Contributions to HiLiftPW-3 Using Structured, Overset Grid Methods - - PowerPoint PPT Presentation

▶
contributions to hiliftpw 3 using structured overset grid
SMART_READER_LITE
LIVE PREVIEW

Contributions to HiLiftPW-3 Using Structured, Overset Grid Methods - - PowerPoint PPT Presentation

Dec 18, 2022 492 likes •873 views

Contributions to HiLiftPW-3 Using Structured, Overset Grid Methods Presented at AIAA SciTech 2018 Kissimmee, FL January 10, 2018 Jim Coder University of Tennessee, Knoxville Tom Pulliam and James Jensen NASA Ames Research Center 1 Ou

slide-1
SLIDE 1

1

Contributions to HiLiftPW-3 Using Structured, Overset Grid Methods

Presented at AIAA SciTech 2018 Kissimmee, FL January 10, 2018 Jim Coder University of Tennessee, Knoxville Tom Pulliam and James Jensen NASA Ames Research Center

slide-2
SLIDE 2

2

Ou Outline

  • Introduction
  • Description of HiLiftPW-3 Geometries and Cases
  • Computational Methodologies
  • Results
  • Conclusion
  • Acknowledgments

AIAA SciTech 2018 2

slide-3
SLIDE 3

3

In Introd

  • duction

ion

  • Two geometries of interest

‒ High-Lift Common Research Model (HL-CRM)

  • Completely predictive

‒ JAXA Standard Model (JSM)

  • Transitional test case
  • Structured, overset grids generated and provided by the
  • rganizing committee
  • Two overset solvers considered in this paper

‒ OVERFLOW (UTK and NASA) ‒ LAVA (NASA)

AIAA SciTech 2018 3

slide-4
SLIDE 4

4

Ou Outline

  • Introduction
  • Description of HiLiftPW-3 Geometries and Cases
  • Computational Methodologies
  • Results
  • Conclusion
  • Acknowledgments

AIAA SciTech 2018 4

slide-5
SLIDE 5

5

HL HL-CR CRM G Geometry

  • Open-source high-lift configuration based on the Common

Research Model (Lacy and Sclafani, 2016)

AIAA SciTech 2018 5

slide-6
SLIDE 6

6

HL HL-CR CRM Ca Cases ( (Ca Case 1 1)

  • Case 1a (requested): Full-Chord Flap Gap grid-refinement

study

  • Case 1b (optional): Full-Chord Flap Gap with grid adaptation
  • Case 1c (optional): Partially Sealed Chord Flap Gap for

medium-resolution grid only

  • Case 1d (optional): Partially Sealed Chord Flap Gap with grid

adaptation

AIAA SciTech 2018 6

Free-stream Mach Number 0.2 Angles of Attack 8° and 16° Mean Aerodynamic Chord (MAC) 275.8 in (full scale) Reynolds Number (based on MAC) 3.26 x 10

6

Reference Static Temperature 518.67 °R (288.15 K) Reference Static Pressure 14.700 psi (760.21 mm-Hg)

slide-7
SLIDE 7

7

JS JSM Geometr try

  • Representative of a 100-person-class transport with a

modern high-lift system (Yokokawa et al., 2006 and 2008)

AIAA SciTech 2018 7

slide-8
SLIDE 8

8

JS JSM Case ses s (Case se 2)

  • Case 2a (requested): Nacelle/Pylon Off
  • Case 2b (optional): Nacelle/Pylon Off with grid adaptation
  • Case 2c (requested): Nacelle/Pylon On
  • Case 2d (optional): Nacelle/Pylon On with grid adaptation

AIAA SciTech 2018 8

Free-stream Mach Number 0.172 Angles of Attack 4.36°, 10.47°, 14.54°, 18.58°, 20.59°, and 21.57° Mean Aerodynamic Chord (MAC) 529.2 mm (model scale) Reynolds Number (based on MAC) 1.93 x 10

