, several of these points are clear. The shock
pressure at the center gage is the full shock pressure, and the other
gages record peak pressures near this value. The steady-'flow pressure is
also seen to be highest at the gage 2.54 cm above center. The fact that
the pressure here is somewhat below the stagnation pressure implies thait
this gage was not located exactly at the stagnation point. It should be
noted that for much smaller angles of impact (greater impact obliquityl
the amplitude of the calculated shock pressure will be less than the
steady-flow stagnation pressure for RTV.
In Figures 34-39, typical pressure histories for the various
materials were presented in order to demonstrate a qualitative under-
standing of the pressures generated during soft body impact. However,
in order to compare the fluid theory with the soft body impact experi-
ments, a mare quantitative presentation of the pressure data is required.
In this section, the experimental results are compared with theoretical
predictions for the various materials.

The variation of shock pressure with impact velocity for normal

impact is given in Figures 40, 41, and 42 for porous RTV, micro-balloonI
gelatin, and birds, respectively. In Figure 40, the theoretical
Hugoniot shock pressure is given for both pure (z = 0.00) and porous
(z = 0.50) RTV. Experimental measurements for the porous RTV tend to
be higher than that predicted. However, as stoted previously, it
appears in Figure 36 that the high peak pressures may have been due to
some transducer overshoot. In fact, the pressures tended to oscillate
about the predicted shock pressures. In Figure 41, the theoretical
Hugoniot shock pressure is given for pure (z - U.00) and porous (z = 0.40)
gelatin. Experimental measurements for the impact of the micro-balloon
gelatin are shown to agree very closely with the predictions.
In Figure 42, the theoretical Hugoniot shock pressure is given
for birds. As stated in Appendix C, this curve is based on the assumption

82
 17

AFML-TR-77-134

350

3001

A:~4 1c:I30 kg/.3

250 Z 0.O
z

-L,oo

0:

I. mP.100

50 o0
0 EXPERIMENTAL DATA
0 - Z=.50
0
0 50 100 150 200 250 300
IMPACT VELOCITY, U. (m/s)

Figure 40. Hugonlot Shock Pressures Measur~ed Duri-:ig Normal

Imp act o5 Porous RTV. (I MN/nL =10 Bars,
109 kg/mn .036 lb/In 3).

83
AFML-TR-77-134

350

300 " 1060 kq/ý

Z250o

~200

S150

0G _EXPERIMENTAL DATA
0- Z a.4%)

0
0 50 100 150 200 250 300
IMPACT VELOCITY, u. (m/s)

Figure 41. Hugonlot Shock Pressores Metsur.d During Nirmra impact

of Porous Gelotin. 0
(1 N/)n - ýO Bars, 10 kg/m -. 036 1b/in3 ).

84
AFML-TR-77-134

350
3
Po 1060 kg/M
ti Z -a00

oo 9 5 0 kg//

a 250-~g
® ' zz:,.Io .,
1060 qw

•=200

1l501~

#4 06 g ID
100.

'-a.' U EXPERIMENTAL DATA

0 4 kg BIRDS
50 0
a,
2I k9
kq BIRDS
1B1RrD.S
0 '
0 .5 L4 WR0S
-N .1 kg BIRDS
0
0 50 100 150 200 250 300
IMPACT VELOCITY, u. (m/$)

Figure 42 Hugoniot Shock Pressures Measured During Normal Impact of

Birds. (I MN/rn - 10 Bars, 103 kg/m 3 * .036 lb/in 4 ).

85
AFML-TR-77-134

that the Hugoniot relationship for birds is similar to that for water
(or gelatin) with an assumed pure density of pf = 1060 kg/m 3 and a
porosity of z a 0.10 (average density pz * 950 kg/m 3). The data shown
here were obtained from tests conducted on small birds at AFML and on
large birds at AEDC (Reference 66). The measured peak pressures fall
below that predicted. The fact that the peak pressure tended to increase
with bird mass, hence with bird size, implies that the low peak Aressure
measurements may be due to the lack of planar impact for birds. As shown
in the theoretical section, the duration of the full shock pressure is a
function of the radius of curvature of the impacting surface, so that the
shock duration at the center of impact is much less for a projectile with
a rounded end than for a right circular cylinder. For small birds with
an irregular impact surface and a small radius of curvature, the shock
duration was probably so small that the transducers could not respond
rapidly enough to record the shock pressure. For larger birds with larger
radii of curvature, the duration of the shock was greater. The peak
pressures measured on several of the 4 kg bird tests approached the
predicted values. Other reasons for the low peak pressures include the
existence of material such as feathers and legs which impacted prior to
the main body and the existence of lower impedance material at the
surface of the main body.
The variation of pressures in the steady flow regime with impact
velocity is presented in Figures 43, 44, and 45. The pressure at the
center stagnation point is plotted for RTV in Figure 43. Theoretical
predictions are given for pure and porous RTV using the compressible
fluid theory. A curve of P 1 1/2 pUo 2 (incompressible fluid theory,
p = 670 kg/m 3 ) is also presented for comparison. The measured stagnation
pressures for porous RTV agree very closely with those predicted by
the compressible theory.
In Figuye 44, the steady-flow, stagnation pressure is given for
both pure gelatin and micro-balloon gelatin impact tests. Theoretical
curves from compressiblc theory arc shown as well as the plot of
2
P = 1/2 Puo (Pz 640 kg/mr). The experimental data is seen to agree
very nicely with the compressible theory for both types of gelatin.

