The Concept behind modeFRONTIER

modeFRONTIER® is a registered product of ESTECO srlCopyright © ESTECO srl 1999-2007

For more information visit:

www.esteco.com or send an e-mail to:

[email protected]

The concept behind modeFRONTIER

Some conceptual fundaments for introducing the modeFRONTIER design environment

mailto:[email protected]




[email protected]

Defining modeFRONTIER

modeFRONTIER is a multi-objective optimization and design environment,

written to allow easy coupling to almost any computer aided engineering

(CAE) tool, whether commercial or in-house.





[email protected]

The concept behind modeFRONTIER

Input Variables:

Entities that define the

design space.

Output Variables:

Measures from the system

The Black Box:

Generates the outputs

accordingly to the inputs





[email protected]

The input variables

Variables:

Variables are the free parameters, i.e. the quantities that the designer can vary or the choices the designer can make.

Continuous variables:

• point coordinates

• process variables

Discrete variables:

• components from a catalogue

• number of components





[email protected]

The Black Box

The black box can be:

• A set of solvers that models and solves in a numerical

manner the design problem (e.g. CAD/CAE tools)

• A set of experiments that produces some data

CAD(CATIA, UGS,

PROE)

CFD(StarCD,

Fluent, CFX)

FEM(Nastran, Ansys,

Madymo, etc)

Others(In-House codes,

MATLAB, Excel)

Multi-disciplinary Scenario





[email protected]

The output variables are a measure of the system response and/or performance, i.e.:

acceleration, speed, consumption, confort,…

deformation, stress, mass, volume,…

lift, drag,…

defects, number of failures, cost,…

….

The output variables





[email protected]

Input Variables:

Entities defining

the design space.Output Variables:

Measures from the

system

The Black Box:

(ANSYS, FLUENT, Workbench,

MatLab, etc…)

Optimization means to find a set of system configurations (input

variables) that meets the objectives and satisfy the constraints

ObjectivesConstraints





[email protected]

The optimization can be

multiobjectives

Many softwares can be

used to describe the

behavior of the system

under exam

CAE1

INPUT

OBJECTIVES

CONSTRAINTS

CAE2

OUTPUT

OPTIMIZER

Black box





[email protected]

Multidisciplinary Optimization - MDO

Discipline 1

Discipline 2

Discipline 3

Discipline 4

Discipline 5

Discipline …x1

x2

…

xn

Objective 1

Objective 2

…

…

Objective k

Multiobjective optimization, Approximation methods, Sensitivity Analysis, Space exploration, Multivariate Analysis





[email protected]

EMO and its fields*

aerospace & defence

appliance

automotive

chemical

construction

electronics

healthcare equipment

industrial equipment

marine & offshore

materials & processes

others

Leisure &

Sport

*On a database of 148 real world applications





[email protected]

Leisure & Sport Applications

• Study of innovative solutions of a cycle wear crotch pad for Campagnolo

• The optimization considered the ergonomic level of the crotch pad function

of both geometry and materials.





[email protected]

Simplified product analysis Decrease in cost

Alternative materials Stiffness and energetic absorption

Ergonomic/Marketing Independence by anthropomorphic features

Ergonomic improvement Shape and thickness modification

Objective: Product innovation for high performances

Innovation – HOW ?

Technologic issue (cost) – appearance/ Marketing – performances





[email protected]

A

B

C

Parametric models carrying outObjective : Ergonomic improvement Shape change, thickness, material

Assessment of the critical zone





[email protected]

Parametric models carrying out

Objective: Ergonomic improvement Shape, thickness, weight change





[email protected]

Geometric parametersSimulation CAD parameters. These parameters have been implemented in modeFRONTIER.





[email protected]

Optimization process

Workflow

Ansys node - WorkBench

Independent variables

Excel node

Dependent variables

Objectives





[email protected]


The optimization process:

• Input variables: Geometry, materials mechanical properties

• Objectives: Minimization of back and front cushions volume, minimization

of pressure distribution and its maximum value.

• DOE: Sobol (50 designs initial population )

• Optimisation algorithm: MOGA II





[email protected]


1810





[email protected]

Assos F13

Optimization – Assos F13

Des. 1810

Comparison

between initial

configuration and

optimal points

http://www.campagnolosportswear.com/





[email protected]

Graph of relative drag difference between a cyclist using a rear wheel with and without a disk in a range of crosswinds

Example of Applications: Biomechanics





[email protected]

Example of Applications: Biomechanics





[email protected]

Bozen

Verona

München

Innsbruck

• Two single track railway tunnels connecting Fortezza (Italy) and Innsbruck (Austria).

