ALTAIR Info-Tag Strukturmechanik 2015-3-24

24/03/2015

1

Innovation Intelligence®

Info-Tag Strukturmechanik

Christian Alscher

Kristian Holm

Copyright © 2012 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.

9:00 Registrierung und Kaffee

9:30 Begrüßung

Überblick: Aufgabenstellungen in der Strukturmechanik

Kaffeepause

Thermisch-mechanische Analyse am Beispiel eines Motorblocks

Materialmodellierung für unterschiedliche Werkstoffe

Strukturberechnung mit MKS-Lastbestimmung

13:00 Mittagessen

Composites: Modellaufbau, Berechnung und Auswertung

Workflow für Schwingungs- und Akustikanalysen

Möglichkeiten zur Strukturoptimierung

Diskussion

Agenda

24/03/2015

2


• High stiffness to weight ratio!

• Tailored properties!

• Fibers can be aligned to provide directional stiffness

• Thermal expansion control• Vibrational damping

• Desirable Material Properties!

• Fatigue and corrosion resistance

• Design and manufacturing flexibility:

fewer parts!

Concrete Carbon

Design Challenge:• High Number of Variables• Orthotropic Material Properties• Orientation of the Ply Angles• Stacking Sequence

Why Composites?

Copyright © 2012 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.Copyright © 2012 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.

• Each layer with at least one solid

• Large model, long CPU time

• High accuracy

• Mixed approach (middle layer thick)

• Shells for top and bottom layer

• Solid, thick shell or coehesive for middle layer

• Several solid layers to provide rotation

• Sandwich shell approach

• One shell element through the thickness

• Multiple layers, with different materials

solid

solid shell

shell

Modeling Techniques for Composites

24/03/2015

3


• One property required for each ply drop/add (i.e. Zone)

• Properties are not linked across zones �� Data duplication, No Direct Ply Shape

• No direction relationship to manufacturing process

• Example:

Plate requires three property tables

with repeating ply definitions

Delete P4 requires three updates

Zone 1 Zone 2 Zone 3 Zone 2 Zone 1

Zone 1 – Property Table

Ply Mat Thk Theta

P7 M1 0.01 45

P4 M1 0.01 0

P1 M1 0.01 45


Ply Mat Thk Theta

P7 M1 0.01 45

P5 M1 0.01 -45

P4 M1 0.01 0

P3 M1 0.01 -45

P1 M1 0.01 45


Ply Mat Thk Theta

P7 M1 0.01 45

P6 M1 0.01 90

P5 M1 0.01 -45

P4 M1 0.01 0

P3 M1 0.01 -45

P2 M1 0.01 90

P1 M1 0.01 45

P1 45P2 90

P3 -45

P4 0

P6 90P7 45

P5 -45

Composite Modeling Intro: Zone Based


• One stack (property) required per part

• A stack defines zones via ply shape definitions �� No Data Duplication

• Direction relationship to manufacturing process

• Example:

Plate requires seven ply definitions,

no repeating plies, each ply shape defined

Stacking plies automatically defines zones

Delete P4 requires only one update

P1 45P2 90

P3 -45

P4 0

Zone 1 Zone 2 Zone 3 Zone 2 Zone 1

P6 90P7 45

P5 -45

Stack Table

Ply Mat Thk Theta

P7 M1 0.01 45

P6 M1 0.01 90

P5 M1 0.01 -45

P4 M1 0.01 0

P3 M1 0.01 -45

P2 M1 0.01 90

P1 M1 0.01 45

Ply Based Composite Modeling

24/03/2015

4


Integrated into HyperMesh via Model Browser Dialogs

Ply Based Composite Modeling


• Objective: Visually verify engineering data associated with math model

• “Live” 3D representation of 1D/2D elements w/ or w/o layers

• Color “by thickness”

• Element normal

• Material direction

• Ply orientations

• Ply expansion

Composite Visualization

24/03/2015

5


• Individual Ply Results

• Strain Tensor Components

• Stress Tensor Components

• Principal Strain/Stress

• Failure Theories (Maximum

Strain, Hoffman, Tsai-Hill, Tsai-

Wu)