86
 AFML-TR-77-134

50

It

1a
330 ko/4,
40

Z .5

i'~~~ a.0 0O - "S

20 b 0.

!3 ou 50I0 5 kg/rn5IPau* 3

0.l IMATVI.CIY (oe

EXPERIMENTAL DATA

0
0 50 100 15O 200 250 300
IMPACT VELOCITY, us (We/)
Figure 43. Steady Flow Pressures Measured at the Center of Impact
During Ngrial Impact of Porous RTV. ( NW 1 as
10J-kg/mJ .036 lb/In 3 ).
I-.

87
AFML-TR-77-134

50

1~U060 kg/"S3
Z R0.

40

10
l60 ko/,,

F•20 0
//
00s UO

SrsS/ Z = .40

sa
520 oJ

do EXPERIMENTAL DATA
- 0I Z ;00
0 . - O-Z .40

0 50 100 15O 200 250 300

IMPACT VELOCITY, u. (rn/a)

Figure 44. Steady Flow Pressures Measured at the Center of act

Dul~n9 NQrmal Impact ol Poous Gelatin..0 MN/m 10 Bars,
10 kg/mJ .036 lb/InPo a (

88
AF-L-TR-77-134

50

40

__ zI o

K MPa I

30 ka 1O060k/

I z*0jo kg/M3
F4a950

200
00 0

IO O O

EXPERIMENTAL DATA

0 O BI(RS (CHICKENS)
.0
0
0 50 1OO 150 200 200 300
IMACT VELOCITY, ue (mr/)

Figure 45. Steady Flow Pressures Measured at the C nter of Impact

Du~tng NQrmal Impact o Birds. (1 MN/mW 10 Bars,
104 kg/m a .036 lbin)

89
AFML-TR-77-134

In Figure 45. the stagnation pressure is given for the impact of

birds. A theoretical prediction is made assuming pf a 1060 kg/m3 and
z a 0.10 and pf - 950 kg/m 3 and z - 0.0. As shown earlier in Figure 35
for the impact of birds, end effects in the bird causes a gradual decrease
in flow pressure, making measurement of the steady flow pressure difficult.
However, the measured data does appear to be in fair agreement with the
theoretical curve based on the compressible flow theory, although there
is more scatter in the data than seen previously for the RTV and gelatin.
Figures 46, 47, 48, and 49 show the variation of pressure in the
steady flow regime with radial distance from the center. The pressure is
again nondimeoisionalized and the radial distance is nondimensionalized
with respect to the initial radius of the projectile, a.

1.6

a.S

0 .5 ID
NONDIMENSIONALIZED 0,
RADIUS, r/q 2.0

Figure 46. Radial Distribution of Steady Flow Pressures for thv

Nora1 Impact of Porous RTV.

90
AFML-TR-77-134

.4

0
0 .5 1.5 20
?4ONDItMNSIONALIZED RADIUS, V/a

Figure 47. Radial Distribution of Steady Fl.* Pressures for the

Normal Impact of Pure Gelatin.

-. 2

0•

10 0
.....
.5 LO
_____

1.5 2.0
NONDIMENSIONALIZED RADIUSt '/a
Figure 48. Radial Distribution of Steady Flow Pressures for the
Normal Impact of Porous Gelatin.

91
AFML-TR-77-1 4
" $. 1.6

I.;

.4

0 .5 tO 1.5 2.0
NONDIMENSIONAIUZED RADIUS, r/0
(ADJUSTED)

Figure 49. Radial Distribution of Steady Flow Pressures for the

Normal Impact of Birds.

The theoretical curves are drawn from the approximate relationship given
by Equation 27. The experimental data are presented in the form of bars
which designAte the data bounds. The large scatter in the data may have
been due to the projectiles hitting off-center and inaccuracies in the
pressure measurement. Because of the data scatter, it is not possible to
determine how accurately the theoretical curves predict the actual pressure
distribution.

b. Oblique Impact
Original plans called for experimental testing of normal impact
only. However, during the course of this research program, it became
evident that some investigation of oblique impact should be performed
in order to test a few of the basic concepts.

As already noted in the Hopkinson bar tests, it was shown that

the impulse imparted to the target was related t: the initial momentum of

92
AFML-TR-77-134

the projectile by the sine of the impact angle; that is, the impulse was
equal to the normal component of momentum. As was demonstrated in
Figure 39, the amplitude of the Hugoniot shock pressure for an oblique
itvact of angle a is simply the shock pressure predicted for the normal
component, uo sin m, of the impact velocity. For steady flow, the
stagnation pressure is the same as for normal impact, although the
stagnation point is no longer located at the center of impact, and the
distrilition of pressure over the surface is no longer axisymmetric.

