Validation of US3D for Capsule Aerodynamics using 05-CA Wind … · 2013-04-10 · 05-CA Overview 05-CA was a wind tunnel test performed in two NASA Ames test sections: ‣9-foot

The University of Minnesota NASA

Validation of US3D for Capsule Aerodynamicsusing 05-CA Wind Tunnel Test Data

Alan Schwing

11/19/12

https://ntrs.nasa.gov/search.jsp?R=20120017946 2020-07-04T11:52:13+00:00Z

Outline of Presentation

Motivation

05-CA Overview

Numerical Inputs - Solver, Method, Grid

Sensitivity Analysis - Spacial, Temporal

Results‣Integrated Aerodynamic Loads‣Pressure Port Comparisons‣Possible improvements

Conclusions

Questions

Test 5-CA Final Report Crew Exploration Vehicle (CEV) Program

EG-CEV-06-19 Final Report Test 5-CA March 2006 James Bell Code AOX, MS 260-1 [email protected], (650)604-4142

National Aeronautics and Space Administration NASA Ames Research Center Moffett Field, California

Motivation

NASA is currently investigating and has a long history of employing capsule shapes for a variety of missions - manned and unmanned.

In order to design the most efficient system possible, correct characterization of the vehicle’s aerodynamics are imperative.‣Parachute Mass ‣Structural Loading ‣Downrange Concerns

Historically, a combination of wind tunnel testing and flight testing has been most useful. Computational Fluid Dynamics (CFD) are becoming an increasingly vital piece of this work, though.

Current models employed at NASA still allow for a large degree of uncertainty in aerodynamic prediction. There is room for new tools to assist in design decisions.

US3D continues to improve the state of the science and is beginning to see use in aerodynamic prediction for capsule shapes. This work is designed to:‣Help understand how applicable the code is to these shapes‣Provide suggestions on best practices‣Outline avenues for improvement and areas of greatest confidence

05-CA Overview

05-CA was a wind tunnel test performed in two NASA Ames test sections:‣9-foot x 7-foot supersonic test section : Mach 1.6 - Mach 2.5‣11-foot transonic test section : Mach 0.3 - Mach 1.4

This work will focus on the 11-foot section and a 7.66% model.‣Model was heavily instrumented and provides the largest wealth of validation data.

• 151 pressure ports with static measurements • 6-axis balance for integrated loads• 12 unsteady pressure transducers

‣Wide range of Mach number and Angle of Attack (α) available for comparison.

A supersonic, transonic, and subsonic case were considered at three angles of attack in order to provide an initial look at how US3D’s performance varied between these three regimes.

0.3 0.5 0.6 0.7 0.8 0.9 0.95 1.05 1.1 1.2 1.4140 x x x145150155 x x x160165170 x x x

α Mach

Flow Solver - US3D

US3D is an unstructured, finite-volume CFD code developed at the University of Minnesota. It is fully-parallel and implicit.‣Steiger-Warming flux vector splitting‣Differed-Correction approach to viscous fluxes‣Includes both RANS and LES turbulence modeling‣Detached Eddy Simulation (DES) also fully supported‣2nd-order implicit Euler time advancement‣Low dissipation numerics (CITE)

High-order and low dissipation allow for superior computations. Additionally, combined RANS/LES simulations (DES) provide much higher fidelity in the wake region - a major driver for this problem’s aerodynamics.

Going to compare three numerical methods in US3D:‣RANS turbulence model‣DES turbulence model with Steiger-Warming dissipative fluxes‣DES turbulence model with low-dissipation kinetic energy conserving fluxes

Grid Generation

All grid generation was performed with GridPro..

Some important notes and features:‣A portion of the sting was included to improve computational fidelity.‣Wind tunnel walls were not modeled.‣Extensive smoothing iterations were applied in order to drive the grid smoother to a

steady-state. This helped maintain similarity between grids of different resolutions.‣Wake refinement was aligned with the mid-α case in order to best capture the range of

the analysis. It extended 6 diameters downstream - a distance selected based on best practices used by NASA for similar computations using OVERFLOW (CITE).‣Similarly, the outer boundaries extended 30 diameters from the vehicle in order to

minimize their effect for the subsonic cases.‣Viscous wall-spacing selected to maintain y+ of less than 1 on entire geometry.