• Tunnel section: 72,4 m2

• Length: 56 km

• Up to 1650 m under the Alps

• Tunnels about 70 apart, connecting galleries every 330 meters

The problem:

• Uncertainities of data on the mechanical behaviour of the rock mass

• Passing from south to north, the tunnel will be drilled throug granite, paragneiss, schist, gneiss, marble, phylite.

• It also crosses the Periadriatic Seam, caused by the collision of the African plate and the European Continent

Constructions - Brenner Railway Base Tunnel





[email protected]

Time and Cost ReductionmodeFRONTIER tasks

• Reliability analysis (general) – and documentation

• Reliability analysis with respect to rock models (behaviour of the mass during borging)

• Senario and decion on the best excavation method.

Examples – Constructions - Transportation





[email protected]

Automotive Application

The aim of this activity is to OPTIMIZE CONFLICTING ASPECTS in terms of

Handling performances as well as Ride&Comfort performances.

The study results in a set of vehicle set-up, concerning suspension vertical and longitudinal

stiffness, elasto-cinematic behavior, optimizing both aspects without forgetting the robustness of

the solution.

Stability and response of the vehicleUndersteer

Side-slip angle

Rolling

Yaw speed

Comfort for driver and passengersPeak accelerations

Time of dissipation after impact

RMS of low frequency accelerations

on uneven road, highway, obstacles

Courtesy of Fiat





[email protected]

Full-vehicle MSC.ADAMS/Car Models

Rear Suspension

Front Suspension

Assembly Comfort

Front Suspension (incl. flexible

subframe)

Rear Suspension

Steering

Antirollbar

Engine

Front&Rear Tires

Rigid Body

Assembly Handling

Front Suspension (incl. flexible

subframe)

Rear Suspension

Steering

Antirollbar

Conceptual Driveline

Front&Rear Tires

Rigid Body





[email protected]

Definition of Input variables

Input variables:

Spring Stiffness and preload

Bumpstop clearance and characteristics

Anti-roll-bar diameter

Damper characteristics

Bushing characteristics

Vehicle parameters able to influence both the Ride-Comfort and Handling performance

VERTICAL STIFFNESS AND

ROLLING STIFFNESS

VERTICAL DAMPING

ELASTO-CINEMATIC

CHARACTERISTICS

LONGITUDINAL

STIFFNESS AND DAMPING





[email protected]

Objectives:

- Key synthesis Handling parameters(understeer, sideslip curve, yaw, rolling - gains, time

delays)

- Key synthesis Comfort parameters (peak

accelerations, time dissipations, RMS/RMF)

Objectives and Constraints

Constraints:

- Ride height in various load conditions

- Feasibility of the components for

.ex. rate between axial and radial bushing

stiffness, damper characteristics, bumpstop

length and characteristics etc.

- Top mount stiffness for damper

efficiency

- Performance constraints

StdA

P.C.Tarature ammortizzatore posteriore 263

confronto con tarature X250

-2000

-1500

-1000

-500

0

500

1000

1500

2000

2500

3000

3500

4000

-3000 -2500 -2000 -1500 -1000 -500 0 500 1000 1500 2000 2500 3000

velocità [mm/s^2]

Fo

rza

[N

]

REAR_MVmuletto validato CRF

ISO SMORZANTE 223

Minimo realizzabile per 263 conammortizzatore attuale

MT104

92

93

94

95

96

97

98

99

100

101

0 200 400 600 800 1000 1200

Rigidezza verticale [N]

Eff

icie

nza a

mm

ort

izzato

re [

%]

frequency response - lateral acceleration / steering angle

80 km/h 0.45g

0.0000

0.0020

0.0040

0.0060

0.0080

0.0100

0.0120

0.0140

0.0160

0.0180

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

AY

/DV

OL

- g

ain

(g

/de

g)

263 Muletto calcolo pneumMule (195 "Stilo")

223 sperim pneumMule (195 "Stilo")

263 target

263 7q rev4 pneumMule (195 "Stilo")

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

AY

/DV

OL

- p

ha

se

(s

)

frequency response - yaw rate / steering angle

80 km/h 0.45g

0.000

0.050

0.100

0.150

0.200

0.250

0.300

0.350

0.400

0.450

0.500

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

PS

IP/D

VO

L -

ga

in (

1/s

)



263 target


-0.14

-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

0

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

PS

IP/D

VO

L -

ph

as

e (

s)

frequency response - side slip angle / steering angle

80 km/h 0.45g

0.0000

0.0100

0.0200

0.0300

0.0400

0.0500

0.0600

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

BE

TA

/DV

OL

- g

ain

(d

eg

/de

g)