For the purose of this limited testing, only the impact of

porous RTV was considered. Impact tests were conducted at angles of 25"
and 450 to the surface of the target plate. In Figure 39, typical pres-
sure histories along the major axis for an obli;-e impact of 456 were
given. In Figures 50 and 61, the calculated Hugoniot shock pressures
are compared with the measured peak pressures for oblique impacts of 450
and 250, respectively. Thr pressure at each of the four transducers is
presented for the 45 impact case. For the 250 impact case, only the
data from the gage nearest the initial impact point is given. For this
case, the other transducers do not see the full shock pressure since the
critical angle has been exceeded. The general trend of the data seems
to agree with the theory. The few high measuremehts may have been due
to electronic overshoot of the center transducer. This belief is
strengthened by the fact that this discrepancy only occurred at the center
transducer for the highest velocity impacts.

In Figure 52, the steady flow pressure measured at the transducer

nearest the stagnation point is presented as a function of velocity.
Date for both 90' and 45' impact of porous RTV are compared with the
stagnation pressure predicted by Equation 27. The measurements for 900
impact (presented previously in Figure 43) compare very favorably with
the predictions. However, the data for 450 impact is less týan that
predicted. This was probably due to the fact that. the trensckdcer was
not located exactly at the stagnation point or that the region of full
stagnation pressure was smaller than the active aea of the gage.

93
AFML-TR-77-134

I EXPERIMENTAL DATA (45e)

O-UL4.m ADWE CErER
I 0-1.2?m ABOVE CENTER
a-CENTER U9N)ORL

145

125

0
0 So 0O 150 2OO 250 300
IMPACT VELOCITY, o4 (mrl)

Figure S0. Hugoniot Shock Pressures Masured at Various Points Along

the Target Surface D0ring the Oblique Impact (a a 45*) of
Porous RTV. (1 9N/rN' - 10 Bars).

100

z
"75 0 - 2.54 cm ABOVE CENTER
o90 (NORMAL)

*50

125
00 50I 1015010253o
IMPACT VELOCITY, u. (rn/a)

Figure 51. Hugoniot Shock Pressures Measured at the Initial Point of

Impact Ruring the Oblique Impact (a - 25*) of Porous RTV.
(1 MN/n a 10 Bars).

94
Anit-TR-77-134

40

aa

aX EQ.(27)

Itoa

EXPERIMNTAL DATA
ý)-9* CENTER
6-4tb "4 cm ASWdE CENTER
(NEAR STAGNATION)
0
0 50 100 150 200 250 300
IMPACT VELOCITY, ue (m/A)

Figure 52. Steady Flow Pressures Measured at or Near Stagnation Point for
Normal 2and Oblique (a - 450) Impact of Porous RTV.
N/m *10 Bars)
M(1

95
 AFI4L-TR-77-134

SECTION VI
K CONCLUSIONS
IBased on the experimental results, the following conclusions wiere
K drawn.
1. An unsteady, hydrod~ynamic model has been developed which
describes the impact behavior of low strength (soft body) projectiles
against a flat target.

2. This model has been verified experimentally for normal and

oblique impacts of porous and nonporous projectile materials at impact
velocities sufficient to cause large deformations or fluid like flow
of the projectile.

3. Based on this model and its subsequent verification, a bird

(chickens) can best be described as a low strength material with the
equation of state of water, an average density of 950 kg/in (,034 lbm/inA)
and 10% porosity.

96
 AFML-TR-77-134

REFERENCES

f 1. E. Hodgkinson, "On the Collision of Imperfectly Elastic Solids,"

ReReorrttofthe Fourth Meeting of the British Association of Science,

2. P. G. Tait, Properties of Matter, Adam and Charles Black,

Edinburgh, Scotland, 1885.
3. J. H. Vincent, "Experiments on Impact," Proceedings f the
FI Cambridge Philosophical Sociejt, Vol. 10, 1898, pp.332-37.
4. R. V. Raman, "The Photographic Study of Impact at Minimal
Velocities," Physical Review, Vol. 12, 1918, pp. 442-447.

5. A. W. R. Allcock and D. N. Collin, "The Development of a Dummy

Bird for Use In Bird Strike Research," National Ga• Turbine
Establishment, London, England, CP-1071, June 1968.

6. A. J. Tudor, "Bird Ingestion Research at Rolls-Royce," Symposium

on the Mechanical Reliajbil_._ nf Turbo-Machinery Blading, Derby
5strIct
bad College of Technology, April, 1968.

7. S. W. Tsai, C. T. Sun, A. K. Hopkins, H. T. Hahn, and T. W. Lee,

Behavior of Cantilevered Beams Under Impact by a Soft Projectile,
Air Force Materials Laboratory, Wright-Patterson Air Force Base,
Ohio, AFML-TR-74-94, Nov. 1974.

8. G. A. MacCauley, "Some Thoughts on the Forces Associated with

Bird Strike," Associate Committee on Bird Hazards to Aircraft,
Ottawa, Canada, Field Note No. 24, Jan. 1965.
9. J. Mitchell, "More Thoughts on Forces Associated with a Bird
Strike," Associate Committee on Bird ht.:vds to Aircraft, Ottawa,
Canada, Field Note No. 37, March 1966.