Grid Generation

Grid Generation

One of the first analysis performed was an investigation into the spacial refinement and its effect on the quantity of interest, aerodynamic loads.

Final grid of 18,217,949 points and 18,113,216 cells.

Grid Resolution Study

-1.4

-1.39

-1.38

-1.37

-1.36

-1.35

-1.34

-1.33

-1.32

1 10 100

CX

Cells (Millions)

Mach 1.4

0.048

0.05

0.052

0.054

0.056

0.058

0.06

0.062

0.064

0.066

0.068

1 10 100

CZ

Cells (Millions)

Mach 1.4

-0.02

-0.019

-0.018

-0.017

-0.016

-0.015

-0.014

-0.013

-0.012

-0.011

-0.01

-0.009

1 10 100

Cm

Cells (Millions)

Mach 1.4

Half-GridFull-Grid

Temporal Resolution Study

Just as significant to the validity of the results is the selection of the proper timestep to use for these simulations.‣One initial consideration was to maintain a CFL of less than 1 throughout the wake

region. Maximum CFL values varied for each Mach number.

-0.678

-0.676

-0.674

-0.672

-0.67

-0.668

-0.666

-0.664

-0.662

4.3 4.35 4.4 4.45

4.5 4.55 4.6 4.65

4.7

C A

Time (s)

/home/work-schwi262/projects/summer12/05-CA/runCase #1: coarse_155/m1.4a155_2oConfig = coarse_155, Mach = 1.4, α = 155, Order = 2, Vinf = 168.32, Rho = 1.43938, Tinf = 281.650

Diverge at 4.31 seconds

Iterations = 200000/200000CPU Time = 5226.2 hours

0.018

0.02

0.022

0.024

0.026

0.028

0.03

4.3 4.35 4.4 4.45

4.5 4.55 4.6 4.65

4.7

C N

Time (s)




1000500250100

2000

-0.0048

-0.0046

-0.0044

-0.0042

-0.004

-0.0038

-0.0036

-0.0034

-0.0032

-0.003

-0.0028

4.3 4.35 4.4 4.45

4.5 4.55 4.6 4.65

4.7

C m

Time (s)




Temporal Resolution Study

Also of concern was the effect on the integrated forces and moments.

Results

Mach 1.40

Results - Mach 1.40

Results - Mach 1.40

Instantaneous Averaged

Solutions are very unsteady and so the results were time-averaged in order to compare to similarly time-averaged wind tunnel data.

Effects of Numerical MethodKEC

SW

RANS

Instantaneous images show qualitative changes between the numerical methods.

Dissipation grows as we move from KEC → SW → RANS

There is also a noticeable difference in the flow features on the windward side.

Averaged KEC

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

140 145 150 155 160 165 170 175

CL

Alpha

Mach 1.4

1.05

1.1

1.15

1.2

1.25

1.3

1.35

1.4

1.45

1.5

1.55

140 145 150 155 160 165 170 175

CD

Alpha

Mach 1.4

-0.05

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

140 145 150 155 160 165 170 175

Cm

CG

Alpha

Mach 1.4

0

0.05

0.1

0.15

0.2

0.25

140 145 150 155 160 165 170 175

CN

Alpha

Mach 1.4

-1.55

-1.5

-1.45

-1.4

-1.35

-1.3

-1.25

-1.2

140 145 150 155 160 165 170 175

CA

Alpha

XCG = 0.680 DZCG = -0.045 D

WTTRANS

SWKEC

Results - Mach 1.40

Comparisons are good‣RANS appears to perform the worst. SW appears to do a bit

better than KEC, but both are very close.‣It is not surprising that our comparisons are this strong for this

Mach number. The integrated loads are dominated by the forebody pressure and not as sensitive to the modeling in the wake.‣Largest error seen in α=170°, we’ll look at the pressure tap data

to investigate a bit more.