• Envelope Ply Results

• Min/Max/Extreme

• Sum/Average/Range (12.0)

• Identify Min/Max/Extreme Layer

Ply-Based Post-Processing


HyperWorks Solvers

Thermal and CFD

Highly Nonlinear

CrashSafety

Forming

Multi-bodyDynamics

OptiStruct RADIOSS MotionSolve AcuSolve

Optimization

Smart Multiphysics

FEKO

Electro-Magnetics

Statics

NVH

Thermal

Nonlinear

24/03/2015

6


Phase 1: Concept

Free-Size

Phase 2: Dimension

Size

Phase 3: Sequence

Shuffle

OptiStruct: Composite Optimization

• Objective: minimize mass (failure index, …)• Constraint: displacement (stress, …)• Manufacturing constraints:

• Min. and max. total laminate thickness• Min. and max. ply thickness• Min. and max. percentage of a fibre orientation• Linkage of thicknesses of plies• Constant thickness for a particular ply orientation


Phase 1: ConceptQ: Which fibre angles do I need where?A: Free-Size-Optimization creates a thickness

distribution for each fibre angle.

Phase 2: DimensionQ: How many plies of each shape do I need?A: Discrete Size-Optimization calculates optimal

number of each ply shape of each fibre angle.

Phase 3: SequenceQ: I which order do I have to stack the plies?A: Shuffle-Optimization finds optimal stacking order

under consideration of ply book rules.

Phase 1: Concept

Free-Size

Phase 2: Dimension

Size

Phase 3: Sequence

Shuffle


24/03/2015

7


Initial Design

Final Design

Composite Free-Size OptimizationWhat are the most efficient ply shapes?

Composite Size OptimizationHow many plies of each ply shape to meet engineering targets?

Composite ShufflingWhat is the exact stacking sequence to meet

manufacturing requirements?

Ply SlicingDetermines ply shapes



Finite element model representation

Optimal shapes/patches: free size optimization

Optimal number and thickness of plies: sizing

optimization

Ply stacking sequence optimization

�Peak failure index reduction: 20%�Displacement constraint satisfied�Minimum reserve factor met

Optimization of a Composite Sports Car Wing

24/03/2015

8


HyperWorks Solvers

Thermal and CFD

Highly Nonlinear

CrashSafety

Forming

Multi-bodyDynamics


Optimization

Smart Multiphysics

FEKO

Electro-Magnetics

Statics

NVH

Thermal

Nonlinear


RADIOSS - DALLARA Race cars design

24/03/2015

9


RADIOSS - DALLARA Race cars design


• Model: Helicopter floor structure

• Objective: Improve survivability of cabin crew under crash landing

• Loading: Crash landing Vi= 3-10 m/s

Helicopter Survivability

24/03/2015

10


Helicopter Survivability

• Model: Helicopter floor structure

• Objective: Improve survivability of cabin crew under crash landing

• Loading: Crash landing Vi= 3-10 m/s


Partner Caterham Composites

Lotus T127 Formel 1 Simulation dynamischer Seitencrash

24/03/2015

11


Gegenstand der Untersuchung

Seitencrash-Absorber aus Kohlenstofffaser-Kunststoff-Verbund

Laminatmaterial

• 2x2 Twill Gewebe

Paul Schischkin


Gegenstand der Untersuchung

Realversuch Seitencrash: Vorher-Nachher-Vergleich

Paul Schischkin

24/03/2015

12


Modellaufbau

Solver: RADIOSS V12.0.210

Vernetzung der Geometrie:

• Grundlage: CAD-Model

• 5x5mm 4-Knoten-Schalenelemente (quads)

• eine Schale pro Laminatdicke

• Anzahl: insgesamt ca. 7500 Elemente

Paul Schischkin


Modellaufbau

Elementeigenschaften (properties):

• Definition des Laminataufbaus durch den

Sandwich Schalen Ansatz

• Jede property erhält Informationen zu Anzahl

der Lagen, sowie ihre jeweilige Position, Dicke,

Orientierung und Material

Paul Schischkin

24/03/2015

13


Modellaufbau

Materialmodell:

Elasto-plastisch-orthotrop (exemplarisch)

Zugrichtung

Druckrichtung

Paul Schischkin

Zugrichtung

Druckrichtung


#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

/MAT/COMPSH/10

Altair composite material

# Init. dens. Ref. dens.

1.5000E-6 0

# E11 E22 NU12 IFLAG E33

42 40 .05 1 0.50

# G12 G23 G31 EPSF1 EPSF2

3.4 3 3 0 0

# ESPT1 EPSM1 EPST2 EPSM2 Dmax

0 0 0 0 .9999

# Wpmax Wpref Ioff

0 0 5

# C EPS ALPHA Icc

0 0 0 0

# sig_trac_1 B_1T N_1T SIGMA_1MAXT C_1T

0.1 25 .10 0 0

# EPS_1T1 EPS_2T1 SIGMA_RST1 Wpmax_trac_1

0 0 0 0


0.1 20 .10 0 0


0 0 0 0

# sig_comp_1 B_1C N_1C SIGMA_1MAXC C_1C

.0050 2000 .5 0 0

# EPS_1C1 EPS_2C1 SIGMA_RSC1 Wpmax_comp_1

0 0 0 0


.0050 2000 .5 0 0


0 0 0 0

# sig_12 B_12T N_12T SIGMA_12MAXT C_12T

.0040 83.0 .31 0 0


0.075 0.085 0.05 0

# GAMMA_INI GAMMA_MAX Dmax

1E31 1E31 .9999

# Fsmooth Fcut

0 0

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

Radioss Composite Material Model

Linear

Nonlinear

24/03/2015

14


#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

/MAT/COMPSH/10



1.5000E-6 0


42 40 .05 1 0.50


3.4 3 3 0 0


0 0 0 0 .9999

# Wpmax Wpref Ioff

0 0 5

# C EPS ALPHA Icc

0 0 0 0


1e+30 0 0 0 0


0 0 0 0


1e+30 0 0 0 0


0 0 0 0


1e+30 0 0 0 0


0 0 0 0


1e+30 0 0 0 0


0 0 0 0


.0040 67.0 .29 0 0


0.00 0.00 0.00 0


1E31 1E31 .9999

# Fsmooth Fcut

0 0

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

Composite Material Model: Yield Stress

Yield Stress

(((( )))) 1====σσσσF

1σσσσ

2σσσσ

t

1σσσσc

1σσσσ

c

2σσσσ

t

2σσσσ

t

12σσσσ

c

12σσσσ


#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

/MAT/COMPSH/10



1.5000E-6 0


42 40 .05 1 0.50


3.4 3 3 0 0


0 0 0 0 .9999

# Wpmax Wpref Ioff

0 0 5

# C EPS ALPHA Icc

0 0 0 0


1e+30 0 0 0 0


0 0 0 0


1e+30 0 0 0 0


0 0 0 0


1e+30 0 0 0 0


0 0 0 0


1e+30 0 0 0 0


0 0 0 0


.0040 67.0 .29 0 0


0.00 0.00 0.00 0


1E31 1E31 .9999

# Fsmooth Fcut

0 0

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

Composite Material Model: Plasticity

Plasticity, Strain Rates

24/03/2015

15


#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

/MAT/COMPSH/10



1.5000E-6 0


42 40 .05 1 0.50


3.4 3 3 0 0


0 0 0 0 .9999

# Wpmax Wpref Ioff

0 0 5

# C EPS ALPHA Icc

0 0 0 0


1e+30 0 0 0 0


0 0 0 0


1e+30 0 0 0 0


0 0 0 0


1e+30 0 0 0 0


0 0 0 0


1e+30 0 0 0 0


0 0 0 0


.0040 67.0 .29 0 0


0.00 0.00 0.00 0


1E31 1E31 .9999

# Fsmooth Fcut

0 0

#---1----|----2----|----3----|----4----|----5----|----6----|----7----|----8----|----9----|---10----|

Composite Material Model: Damage and Failure

Damage & Failure

plastic work

Wp


Failure models can be coupled with compatible material laws

Composite Material Model: Advanced Failure Models

24/03/2015

16


Puck failure criteria


Ergebnisse Paul Schischkin

24/03/2015

17





24/03/2015

18


CAE Process Chain

Phase 1:Free-Size

Optimisation

Phase 2:Sizing

Optimisation

Phase 3:Shuffle

Optimisation

Manufacturing and Design

ManufacturingSimulation

What is known ?

