Download this Presentation - Altair Technology Conference 2013

Cassidian Air Systems 

Integrated CAE Solutions 

Multidisciplinary Structure Optimisation with 

CASSIDIAN LAGRANGE 

2011 European HyperWorks Technology Conference 

Dr. Fernass Daoud, Head of Design Automation and Optimization

Contents 

Dr. Gerd Schuhmacher 

Contents 

• Introduction: Cassidian Air Systems 

• Motivation: Challenges & Opportunities of the Airframe Design Process 

• Multidisciplinary Airframe Design Optimization at Cassidian Air Systems 

� Traditional Airframe Design vs. automated Multidisciplinary Design Optimization 

� Multidisciplinary Airframe Design Optimization Procedure LAGRANGE 

• Applications 

� Overview on past applications 

� Application to the Unmanned Aerial Vehicle Talarion 

• LAGRANGE profile in Hypermesh 

• Benefits of the Automated Airframe Design Process 

© 2010 CASSIDIAN - All rights reserved Page 2

Introduction 

Airbus 


Introduction: EADS and Cassidian Structure 

Tom Enders (CEO) 

Fabrice Brégier (COO) 

Airbus Military 

Domingo Ureña-Raso 

MBDA 

A. Bouvier 

Eurocopter 

Lutz Bertling (CEO) 

Cassidian 

Systems 

H. Guillou 

Cassidian 

Stefan Zoller 

(CEO) 

Cassidian 

Electronics 

B. Wenzler 

© 2010 CASSIDIAN - All rights reserved Page 3 

EADS Astrium 

François Auque (CEO) 

Cassidian 

Air Systems 

B. Gerwert 

EADS 

Divsions 

Cassidian 

Business 

Units

Contents 


Contents 












Motivation 


Motivation 

Challenges of the Multidisciplinary Airframe Design Process 

• The aircraft design process requires the combination of a broad spectrum of 

commercial as well as company specific analysis and sizing methods: 

– a broad spectrum of A/C (company-) specific strength and stability analysis 

methods 

– company specific aerodynamic and aero-elastic / loads analysis methods 

– company specific composite analysis, design and manufacturing methods. 

• The aircraft design is driven by a huge number of multidisciplinary design 

criteria (manoeuvre, gust and ground loads, aeroelastic efficiencies, flutter 

speeds, strength and stability criteria, manufacturing requirements etc.) 

resulting from different disciplines (loads, flight controls, dynamics, stress, 

design, etc.) 

• The design process needs to consider and meet all these design driving 

criteria simultaneously, in order to determine an optimum compromise 

solution, i.e. all disciplines and multidisciplinary design criteria driving the 

airframe structural sizes and the composite lay-up need to be combined and 

have to interact within an integrated airframe design process. 


Contents 


Contents 












Traditional Airframe Design vs. automated Multidisciplinary Design Optimization 

Automation of the Global Airframe Development Process 


1. Aerodynamic Analysis 

and Loft Optimization 

2. Loads Analysis 

and Aeroelastics 

Load Loops 

3. Structural Analysis 

and Sizing 


Aerodynamics Optimization 

LAGRANGE: 

Automation of Loads 

and Sizing Loop 

Structures + Loads (2007-11) 

Structural Optimization (until 2006) 

Aerodynamic Shape + Structural Sizes (2012+)


Structural 

Concept 

Material 

Selection 


Key- 

Diagram 

Property Selection 

1. Loop: (guess / engineering judgement) 

Traditional Sizing Process 

Aerodynamic Loft, 

Inboard Profile 

Panel Model 

Mass Loads 

Loads Loop 

GFEM*) 

Property Update 

for follow up Loops 

modified for Dyn. 