10. W. Goldsmith, Impact, Edward Arnold Ltd., London, England, 1960.

11. H. Kolsky, Stress Waves in Solids, The Clarendon Press, Oxford,
England, 1953.-
12. _Stength of Materials, Edward Arnold Ltd.,
W. Johnson, Impact
London, England, 1972,
13. H. G. Hopkins, and H. Kolsky, "Mechanics of Hypervelocity Impact
of Solids," Proceedings of the Fourth Symposium on Uperveloctty
Impac, Air Proving Ground"fenter-, Egin Air Force Base, Florida,
T9HU, paper no. 12.

14. J. P. Barber, (Private Communication).

I 97
AFIML-TR-77-134

REFERENCES (CONTINUED)
15. L. Pochhammer, "Uber die Fortpflanzungsgeshwindigkeitenkleiner
Schwingungen in Einem Unbegrenzten Isotropen Kreiszy1inder,"
Journal fur die Reine und Angewandte Mathematik, Vol. 81, 1876,
pp. 32-36
16. C. Chrem, "The Equation of an Isotropic-Elastic Solid in Polar
and Cylindrical Coordinates, Their Solution and Application,"
Transactions of the Cambridoa Philosophical Socjetp, Vol. 14,
1889, p. 250.

17. A. E. H. Love, The Mathematical Theory. of Elasticity, Cambridge

University Press, England, 1906, p. 551.
18. J. W. S. Rayleigh, Theory of Sound, 1 and 2, Dover Publications,
New York City, New or Ei5 7.
19. M. Redwood, Mechanical Waveguides, Pergamon Press, London, England,
1960, p. 300.
20. H. D. Conway, and M. Jakubowski, "Axial Impact of Short
Cylindrical Bars," Journal of Applied Mechanics, Vol. 36, 1969,
p. 809.
21. L. H. Donnell, 'Longitudlnal Wave Transmission and Impact,"
Transactions of V-he American Society for Mechanical Engineen ,
Vol. 52, 1930 p.T5-.T
22. G. I. Tayior, "Testing of Materials at High rates of Loading,"
Journal of the Institution of Civil Eninaers Vol. 26, No. 8,
October,-94,7 pp. 486-519.
23. K. A. Rakhmatulin, "Propagation of a Wave of Unloading,"
Prika~dnaya Mathematika i Mekhanika, Vol. 9, No. 1, 1945,
pp. 91-100.
24. Th. Karman, and P. Duwez, "The Propagation of Plastic Oeformation
in Solids," Journal of Applied Physics, Vol. 21, 1950, pp. 987-994.
25. G. 1. Taylor, "The Use of Flat Ended Projectiles for Determining
Dynamic Yield Stress. Part I, Theoretical Considerations,"
Proceedings of the Ral.Society of London, Series A, Vol. 194,
198, pp. 289-299.
26. A. G. W•liffen, "The Use of Flat Ended Projectiles for Determining
Dynamic Yield Stress. Part 1I, Tests on Various Metallic
Materials," Proceedings of the R1gal Society of London, Series A.
Vol. 194, 1948, pp. 300-322.

98
 AFML-TR-77-134

REFERENCES (CONTINUED)
27. J. B. Hawkyard, "Mushrooming of Flat Ended Projectiles Impinging on
a Flat Rigid Anvil," International Journal of Mechanical Sciences,
Vol. 11, 1969, pp. 313-333.

28. G. I. Barenblatt and A. 1. Ishlinskii, "On the Impact of a

Viscoplastic Bar on a Rigid Wall," Prikladntya Matematika i
Mekhanika, Vol. 26, 1962, pp. 497-502.
29. T. C. T. Ting, "Impact of a Nonlinear Viscoplastic Rod on a Rigid
Wall," Transactions of the ASME, Vol. 33, Sept. 1966, pp. 505-513.
30. B. Hopkinson, "A Method of Measuring the Pressure Produced in the
Detonation of High Explosive or by the Impact of Bullets,"
23194, pp. Transactions
Philosophical 437-466. of the Royal Society of London, A, Volo

31. G. Birchoff, D. P. MacDougall, E. M. Pugh, and G. I. Taylor,

"Explosives with Lines Cavities," Journal of Applied Physics,
Vol. 19, No. 6, June 1948, pp. 563-5-P.
32. B. Chakrapani and J. L. Rand, "An Analytical and Experimental
Study of the Behavior oF Semi-infinite Metal Targets Under
Hypervelocity Impact," Texas Engineering Experiment Station,
Texas A & M University, College Station, Texas, TEES-9075-CR-71-02,
Oct. 1971.
33. F. J. Heymann, "A Survey of Clues to the Relationship Between
Erosion Rate and Impact Parameters," Proceedings of the Second
Meersburl Conference on Rain Erosion and Allied Phenomena,
Fo'aTAircraft Establis - ,Vol.,-'A'u-9.-9-, pp. 683-740.
34. S. S. Cook, "Erosion by Water-Hammer," Proceedings of the Royal
Society of London, Series A, Vol. 119, 1928, pp. 481-48
35. F. J. Heymann, "On the Shock Wave Velocity and Impact Pressure
in High-Speed Liquid-Solid Impact," Transations of the ASME,
(Journal of Basic Engineering), Series 6 o'. 90, TuTy 1968,
• • p. 400.