• Manufacturing process• Design idea• Available material (fabric, UD-layers, fiber/matrix)• Possible fiber orientations

In addition, we can answer:

• Is it possible to manufacture my design at all?• What is the influence of the manufacturing on the fiber orientation?• Do I have local thickenning or thinning?

What is not known ?

• Distribution of material and fiber orientation • Final lay-up of the laminate


HyperMesh: Drape Estimator for Composite Fibers

• Calculate draping angles and thickness variation

• Geometry based, very fast (seconds)

• Good estimation for deviation of the fiber oriention,

but feedback on manufacturability only indirect

material/fiberdirection

24/03/2015

19


Radioss: Draping Simulation

Sandwich method

1 Component1 Material1 Property

Sandwich material law

• All plies defined in the same property

Independant layers method

1 Component for each Ply1 Material for each Ply1 Property for each Ply

Fabric material law

• All layers definedindependently

• Contact interface betweenthe layers

Setup in HyperFormuser process


Sliding

• Sliding effect between layers can be modeled by contact interface

• Each layer is free to slide over other layers

• In this example, fiber orientaions are 0, +45, 90 and -45°regarding X axis

• The behavior during stamping is different for each layer

Radioss: Independent Layers Simulation

24/03/2015

20


• Fiber orientation

• The rotation of fibers for each layer can be displayed in HV

Sandwich Method

Composite Forming


Radioss: Independent Layers with Resin

Independant layers + resin method

Connect material is used to model resin :

• A plasticity domain is reached after a small yield stress value• Strain rate dependancy available for user defined curves

Independent FiberLayers

Shell elements

Warm ResinSolid elements

Fibers + ResinShell + solid withcoincident nodes

24/03/2015

21


Altair Partner Alliance: Composites Offering

The composite offering from the Altair Partner Alliance reaches across many applications.

Working in tandem with OptiStruct and RADIOSS, material modeling, failure modeling,

optimization with failure constraints, composite panel analysis and composite panel detail

stress analysis can all be performed effectively.



9:30 Begrüßung


Kaffeepause




13:00 Mittagessen




Diskussion

Agenda

24/03/2015

22


Workflow für Schwingungs-und Akustikanalysen

HyperWorks Solver Technology


Noise and Vibration - workflow

Thermal and CFD

Highly Nonlinear

CrashSafety

Forming

Multi-bodyDynamics


Optimization

Smart Multiphysics

FEKO

Electro-Magnetics

Statics

NVH

Thermal

Nonlinear

24/03/2015

23


Model Build

Analysis Root Cause Analysis

SensitivityFocused

Optimisation

MDO

ValidationDelivery

- Modular Construction

- Flexible Assembly

- Parametric

- Reporting- Targets / Metrics- Model reduction- Rating /

Appraisal

- TPA- Participation- Force vs TFs- Test forcing

- What if- Sensitivity- Robustness- DOE

- Correlation- Performance- Issue Resolution- Mass / Cost

- Size- Gauge- Shape- Topography- ERP

- Common Model Parameters

- Design Constraints



Efficiency gains through integration

61

24/03/2015

24


Improve simulation efficiency

CMS super element method

� Reduction of degrees of freedom

� Structural/Fluid or Combined Fluid-Structural SE

� Both Craig-Chang and Craig-Bampton methods

� General Method (combined free and fixed

attachment points)

� Participation factors (PFCMS)

FRF based super element

� Compliments CMS based approach

� Faster residual run solution in the medium

frequency range (up to 500 Hz)

� component dynamic stiffness (CDS) - OptiStruct

AMSES (Automatic Multilevel Substructuring

Eigen Solver) - OptiStruct



• CDS superelement technology in OptiStruct

• Reduces runtimes for variants to seconds

NVH parametric study (e.g. bushings)

overnight


24/03/2015

25


Model Build


SensitivityFocused

Optimisation

MDO

ValidationDelivery


- Flexible Assembly

- Parametric


Appraisal









Problem Response

What’s Moving?

