Computational Fluid Dynamics Flow Field Solutions for a

Kinetic Energy (KE) Projectile With Sabot

by Karen R. Heavey, James DeSpirito, and Jubaraj Sahu

ARL-MR-572 September 2003

Approved for public release; distribution is unlimited.

 NOTICES

Disclaimers

The findings in this report are not to be construed as an official Department of the Army position unless
so designated by other authorized documents.

Citation of manufacturer’s or trade names does not constitute an official endorsement or approval of the
use thereof.

Destroy this report when it is no longer needed. Do not return it to the originator.
Army Research Laboratory
Aberdeen Proving Ground, MD 21005-5066

ARL-MR-572 September 2003

Computational Fluid Dynamics Flow Field Solutions for a

Kinetic Energy (KE) Projectile With Sabot

Karen R. Heavey, James DeSpirito, and Jubaraj Sahu

Weapons and Materials Research Directorate, ARL

Approved for public release; distribution is unlimited.

 Report Documentation Page Form Approved
OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering
and maintaining the data needed, and completing and reviewing the collection information. Send comments regarding this burden estimate or any other aspect of this collection of
information, including suggestions for reducing the burden, to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188),
1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty
for failing to comply with a collection of information if it does not display a currently valid OMB control number.
PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS.
1. REPORT DATE (DD-MM-YYYY) 2. REPORT TYPE 3. DATES COVERED (From - To)
September 2003 Final January 2002 – May 2003
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
Computational Fluid Dynamics Flow Field Solutions for a Kinetic Energy (KE)
Projectile With Sabot 5b. GRANT NUMBER

5c. PROGRAM ELEMENT NUMBER

6. AUTHOR(S) 5d. PROJECT NUMBER

Karen R. Heavey, James DeSpirito, and Jubaraj Sahu 1L162618AH80
5e. TASK NUMBER

5f. WORK UNIT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION

REPORT NUMBER
U.S. Army Research Laboratory
ATTN: AMSRL-WM-BC ARL-MR-572
Aberdeen Proving Ground, MD 21005-5066
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR'S ACRONYM(S)

11. SPONSOR/MONITOR'S REPORT

NUMBER(S)

12. DISTRIBUTION/AVAILABILITY STATEMENT

Approved for public release; distribution is unlimited.

13. SUPPLEMENTARY NOTES

14. ABSTRACT
A study was undertaken to investigate and analyze the flow field results produced by various computational solvers for a
projectile of interest to the U.S. Army. Computational fluid dynamics (CFD) techniques were used to obtain numerical
solutions for the flow field of a kinetic energy projectile with the original and a modified (experimental) sabot. Computed
results were obtained at Mach 4.5 and a 0° angle of attack. Qualitative flow field features showed the pressure on the surface
of the model as well as pressures in the flow field. The surface pressure data on the projectile were extracted from the
solution files and compared. In all cases, the results were comparable. These results show the predictive capabilities of CFD
techniques in the analysis of supersonic flow over projectiles with sabots. They also provide an insight into the software
capabilities of several of the many tools available to research scientists in the field of CFD.

15. SUBJECT TERMS

computational fluid dynamics (CFD), projectile aerodynamics, projectile with sabot, structured and unstructured grids
17. LIMITATION 18. NUMBER 19a. NAME OF RESPONSIBLE PERSON
16. SECURITY CLASSIFICATION OF: OF ABSTRACT OF PAGES
Karen R. Heavey
a. REPORT b. ABSTRACT c. THIS PAGE 19b. TELEPHONE NUMBER (Include area code)
UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED UL 30 410-278-2916
Standard Form 298 (Rev. 8/98)
Prescribed by ANSI Std. Z39.18

ii
Contents

List of Figures iv

List of Tables v

Acknowledgments vi

1. Introduction 1

2. Flow Solvers 1
2.1 ZNSFLOW Code...............................................................................................................1
2.2 CFD++ Code .....................................................................................................................2
2.3 FLUENT Code ..................................................................................................................3

3. Projectile and Model Geometry 4

3.1 Projectile System ...............................................................................................................4
3.2 Computational Models ......................................................................................................4

4. Computational Meshes 5
4.1 ZNSFLOW Grids ..............................................................................................................5
4.2 CFD++ Grids.....................................................................................................................6
4.3 FLUENT Grids..................................................................................................................7