Sizing Loop 


SDC Requirements, 

Parameter envelope 

Manual Sizing 

(global / local 

strength & stability checks) 

Check Stress 

*) No mass application => no correlation between stiffness & masses 

**) Dynamic landing and dynamic gust loads. 

NOT SYNCHRONISED with static loads 

**) 

FCS 

Dynamic 

Model 

Dynamic 

Loads 

Flutter 

not 

fulfilled 

Aerodynamics, 

Preliminary Design 

Design 

Stress 

Loads 

Dynamics 

Mass



Multidisciplinary Airframe Design Optimization Process 

Structural 

Concept 

Key- 

Diagram 

Material 

Selection 

Mass 

Property Selection 

Aerodynamic Loft, 

Inboard Profile 

Panel Model 

Dynamic 

Nodal loads 

GFEM**) 

Prelim. Dynamic 

Checks 

1. Loop : (acc. to mass breakdown) 

Flight Physics 

Loads 

select design 

driving manoeuvres 

*) Within Block 0 dynamic nodal loads will be 

provided externally. Integration into process ongoing 

**) stiffness & masses in correlation 

***) 1st Global Sizing Loop performed 

MDO 

(Loads, 

Stress, 

Dynamics) 

Balanced 

GFEM 

***) 


Dynamic 

Check 

Property Update 

Manual Sizing 

(global / local 

for prob. follow up Loops 

Check Stress 

strength & stability checks) 

Loads 

Check 

Checks with tools 

from traditional 

Sizing Process 

Aerodynamics 

Design 

Stress 

Loads 

Dynamics 

Mass

Contents 


Contents 












Multidisciplinary Airframe Design Optimization Procedure LAGRANGE 

Eurofighter 

Composite Wing & Fin 

Talarion 


A400M Rear Fuselage 

& Cargo Door 


Large Spectrum of applications in Military and Civil Aircraft Design 

X31 CFC Wing 

Multidisciplinary Analysis and Optimization Software Tool 

FEM 

Statics 

Aeroelasticity 

Steady Flutter Gust 

Dynamics 

Stability 

Aerodynamics 

Doublet 

Lattice 

Higher 

Order 


Skill Tools 

Strength 

Buckling 

Postbuckling 

Design for 

Manufacturing 

Composite Design 

Models & Manufact. 

Constraints 

• Developed by CASSIDIAN Air Systems since 1984 

• More then 140 man-years of development 

A380 Leading Edge Rib 

A350 Wing 

A350 Fuselage


Multidisciplinary Analyses within the Optimization Process 

• Simultaneous consideration of aiframe design driving disciplines during 

analysis and optimisation: 

Stress 


Optimisation 

Optimisation model 

Criteria model 

Optimisation runs 

Manufacturing 

Minimum & maximum 

thickness / dimensions 

Composite manuf. rules 

Thickness jumps, etc. 

Dynamics 

Frequency 

requirements 

Flutter speed 

DLM model 

for unsteady 

aeroelasticity 


Stressing criteria 

(strength & stability) 

On basis of GFEM 

(with mass data) 

Aeroelasticity 

Aeroelastic 

efficiencies 

Flutter speed 

Gust 

Aeroelastic 

requirements 

Loads 

Selected design 

driving manoeuvres 

including flight 

conditions for each 

manoeuvre 

Aerodynamic 

HISSS model 

Coupling model 

(Beaming)


hh 

bb 

Shear walls & longerons 

• Cross-sectional dimensions 

• Skin thicknesses 


Structural Components to be optimized 

bb 

ee 

cc 

DD 

Composite & Metallic Stringer 


• Ply thicknesses 

Stringer-stiffened panels: 


• Skin thicknesses 


bb 

ee 

11 

Composite & Metallic Skin: 

• Ply thicknesses / fibre orientation 

e.g. composite skin (wing, fuselage, 

taileron) 

Metallic frames: 

• Cross-sectional dimensions



Multidisciplinary Analysis Types 

• Multidisciplinary structure optimisation (variable structure & variable loads) 