What If…?

Which Mode?

Which Panel?

Which Paths?

What Power Flow?

What DSA?What

Energy?


24/03/2015

26


Diagnostic tools

Vib

ratio

nN

oise

Low Frequency Mid Frequency

Transfer Path AnalysisGrid participation

Energy consideration

Power Flow

Design Sensitivity Analysis

Modal participation

N&

VNoise and Vibration - workflow



Sensitivity filtering

� Optimization on all possible parameters in full vehicle model is to large

� Detection of sensitive components or parameters (use normalized variables)

� Size variables; gauge, stiffness, damping, materials

24/03/2015

27


Model Build


SensitivityFocused

Optimisation

MDO

ValidationDelivery


- Flexible Assembly

- Parametric


Appraisal









• Frequencies

• FRF

• Forces, Stress, von-Mises Stress

• Functions: max, average, sum, …

• Normal displacement, velocity, acceleration,

• Frequency sub-ranges

• Mode shape

• PSD and RMS of displacement, velocity, acceleration, pressure, stress, strain

• Acoustic pressure

• ERP (Equivalent Radiated Power)

• Equations

• External routines (DRESP3 -> Fortran, C, HyperMath)


24/03/2015

28


� Min(max)

� Min (max) with frequency sub ranges

� System identification

mjg

T

Txfwwith

or

mjg

T

Txfw

j

i

iii

j

n

i i

iii

,,10)(: Subject to

)( : Minimize

,,10)(: Subject to

)(: Minimize

1

2

K

K

=≤

≤−

=≤

−∑

=

x

x

ββ

[ ]mjg

ffixd

j

ni

,,10)(: Subject to

,...,)(max: Minimize 0

K=≤

=

x

[ ]mjg

ffixd

j

subsubi

,,10)(: Subject to

,...,)(max: Minimize 21

K=≤

=

x

Curve Fitting



� Trimmed body optimisation using screening techniques

TPA Result (Vehicle Model)

+

Full FE Tbdy Model

Panel Participation Plot at peak frequencyModal Participation

Identify peaks

Component Gauge + Topography optimisation

Panel Set as design Variables


24/03/2015

29


� Trimmed body optimisation using screening techniques

Response Curves

� Minimise Mass with constraint on Response (157Hz)

� 3dB improvement at peak of interest

� 40% reduction in mass of panel set

� Topography result to lead bead pattern design loop

� 35hr Run time for 30 iterations (12 cpus, AMSES)

Gauge Results

Topography Results



• Use topology optimisation to design an efficient

engine cover

• Minimise radiated noise for a given volume of

material:

• Equivalent radiated power

• Σ area * normal velocity2 * constant

• Use CMS superelement of non-designable

structure

Designable Material

12400 nodes included in ERP response


24/03/2015

30


• Concept design for internal reinforcement

structure

• Optimised structure reduces ERP by 30%

• Run time 2.5hrs (laptop)