5. Results 8
5.1 Original Sabot....................................................................................................................8
5.2 Modified Sabot ................................................................................................................12
5.3 Force Comparison ...........................................................................................................16
5.4 Solver Differences ...........................................................................................................19

6. Conclusion 20

7. References 21

iii
List of Figures

Figure 1. Drawing of an M829A2 round. .......................................................................................4

Figure 2. Schematic of computational model. ................................................................................4
Figure 3. Computational model of original sabot. ..........................................................................5
Figure 4. Expanded view of sabot cup: original (left) and modified (right)..................................5
Figure 5. The 2-D structured mesh for original sabot.....................................................................6
Figure 6. Computational mesh for 3-D model with modified sabot. ..............................................7
Figure 7. Mach contours for 2-D ZNSFLOW. ...............................................................................9
Figure 8. Mach contours for 2-D CFD++. ......................................................................................9
Figure 9. Mach contours for 2-D FLUENT. .................................................................................10
Figure 10. Pressure contours for 2-D ZNSFLOW. .......................................................................10
Figure 11. Pressure contours for 2-D CFD++...............................................................................11
Figure 12. Pressure contours for 2-D FLUENT............................................................................11
Figure 13. Comparison of Cp along centerline for 2-D cases.......................................................12
Figure 14. Mach contours for 3-D ZNSFLOW. ...........................................................................13
Figure 15. Mach contours for 3-D CFD++. ..................................................................................13
Figure 16. Mach contours for 3-D FLUENT. ...............................................................................14
Figure 17. Pressure contours for 3-D ZNSFLOW. .......................................................................14
Figure 18. Pressure contours for CFD++......................................................................................15
Figure 19. Pressure contours for 3-D FLUENT............................................................................15
Figure 20. Surface pressure for 3-D CFD++. ...............................................................................16
Figure 21. Surface pressure for 3-D ZNSFLOW..........................................................................17
Figure 22. Surface pressure for 3-D FLUENT. ............................................................................17
Figure 23. Comparison of Cp along the projectile centerline for 3-D cases.................................18
Figure 24. Comparison of Cp along 60° plane for modified sabot...............................................18
Figure 25. Comparison of centerline Cp for original and modified sabot, CFD++ solutions. .....19

iv
List of Tables

Table 1. Dimensions for structured grids........................................................................................6

Table 2. Elements for unstructured grids........................................................................................7
Table 3. Axial forces.....................................................................................................................19

v
Acknowledgments

The authors would like to thank Harris Edge, U.S. Army Research Laboratory (ARL) Weapons
and Analysis Branch, Aberdeen Proving Ground, MD, for his contributions to this work. As a
former member of the computational fluid dynamics team, he provided the initial computational
mesh for this study.
This work was supported by a grant of computer time from the Department of Defense High
Performance Computing Major Shared Resource Center at ARL.

vi
1. Introduction

The use of computational fluid dynamics (CFD) in the areas of projectile design and
development is a continuously evolving field. Improved computer technology and
state-of-the-art numerical procedures enable scientists to develop solutions to complex,
three-dimensional (3-D) problems associated with projectile and missile systems of interest to
the U.S. Army. Many geometrically complex single-body and multibody systems have been
investigated using commercial as well as in-house CFD software packages (1–7). The focus of
this study is a kinetic energy (KE) projectile with an experimental sabot design. The
computational problem involves two-dimensional (2-D) and 3-D flow computations at a Mach
number of 4.5 and at a 0° angle of attack for both the original and the modified sabots. The sabot
petals separate from the projectile after the sabot leaves the gun barrel. So the results from this
study are only valid for an instant at muzzle exit. The original purpose of the study was to get a
quantitative estimate of the surface pressure values on the sabot at that instant. This report
describes the application of several CFD solvers to this problem.