6

Reference Static Temperature 551.79 °R (306.55 K) Reference Static Pressure 14.458 psi (747.70 mm-Hg)

slide-9
SLIDE 9

9

Ou Outline

  • Introduction
  • Description of HiLiftPW-3 Geometries and Cases
  • Computational Methodologies
  • Results
  • Conclusion
  • Acknowledgments

AIAA SciTech 2018 9

slide-10
SLIDE 10

10

Fl Flow S Solvers a and A Approach ch

  • OVERFLOW 2.2 (UTK and NASA)

‒ Node-centered, finite-difference ‒ RHS discretization: 3rd-order MUSCL w/ Roe fluxes ‒ LHS algorithm: ARC3D scalar pentadiagonal solver ‒ Turbulence model: Spalart-Allmaras SA-noft2-RC-QCR2000 ‒ Transition model: Coder AFT2017b (SA-RC-QCR2000-AFT2017b)

  • Turbulence model variant and inclusion of transition

modeling studied

  • Time accuracy effects studied

‒ BDF2 implicit scheme ‒ Timestep chosen to give 2 orders of magnitude drop in unsteady residual in 10-20 subiterations

AIAA SciTech 2018 10

slide-11
SLIDE 11

11

Fl Flow S Solvers a and A Approach ch

  • LAVA (NASA)

‒ Node-centered, finite-difference ‒ RHS discretization: 2nd-order MUSCL w/ Roe fluxes ‒ Van Albada limiter ‒ Turbulence model: Spalart-Allmaras SA-noft2-RC-QCR2000

  • ”Cold starts” used for all cases

AIAA SciTech 2018 11

slide-12
SLIDE 12

12

Co Computational R Resources

  • All simulations run on NAS Pleiades

‒ SGI ICE system ‒ Over 11,000 nodes with over 245,000 cores ‒ Intel Xeon (Broadwell, Haswell, Ivy Bridge, Sandy Bridge)

  • OVERFLOW simulations run on 420 cores (fully turbulent)

and 560 cores (transitional)

‒ 24-48 hours of wall-clock time to convergence

  • LAVA required 2000 cores with 48 hours of wall clock time

AIAA SciTech 2018 12

slide-13
SLIDE 13

13

Ou Outline

  • Introduction
  • Description of HiLiftPW-3 Geometries and Cases
  • Computational Methodologies
  • Results
  • Conclusion
  • Acknowledgments

AIAA SciTech 2018 13

slide-14
SLIDE 14

14

Ca Case 1 1: Su Surface Sm Smoothness I Issues

  • Original HL-CRM overset grids were projected onto a surface

triangulation rather than the smooth CAD

‒ Leads to oscillatory pressure behavior

  • New grids generated with projection directly to CAD

AIAA SciTech 2018 14

slide-15
SLIDE 15

15

Ca Case 1 1: T Turbulence M Modeling E Effects

  • Use (or exclusion) of QCR had a prominent effect on the flow

behavior around the flap gap

‒ QCR typically regarded as primarily affecting juncture flows

AIAA SciTech 2018 15

slide-16
SLIDE 16

16

Ca Case 1 1: T Turbulence M Modeling E Effects

AIAA SciTech 2018 16

slide-17
SLIDE 17

17

Ca Case 1 1: G Grid R Refinement St Study

  • Lift

AIAA SciTech 2018 17

α = 8° α = 16°

slide-18
SLIDE 18

18

Ca Case 1 1: G Grid R Refinement St Study

  • Drag

AIAA SciTech 2018 18

α = 8° α = 16°

slide-19
SLIDE 19

19

Ca Case 1 1: G Grid R Refinement St Study

  • Pitching Moment

AIAA SciTech 2018 19

α = 8° α = 16°

slide-20
SLIDE 20

20

Ca Case 1 1: G Grid R Refinement St Study

  • Representative behavior (η = 0.151, α = 16°)

AIAA SciTech 2018 20

α = 8° α = 16°

slide-21
SLIDE 21

21

Ca Case 1 1: E Effect o

  • f F

Flap G Gap Se Seal

  • Gap seal reduces separation near the gap, but induces

separation inboard

AIAA SciTech 2018 21

slide-22
SLIDE 22

22

Ca Case 2: : Na Nacelle/Pylon Off

  • Strong effect of turbulence/transition modeling
  • Multiple possible solutions depending on initial condition

