Parallel CFD Supporting NASA's Space o p e 0 Mission Directorate I

I

Reynaldo J. Gomez Ill EGIAeroscience & Flight Mechanics NASA Johnson Space Center Houston,Texas April 2 1 ,2008

https://ntrs.nasa.gov/search.jsp?R=20100017498 2018-05-24T03:52:53+00:00Z

Design Development ARC wind tunnel tests

Operations NAS+CFD

Retirement FY 2010

I

STS-70 & subs

.

1982 .C.t..Ct.C+19~L).I ..... *I.*.t. 1994............+.L. 3000

SSME *a

Space flight

Shuttle engines

advanced Main - . engine Englue I shase II + redesian I

* OVERFLOW t .8 - Buning

OVERFLOW 1.6 * OVERFLOW Z O 2002.

* CGT 0. l a . PEGASUS 5

+ OVERFLOW-D 5 .

OVERFLOW 2.1

1980 1985 1990 1995 2000 2005 20 I0 300,000 grid points Proof of concept

16.2 million grid points 96.4 million grid points Matched flight derived wing Detailed airloads used for loads within 5% redesigns & flight support

Launch Complex Geometry

- wQmr?@--

Space Shuttle Mission Overview (Video)

Parallel computing from prelaunch to landing On-orbit Assessments

Hypervelocity Orbital Debris

I ransonic airloads

Gap F~ller t21-OBSS-2#01

A = 1 28 inches +/-0 10 I3 = 0 37 ~nches +/4110 C=024knches-cS 0 = 0.58 mnc@ 05 05 .

E = 120 ~nch 1-0.10 F = O 3p inches +I-0 05 C3 = 2 30 ImUes %&0> 15 '"- H 7 o.q~&&i-@g@ 2. - '.

AIM-2003- 1248 Contingency Abort

RCC panels

Many External Tank redesign assessments used solutions run on Columbia

-- Modified Aft Longeron --I

Multiple icelfrost ramp redesigns Ascent & entry windows airloads Discrete airloads data book updates Venting database updates Aerothermal support & others

RCS TyveP covers

Bipod Ramp Removal

+Z Aerovent Modification

LO, feedline bracket redesigns

Debris transport aerodynamic models & prediction tools developed - - ' A

NSTS 08303 day of launch ice ball launch commit tool developed by Stuart RogersIARC NAS-07-004

NSTS 08303 Rev D, Change 13 Iceball Allowable

Rarefied Direct Simulation Monte Carlo Simulations

4 F. E. Lumpkin et al . - 27th JANNAF

Conference 2003

HubbIe Servicing Mission Support

Venting of the airlock into the payload bay can overload the HST solar arrays. DSMC simulations are used to redesign vents and specify vent rates to minimize risks to the hardware. rn

LeBeau, G.J., and Lumpkin, F. E., Ill, "Application highlights of the DSMC Analysis Code (DAC) software for simulating rarefied flows", Computer Methods in Applied Mechanics and Engineering, 1 9 1,595-609,200 I.

Orbital debris hypervelocity impacts > 3 kmlsec

=IL 1 nrrr-m-mpgy

1 i - 2m 'am-mm

v

Figure 1. Stuffed Whipple shield schematic, and a Stuffed Whipple shield impact test (NASA JSC photo).

Shuttle support CFD tools developed on vector machines and transitioned to parallel architectures

Parallel Development

Engineering capability (DOF) doubling every 2 years

1989: O(1 05) DOF 2009: 0( 1 08) DOF

STS- I 18 Tile Damage

W e went to the moon without CFD or parallel computers.Why do we need them now?

Reduce number of physical testsL and improve relevance when you do test

Nearly 1 00,000 hours ( l I years) of Shuttle wind tunnel testing Many facilities have shut down or been mothballed

Provides flight increments1 environments that cannot be obtained otherwise

Verification &Validation requirements are

Guide forverification andvalidation of Computational Fluid Dynamics Simulations, AlAA G-077, 1 998.

Columbia Accident lnvestigation Board Report,Volume I, August 2003.

A Renewed Commitment to Exce1lence:An Assessment of the NASA Agency-wide Applicability of the Columbia Accident lnvestigation Report, PB2005- 100968, January 30,2004.

Standard for Models and Simulations, NASA-STD-7009, July, I I,

. . . and there is still more work to do.

Some STS- I flight anomalies are still beyond current CFD tool capabilities, e.g.

Acoustics and heating on complex configurations with strong shock wave-boundary layer interactions

RCSIaerolthermal interactions

Physical models (turbulence, chemistry, ...) are key limitations that need to be improved.

Uncertainty quantification (where are the CFD error bars?)