麻省理工学院：《Multidisciplinary System》Peter A. Fenyes General Motors R and Planning

Multidisciplinary Analysis and Optimization of vehicle Architectures Peter A Fenyes General Motors R&D and Planning nicle Development Research Laboratory Presented at MIT 4/14/04 for 16.888/EsD.77J Multidisciplinary System Design Optimizatie MSDO) Prof. olivier de Weck and Prof. Karen Willcox …… Designing Great GM Cars and Trucks! ◎

1 4/14/04 Fenyes Multidisciplinary Analysis and Optimization of Vehicle Architectures Peter A. Fenyes General Motors R&D and Planning Vehicle Development Research Laboratory Presented at MIT 4/14/04 for 16.888/ESD.77J Multidisciplinary System Design Optimization (MSDO) Prof. Olivier de Weck and Prof. Karen Willcox 4/14/04 Fenyes …. Designing Great GM Cars and Trucks !

Multidisciplinary Analysis and Optimization of vehicle Architectures Outline The vehicle design process Early stage vehicle design Vehicle architecture Parametric vehicle representation to support MAO Multidisciplinary analysis and optimization Engineering, manufacturing, business Example vehicle design problem MAO challenges GM RD General Motors r&D http://www.gm.com/company/careers/career_paths/rmd/lab.html erials and Pre

2 4/14/04 Fenyes Multidisciplinary Analysis and Optimization of Vehicle Architetctures Outline • The vehicle design process – Early stage vehicle design – Vehicle architecture • Parametric vehicle representation to support MAO • Multidisciplinary analysis and optimization – Engineering, manufacturing, business • Example vehicle design problem • MAO challenges 4/14/04 Fenyes General Motors R&D http://www.gm.com/company/careers/career_paths/rnd/lab.html Chemical & Environmental Sciences Reaffirming GM's environmental principles Electrical & Controls Integration Enabling electronics-based, driver￾friendly, smart GM cars and trucks Manufacturing Systems Research Transforming the enterprise through math-based processes and systems Materials and Processes Adapting tomorrow's materials for today's automobiles Powertrain Systems Research Developing high-value, energy-efficient, high￾performance powerplants and drivetrains Vehicle Development Research Speeding and enhancing engineering design and development

Vehicle Development Research Lab What do we work on? Safety Syste ightweight Materials Work Systems CAE Methods Virtual Reality Mechametronics GM RD Vehicle desi gn and Analysis Timeline System Analysis 悬平 Manufacturing desig Conceptual design Detail design

3 4/14/04 Fenyes Vehicle Development Research Lab • What do we work on? – Safety Systems – Lightweight Materials – Work Systems – CAE Methods – Vehicle Design – Virtual Reality – Mechametronics – Quality Methods 4/14/04 Fenyes Vehicle Design Design and Analysis Timeline Design freedom Design detail (# design parameters) Conceptual design Detail design HIGH ARM S LA NON-DRIVEN STRUTS W/COILS Component selection Topology design Parametric geometry WB L FO RO GC FH RH LA LF LB LR LD LH h b Structures Aero dynamics Solar Load Occupant Dynamics Ride & Handling Fuel Economy Crash worthiness Other Analyses* Representation System Analysis Time Balance requirements Detailed Analysis Part design Manufacturing design

How Can MAo be Applied in Vehicle Design? Most effective application during early vehicle development Vehicle "balance and integration This will be the focus of our discussion Can be applied in more detailed phases of design Example: tradeoffs for aerodynamics and structure Example: tradeoffs in component design Example: Architecture Selection Problem Fuel economy New vehicle Multiple derivatives/multiple requirements requirements Vehicle performance Material process strategy Bending stiffness Profit by Architecture Ar