• Memory 8Gb



Fast and Accurate Analysis

Model Quality – Flexibility – Time Reduction

Root Cause Understanding

Problem Diagnostics – Transfer Path – Sensitivity

Robust Optimization Algorithms and Processes

Problem Definition – Method selection – Reduction Techniques –

Multi Disciplinary


SensitivityFocused

Optimisation

MDO

ValidationDeliveryModel

Build

OptiStructHyperStudy


24/03/2015

31



9:30 Begrüßung


Kaffeepause




13:00 Mittagessen




Diskussion

Agenda



HyperWorks Solver Technology

24/03/2015

32


Cost

Weight

Performance

Performance / Weight / Cost Balancing


Optimization enhances the comprehensible Design Space

New Weight Target

PhysicalDesign Limit

Compreh.Design Limit

Current DesignPoint

Weight

Perf

orm

an

ce

Effort / Cost

Weight

Optimization

24/03/2015

33


Structural Optimization

Thermal and CFD

Highly Nonlinear

CrashSafety

Forming

Multi-bodyDynamics


Optimization

Smart Multiphysics

FEKO

Electro-Magnetics

Statics

NVH

Thermal

Nonlinear


Structural Optimization

OptiStruct

HyperStudy

Syst

em

Le

ve

l

C

om

po

ne

nt

Leve

l

Concept Design Detailed Design

INSPIRE

24/03/2015

34


NVH Optimisation Methodology C

once

pt P

hase

Det

aile

d D

esig

n

System Level Subsystem/Component Level

TopologyHigh number of DV

Global stiffness responsesDesign Space

SizeMedium number of DV

NVH responsesConcept Model

Topology/TopographyHigh number of DV

NVH responsesReduced Model

Size/ShapeLow number of DV

NVH responsesReduced ModelSize/Shape

Medium number of DV NVH responsesDetailed Model

Body Gauge Reduce mass

Topography Panel Acoustic Response

Size/Shape Mount LocationsMode Tracking

Topology Concept Structure and

loadpath definition

FreeSizeHighlight reinforcement

patterns


Optimization – State of the Art Technology Today

..... ..... .....

15% 20%.....

23%

Calculating weight saving Calculating weight savingCalculating weight saving

FE-Based optimization technology

topology, topography, sizing, shape,

staking, free-shape, multi-disciplinary...

24/03/2015

35


Moving in Direction of Strategic Optimization

What is strategic optimization?

1. Simultaneous consideration of all important functional

requirements

2. Application on subsystems, systems and global level

3. Deployment of optimization early in development

(Optimization drives Design)

4. Direct connection to development program

(and mass)

(management)


Optimization of Assemblies – getting Strategic!

.....Calculating weight saving .....Calculating weight saving

20% 45%

FE-Based optimization technology

topology, topography, sizing, shape,

staking, free-shape, multi-disciplinary...

24/03/2015

36

86

Optimization of Assemblies

• Multi-model optimization (MMO) aims to simultaneously optimize

multiple parts or configurations with common design variables

• Simplify the iterative design process associated with optimizing multiple

parts, especially when conflicting requirements exist

• Models may be entirely different or share identical parts

• Models may be assigned individual objectives and constraints, while

global responses may also be defined

• Performance and scalability are taken into consideration

87


• Similar models with different meshes

(e.g. coarse and fine)

• Similar models with subcase-dependent

configurations or characteristics (e.g.

damping)

• Different models sharing identical

designable parts

• Different models connected at

designable locations

• Different models with combined

objectives or constraints

• Any combination of the above

W1 W2

min (W1+W2)

24/03/2015

37


Single Model Optimization Multi-Model OptimizationOptimization Problem

OptiStruct: Multi-Model Optimization



95%

100%

105%

110%

115%

120%

125%

130%

135%

140%

93% 94% 95% 96% 97% 98% 99% 100% 101% 102%

Tors

ion

Mo

de

(H

z)

BIW Weight (Kg)

Modal Performance v BIW Weight

Target

Optimum Modal

Performance versus

Weight


Require Major Architectural

Changes

Risk

• Modal Performance versus Weight Study

• Study creates visibility on relations between performance and weight

• Example is showing a design limit not known to customer

24/03/2015

38


Optimization - Fundamentals

Robust and Repeatable Analysis

Predict the Trend

Fast and Accurate Analysis

Performance and Scalability

Robust Optimization Algorithms and Processes

Convergence and Efficiency


Altair Solver Technology

Multiphysics Analysis and Optimization

Structural Analysis

Manufacturing Simulation

Systems Simulation

Fluid Dynamics

ThermalAnalysis

Crash, Safety, Impact & Blast

Electro-Magnetics

24/03/2015

39


Altair HyperWorks: Performance Enhancing Software

Documents

ALTAIR Info-Tag Strukturmechanik 2015-3-24