optislang-inside-ansys.pdf

8/10/2019 optislang-inside-ansys.pdf

1/32

1

optiSLang inside ANSYS Dynardo GmbH 2013

"optiSLang inside ANSYSWorkbench"

efficient, easy, and safe to use

Robust Design Optimization (RDO)

Dynardo GmbH


2/32

2

Optimization and Robust Design toolbox for arbitraryCAE-processes

- for sensitivity study, optimization, robustnessevaluation, reliability analysis and robust designoptimization having > 10 parameter, design failure,noise at result values, strong non-linearities, ..

Entry Solution: ANSYS DesignXplorer

- for Design Exploration, Optimization and RDO forsmall problems (recommended up to 5..10 variables)

High End Solution

RDO inside ANSYS Workbench

ANSYS User Conferences Dynardo GmbH October 2012


3/32

3


Excellence of optiSLang

optiSLang is an algorithmic toolbox for

sensitivity analysis, optimization,robustness evaluation, reliability analysis

and robust design optimization

optiSLang is the commercial tool that has

completed the necessary functionality of

sensitivity analysis, optimization and

stochastic analysis to run real worldindustrial applications in CAE-based robust

design optimizations

optiSLang offers the beginner and

expert users easy and safe to use

predefined workflows of best practice

Since 2009 interfacing via optiPlugsupport ANSYS workbench-optiSLang

applications


4/32

4

ParameterManager

Parameter &

Responses

optiPlug - ANSYS Workbench optiSLang Interface

OptiSLang-Plugin:

just click to write all

necessary files foroptiSLang coupling



5/32

5

Safe to use.

automate best practice to black box flows minimize the risk to miss better designs (optimization) minimize the risk to estimate misleading measures for robustness,

safety and reliability offer easy to use measurements of prognosis quality

That task requires sophisticated technology with carefully balance between

number of solver calls and safety to reach the RDO goal.

Technology takes care that non expert can use it!

optiSLang Flows of best Praxis



6/32

6


Modules Sensitivity+MOP, Optimization and Robustness+MOP provide

best practise optiSLang functionality

optiSLang inside ANSYS Workbench v14


7/32

7

ANSYS Workbench

parametric set up ofcomplex simulations

optiSLang inside ANSYS Workbench


Easy to use:

minimize user input offer best practise

defaults for modules

offer pre defined postprocessing modes


8/32

8

What do we mean with that? classic DOE+RSM technology ask user to reduce number of

variables, choose a suitable DOE with a suitable regression functionand check the quality of the resulting response surface (RS) and theoptima on the RS.

optiSLang provides a automatic flow to reduce variables and generate

the best possible response surface for every response with a givennumber of solver calls [Meta model of optimal Prognosis (MoP)] andchecks MoP Prognosis quality and optima in real space.

optiSLang inside ANSYS WorkbenchEasy and safe to use!



9/32

9

Sensitivity Analysis

(Design Exploration)



10/32

10

Sensitivity Analysis

Dynardo GmbH 2013


11/32

11

Sensitivity Module

Minimal required user input:

Definition of parameter variation



12/32

12

The Meta Model of optimal Prognosis (MOP) is automatically created out ofthe DOE-Sampling

Minimal required user input: nonAdditional features:

supports removing designs out of DOE Post Processing

Sensitivity Module



13/32

13

Optimization



14/32

14

4) Run an ARSM, gradient based or biological based optimization Algorithms

1) Start with a sensitivity study usingthe LHS Sampling

Sensitivity Analysis and Optimization

3) Use MOP+gradient solver for fastDesign Improvement

Understand theProblem using

CoP/MoP

Search for Optima

Scan the whole Design Space

optiSLang

2) Identify the important parametersand responses

- understand the problem- reduce the problem



15/32

15

Optimization using MOP

After sensitivity analysis optimization

using MOP is supported.

Minimum required user input: drop the optimization module

onto MOP defining objective and

constraints

Optima which are based on metamodels need to be verified!

Proof optima: Automatic verification with

real ANSYS call Check differences in post

processing



16/32

16


Optimization Algorithms

Gradient-based Response surface method

Nature inspired optimization algorithms:- Genetic algorithms,- Evolutionary strategies- Particle Swarm Optimization

Start

Pareto Optimization

Adaptive RSM


17/32

17

When to use which Algorithm

Gradient-Based

Algorithms

EvolutionaryAlgorithm

ParetoOptimization

AdaptiveResponse

Surface

global

ResponseSurface

OptimizationAlgorithms:

Sensitivity Analysisallows best choice!