36. F. P. Bowden and J. H. Brunton, "The Deformation of Solids by

Liquid Impact at Supersonic Speeds," Proceedings of the Royal
Society . Londo.n, Series A, Vol. 263, Oct. 1961, pp. 433-450.

37. F. P. Bowden and J. E. Field, "The Brittle Fracture of Solids by

Liquid Impact, by Solid Impact, and by Shock," Proceedings of the
Royal Society of London, Series A, Vol. 263, l9W",-2p.'- -
38. J. H. Brunton, "The Physics of Impact and Deformation: Single
Impact," Philosophical Transactions of the Royal Society of
London, Ano _260, JuTy'- .- p. 73-437.

99
AFkL-TR- 77 -134

REFERENCES (CONTINUED)
39. R. Kinslow, 0. G. Smith, and V. Sahai, "High-Velocity Liquid
Impact Damage," Tennessee Technological University, Cookeville,
Tennessee, TTU-ES-74-1, Jan. 1974.
40. J. E. Field. J. J. Calus, D. A. Gorham, and K. G. Rickerby,
"Impacc Damng Produced by Large Water Drops," Proceedings of the
Fourjt (nrrational Conference M Rain Erosion and Associated
Phlknomnia-;'Rues Schloss, Neer&sbu,0g7, Msa Geran4TWy, ay174,T
Vol. i, pp. 395-420.
41. 0.
Journal of e"Waterdrop
G. Engel, Collisions
of the National With
Bureaq of Surfaces,"
Solid Standards, Vol. 54,
5,pp 1o5
42. P. C. Jenkins and J. D. Booker, "Photographic Study of the Impact
Betweor Water Drops and a Surface Moving at High Speed," Royal
Aircraft Establishment, Fernborough, England, TN-ME 275, Nov. 1958.
43. R. Skalak and 0. Felt, "Impact on the Surface of a Compressible
Liquid," Transactions of the ASME, (Journal of Engineering for
Industry)--.o.88;17, 519;6,_ pp. 325-33
44. F. J. Heymann, "High-Speed Impact Between a Liquid Drop and a
Solid Surface," Journal of Appied_ Physics, Vol. 40, No. 13,
Dec. 1969, pp. 511T-5122-.
45. W. Johnson and G. W. Vickers, "Transient Stress Distribution
Caused by Water-Jet Impact," Journal of Mechanical EngineeringL
Science, Vol. 15, No. 4.4973.
46. M. C. Rochester and J. H. Brunton, "Surface Pressure Distribution
During Drop Impingement," Proceedings of the Fourth International
Conference on Rain Erosion and ssoc•aEted nomena, Neues Schloss,
?aTOW
Meersburg, West Germianylk o. pp '71-93
47. A. A. Fyall, "Single Impact Studies of Rain Erosion," Shell
Aviation News, Vol. 374, 1969.
48. J. H. Brunton and J. J. Camus, "The Flow of a Liquid Drop During
Impact," Proceedinqs of the Third International Conference on
Rain Erosion and Associated eIi-inom'na, luthem Hall, Englan,
Sg TM. pp.- 27-3 2.
49. A. A. Fyall, 'Single Impact Studies with Liquids and Solids,"
Proceedinas of the Second Mftersburg Conference on Rain Erosion and
pp.--66 Phenona,
"Allied -586. RoyalA- ircraf Establishment,-VoF-T, Aug. 19•7,

50. 1. I. McHaughtan, S. W. Chisman, and J. D. Booker, Hail I

Studies, Unpublished Report, Engineering Physics Department, Royal

1
Tr-c-rat Establishment.

100
 AFML-TR-77-134

REFERENCES (CONTINUED)
51. Y. C. Huang, F. G. Hammitt, and W. J. Yang, "Hydrodynamic
Phenomena During High-Speed Collision Bet wee Liquid Droplet and
Rigid Plane," Transactions of the ASNE, Jgurnal of Fluids
Engineering), Vol. 94, June973, pp. 27b-294.
52. L. A. Glenn, "On the Dynamics of Hypervelocity Liquid Jet Impact
on a Flat Rigid Surface," Jotkrel of Alted 1lMathematics and
Physics, (ZAMP), Vol. 25, 1974, pp. 383-398.
53. N. B. Lesser and J. E. Field, "The Fluid Mechanics of Compressible

54.
IeRs-'r9,
Liquid Impact," Proceeding
Germany, %
fh Fourth International Conference
on Rain Erosion and Associated Phenomena, Neues Schloss,
,o
Vol9 pp. 235-269.
J. W. Prichett and T. D. Riney, "Analysis of Dynamic Stresses
Imposed on Rocks by Water Jet Impact, eco International
Symposium on Jet Cutting Technolov, St7John's College, Cambridge,
England, ril 1974, Section B2, pp. 15-36.
55. L. N. Milne-Thompson, Theoretical Hydrodynamics, MacMillan
Company, New York, 1960.