4 4/14/04 Fenyes How Can MAO be Applied in Vehicle Design? • Most effective application during early vehicle development – Vehicle “balance and integration” – This will be the focus of our discussion • Can be applied in more detailed phases of design – Example: tradeoffs for aerodynamics and structure – Example: tradeoffs in component design 4/14/04 Fenyes Example: Architecture Selection Problem New vehicle requirements #1 #2 #3 #4 #5 Populated Databases - Architectures - Materials - Processes 0 1 2 3 4 Arch #1 Arch #3 Arch #5 Concept Description - Fuel economy - Wheelbase - Interior volume - Bending stiffness … Profit by Architecture Design Representation (Unigraphics) Database (MS Access) Multidisciplinary Optimization (iSIGHT) Structural Analysis (NASTRAN) Fuel Economy (EconoSTRAT) Market Share (InSight) Aerodynamics (ASD) Interior Roominess (Excel) BIW Cost (IBIS) Facilities Investment (QuickStudy) Profitability (Excel) Summary of Results (Excel) Design for profit Considerations • Multiple derivatives/multiple requirements • Vehicle performance • Material/process strategy • Financial/marketing Parametric Architectures

Consider the Key Components of an MDO System Elastic Structures Templates aEternalAerodynamics Solar load User Input Vehicle Database Databas ccupant Dynamics g Suspension Loads GM RD Parametric Design Representation The shared common representation is used to generate analysis input for ALL disciplines Need a single consistent representation for architecture and derivatives Combine UG geometry with non-geometric database UG parametric model BOM data-availablelallowable components

5 4/14/04 Fenyes Consider the Key Components of an MDO System Elastic Structures External Aerodynamics Solar Load Occupant Dynamics Suspension Loads Fuel Economy Crashworthiness Other Analyses* Results Database Vehicle Database Parametric Modeler Discipline Modelers Library Data Library Data Templates Templates User Input Multidisciplinary Optimization Discipline Analysis & Sensitivity Calculations Parametric Representation 4/14/04 Fenyes Parametric Design Representation • The shared common representation is used to generate analysis input for ALL disciplines • Need a single consistent representation for architecture and derivatives • Combine UG geometry with non-geometric database – UG parametric model – BOM data - available/allowable components – Marketing data

Consistent Parametric Design Representation The shared common representation used to generate analysis input for ALL disciplines Geometric, non-geometric data Template Data to be vaned parametrically Library da Data remains fixed(components)-location, orientation/location may change Parametric CAD system Selection and location of components, settings of dimensional parameters GM RD Parametric Design Representation Combines Parametric geometry(Unigraphics) on-geometric configuration data(MS Access Many parameters available Detail level component variables )

6 4/14/04 Fenyes Consistent Parametric Design Representation The shared common representation used to generate analysis input for ALL disciplines • Geometric, non-geometric data – Templates • Data to be varied parametrically – Stored in parametric CAD system – Library data • Data remains fixed (components) - location, orientation/location may change – Parametric CAD system – Auxiliary database information – User input • Selection and location of components, settings of dimensional parameters HIGH ARM SLA NON-DRIVEN P=125 P=63 4/14/04 Fenyes Parametric Design Representation • Combines – Parametric geometry (Unigraphics) – Non-geometric configuration data (MS Access) • Many parameters available – High level architecture variables – Detail level component variables WB L FO RO GC FH RH LA LF LB LR LD LH h b W T H WR BL

Parametric Geometry Driving Discipline Analysis Wheel base Occupant, section dimensions, BOM items Overhangs Measures Axle positions Areas, lengths, volumes, angles, Width Aerodynamics FYD FWD DRI VER VISION DCw GM RD Parametric Geometry Parametric Dimensions and Occupant Placement

7 4/14/04 Fenyes Parametric Geometry Driving Discipline Analysis High level input parameters Wheel base Overhangs Axle positions Width ... Packaging Occupant, section dimensions, BOM items, … Measures Areas, lengths, volumes, angles, … Aerodynamics …. Resulting geometry used for: 4/14/04 Fenyes Parametric Geometry Parametric Dimensions and Occupant Placement