View From φ = 270 °

m1.40 a155

V∞

Page 3

M∞ = 1.390, α = 155.22˚, β = 0.14˚, ReD = 5.32 x106, P = 30507.37 Pa, q = 41265.93 PaM∞ = 1.390, α = 155.22˚, β = 0.14˚, ReD = 5.32 x106, P = 30507.37 Pa, q = 41265.93 PaM∞ = 1.390, α = 155.22˚, β = 0.14˚, ReD = 5.32 x106, P = 30507.37 Pa, q = 41265.93 PaM∞ = 1.390, α = 155.22˚, β = 0.13˚, ReD = 5.32 x106

RANS conditions: SW conditions:

KEC conditions: Run429 conditions:

View From Apex

m1.40 a155

Page 3

φ = 270 °

φ = 0 °

φ = 90 °

φ = 180 °




-0.5

0

0.5

1

1.5

2

60 90 120 150 186 150 120 90 60X (in)

m1.40 a155

Page 3

ConeShoulder

Heat ShieldShoulder

ConeCp




RANS - phi 0SW - phi 0

KEC - phi 0Run_429


KEC - phi 180

Results - Mach 1.40 α = 155°

Results

Mach 0.95

Results - Mach 0.95

Results - Mach 0.95

0.1

0.2

0.3

0.4

0.5

0.6

0.7

140 145 150 155 160 165 170 175

CL

Alpha

Mach 0.95

0.8

0.9

1

1.1

1.2

1.3

1.4

140 145 150 155 160 165 170 175

CD

Alpha

Mach 0.95

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

0.03

0.04

0.05

140 145 150 155 160 165 170 175

Cm

CG

Alpha

Mach 0.95

0

0.02

0.04

0.06

0.08

0.1

0.12

140 145 150 155 160 165 170 175

CN

Alpha

Mach 0.95

-1.4

-1.35

-1.3

-1.25

-1.2

-1.15

-1.1

-1.05

-1

-0.95

140 145 150 155 160 165 170 175

CA

Alpha

XCG = 0.680 DZCG = -0.045 D

WTTRANS

SWKEC

Comparisons are good‣RANS is clearly not as capable as the KEC DES for this regime.

SW was not run.‣Examination of the pressure ports for these cases reveals why

RANS has such trouble.

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

140 145 150 155 160 165 170 175

CL

Alpha

Mach 1.4

1.05

1.1

1.15

1.2

1.25

1.3

1.35

1.4

1.45

1.5

1.55

140 145 150 155 160 165 170 175C

D

Alpha

Mach 1.4

-0.05

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

140 145 150 155 160 165 170 175

Cm

CG

Alpha

Mach 1.4

0

0.05

0.1

0.15

0.2

0.25

140 145 150 155 160 165 170 175

CN

Alpha

Mach 1.4

-1.55

-1.5

-1.45

-1.4

-1.35

-1.3

-1.25

-1.2

140 145 150 155 160 165 170 175

CA

Alpha

XCG = 0.680 DZCG = -0.045 D

WTTRANS

SWKEC


m0.95 a140

V∞

Page 3

M∞ = 0.950, α = 142.32˚, β = 0.11˚, ReD = 5.31 x106, P = 54028.05 Pa, q = 34158.75 PaM∞ = 0.950, α = 142.32˚, β = 0.11˚, ReD = 5.31 x106, P = 54028.05 Pa, q = 34158.75 PaM∞ = 0.950, α = 142.32˚, β = 0.09˚, ReD = 5.31 x106

RANS conditions: KEC conditions:

Run424 conditions:

View From Apex

m0.95 a140

Page 3

φ = 270 °

φ = 0 °

φ = 90 °

φ = 180 °



Run424 conditions:

-1.5

-1

-0.5

0

0.5

1

1.5

60 90 120 150 186 150 120 90 60X (in)

m0.95 a140

Page 3

ConeShoulder

Heat ShieldShoulder

ConeCp



Run424 conditions:

RANS - phi 0KEC - phi 0

Run_424RANS - phi 180

KEC - phi 180



m0.95 a140

V∞

Page 3



Run424 conditions:

View From Apex

m0.95 a140

Page 3

φ = 270 °

φ = 0 °

φ = 90 °

φ = 180 °



Run424 conditions:

-1.5

-1

-0.5

0

0.5

1

1.5

60 90 120 150 186 150 120 90 60X (in)

m0.95 a140

Page 3

ConeShoulder

Heat ShieldShoulder

ConeCp



Run424 conditions:



KEC - phi 180

Results - Mach 0.95

RANS Averaged KEC

α = 140°



m0.95 a155

V∞

Page 3



Run424 conditions:

View From Apex

m0.95 a155

Page 3

φ = 270 °

φ = 0 °

φ = 90 °

φ = 180 °



Run424 conditions:

-1.5

-1

-0.5

0

0.5

1

1.5

60 90 120 150 186 150 120 90 60X (in)

m0.95 a155

Page 3

ConeShoulder

Heat ShieldShoulder

ConeCp



Run424 conditions:



KEC - phi 180



m0.95 a170

V∞

Page 3



Run424 conditions:

View From Apex

m0.95 a170

Page 3

φ = 270 °

φ = 0 °

φ = 90 °

φ = 180 °



Run424 conditions:

-1.5

-1

-0.5

0

0.5

1

1.5

60 90 120 150 186 150 120 90 60X (in)

m0.95 a170

Page 3

ConeShoulder

Heat ShieldShoulder

ConeCp



Run424 conditions:



KEC - phi 180

Results

Mach 0.5

Results - Mach 0.50

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

140 145 150 155 160 165 170 175

CL

Alpha

Mach 0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0.95

1

140 145 150 155 160 165 170 175

CD

Alpha

Mach 0.5

-0.02

-0.01

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

140 145 150 155 160 165 170 175

Cm

CG

Alpha

Mach 0.5

-0.1

-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

0.1

0.12

140 145 150 155 160 165 170 175

CN

Alpha

Mach 0.5

-1

-0.95

-0.9

-0.85

-0.8

-0.75

140 145 150 155 160 165 170 175

CA

Alpha

XCG = 0.680 DZCG = -0.045 D

WTTRANS

SWKEC

Results - Mach 0.50

Comparisons reveal greater errors in modeling‣Comparisons are definitely varied with no clear winner. RANS and

KEC appear to grab the trend across the forces/moment.‣For a flow this unsteady and with this low of a dynamic pressure,

there is a significant contribution from the base flow. Also of interest are large magnitude pressure near the shoulder.‣Pressure port measurements will help illustrate where each model

is having troubles.

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

140 145 150 155 160 165 170 175

CL

Alpha

Mach 1.4

1.05

1.1

1.15

1.2

1.25

1.3

1.35

1.4

1.45

1.5

1.55

140 145 150 155 160 165 170 175C

D

Alpha

Mach 1.4

-0.05

-0.04

-0.03

-0.02

-0.01

0

0.01

0.02

140 145 150 155 160 165 170 175

Cm

CG

Alpha

Mach 1.4

0

0.05

0.1

0.15

0.2

0.25

140 145 150 155 160 165 170 175

CN

Alpha

Mach 1.4

-1.55

-1.5

-1.45

-1.4

-1.35

-1.3

-1.25

-1.2

140 145 150 155 160 165 170 175

CA

Alpha

XCG = 0.680 DZCG = -0.045 D

WTTRANS

SWKEC

Results - Mach 0.50


m0.50 a140

V∞

Page 3




View From Apex

m0.50 a140

Page 3

φ = 270 °

φ = 0 °

φ = 90 °

φ = 180 °




-3.5

-3

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.560 90 120 150 186 150 120 90 60

X (in)

m0.50 a140

Page 3

ConeShoulder

Heat ShieldShoulder

ConeCp





KEC - phi 0Run_418


KEC - phi 180

α = 140°


RANS

Results - Mach 0.50 α = 140°α = 140°

SW

Results - Mach 0.50 α = 140°α = 140°

KEC

Unsteady Frequency Response

Wind tunnel model had 11 taps arranged on the surface of the backshell of the model. These probes recorded data at a rate of 6400 Hz for 6 seconds.‣All cases examined in this work included unsteady data.‣Comparisons revealed that the computational expense necessary to obtain results that

were ‘statically converged’ required significant wall time. For this reason, only one case was examined thoroughly.

The Mach 0.50 case at α = 140° was selected for comparison for two reasons:‣It was a massively separated case that had a strong response in the wind tunnel data.‣Integrated loads showed a large spread between the methods examined here and might

highlight differences between then.Data was sampled from the CFD at tap locations‣Downsampled to 6400 Hz using Welch’s Method with 50% overlap‣0.2 second intervals (wind tunnel test used 0.2s)‣Tukey Window used on each interval

Frequency response from integrated loads alsoconsidered 4 5 6

210

1 9 3

8117

Unsteady Frequency Response - TapsMonitors - r0418s0001Monitors - r0418s0001Monitors - r0418s0001Monitors - r0418s0001Monitors - r0418s0001Monitors - r0418s0001Monitors - r0418s0001Monitors - r0418s0001Monitors - r0418s0001Monitors - r0418s0001Monitors - r0418s0001

