8 Fenyes4/14/04 Parametric Model Structure Representation Exterior Representation Fenyes4/14/04 Parametric Geometry Changes Parametric animation 9 Fenyes4/14/04 Challenges in Parametric Representation ? Must combine geometric, non-geometric data ? Robust parameterization of points, curves, surfaces, solids – Maintain robust associativity across parts and assemblies ? Must be able to flexibly modify – Relationships (independent, dependent) - constraint management – Geometry and parts (add / remove / modify) ? Manage coarse-to-fine strategy ? Share parametrics with – Other CAD systems - STEP/IGES are inadequate – CAE applications, MDO systems Structures Aero dynamics Solar Load Occupant Dynamics Ride & Handling Fuel Economy Crash worthiness Other Analyses* Representation Fenyes4/14/04 Multidisciplinary Analysis and Optimization Required Functionality ?Automatically run analysis and optimization from the shared representation ? Coordinate the analyses ?Dataflow ?Distributed, multiplatform ? Share design variables and responses ?Database is the common repository ? Improve the design ?Optimization, DOE, manual, … Our Goal: Multidisciplinary analysis and optimization for early vehicle development – coarse balance and integration 10 Fenyes4/14/04 Key MAO Framework Concepts – Modular system ? Easily add, modify, replace, analysis tools, modify framework – Flexible, parametric design representation and database – Discipline analysis tools tightly coupled to design representation ? Automated generation of inputs ? Automated capture of results (responses, sensitivities, histories, etc) – Analysis and design shell ? Coordinate and execute discipline analyses ? Iterative design improvement Fenyes4/14/04 Commercial and University MAO Systems ?DOME –MIT ? VADOR – Waterloo ? iSIGHT/FIPER – Engineous ? Model Center - Phoenix Integration ? AML/TIE - TecnoSoft ? Optimus – LMS And growing …. 11 Fenyes4/14/04 Approaches to Multidisciplinary Design and Improvement ? Flow down, Target Cascading ? Hierarchical methods – COO, coordination methods ? Multi Objective methods – Rankings, Pareto optimality ? Preference specification via combined objective ? “Natural” objective – Formulate a objective – e.g. profit – that combines the discipline interactions in a natural way – This is the approach presented Fenyes4/14/04 GM’s Framework for Architecture Design Design Representation (Unigraphics) Database (MS Access) Multidisciplinary Design (iSIGHT) Structural Optimization (NASTRAN) Aerodynamics Interior Roominess (Excel) Business Summary of Results (Excel) Energy (Custom) (Custom) (Custom) 12 Fenyes4/14/04 Discipline Analysis to Support Tradeoffs ? Focus on key disciplines ? Provide consistent information to all discipline analyses – Tight coupling to representation – Automated discipline modeling ? Balance analysis detail against design knowledge and required analysis speed Structures Aero dynamics Solar Load Occupant Dynamics Ride & Handling Fuel Economy Crash worthiness Other Analyses* Representation Fenyes4/14/04 Vehicle Design – Many Disciplines Examples of key drivers and responses Parameters: p 1 ,..p n BOM BOP WB L FO RO GC FH RH LA LF LB LR LD LH h b Database/ Parametric Architecture Representation Engineering Business Decision Process Packaging Structures (frequency) drivers: overall length, width, component mass responses: bending, torsion frequency Fuel Economy drivers: C d , powertrain, 0-60 performance, mass responses: city & highway economy Aerodynamics drivers: backlight angle, tumblehome responses: frontal area, C d Profitability drivers: sales, components, assembly, physical plant responses: component cost, investment, revenue Packaging drivers: b-pillar size, overall width, height, tumblehome responses: aero, fuel economy, structures, piece cost Geometry drivers: topology, layout, proportions responses: load paths, aesthetics 13 Fenyes4/14/04 Example Tools to Support Early Vehicle Design ? Analysis/Communication Framework – iSIGHT (Engineous) ? Database – MS/Access ? Geometry engine, parametric CAD & CAE model creation – Unigraphics ? Discipline Analysis – Structures: NASTRAN (MSC) – Fuel Economy: Proprietary (GM R&D) – Financial - Piece Cost: Technical Cost Modeling (J. Clark, MIT) – Manufacturing - Investment Cost - Proprietary (GM R&D) – Packaging: UG, Spreadsheet, – Aerodynamics: Proprietary (GM R&D) – Safety: Proprietary (GM R&D) – Business (marketing, revenue): Proprietary (GM R&D) ? Decision Engine – iSIGHT (Engineous) - Optimization, DOE Simple, fast-running analyses – run in minutes! Fenyes4/14/04 Database/Architecture Representation ? Single consistent representation for architecture and derivatives – Must comprehend data used by all disciplines ? Combine geometric and non-geometric; combine inputs and responses – UG parametric data – BOM data - available/allowable components – Marketing data – Responses - analysis results ….