2. Flow Solvers

2.1 ZNSFLOW Code

The ZNSFLOW solver is a product of a common high performance computing software support
initiative (CHSSI) project (8). It is a descendant of F3D, a code used successfully for many
years on Cray vector processors such as the C90. The complete set of 3-D time-dependent,
generalized-geometry, Reynolds-averaged, thin-layer Navier-Stokes equations is solved using
the finite volume method in general spatial coordinates using an implicit, approximately factored
scheme with an upwind scheme in the stream-wise direction and central differencing in the other
two directions. The complete set of 3-D, time-dependent, generalized-geometry,
Reynolds-averaged, thin-layer Navier-Stokes equations is solved numerically to obtain a solution
to this problem and can be written in general spatial coordinates ξ, η, and ζ as follows:

∂ τ qˆ + ∂ ξ Fˆ + ∂ η Gˆ + ∂ ζ Hˆ = Re −1 ∂ ζ Sˆ , (1)

where

• ξ = ξ(x, y, z, t)—longitudinal coordinate,

• η = η(x, y, z, t)—circumferential coordinate,

1
 • ζ = ζ(x, y, z, t)—nearly normal coordinate, and

• τ = t—time.
Under CHSSI, the code was rewritten to provide scalable performance on a number of computer
architectures. Programming enhancements include the use of dynamic memory allocation and
highly optimized cache management. ZNSFLOW features a graphical user interface to facilitate
problem setup. The flow solver can be used for problems using structured zonal grids as well as
Chimera overset grids for complex geometrical configurations. For turbulent flows, the
algebraic eddy viscosity turbulence model developed by Baldwin and Lomax (9), a one-equation
Rt model and a two-equation k-ε model are available. For the original sabot, the k-ε model was
used, while the Baldwin-Lomax model was used for the modified sabot. (The reason for this will
be discussed later.) Second-order spatial discretization was used in both cases.

2.2 CFD++ Code

CFD++ (10, 11) is a commercially available code that can be used for either steady-state or time-
accurate unsteady CFD problems. The basic numerical framework of this code is one that
implements a unified grid, unified physics, and unified-computing approach. The 3-D, time-
dependent Reynolds-averaged Navier-Stokes (RANS) equations are solved using the finite
volume method:
∂
WdV + ∫ [F − G ]⋅ dA = ∫ HdV ,
∂t V∫
(2)
V

where W is the vector of conservative variables, F and G are the inviscid and viscous flux
vectors, respectively, H is the vector of source terms, V is the cell volume, and A is the surface
area of the cell face. Second-order spatial discretization was used.

The numerical framework of CFD++ is based on the following general elements:

• unsteady compressible and incompressible Navier-Stokes equations with turbulence
modeling (unified-physics),
• unification of Cartesian, structured curvilinear, and unstructured grids, including hybrids
(unified-grid),
• unification of treatment of various cell shapes including hexahedral, tetrahedral, and
triangular prism cells (3-D), quadrilateral and triangular cells (2-D), and linear elements
(one-dimensional [1-D]) (unified-grid),
• treatment of multiblock patched aligned (nodally connected), patched-nonaligned and
overset grids (unified-grid),
• total variation diminishing discretization based on a new multidimensional interpolation
framework,

2
 • Riemann solvers to provide proper signal propagation physics, including versions for
preconditioned forms of the governing equations,
• consistent and accurate discretization of viscous terms using the same multidimensional
polynomial framework,

• point-wise turbulence models that do not require knowledge of distance to walls,

• versatile boundary condition implementation that includes a rich variety of integrated
boundary condition types for the various sets of equations, and
• implementation on massively parallel computers based on the distributed memory message
passing model using native message-passing libraries or message passing interface, parallel
virtual machine, etc. (unified-computing).

2.3 FLUENT Code

FLUENT (12) is another commercially available code that solves steady-state and time-accurate
CFD problems on unstructured grids. The FLUENT software suite has several solvers for both
incompressible and compressible flows in both an implicit and explicit numerical framework.
The implicit, compressible (coupled) CFD solver was used for this study. The 3-D,
time-dependent, RANS equations are solved using the finite volume method as defined in
equation 2.
The inviscid flux vector F is evaluated by a standard upwind flux-difference splitting. In the
implicit solver, each equation in the coupled set of governing equations is linearized implicitly
with respect to all dependent variables in the set, resulting in a block system of equations. A
block Gauss-Seidel, point implicit linear equation solver is used with an algebraic multigrid
method to solve the resultant block system of equations. The coupled set of governing equations
is discretized in time and time marching proceeds until a steady state solution is reached. In the
implicit scheme, which was used in this study, an Euler-implicit discretization in time is
combined with a Newton-type linearization of the fluxes. Second-order spatial discretization
was used.
Several turbulence models are available; the two-equation realized k-ε turbulence model was
used for these calculations. Due to the complexity of the model geometry, the solution was
started with first-order discretization. When the solution was stabilized, second-order
discretization was used for the final solution.

