ST STAR AR-CCM+ / Optimate Kynan Maley ey Optimate Descr crip - - PowerPoint PPT Presentation

ST STAR AR-CCM+ / Optimate

Kynan Maley ey

Descr crip iptio ion Proce cess ss Study exampl ples Toke ken based licensi sing Summary

Optimate

STAR-CCM+ M+ add-on

• n

Easy y to use! Pan anel to setup multip iple ST STAR-CCM+ + analys yses

– Scriptless Automation – Leverage STAR-CCM+ design space

Five ava availab able study y types

– Design Exploration – Design of Experiment – Robustness – Optimization – Pareto Optimization

Setup is all done within in STAR-CCM+ + envir vironment Optimat ate performs the fo followin ing

– Creates all necessary scripting – Submits and monitors jobs – Collects the simulation data – Post-process the study

Optimate

Desig sign Explor loration

• n

– Supplied inputs – Test replication

Optimization

• n

– Global and local design space exploration – Identify important variables – Pareto optimization

• Identify set of optimal designs

– Uses SHERPA algorithm

Desi sign of Experiment

– Local design space exploration – Response surface models

Robust stness s Analysis lysis

– Characterize design due to input variation – User supplies probability distribution

Study dy Type pes

3D 3D-CAD AD Design Parameters

– Any scalar quantity – Translation Components

Motion

• ns

– Rotation rate – Translational velocity

Part Swapp pping

– CAD parts

Region ion Physics ics Values

– Volumetric heat source

Mesh Settings

– Base size

Field ld Function

• ns

– User scalar field functions

Boundary Physics cs Values

– Velocity magnitude – Mass flow rate – Static temperature – Static pressure – Total temperature – Total pressure – Mach number – Flow angles (modifies ‘wind’ CS) – Mass fractions – Wall roughness – Turbulence scalars – Thermal scalars

Curr rren ently y Suppor

• rted

ed Variabl bles es

Proble blem External aerodynamics Imported CAD combined with 3D-CAD model Goals ls Investigate mirror placement effects on vehicle aerodynamics Modify position of side view mirrors

Design gn Explora

• ration
• n Exampl

ple

Design gn Explora

• ration
• n Exampl

ple

Design gn Explora

• ration
• n Post-pr

processi essing

Joint Common Missile

Extern ernal Ae Aerodyn rodynamics

Design gn Explora

• ration
• n

Pipe Flow

• w

– Investigate blockage size on pressure drop

Bloc

• cka

kage created in 3D-CAD AD

– Thickness – Depth

Latin in hypercu cube be sampli pling Ran 10 cases

Design gn of Experi erimen ent Exampl ple

Design gn of Experi erimen ent Exampl ple

Propr prie ietary SHERPA A / MO-SHE HERPA algor

• rith

ithm Effici icient

– Fewer evaluations required – Rapid setup – No iterating to find best method – Solution first time more often

Easy sy to use:

– Simply define problem and number of evaluations – No tuning – No optimization expertise required

Robust st

– Finds better solutions more often for broad classes of problems

Optimization

• n Algori
• rithm

Util iliz ize pow

• wer of MO-SHE

HERPA A algor

• rith

ithm

– Better solutions in fewer iterations

Elect ctron

• nics

cs cooling

– Conjugate heat transfer

Pin config iguratio ion parameteriz ized

– Spacing – Height – Diameter – Taper

Optim imiz ization ion object ctiv ives

– Minimize pressure drop – Minimize chip temperature

Ran 100 cases

Paret eto

• Optimization
• n Exampl

ple

Paret eto

• Optimization
• n Exampl

ple

Enable bles many concurrent jobs s all derived from same parent sim file ST STAR AR-CCM+/ +/Opt ptim imate

– Design Exploration, Design of Experiment, Robustness modes

ST STAR AR-CCM+/Optim ptimate+

– All modes (including Optimization and Pareto Optimization)

ST STAR AR-CCM+/ +/Power Toke kens

– Used only during run – Counts as STAR-CCM+ Session or STAR-HPC – ex. 100 tokens

• 1 concurrent job on 100 cpus, 10 concurrent jobs on 10 cpus, 100 concurrent jobs
• n 1 cpu

Optimate e Licensi sing

Know

• wled

ledge Databa base se

– suppor

• rt.cd-ada

dapco.

• .com
• m

– Evaluation Mode Demonstration – Pareto Optimization Mode Demonstration

Learn rn More

Optim imate is a ST STAR-CCM+ + add-on

• n

Five avail ilable ble study types

– Design Exploration – Design of Experiment – Robustness – Optimization – Pareto Optimization

Optim imate licensin sing scheme

– Designed to run lots of similar jobs concurrently

Tutor

• rial

l videos

• s on Know
• wled

ledge Base se Av Avail ilable le now

• w

Summary

Thank yo you