Which one is thebest?



18/32

18


Optimization Wizzard

optiSLang helps you to select a suitable optimization algorithm. Support

the underlying (automatic) selection process with some additionalinformation about the solver and the problem itself.

Exampel for using MOP and best_design_Sensitivity:

1. Set the analysis status as Preoptimized (best design from Sensitivity)

2. Set the constraint violations to Seldom

3. Set failed designs to None (MOP gives always response values)

4. Set solver noise to None (MOP gives a smooth surface)


19/32

19


Optimization Wizzard using MOP

Suggested algorithm is NLPQL

Start point is automatically selected

Press Next

1.

3.2.


20/32

20

Optimization with real design calls

After sensitivity und optimization on MOP the

user can continue with gradient-based, NOA-based optimization or ARSM optimization.

Minimum required user input: drag & drop the optimization module define objectives and constraints

Use of wizard to choose adequate optimizer Gradient based optimizer module NOA: Nature inspired optimization

containing evolutionary, genetic, particleswarm optimization

ARSM: Adaptive Response Surface Method



21/32

21

Robustness Design Optimization



22/32

22

Robustness = Sensitivity of Uncertainties



23/32

23

Robustness Evaluation

Minimum required user input: definition of input variation /scatter definition of robustness criteria number of samples for ALHS



24/32

24

Robust Design Optimization

P a r e t o

O p t i m i za t i o n

A d a p t i v e

R e s p o n s eS u r f a c e

E v o l u t i o n a r y

A lg o r i t h m



25/32

25

Update Functionality optiSLanginside ANSYS v14.5



26/32

26

optiSLang saves every design which was successfully calculated

using update at optiSLang container continue or reset can be chosenusing continue optiSLang only send unsolved designs

Continue crashed session option inside ANSYS


ANSYS Conference & 30th CADFEM Users Meeting,Kassel, October 24.-26, 2012


27/32

27

Recalculate failed designs

Due to different reasons design evaluations may fail With Recalculate Failed Design Points you can start them again




28/32

28

Interrupt, save, send & continue

Stop your analysis

At the end of the day

If licenses are not available

Save the Workbench project

Continue later

saves every successful design run!

external optimization using MOP possible!




29/32

29

Parallel evaluation using Ansys RSM

ANSYS RSM is the powerful tool to distribute jobs

optiSLang can fill the Workbench design table with a predefined numberof designs

ANSYS RSM organizes distribution of jobs

If ANSYS RSM is installed you only need to:

Choose RSM Mode

Set max. number of parallel jobs


ANSYS Conference & 30th CADFEM Users Meeting,

Kassel, October 24.-26, 2012


30/32

30

Update via Python scripting

ANSYS initialize per default an update mechanism, which updates

a complete ANSYS Workbench project

Mechanism can be overridden via python file

optiSLang provides this feature for optiSLang design evaluations

user has full access to his ANSYS model update


ANSYS Conference & 30th CADFEM Users Meeting,

Kassel, October 24.-26, 2012


31/32

31


ANSYS HPC Parametric PackoptiSLang inside Ansys Workbench v14.5

optiSLang Algorithm Settings

Select Use RSM Mode to enable parallel

design point submission

Set the Preferred Number of Design

Points in Parallel to the intended RSM jobsize


32/32

32

Flow simulation of LCDmanufacturing process

Problem Description Identify model inflow parameter to match outflow

Identified output velocities have to be in 10% error

ranges comparing with reference values

Input parameter: 10 pressure areas (100 design points)

Detail:

Pressure-Based solver, K-Epsilon Model with Standard Wall Functions

78.800 nodes, 66.900 hexahedral elements

Hardware: - Workstation with dual Intel Xeon E5645 (2,4 GHz,

12 Cores), 96 GB - all jobs running 1 core

Licensing Solution

1 Ansys Fluent

2 Ansys HPC Parametric Packs

Result/Benefit

~6,2x speedup over sequential execution

optislang-inside-ansys.pdf

Documents