56. R. B. Banks and D. V. Chandrasekhara, "Experimental Investigation

of the Penetration of a High-Velocity Gas Jet Through a Liquid
A, Surface," Journal of Fluid Mechanics, Vol. 15, Jan. 1963, pp. 13-16.

57. A. H. Gibson, Hydraulics and Its Applications, D. Van Nostran

Company, Inc. ,New ork, Md0.
58. S. J. Leach and G. L. Walker, "The Application of High Speed
Liquid Jets to Cutting," Philosophical Transactions of the Royal
Society of London, A, Vol. 260, July 1966, pp. 295-109.-

59. G. I. Taylor, "Oblique Impact of a Jet on a Plane Surface,"

Philosophical
Vol. 250, July Transactions
1966, p.9-I0. of the Roal Society of London, A,

60. V. N. Cogolev, V. G. Mirkin, and G. J. Yablokova, "Approximate

Equation of State for Solids," Zhurnal Prikladnoi Mekhaniki i
Teknickeskoi Fiziki, No. 5, 1963, pp. 9T-3'.T

61. A. L. Ruoff, "Linear Shock-Velocity-Particle-Velocity Relationship,"

Journal of Applied Physics, Vol. 38, No. 13, Dec. 1967,
pp -4980o.
62. M. H. Rice, R. G. McQueen, and J. M. Walsh, "Compression of
Solids by Strong Shock Waves", Solid State Physics (Sietz and
Turnball, eds.), Academic Press,--c.7--w York, ol. 6, 1958.
pp. 1-63.

101
AFML-TR-7"7-134

REFERENCES (CONCLUDED)
63. G. E. Duvall and G. R. Fowles, "Shock Waves", igh Pressure
Physics and Chemistry, Vol. 2, Ed. R. S. Bradley, Academic
Press, Inc., London, 1963, pp. 209-291.

64. P. J. Torvik, "A Simple Theory for Shock Propagation in Homogeneous

Mixture % Air Force Institute of Technology, Wright-Patterson Air
Force Base, Ohio, AFIT-TR-70-3, MY 1970.

65. J. P. Barber, H. R. Taylor, and J. S. Wilbeck, Characterization of

lirg I pcts o2 a Rigld Plate: Part 1,Air Force Flight Dynamics
Laboratory, Wright-Patterson Air Force Base, Ohio, AFFDL-TR-75-5,
Jan. 1975.
66. R. L. Peterson and J. P. Barber, Bird Impact Forces in Aircraft
Vtndshield beign, Air Force Flight Dynamics Laboratory, Wright-
Patterson Air Force Base, Ohio, AFFDL-TR-75-150, March 1976.
67. F. P. Bowden, and J. H. Brunton, 'Damage to Solids by Liquid
Impact at Supersokitc Speeds," Nature, Vol. 181, 1958, pp. 873-875.
68. W. E. Deal, "Shock Hugonoit of Air," Journal of ADDlied Physics,
Vol. 28, No. 7, July 1957, pp. 782-78T.

102
AFI4L-TR-77-l 34

APPENDIX A
RADIAL PRESSURE DISTRIBUTION

The pressure distribution along tha impact surface during steady

flow may be approximated by Equation 26
P a Ps exp (-C (M.)2) (A-1)
Ia

or by Equation 27
P 3P. (1 - 3 (-Z)2 + 2(-E-) 3) (A-2)

These two expressions were developed by Banks and Chandrasekhara

(Reference 56) and Leach and Walker (Reference 58) respectively. Both
relations imust satisfy momentum considerations, given here by Equation 2Z

The~, tth srac ~ foce (A-3)

atthesuraceis simply the integrated pressure, such that
TheForc,
F a J~dA (A-'4)

For a cylindrical projectile, Equations A-3 and A-4 may be combined to

obtainj
wpa u 2 2wf; rdr~ AS
0 0o(

The constant C,may be obtained by substituting Equation A-i into

Equation A-5 such that

o u0 2 2PJ e/dz (A-B)

Solving this expression gives

P
PU T (A-7)

103
AFML-TR-77-134

The constant c2 may be obtained by substituting Eauatior. A-2 ik.

Equation A-5
a "{.l( ) 2 (..))z}4r (A-S)
IU 2 P. (1- 4

Solving this expression gives

S2 " (A-9)

104
AF-L-TR-77- 134

APPENDIX 5
GENERAL PRESSURE - DENSITY RELA(IONSHIPS

In this section, the simple mixture theory given by Torvik

(Refrenmce 64) is used to develop the pressure-density relationships for
a porous material under both itentropic and shock compression. Torvik's
theory predicted the pressure-dinsity relationship acrost a shock for a
homageneous mixture based on the Hugoniot relationships of the constituents.
In this theory several fundamental assumptions are made as follows:

1. The mixture is macroscopically homogeneous and isotropic.

2. The density of the mixture is the sum of the products of the
volume fraction and density of each component.
3. The density of each component is the same as that predicted
for a homogeneous sample tf the individual material at the same
pressure.
4. The size of individual particles of each component is su:h that
the shock pressure is the same in each component.