… 

Criteria Model 

HISSS 

Analysis Model 

Linear Statics 

Linear Dynamics 

Steady Aeroelastics 

Unsteady Aeroelastics 


FEM 

Optimisation Model 

…



Design Variables and Design Criteria 


… 


HISSS 







hh 

bb 

bb 

ee 

cc 

DD 

• Omega profile 

• T profile 

• Box profile 

• C profile 

• LZ profile 

bb 

ee 

� Geometric Sizes + 

Composite-Lay-up 

FEM 

11 


Parametric model defining the design 

variables: 

• Cross-sectional area of bars (sizing) 

• Thickness of shell or membrane 

elements (sizing) 

• Ply thickness of composites (sizing) 

• Fibre orientation in composite stacks 

(angles) 

• Coordinates of FE nodes (shape) 

• Coordinates of control points (CAD, 

NURBS) 

• Trimming variables (angles of attack)





… 


HISSS 







FEM 



variables: 






(angles) 



NURBS) 






… 


HISSS 







FEM 



variables: 






(angles) 



NURBS) 







Strength (diverse models 

depending on failure criteria) 

Analytical buckling analysis (also 

theory 2nd order) 

Local stability analysis (for critical 

parts of cross-section), crippling 

Displacements (stiffness 

requirements) 

Constraints for natural frequencies 

aeroelastic requirements (steady, 

unsteady): efficiencies, flutter, etc. 

Manufacturing constraints 

HISSS 

Constraints for trimmed flight and 

landing manoeuvres 






comp 

� � �xxxx�s 


allow 

tens 

� � �yyyy�s 

allow 

FEM 

Damage Tolerance & Repairability 

• Yamada-Sun 

• Puck 

• Tsai-Hill 

• …. 


… Parametric model defining the design 

variables: 






(angles) 



NURBS) 














requirements) 





HISSS 









w eff 

• Skin & Column Buckling 

for isotropic, orthotropic 

and anisotropic skins 

FEM 

• Post-Buckling for isotropic 

and composite structures 



variables: 






(angles) 



NURBS) 














requirements) 





HISSS 









FEM 



variables: 






(angles) 



NURBS) 


Contents 


Contents 












Overview on past applications 

Eurofighter (�1985) 

Composite Wing & Fin 

Trainer Wing (2000) 

Composite Wing& Fin 


Overview on past applications 

X-31A Wing (1990) 

Composite Wing 

A400M (2004-2006) 

Rear Fuselage Skin+Frames 


Stealth Demonstrator (1995) 

Full A/C Design 

Advanced UAV (2006 + ) 

Composite Wing

Application of MDO at Cassidian (2) 

A350 XWB VTP Optimisation 

A30X Wing Optimisation 

Topology & Sizing Optimization 

of Sec. 19, A350 

• > 20 % weight reduction ! 

Page 27 


Overview on selected Applications 

Optimum Composite 

Sizing Layout within 2 

Month (MAS- 

Acquisition phase) 

• Optimum Composite Sizing 

of several variants with 2 

FTE * 12 Month (AI UK) 

• > 35 % weight saving 

(compared to AL-design)! 


A350 XWB Wing Optimisation 

ca. 3000 DV 

250 000 Constraints 

Aeroelastics 

• Optimum Composite 

Sizing of 40 Variants with 

3 FTE * 6 Month for AI 

Toulouse 

• ca. 20 % weight saving ! 

A350 XWB Fuselage Optimisation Sec. 13-14 

ca. 15000 DV 

1000 000 Constraints 

A380 Leading Edge Rib Optimization 

• Optimum Composite Sizing 

with 2 FTE * 5 Month 

• Feasible Design without 

weight increase ! (PAG) 

• > 40 % weight reduction !