3
3. Projectile and Model Geometry

3.1 Projectile System

The projectile of interest for this study is the M829A2. It is a long-rod KE round that is one of a
family of weapons used with the M1 Abrams tank. Figure 1 is a drawing of the round, showing
the projectile, sabot, and cartridge.

Figure 1. Drawing of an M829A2 round.

3.2 Computational Models

The computational model consists of a 635-mm-long projectile with sabot. The diameter of the
projectile is ~24 mm. The sabot begins 75 mm from the tip of the projectile. For purposes of
this study, the section of the projectile after the sabot (i.e., the tail fin) was not modeled. The
schematic shown in Figure 2 depicts an outline of the model along the axis of symmetry. This
configuration represents the projectile for an instant at muzzle exit, prior to the removal of the
sabot petals via the aerodynamic forces.

Figure 2. Schematic of computational model.

4
Two computational models were used, one for the original sabot design and one for the modified
(experimental) sabot design. For the most part, the two models are quite similar; the one
difference is that the experimental model has an insert in the front cup of the sabot. Figure 3
shows the overall model for the original design. An expanded view of the sabot cup for both
the original and modified sabots is shown in Figure 4.

Figure 3. Computational model of original sabot.

Figure 4. Expanded view of sabot cup: original (left) and modified (right).

4. Computational Meshes

4.1 ZNSFLOW Grids

For the ZNSFLOW cases, two structured zonal grids were created. Because of the symmetry of
the original sabot, a 2-D mesh was suitable for the computation of the flow field. For the
experimental sabot, a 3-D mesh was required. In both cases, one large grid provides the mesh
around the projectile and the original sabot. For the experimental model, a single smaller grid
provides the mesh for the insert in the front cup of the sabot. Table 1 shows the specific details
for each of these configurations.

5
 Table 1. Dimensions for structured grids.

Original Sabot Modified Sabot Sabot Insert

Longitudinal 188 222 30
Circumferential 3 36 25
Radial 59 59 21
Total points 33,276 471,528 15,750

The computational meshes were created using the various tools found in the OVERGRID Grid
Generation software package (13). Figure 5 shows a single-plane 2-D mesh for the original
sabot. The boundary conditions consist of a collapsed axis at the nose of the projectile,
supersonic outflow at the rear of the mesh, no-slip wall conditions on the projectile surface, and
far field (pressure- and temperature-based inflow/outflow) at the outer edge.

Figure 5. The 2-D structured mesh for original sabot.

Figure 6 shows a cross-sectional view of the 3-D mesh for the modified sabot configuration.
Note that due to the periodicity of the modified design, it was only necessary to model a
60° wedge for the computation. In addition to the boundary conditions for the 2-D case,
symmetry plane conditions for the circumferential boundaries and isothermal wall conditions for
the surfaces of the sabot insert were included in the parameters of the solution.

4.2 CFD++ Grids

Using the grid tools available in the CFD++ software package, the structured grids created for
the ZNSFLOW cases were converted into unstructured grids for compatibility with the CFD++
solver. Physically, the mesh retained its hexagonal cell structure. Logically, it is now referenced
by cells, nodes, and faces instead of longitudinal, circumferential, and normal gridpoint indices.
Any zonal boundaries are merged so that the resulting mesh is one large entity. External
boundary conditions remain as before. Table 2 shows the size of the unstructured grids for both
the original 2-D and modified 3-D sabot designs.

6
 Figure 6. Computational mesh for 3-D model with modified sabot.

Table 2. Elements for unstructured grids.

Original Sabot Modified Sabot

Cells 10,730 461,854
Nodes 10,974 485,778
Faces 21,703 471,800

4.3 FLUENT Grids

As with the CFD++ solver, FLUENT requires the computational mesh to be unstructured.
However, it must also be in a FLUENT specific format, as there are no conversion tools
available in the software package. GRIDGEN (14) is a grid generation software package that has
the capability of reading and writing various file formats, including PLOT3-D and FLUENT. It
was used to convert the 2-D grid. However, due to a bug in the GRIDGEN code, it could not
correctly output the collapsed axis boundary in the 3-D grid to a FLUENT mesh file. The mesh
had to be slightly modified by patching the grid at the nose of the projectile. This eliminated the
need for a boundary condition specification along the symmetry axis in front of the projectile.
The model geometry was not changed, and the remainder of the grid retained its original
structure. The size of the grid only changed by 3366 cells (or <1%).