263 target


-0.4

-0.35

-0.3

-0.25

-0.2

-0.15

-0.1

-0.05

0

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

BE

TA

/DV

OL

- p

ha

se

(s

)

frequency response - lateral acceleration / steering angle

80 km/h 0.45g

0.0000

0.0020

0.0040

0.0060

0.0080

0.0100

0.0120

0.0140

0.0160

0.0180

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

AY

/DV

OL

- g

ain

(g

/de

g)



263 target


-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

AY

/DV

OL

- p

ha

se

(s

)

frequency response - yaw rate / steering angle

80 km/h 0.45g

0.000

0.050

0.100

0.150

0.200

0.250

0.300

0.350

0.400

0.450

0.500

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

PS

IP/D

VO

L -

ga

in (

1/s

)



263 target


-0.14

-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

0

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

PS

IP/D

VO

L -

ph

as

e (

s)


80 km/h 0.45g

0.0000

0.0100

0.0200

0.0300

0.0400

0.0500

0.0600

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

BE

TA

/DV

OL

- g

ain

(d

eg

/de

g)



263 target


-0.4

-0.35

-0.3

-0.25

-0.2

-0.15

-0.1

-0.05

0

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

frequency (Hz)

BE

TA

/DV

OL

- p

ha

se

(s

)


100 km/h - 0.4 g

0.0000

0.0100

0.0200

0.0300

0.0400

0.0500

0.0600

0.0700

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5frequency (Hz)

BE

TA

/DV

OL

- g

ain

(d

eg

/de

g)

Option 1

Option 2





[email protected]

Coupling modeFRONTIER & ADAMS/Car

Example Process includes modifying and launching

3 models (K&C, Assembly Handling e Comfort),

7 analysis (4 K&C, 2 Handling e 1 Comfort).

Every run requires approx. 5min => weekend 2.5ggr = about 800 run

MSC.ADAMS CarmodeFRONTIER

INPUTS

CONSTRAINT CHECK - ANALYSIS, POST-PROCESSING

INPUT MODEL, ANALYSIS,

POST-PROCESSING -COMMAND FILES ADAMS

OUTPUTS => COMPARISONS OBJECTIVES AND

CONSTRAINTS

Model modifications

Results simulation





[email protected]

DOE Study and Optimization Method

Influence Study: 8 Input variables, 4 Targets for example

DOE Study Sobol/Full-factorial => Excluding input variables (and constraints/objectives) + Adapting range of study

Optimization with limited numbers of variables, objectives, constraints - real or virtual response surfaces

Pareto FRONTIER => Selection of “optimum” solutions related to the particular project vehicle target setting

Verification of optimum solutions belonging to Pareto FRONTIER





[email protected]

The geometric model built-up in Catia V5 has been imported into ANSYS

WorkBench 11.0.

Centine (Ribs)

Superfici Alari (Skins)

Longheroni a C (Spars)

Aerospace Application

Courtesy of Alenia Aeronautica





[email protected]

Subdivision of the wing into 6 parts

In this way it is possible to reduce the number of skins while getting closer

to the tip of the wing (more efficient optimization process)

Different values of skin and caps

structural parameters between the

wing underside and the top surface

With the aim to get the best

material performances

The input variables values (thickness, skins number, …) are constant within every 6

parts

Geometric model set-up





[email protected]

Optimization strategy

The optimization process has been sub-divided into the 2 following phases:

1. Firstly, the whole design space has been explored with the scope to get the

global optimal solution. This initial search exploited the MOGA-II

2. In a second step, the more important input parameters have been further

investigated, while the remaining ones have been fixed to constant values.

This approach enabled to get more accurate solutions. In this phase both

MOGA-II and B-BFGS (hybrid approach) have been used.





[email protected]

Workflow modeFRONTIER

DATA FLOW

17 input variables + 67 constants

29 output variables

1 objectives

30 constraints

CPU TIMEAround 680 analyses

10’ per run

around 4.5 days

LOGIC FLOW

DOE: 4 best designs fase 1

Optimizer: B-BFGS

DOE: 16 best designs B-BFGS

Optimizer: MOGA-II





[email protected]

The flexural and the total deformation (torsion and flexural) are depicted. In

both cases (positive and negative nz) the maximum deformation values belong to

the feasibilty domain.

Wing tip deflection (1989.3 mm)

“positive nz”

Total deformation

(torsion < 7°)

Results


The Concept behind modeFRONTIER

Documents

Transcript of The Concept behind modeFRONTIER