5. The constituents do not react or change phase during the

compression process.

From the second assumption it is clear that

K fV.
•AVG = £ - (BA)

where pAVS is the average density of the mixture, fv, is the volume
fraction of the ith component, pi is the density of the ith component,
and N is the number of co enents. Rewriting Equation 8-1 in term of
the mass fraction
Z (B-2)
PAVG 1=L P:i

whe fm i is the mass fraction of the ith component.

105
 !I
I

AFML-Th-77-134

Assuming that the mass fraction of the components is unchanged

after the mteril compression, it is seen that

N __ I
PIAVG 1 P2L

dkAVG z S

where the subscript I refers to the Initial (unstressed) state and 2 the
final (stressed) state. Equation B-3 may be simplifitA' by realizing that

P21 Pli 21

and

Plk

Thus, Equation 8-3 can be rewritten

P1AVG' N
--- Z fv,
9 2AVG i=1

Although Equation 8-6 was derived by Torvik for density changes across a
shock, it should hold true for any compression process in which all the
assumptions are satisfied and the pressure-density relationships are
known for each constituent of the mixture. Therefore, Equation 8-6
will be used to derive the pressure-density relationships for both
shock and isentropic compression.

106
AFML-TR-77-134

1. ISENTROPIC
The pressure-density riatlonship of the soft body materlal will be
represented by Equation 13

P a A ((-)
P2
-1)

Assuming the material can be represented by a linear Hugoniot (Equation 48)

Uuaa •o
a ÷* ku
(p

the constants can be expressed by Equation 54

A = P1 Co2 /(4k-1)

B = 4k-1

Equation 53 may be rearranged to obtain

P2 = I (B-7)

The pressure-density relationship of air for an isentropic

compression may be approximated by the relationship for a perfect gas

P2 ( 2 1Y(B-8)

where y is the ratio of specific heats (y - 1.4 for air). Thus, sub-
stituting Equations 8-7 and A-6 into Equation B-6, the pressure-density
relationship under isentropic compression for a porous material is seen
to be
P1AVG /(P
P B+ /
P2AVG 9 *

where subscript m represents the soft body material and n represents the
air.

107 4
AFML-TR-77-134

2. SHOCK
As stated in the section on equations of state, for pressures in
the range of interest, the pressure-density relationships under lsen-
tropic and shock compression are approximately equal. Thus, Equation B-7
will be used to represent the shock pressure-density relationship for
soft body materials.

For air the shock relation is significantly different from the

isentropic relation. The shock relation for a perfect gas is

I + If
"!2 ey-1 -I
P y+1 + ?2
Y-1 PI

This equation is plotted in Figure 53 along with experimental values for

real air obtained by Deal (Reference 68). As can be seen from the
figure, the density ratio for a perfect gas approaches a fixed value
(6.0) as the pressure increases. This is not true with the experimental
data for air.

In order to develop a more realistic equation of state for air,

the data of Deal's work was studied in more detail. Figure 54 is a
plot of the shock velocity versus particle velocity for air. From this
data, it can be seen that air can be represented by the linear Hugoniot
relationship

u aCs o +ku p

4here k - 1.03. In the section on Equations of State it was noted that

materials which could be represented by the linear Hugoniot relation
could also be represented by the pressure-density relationships given
in Equations 52 or 53. Upon examination, air was best represented by
Equation 52 which has the form

P2 p1 co 2 q/(l-kq)
0 2

108
 I AFlhL-TR-77.
13II
4

i•9
.o.0. ...-
r 0,

-
-

S-0

0o0o DEAL (Referente

WI

S- " EQ. (8-10)

0 EQ. (811
50 100 150 200 250
PRESSURE RATIO,
Pw /Po
Figure 53. Shock Hugoniot for
Air.

03

i I•" SLOPE 1.03

I i
o~00 DECAL (R eferencet
S~0
0 1
SPARTICLE 3 4
VELOCITY, uo, (IOSM/4)
'109 5.
Figue Epertnta Vaues Of Shock Velocity3
vs. Particle Velocity.
AFNL-TR-77-134

A plot of Equation S-11 in Figure 53 was seen to fit the experimental date
very well.

Rearranging Equation B-11 into a fore compatible with Equation B-6

give s (1-kkq)2(5-12)
S~which
whc ivs(-u) Po2 q - 0 (5-13)

Solving by the quadratic equation, it is seen that

q q11 ± q 2 (B-14)

where 2Pk +÷ P

2Fk2

and 2
q2 " {(2Fk + -2 2
0l22 2
2 P0

It can be noted from the definition of q that

In order for to be positive, q must be less then unity. For this to

be true, Equation 5-14 must take the form

q q "q 2 (O-1 ?)

110
AFNL-TR-77-1 34

Thus* the final expression for the shook pressure-density relationship

of air is given by Equation 6-16 where q is given by Equation 8-17 and
q, and qZ by Equation B-16.