7
5. Results

Several computations were performed using the three flow solvers and the two different sabot
models described previously. Study parameters include Mach number of 4.5, 0° angle of attack,
and standard atmospheric flight conditions. For each solver, boundary conditions (as described
in the previous section) were selected to provide as reasonable a comparison as possible, given
the choices and constraints of each code. A modified two-equation k-ε turbulence model was
selected in each case. However, after over 100,000 iterations, the ZNSFLOW (3-D) modified
sabot case was not converged; the Baldwin-Lomax model was used instead for the final solution.
Initially, the free-stream was set to zero near the wall boundary for CFD++ and FLUENT; this
option is not available in ZNSFLOW. All computations were performed on the SGI suite
computers at the ARL Major Shared Resource Center (MSRC). Resources required for the 2-D
computations were minimal. For the 3-D calculations, resources used differed depending on the
solver used, but were also minimal.
Each of the CFD software packages used in this study has internal graphical capabilities for
viewing the flow field solution while the computation is in progress. Additionally, the files can
be reformatted and used with other available postprocessors. For all cases in this study, the grid
and solution files were reformatted and then imported into FIELDVIEW (15), a scientific
visualization software package.

5.1 Original Sabot

Contour plots in the symmetry plane are used to compare the 2-D solutions for the original sabot.
Figures 7 and 8 show the Mach contours to be quite similar, though the shock waves off the
sabot cup are more defined in the CFD++ solution (Figure 8). Figure 9 shows the same
characteristics, with slightly more definition of the flow field in front of the sabot cup.
Similarly, all three pressure contours (Figures 10–12) show similar flow characteristics. The
area of high pressure in front of the sabot cup and the shock wave at the edge of the cup are
visible in all three flow fields. The FLUENT solution (Figure 12) shows a more defined area of
low pressure in front of the sabot cup.
A comparison of the pressure coefficient (Cp) along the centerline of the projectile, Figure 13,
shows the solutions to be qualitatively similar. The differences in the curves directly correspond
to some of the differences in the flow field in the previous figures. The increase in Cp at ~0.06 m
coincides with the oblique shock wave formed at the end of the conical nose of the projectile.
This shock wave forms a small distance forward of the conical-cylindrical intersection in the
FLUENT case (Figures 9 and 12), while in the other cases it forms closer to the intersection.
Both ZNSFLOW and CFD++ show a sharp increase at ~0.1 m, where the flat face of the sabot
petal begins, while FLUENT shows a very small increase. CFD++ shows the largest increase

8
Figure 7. Mach contours for 2-D ZNSFLOW.

Figure 8. Mach contours for 2-D CFD++.

9
 Figure 9. Mach contours for 2-D FLUENT.

Figure 10. Pressure contours for 2-D ZNSFLOW.

10
Figure 11. Pressure contours for 2-D CFD++.

Figure 12. Pressure contours for 2-D FLUENT.

11
 Figure 13. Comparison of Cp along centerline for 2-D cases.

in pressure, 21% higher than the ZNSFLOW solution at ~0.11 m, while the FLUENT solution
gives the lowest peak pressure, 18% lower than the ZNSFLOW solution. All three solutions
show a qualitatively similar reduction in pressure as the flow expands past the first conical-
cylindrical section of the sabot at ~0.14 m. The largest expansion was in the FLUENT solution,
which gave a minimum pressure 31% lower than the (first) maximum. The reduction in pressure
was 24% and 15% in the ZNSFLOW and CFD++ solutions, respectively. This expansion is
more evident in the flow field plots for the FLUENT case (Figures 9 and 12) due to the plotting
scales, but it is present in all cases. The pressure in the sabot cup is qualitatively similar in all
three cases. The CFD++ and ZNSFLOW solutions show a large reduction in pressure and a
monotonically increasing pressure between the cup and the obturator, while the FLUENT
solution shows a low, constant pressure.

5.2 Modified Sabot

Multiplane contour plots are used to represent the 3-D solutions for the modified (experimental)
sabot. The Mach contours shown in Figures 14–16 show the solutions to be similar. However,
the contours in Figures 15 and 16 (CFD++ and FLUENT, respectively) are more clearly defined,
especially along the nose of the projectile. Figures 17–19 show pressure contours in the
symmetry planes of the flow fields. All three solutions show the shock at the edge of the sabot
cup, but the contours for the ZNSFLOW solution (Figure 17) are much less defined. The final
three figures show the surface pressure on the projectile model for each of the calculations. The

12
Figure 14. Mach contours for 3-D ZNSFLOW.

Figure 15. Mach contours for 3-D CFD++.