Contents 


Contents 












Application to the Unmanned Aerial Vehicle Talarion 



Unmanned surveillance and reconnaissance aircraft 

Appr. Dimensions: Length: 14 m; Height: 4,5 m ; Span 26 m 

Take-off weight : 8000 kg class 

Performance 

Loiter Speed: >200 ktas 

Ceiling: > 43 kft 

Endurance: > 20 h class 



• Objective: 


Talarion Rapid Fuselage Design Study 

– Mass estimation 

• In consideration of 

– Flight-Physical constraints 

– Strength 

– Stability 

– Manufacturing 

FE Model 

Analysis Model: 

Fully coupled 

structure-aerodynamic 

(Aeroelastic) 


Automation of both loops: 

Loads & Sizing Loop 

Aerodynamic Model



Steady manoeuvre loads analysis within LAGRANGE 

Phase 1: Manoeuvre load simulation 

• Requirements: 

– (full aircraft) finite element model 

– (full aircraft) aerodynamic model 

HISSS (Higher Order Sub- and 

Supersonic Singularity Method) or 

DLM (Doublet Lattice) panel model 

– Coupling model: 

Beaming & Splining Methods used in order to 

transfer aerodynamic loads to the FE-Model 

and structural deflections to the Aero-Model 

i.e. fully coupled analysis models, without 

condensation 


FE model 

aerodynamic 

model




Manoeuvre Load Simulation 

• Based on the Mission and Structural Design Criteria the flight envelope is 

established and scanned (10 3 - 10 5 manoeuvres) in order to determine the 

design driving, steady manoeuvres with maximum loads 

� Down selection of design driving steady manoeuvres (~10 2 manoeuvres). 





Manoeuvre Load Simulation 

• Each design driving, steady manoeuvre can be described by the 

– Mass configuration and CoG position 

– Altitude 

– Mach number 

– Accelerations and rotational speeds (up to 9 values) 

• Global Equilibrium of Forces and Moments has to be achieved for these 

pre-scribed manoeuvres of the elastic aircraft: 

Flight Parameter 


F y(inertia) + F y(aerodynamic) = 0 

F z(inertia) + F z(aerodynamic) = 0 

M x(inertia) + M x(aerodynamic) = 0 

M y(inertia) + M y(aerodynamic) = 0 

M z(inertia) + M z(aerodynamic) = 0




Trimming Process 

• For each steady manoeuvre (Mass, CoG, Altitude, Mach, Accelerations) 

the angles of attack (pitch angle, yaw angle and the AoA of the control 

surfaces are determined by minimizing the residual forces. 

β: 

mainly trim sideslip 

antimetric ailerons: 

mainly trim roll axis 

β 

α 

α: 

mainly trim lift 

Flight Parameter 

sym. elevators: 

mainly trim pitch axis 

rudder: 

mainly trim yaw axis 






F y(inertia) + F y(aerodynamic) = 0 

F z(inertia) + F z(aerodynamic) = 0 

M x(inertia) + M x(aerodynamic) = 0 

M y(inertia) + M y(aerodynamic) = 0 

M z(inertia) + M z(aerodynamic) = 0 

alfa = - xx ° 

beta = - yy ° 

delta aileron = zz ° 

delta elevator = hh ° 

delta rudder = kk °


Under Development 


Integration of Transient Gust into Optimisation 

A “gust case” is defined as a combination of: 

a) steady manoeuvre: flight condition and mass configuration (c.o.g. position !) 

(altitude & aircraft speed; usually 1g cruise) 

b) gust condition: wave-length and up- or down wind gust velocity and 

incidence angle (usually sinusoidal shaped) 

� leading to huge amount of different gust cases 

(up to ~10000), which have to be considered ! 

example: 


~ 

evaluated 

time steps 

(approx. 1000) 

flight condition 

(incl. mass configuration) 

gust wave 

length 

gust up- 

wind profile




Many gust blocks 

gust block for selected mass configuration 

Many gust cases 

gust case = 

Gust Process 

incremental gust analysis 

• specific mass configuration, 

• specific incidence angle, 

• specific wave length, 

• specific speed, 

• specific altitude 

+ 

basic flight attitude 

• trimmed aeroelastic steady manoeuvre 

• for specific mass configuration, 

• specific altitude, 

• specific speed 

• to be superimposed to an incremental gust 


Database 

(HDF5) 

evaluation of all 

gust cases for 

selected mass 

configuration 

• Implementation of the Incremental Gust Response and the Sensitivities is completed. 