AIAA SciTech 2018 22

slide-23
SLIDE 23

23

Ca Case 2: : Na Nacelle/Pylon Off

  • Selected pressure distribution (4.36 deg)

AIAA SciTech 2018 23

Main element, η = 0.89

slide-24
SLIDE 24

24

Ca Case 2: : Na Nacelle/Pylon Off

  • Selected pressure distribution (18.58 deg)

AIAA SciTech 2018 24

Main element η = 0.89 η = 0.77 η = 0.56

slide-25
SLIDE 25

25

Ca Case 2: : Na Nacelle/Pylon On

  • Strong effect of turbulence/transition modeling
  • No evidence of multiple solutions

AIAA SciTech 2018 25

slide-26
SLIDE 26

26

Ca Case 2: : Na Nacelle/Pylon On

  • Surface flow patterns (α = 18.58°)

AIAA SciTech 2018 26

LAVA

slide-27
SLIDE 27

27

Ca Case 2: : Na Nacelle/Pylon On

  • Surface flow patterns (α = 18.58°)

AIAA SciTech 2018 27

OVERFLOW (fully turbulent)

slide-28
SLIDE 28

28

Ca Case 2: : Na Nacelle/Pylon On

  • Surface flow patterns (α = 18.58°)

AIAA SciTech 2018 28

OVERFLOW (transitional)

slide-29
SLIDE 29

29

Ca Case 2: : Na Nacelle/Pylon On

  • Transition patterns (α = 18.58°)

AIAA SciTech 2018 29

OVERFLOW (turbulent index) Experiment (China clay)

slide-30
SLIDE 30

30

Ca Case 2: : Na Nacelle/Pylon On

  • Transition patterns (α = 18.58°)

AIAA SciTech 2018 30

OVERFLOW (turbulent index) Experiment (China clay)

slide-31
SLIDE 31

31

Ou Outline

  • Introduction
  • Description of HiLiftPW-3 Geometries and Cases
  • Computational Methodologies
  • Results
  • Conclusions
  • Acknowledgments

AIAA SciTech 2018 31

slide-32
SLIDE 32

32

Co Conclusions ( (HL-CR CRM)

  • Fully predictive, so no experimental data available for

comparison

  • Surface smoothness had an impact on surface pressure

distributions

‒ Grid should be projected to smooth CAD rather than triangulated surfaces

  • Use of QCR had a strong influence of flap separation patterns

with the unsealed flap gap

AIAA SciTech 2018 32

slide-33
SLIDE 33

33

Co Conclusions ( (JSM SM)

  • Evidence of multiple solutions observed for nacelle/pylon off

‒ “Warm” versus “cold” starts influenced final solution ‒ Time accurate results more consistent with warm starts ‒ Phenomenon not observed with nacelle/pylon on

  • Excluding QCR had an impact, but not a consistent shift

‒ Nacelle/pylon off: Excluding QCR delays stall with AoA ‒ Nacelle/pylon on: Excluding QCR accelerates stall with AoA

  • Transition modeling had an overall positive impact

‒ Better agreement in aerodynamic coefficients ‒ Predicted transition patterns consistent with experiment ‒ Not a panacea – separation patterns still have discrepancies

AIAA SciTech 2018 33

slide-34
SLIDE 34

34

Ou Outline

  • Introduction
  • Description of HiLiftPW-3 Geometries and Cases
  • Computational Methodologies
  • Results
  • Conclusion
  • Acknowledgments

AIAA SciTech 2018 34

slide-35
SLIDE 35

35

Ac Ackno nowledgm dgments ts

  • J.G. Coder thanks Cetin Kiris of NASA Ames Research Center

for providing access to the NASA Advanced Supercomputing (NAS) Pleiades cluster

35 AIAA SciTech 2018

slide-36
SLIDE 36

36

Qu Ques estion ions?

36 AIAA SciTech 2018