13
Figure 16. Mach contours for 3-D FLUENT.

Figure 17. Pressure contours for 3-D ZNSFLOW.

14
Figure 18. Pressure contours for CFD++.

Figure 19. Pressure contours for 3-D FLUENT.

15
pressure on the edge of the front cup is more prominent in Figure 20 (FLUENT) than in Figures
21 and 22. All three figures, however, show the same overall qualities of higher pressure in front
of the cup and lower pressure along the rest of the sabot.
As seen previously in the original sabot, the Cp values along the centerline of the projectile
(Figure 23) are quite similar for the CFD++ and FLUENT solutions. For the ZNSFLOW results,
there is quite a difference in the values in front of the sabot cup; the pressure is slightly higher at
the obturator as well. This is most likely due to the fact that a converged solution could not be
achieved for this case with the two-equation turbulence model. Similar trends appear along the
60° plane. As seen in Figure 24, all three solutions are similar, except in the area in front of the
insert. Figure 25 shows a comparison of the Cp values for the original and modified sabots,
using the CFD++ solver. The pressure on the modified sabot is slightly higher in the area in
front of the sabot cup (near 0.14 m) and ~20% higher in the cup itself, in the area between
0.14 m and 0.17 m. Otherwise, the two sabot models show nearly identical values.

5.3 Force Comparison

In addition to the qualitative comparisons of the flow fields, it is useful to compare some
quantitative variables. For ZNSFLOW, these variables must be extracted in the post-processing
procedure; CFD++ and FLUENT provide the option to have them calculated as the solution
progresses. Table 3 shows the axial force in Newtons for a 60° wedge of the actual projectile.
For the original sabot, the forces from all three solvers are only slightly different. For the
modified sabot cases, using the ZNSFLOW solution as the mean, FLUENT and CFD++ are
~10% lower and higher, respectively. The difference in the axial force follows directly from
differences in the predicted maximum pressures (Figure 13).

Figure 20. Surface pressure for 3-D CFD++.

16
Figure 21. Surface pressure for 3-D ZNSFLOW.

Figure 22. Surface pressure for 3-D FLUENT.

17
Figure 23. Comparison of Cp along the projectile centerline for 3-D cases.

Figure 24. Comparison of Cp along 60° plane for modified sabot.

18
 Figure 25. Comparison of centerline Cp for original and modified sabot, CFD++
solutions.

Table 3. Axial forces.

Original Sabot Modified Sabot

FLUENT 1777 N 1863 N
ZNSFLOW 1722 N 2011 N
CFD++ 1799 N 2161 N

5.4 Solver Differences

This model problem is a fairly difficult test case and all three solvers performed reasonably. It is
difficult to quantify for this case which solver gave the most accurate results because there is no
test data available for validation. In fact, the primary purpose of the study was to get an estimate
of the surface pressure on the sabot petals because they cannot be experimentally determined.
A statement on the performance of the solvers in getting the final answers can be made.
Although the details provided are specific to the 3-D solutions, the comments in general also
apply to the 2-D cases.
CFD++ was the most robust and obtained a final solution in the least time and least number of
iterations. It took ~2000 iterations to converge the axial force to a constant value, and the
residuals dropped about three orders of magnitude.
ZNSFLOW took over 20,000 iterations to achieve a converged solution, with many iterations
used to get the solution “started” with smaller Courant (CFL) numbers and added dissipation.
The Courant number and smoothing factors are increased manually until the residuals stabilize.

19
FLUENT took ~24,000 iterations, primarily because a first-order spatial discretization was run
first, followed by the second-order case. The case was hard to start from the second-order
solution directly. Also, the CFL number was only increased to one. The residuals dropped from
two orders of magnitude (the cross flow velocities) to four orders of magnitude (the turbulent KE
and diffusivity).
Advantages of the CFD++ solver are its automatic CFL number ramping and a capability for
adding dissipation parameters in the governing equations. In FLUENT, one must ramp the CFL
number manually and there are no added dissipation parameters in the governing equations. This
difficult test case is the first in which FLUENT has been difficult to run at higher CFL numbers.
In previous cases (1, 2), simulations were started with the second-order spatial discretization
solver, at CFL numbers of 1 or 2, and with a maximum CFL number between 5 and 10.