• Implementation process for the fully automated determination of the design driving 

time steps and the superposition to manoeuvre load cases is ongoing.


Currently under 

development 



Summary for Phase 1: Manoeuvre, Gust and Landing Loads Analysis 

• The manoeuvre load simulation of elastic aircraft (fully coupled aerodynamicstructure 

model) is combined with a trimming process (optimisation task) in 

order to provide the distributed, elastic aircraft manoeuvre loads. 

• The distributed aerodynamic and inertia loads are directly applied to the global, 

non-condensed FE model, providing the stresses and displacements for the 

subsequent strength and stability analysis. 

• Gust loads are determined as incremental dynamic response. The time steps 

resulting in maximum local stresses are determined and the resulting 

deflections are superimposed to the corresponding steady manoeuvres. 

• Landing Gear Loads are determined by an external Multi-Body-Analysis and 

then applied to the global full aircraft model in order consider them in the sizing 

process. 

• By incorporating the loads analysis into the optimisation platform LAGRANGE 

both very time consuming loops (loads & sizing) are automated. 





Phase 2: Multidisciplinary Sizing optimisation 


• Strength analysis: 

• Damage tolerance 

• Von Mises 

• Skin buckling (for composite skin and 

metallic shear walls) 

• Column buckling 

• Flight-Physics: 

• Force equilibrium in Y, Z 

• Moment equilibrium around X, Y, Z 



Aeroelastic analysis 

Manoeuvre Simulation 

trimming Process Optimisation Model 

• Ply thickness of composite skin 

• Thickness of metallic shear walls 

• Cross-sectional areas of stringers 

• Trimming variables


• Optimisation Model: 


Stacking sequence: 


2312 elements linked to 43 patches 

45° 

-45° 

90° 

0° 

45° 

-45° 

90° 

0° 

0° 

90° 

-45° 

45° 

0° 

90° 

-45° 

45° 

Composite skin Composite/Metallic Stringer 

16 layers 

linked to 3 design variables 

129 design variables 

(3 * 43 patches) 

4212 elements linked to 174 patches 


(1 * 174 patches) 

Total: 1015 design variables 


Metallic Shear Walls 

8087 elements linked to 712 

patches 


(1 * 712 patches)

Analysis Model: Steady Aeroelasticity 

• Optimisation Model: 


β: 

mainly trim sideslip 




β 

α 



α: 

mainly trim lift 

5 design variables for each load case 

~ 50 Load cases 

________________________________ 

~250 Design variables 

rudder: 

mainly trim yaw axis 







• Criteria Model: 

Composites: 



Maximum strain (Damage Tolerance) 

36992 constraints * 34 load cases 

Metallic: 

Von Mises stress 

Strength Stability 

12299 constraints * 34 load cases 

1.675.894 strength constraints 

Skin & shear wall buckling: 


1080 constr. * 34 load cases 

Total: 1.726.860 constraints 

Stringer Column Buckling: 

419 constr. * 34 load cases 

50966 buckling constraints

Optimisation at Cassidian Air Systems 

• Criteria Model: 

� 

� 

� 


Flight-Physics 

M 

M 

M 

x 

y 

z 

� 0 

� 0 

� 0 

� 

� 

5 constraints / load case 

~ 50 load cases 

_________________ 

~ 250 constraints 


Y � 0 

Z � 0 


Flutter Displacement 

Manufacturing 

_________________ 

1 flutter constraint / 

mass configuration 

U 

� 

max 

max 

~ 5 Displacement 

constraints / load case 

50 load cases 

________________ 

~ 250 constraints 

hh 

bb 

�ttot �ttot 

ttot1 ttot1 

bb 

ee 

Layer thickness fitting: 

8layer * 47 steps = 376 

Minimum relative group 

thickness: 147 

Minimum absolute stack 

thickness: 49 

______________ 

572 constraints 

ttot2 ttot2 

cc 

DD 

bb 

ee 

11




• Skin Thickness & overall Reserve Factor: 

Total Thickness 

Composite Skin 


Overall R f




• Thickness of Metallic Shear Walls, Frames, Floors & overall Reserve 

Factor: 

Total Thickness 

Metallic Shear Walls, Frames, Floor 


Overall R f 

All Results are available in different formats (colour 

plots, Excel Tables, Database etc.)