4 5 6

210

1 9 38117

WTTRANSKECSW

Unsteady Frequency Response - CmCm - r0418s0001Cm - r0418s0001Cm - r0418s0001Cm - r0418s0001Cm - r0418s0001Cm - r0418s0001Cm - r0418s0001Cm - r0418s0001Cm - r0418s0001Cm - r0418s0001Cm - r0418s0001

4 5 6

210

1 9 38117

WTTRANSKECSW

Varying DES Model

There is a sensitivity to the flavor of DES, but the differences are not substantial and in some cases are negligible.

0.08

0.09

0.1

0.11

0.12

0.13

0.14

0.15

0.16

Mach 0.5 Mach 1.4 0.215

0.22

0.225

0.23

0.235

0.24

0.245

0.25

0.255

0.26

0.265CL

DES Sensitivity

0.796 0.8

0.804 0.808 0.812 0.816 0.82

0.824 0.828 0.832 0.836 0.84

0.844 0.848 0.852 0.856 0.86

0.864 0.868

Mach 0.5 Mach 1.4 1.36

1.37

1.38

1.39

1.4

1.41

1.42

1.43

1.44

1.45CD

DES Sensitivity

-0.12

-0.08

-0.04

0

0.04

Mach 0.5 Mach 1.4-0.46

-0.45

-0.44

-0.43

-0.42

-0.41

-0.4

-0.3910*CmCG

DES Sensitivity

-0.008-0.004

0 0.004 0.008 0.012 0.016 0.02

0.024 0.028 0.032 0.036 0.04

0.044 0.048 0.052 0.056 0.06

0.064 0.068 0.072 0.076

Mach 0.5 Mach 1.4 0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045CN

DES Sensitivity

-0.876-0.872-0.868-0.864-0.86

-0.856-0.852-0.848-0.844-0.84

-0.836-0.832-0.828-0.824-0.82

-0.816-0.812-0.808

Mach 0.5 Mach 1.4-1.48

-1.47

-1.46

-1.45

-1.44

-1.43

-1.42

-1.41

-1.4

-1.39

-1.38CA

XCG = 0.680 DZCG = -0.045 D

DESIDES

DDES

Conclusions and Further Work

RANS is ill-suited for analysis of these problems. For transonic and supersonic cases, US3D shows fairly good agreement using DES across all cases.

Separation prediction and resulting backshell pressure are problems across all portions of this analysis. This becomes more of an issue at lower Mach numbers:‣Stagnation pressures not as large - wake and backshell are more significant‣Errors on shoulder act on a large area - small discrepancies manifest as large changes

Subsonic comparisons are mixed with regard to integrated loads and merit more attention. Dominant unsteady behavior (wake shedding) resolved well, though.

Further work:‣Evaluate the effect of other turbulence models on separation prediction‣Use more advanced tools to look into improvements that might be made with grid quality‣Investigate transition and its effect on separation

BACKUP

-0.765

-0.76

-0.755

-0.75

-0.745

-0.74

-0.735

-0.73

-0.725

-0.72

1 10 100

CX

Cells (Millions)

Mach 0.5

0.08

0.082

0.084

0.086

0.088

0.09

0.092

0.094

0.096

0.098

0.1

1 10 100

CZ

Cells (Millions)

Mach 0.5

0.03

0.031

0.032

0.033

0.034

0.035

0.036

0.037

1 10 100

Cm

Cells (Millions)

Mach 0.5

Half-Grid

-1.21

-1.205

-1.2

-1.195

-1.19

-1.185

-1.18

1 10 100

CX

Cells (Millions)

Mach 0.95

0.067

0.068

0.069

0.07

0.071

0.072

0.073

0.074

0.075

0.076

0.077

1 10 100

CZ

Cells (Millions)

Mach 0.95

-0.006

-0.005

-0.004

-0.003

-0.002

-0.001

0

1 10 100

Cm

Cells (Millions)

Mach 0.95

Half-Grid

Documents

Validation of US3D for Capsule Aerodynamics using 05-CA Wind … · 2013-04-10 · 05-CA Overview 05-CA was a wind tunnel test performed in two NASA Ames test sections: ‣9-foot