6. Conclusion

A computational study was undertaken to investigate the application of several CFD solvers to a
computational model of a projectile system with two different sabots. Flow field computations
were performed at Mach number 4.5 for 0° angle of attack, under standard atmospheric
conditions. The computational results show the predictive capabilities of various CFD
techniques in the analysis of supersonic flow for both a 2-D axisymmetric model and a 3-D
model. They also provide a look at some of the software tools available to research scientists in
the field of CFD. All three solvers provided qualitatively similar results. CFD++ provided the
solution in the most efficient manner.

20
7. References

1. Sahu, J.; DeSpirito, J.; Edge, H.; Silton, S.; Heavey, K. Recent Applications of Structured
and Unstructured Grid Techniques to Complex Projectile and Missile Configurations.
Proceedings of the 8th International Grid Generation and Computational Field Simulations,
Honolulu, HI, June 2002.
2. DeSpirito, J.; Vaughn, Jr., M. E.; Washington, W. D. Numerical Investigation of Canard-
Controlled Missile Using Planar and Grid Tail Fins, Part I: Supersonic Flow;
ARL-TR-2848; U.S. Army Research Laboratory: Aberdeen Proving Ground, MD, 2002.
3. Sahu, J.; Heavey, K. R.; Edge, H. L. Numerical Computations of Supersonic Flow Over
Elliptical Projectiles; ARL-TR-2589; U.S. Army Research Laboratory: Aberdeen Proving
Ground, MD, 2001.
4. Sahu, J.; Heavey, K. R. Computational Fluid Dynamics Modeling of a 40-mm Grenade With
and Without Jet Flow; ARL-TR-2572; U.S. Army Research Laboratory: Aberdeen Proving
Ground, MD, 2001.
5. Graham, M. J.; Weinacht, P.; Brandeis, J.; Angelini, R. A Numerical Investigation of
Supersonic Jet Interaction for Finned Bodies; ARL-TR-2312; U.S. Army Research
Laboratory: Aberdeen Proving Ground, MD, 2000.
6. Sahu, J.; Heavey, K. R.; Pressel, D.; Dinavahi, S. Parallel Numerical Computations of
Projectile Flow Fields; ARL-TR-2019; U.S. Army Research Laboratory: Aberdeen Proving
Ground, MD, 1999.
7. Sahu, J.; Edge, H. L.; Heavey, K. R.; Ferry, E. N. Computational Fluid Dynamics Modeling
of Multi-body Missile Aerodynamic Interference; ARL-TR-1765; U.S. Army Research
Laboratory: Aberdeen Proving Ground, MD, 1998.
8. Edge, H. L.; Sahu, J.; Sturek, W. B.; Pressel, D. M.; Heavey, K. R.; Weinacht, P.;
Zoltani, C. K.; Nietubicz, C. J.; Clarke, J.; Behr, M.; Collins, P. Common High Performance
Computing Software Support Initiative (CHSSI) Computational Fluid Dynamics CFD-6
Project Final Report: ARL Block-Structured Gridding Zonal Navier-Stokes Flow
(ZNSFLOW) Solver Software; ARL-TR-2084; U.S. Army Research Laboratory: Aberdeen
Proving Ground, MD, 2000.
9. Baldwin, B. L.; Lomax, H. Thin Layer Approximation and Algebraic Model for Separated
Turbulent Flows; AIAA 78-257; American Institute of Aeronautics and Astronautics,
Reston, VA, 1978.

21
10. Goldberg, U. C.; Peroomian, O.; Chakravarthy, S. A Wall-Distance Free K-E Model With
Enhanced Near-Wall Treatment. American Society of Mechanical Engineers Journal of
Fluids Engineering 1998, 120, 457–462.

11. CFD++ User Manual, version 2.6.5; Metacomp Technologies, Argoura Hills, CA.

12. FLUENT 5.0 User’s Guide, volume 2; Fluent, Inc., Lebanon, NH, 1998.
13. Chan, W. M. OVERGRID Software Documentation. U.S. Army Research Laboratory-
Major Shared Resource Center: Aberdeen Proving Ground, MD, April 1999.