Contents 


Contents 













Four distinct decks of the Lagrange input file to be considered 


LAGRANGE Input deck 

LAGRANGE Profile in HyperMesh 

Design Control Deck 

(control commands) 

NASTRAN- 

Deck 

Case Control 

Deck 

Optimization Data Deck 

(optimization data) 

LAGRANGE Profile I/O 

Bulk Data Deck 

� loads reduced NASTRAN template 

... 

... 

... 

ENDCONTROL 

... 

... 

... 

BEGIN BULK 

... 

... 

... 

ENDDATA 

... 

... 

... 

ENDE 

� uses standard NASTRAN I/O- procedures (CCD & BDD) 

� provides extended I/O- features (DCD and ODD) 


Chap. 3 

Chap. 4 

Chap. 5 

to 

Chap. 8 

HM user-page



Automatic Generation of DVs and BFs (defopt) 

External defopt- usage fully supported 


manage config-files 

execution of external binary 

� defopt writes files to disk 

� data to be imported 

� generic process extendible for further external executables



Manual Generation / Editing of DVs and BFs 

=> Analysis => ODD 

=> Model browser => OutputBlocks 


Contents 


Contents 












Summary 


Benefits of the Automated Airframe Design Process 

• The optimization assisted airframe design process has been established and 

applied within all design phases of a broad range of A/C projects (civil and military 

applications; components, large assemblies & full A/C). 

• The multidisciplinary design optimisation with LAGRANGE leads to a feasible 

airframe design which satisfies the requirements of all relevant disciplines with 

minimum weight. 

• The automation of both loops: structural sizing and loads loop results in an 

drastic reduction of development time and effort. 

• The strategic decision for an continued development the in-house MDO tool 

LAGRANGE is due to the specific aerospace design criteria on one hand (no 

Commercial Of The Shelf tool available) and the tremendous benefits and 

competitive advantages on the other hand. 

• The in-house software availability allows the fast adaption to advanced analysis 

methods as well as to new technological product and customer requirements. 

• Further Applications and Co-Operations are welcome ! 


Thank you for your attention! 

The reproduction, distribution and utilization of this document as well as 

the communication of its contents to others without express authorization 

is prohibited. Offenders will be held liable for the payment of damages. 

All rights reserved in the event of the grant of a patent, utility model or design. 



Optimisation at Cassidian Air Systems 


Benefits of Design Automation 

• Multidisciplinary optimisation with LAGRANGE provides a structural 

design which satisfies the requirements of all relevant disciplines with 

minimum weight 

• Automation of both loops: structural sizing and load analysis (through full 

coupling to aerodynamic analysis tools) 

• Huge multidisciplinary criteria model covering requirements of all structure 

design driving disciplines 

• Drastic reduction of development time and effort (>50%) 

• Reduced cost due to automation of the design process 

– Estimated Saving for Talarion Development Process: 

• Traditional Manual Process: 

e.g. 6 Load Loops (LL) * 1 J / LL * 40 FTE (Stress, Loads, Dynamics, Mass, Design, 

etc.) = 240 Man-Years 

• With MDO Process: 2 Load Loops * 0,65 J / LL * 40 FTE = 52 Man-Years 

• Estimated Saving: 188 Man-Years (37,6 Mill. €) !!! 

• Product/project specific requirements can be integrated (if required) in the 

platform easily (Property of CASSIDIAN Air Systems) 

– Over 3 Million lines of code, 140 Man-Yeas development

Download this Presentation - Altair Technology Conference 2013

Create successful ePaper yourself

Delete template?

Save as template?