SubstitUtion of Equations 87 and 8-16 into Equation'A-4 gives the

pressure-density relationship under shock compiresSion fora poriusL
Material of the form

P2AV( A

where subscript m represents the soft body material and n represents

real air, and q is defined by Equations BI1 and B-lS.

I
I
1
I

II
i

I
I

1!I

11,1
AFM4L-TR-77-134

APPENDIX C
MATERIAL PROPERTIES
The properties of the various materials studied in this work are
presented In this section.

1. WATER
For water, Heymann (Reference 35) showed that the linear Hugoniot
takes the form
U C +2.0Ju

where c 0 is the normal sound speed in water. This relation was shown
to provide a quite accurate fit to the exact data for water up to
Mach - 1.2. The density and wave speed for water at a temperature of
200C and pressure of 1 atmosphere (Reference 62) are

P = 998.23 kg/m3 (.0361 Wb/in )

C a 1482.9 m/s (4865 £t/s)

2. GELATIN
For this program, commercial gelatin with a strength of 250 Bloom
was tested. Historically, gelatin has been used to simulate flesh in
bullet impact tests. For this reason, and because of its low cost and
ease of fabrication, this material is currently used by several aircraft
engine blade manufacturers to simulate birds in blade impact tests.
During this test program, it exhibited sufficient shear strength to
withstand the high acceleration loads attained during projectile launch

in the gun range. However, difficulty in launch did occur above velo-
cities of 250 rn/s. Porosity was attained by the addition of phenolic
micro-balloons.

oecause of its unlikely involvement in impact, apparently no

compression data has been generated for gelatin. However, the assumption
made by previous researchers has been to approximate gelatin with the

112
AFNL.-TR-77-134

equation of state of water. The only major dr*wback to this ts the fact
that the normal equation of state of water does not incorporate shear
strength. The linear Hugoniot is taken to be that of water

U8 °o + 2.0u%
with

Of 1000kg/u3 (.0383 Ub/in 3 )

U 1482.9 m/a (4865 Wt/e)

The micro-balloon gelatin was treated as simply a mixture of gelatin and

air. For these tests, the material had the properties

3= .40

P5 64 i~ 2
k'.o (.0231, 1lbin3

3. RTV RUBBER
RTV (Room Temperature Vulcanized) rubber is a silicone elastomer
which is presently used by several organizations as a bird substitute
material, In previous work conducted at the AFNL Impact Facility, it
was determined that the linear Hugoniot for RTV-11 could be approximated
by the relationship
u % 4o+3.62 u

For the current study, a mixture of RTV-560 and RTV-921 was tested.
Porosity wasobtainted by the use of a blowing agent. Since no Hugonlot
date existed for this material, the relationship for RTV-11 was used.
It was understood that use of this relationshitp eight cause som
inaccuracies in the resulting calculations. The original 560/921
material had a density end wave speed of

Of a 1,330 kg/s3 (.0481 lb/in3 )

00 a 630 m/a (2743 ft/#)

113
 AFML-TR-77-134

With the blowing agent added, the porous RTV had the properties

='670 k/rM3 (.0242 lb/in 3)

4. BIRDS
All bird tests conducted in this program used chickens as the
projectiles. However, many researchers feel that chickens are not
representative of flying birds. The major point of contention is that
flying birds have much lower densities than chickens. If this is the
case, then the equation of state derived in this section for chickens

I will not be adequate for other birds, although the only difference will
be the amount oi. assumed porosity.
A large number of tests in which baby chickens have been impacted

against a rigid plate have been conducted in the Impact Mechanics

Facility over the past several years. This work was done in conjunction
with FOD (Foreign Object Damage) studies of aircraft engine blading

and aircraft windshields. Similar testing with large chickens has been
conducted at the Arnold Engineering Development Center. During these
studies, the birds have been accelerated to velocities exceeding 300 m/s.

Although birds are inhomogeneous, anisotropic, and not axisymmetric

it was felt that to some degree, they could be approximated by a cylinder
of material which was both homogeneous and isotropic. Thus, the
assumption made in this program is that they may be approximated by
a cylinder of porous water. Here the shear stwength of the bird has
been neglected.
Thus, the linear Hugoniot is taken to be that of water

u c + 2.0u

with

c 1482.9 m/s (4865 ft/s)

114
AFML-TR-77-134

The density measured for several chickens of various sizes was found
to be
0 a 950 kg/u3 (.0342 lb/n 3)

Several researchers, including Tudor, et al. (Reference 6) have measured

the density of various parts of the chicken to be approximately
3
p - 1,060 kg/m . If this is actually the case, then chickens can be
considered to have a full density of Pf - 1,060 kg/m 3 with 10% porosity
3
(Pz a 950 kg/m ). However, many researchers feel that chickens can
be better approximated by a material with an original density of
Pf - 950 kg/mr3 with no porosity. Both of these models will be checked
against experimental results.

I
I

*J,.m.6sesunment •gnIntlt Ommlw a1,a -- l.10IN