14. GRIDGEN, version 14.3 user’s manual; Pointwise Inc., Fort Worth, TX, 2002.

15. FIELDVIEW, version 8 user’s manual; Intelligent Light, Lindhurst, NJ, 2001.

22
NO. OF NO. OF
COPIES ORGANIZATION COPIES ORGANIZATION

2 DEFENSE TECHNICAL ABERDEEN PROVING GROUND

INFORMATION CENTER
DTIC OCA 2 DIR USARL
8725 JOHN J KINGMAN RD AMSRL CI LP (BLDG 305)
STE 0944 AMSRL CI OK TP (BLDG 4600)
FT BELVOIR VA 22060-6218

1 COMMANDING GENERAL
US ARMY MATERIEL CMD
AMCRDA TF
5001 EISENHOWER AVE
ALEXANDRIA VA 22333-0001

1 INST FOR ADVNCD TCHNLGY

THE UNIV OF TEXAS AT AUSTIN
3925 W BRAKER LN STE 400
AUSTIN TX 78759-5316

1 US MILITARY ACADEMY
MATH SCI CTR EXCELLENCE
MADN MATH
THAYER HALL
WEST POINT NY 10996-1786

1 DIRECTOR
US ARMY RESEARCH LAB
AMSRL D
DR D SMITH
2800 POWDER MILL RD
ADELPHI MD 20783-1197

1 DIRECTOR
US ARMY RESEARCH LAB
AMSRL CS IS R
2800 POWDER MILL RD
ADELPHI MD 20783-1197

3 DIRECTOR
US ARMY RESEARCH LAB
AMSRL CI OK TL
2800 POWDER MILL RD
ADELPHI MD 20783-1197

3 DIRECTOR
US ARMY RESEARCH LAB
AMSRL CS IS T
2800 POWDER MILL RD
ADELPHI MD 20783-1197

1
NO. OF NO. OF
COPIES ORGANIZATION COPIES ORGANIZATION

1 COMMANDER 1 COMMANDER
NAVAL SURFACE WARFARE CTR US ARMY TACOM ARDEC
ATTN CODE B40 BLDG 162S
W YANTA AMCPM DS MO
DAHLGREN VA 22448-5100 P J BURKE
PICATINNY ARSENAL NJ
1 COMMANDER 07806-5000
NAVAL SURFACE WARFARE CTR
CODE 420 1 METACOMP TECHNOLOGIES INC
A WARDLAW S R CHAKRAVARTHY
INDIAN HEAD MD 20640-5035 28632-B ROADSIDE DRIVE
SUITE 255
1 AIR FORCE ARMAMENT LAB AGOURA HILLS CA 91301
AFATL/FXA
D BELK 2 FLUENT INC
EGLIN AFB FL 32542-5434 10 CAVENDISH COURT
CENTERRA RESOURSE PARK
3 COMMANDER G STUCKERT
US ARMY TACOM ARDEC T SCHEIDEGGER
AMSTA AR FSF T LEBANON NH 03766-1442
BLDG 382
H HUDGINS
J GRAU ABERDEEN PROVING GROUND
W KOENIG
PICATINNY ARSENAL NJ 1 DIR USARL
07806-5000 AMSRL WM
J SMITH
1 COMMANDER BLDG 4600
US ARMY TACOM APG MD 21005-5069
AMSTA AR CCH B
P VALENTI 18 DIR USARL
BLDG 65-S AMSRL CI HC
PICATINNY ARSENAL NJ R NOAIC
07806-5001 AMSRL SL BE
A MIKHAIL
1 COMMANDER AMSRL WM B
US ARMY ARDEC A W HORST JR
SFAE FAS SD AMSRL WM BA
M DEVINE D LYON
PICATINNY ARSENAL NJ AMSRL WM BC
07806-5001 P PLOSTINS
J DESPIRITO
1 NAVAL AIR WARFARE CENTER B GUIDOS
MS 3 BLDG 2187 K HEAVEY (5 CPS)
D FINDLAY J NEWILL
PATUXENT RIVER MD 20670 J SAHU
S SILTON
1 UNIV OF ILLINOIS AT URBANA P WEINACHT
CHAMPAIGN WM BD
DEPT OF MECHANICAL AND B FORCH
INDUSTRIAL ENGINEERING WM BF
J C DUTTON S WILKERSON
URBANA IL 61801