SimulationSupported Decision Making仿真支持决策

Simulation-Supported Decision MakingGene AllenPresidentDecision Incite Inc.Simulation A Tool for Decision Making Quickly Identify and Understand How a Product Functions: What are the major variables driving functionality? What are the combinations of variables that lead to problems in complex systems? Ability Exists Today Due to advances in compute capabilityCorrelation MapsGeneration of Correlation MapsCorrelation Map a 2-D view of a Results Data generated from Monte Carlo AnalysisIncorporates Variability and UncertaintyUpdated Latin Hypercube samplingIndependent of the Number of VariablesResults with 100 runsDoes Not Violate Physics No assumptions of continuity“Not elegant, only gives the right answers.”Correlation Maps to Understand Cause & Effect InputVariablesOutputVariables Ranks input variables and output responses by correlation level Follows MIT-developed Design Structure Matrix model format Filters Variables Based on Correlation LevelUpper right positive correlationLower left negative correlationA Correlation MapMeta ModelofDesign AlternativesCorrelation Map: - Includes All Results - Highlights Key VariablesStochastic Simulation Template 100 MCS runsGeneration of Correlation Maps Monte Carlo AnalysisSolution:Solution:Establish tolerances for the Establish tolerances for the input and design variables.input and design variables.Measure the systems Measure the systems response in statistical terms.response in statistical terms.Sources of VariabilitySources of Variability Material Properties Loads Boundary and initial conditions Geometry imperfections Assembly imperfections Solver Computer (round-off, truncation, etc.) Engineer (choice of element type, algorithm, mesh band-width, etc.)x1x2x3y1y2The Fundamental Problem VariabilityStructural Material ScatterMATERIALCHARACTERISTICCVMetallicRupture8-15%Buckling14%Carbon FiberRupture10-17%Screw, Rivet, WeldingRupture8%BondingAdhesive strength12-16%Metal/metal8-13%HoneycombTension16%Shear, compression10%Face wrinkling8%InsertsAxial loading12%Thermal protection (AQ60)In-plane tension12-24%In-plane compression15-20%Source: Klein, M., Schueller, G.I., et.al.,Probabilistic Approach to Structural Factors of Safety in Aerospace, Proceedings of the CNES Spacecraft Structures and Mechanical Testing Conference, Paris, June 1994, Cepadues Edition, Toulouse, 1994.The Deception of Precise GeometryGeometry imperfections should be described as stochastic fields.Monte Carlo Results show RealityUnderstanding the physics of a phenomenon is equivalent to the understanding of the topology and structure of these clouds.Singlecomputerrun =AnalysisCollectionof computerruns =SimulationUnderstanding MCS Results Simulation generates a large amount of data. A typical simulation run requires around 100 solver executions. Each combination of hundreds to thousands of variables produces a point cloud. In each cloud:POSITION provides information on PERFORMANCESCATTER represents QUALITYSHAPE represents ROBUSTNESSKEY: REDUCE the Multi-Dimensional Cloud to EASILY UNDERSTOOD INFORMATION Condense into a CORRELATION MAP Variables are sorted by the strength of their relationshipMonte Carlo Simulation Results12 of the 782D views that resulted from a simulation with6 outputs froma scan of 7 inputs with uniformdistributions.Number of 2D Views of Results = Sum of all integers from 1 to (Number of Variables -1) Displays condensed information from hundreds of analysis runs. Correlation Map = Structured Information = Knowledge A Correlation Map helps an engineer: Understand how a system works. How information flows within the system. how variables and components correlate. Make decisions on how a design may be improved. Identify dominant design variables. Use as input for stochastic design improvement. Find the weak points in a system. Find redundancies in a design. Identify rules that govern the performance (“if A and B then C”).There are NO algorithms to learn. The engineer concentrates on engineering, not on numerical analysis.Correlation Maps: Understanding Cause and EffectDesign Improvement Process1234TargetPerformanceIteration Automotive and Aerospace companies have continued to expand use of process since 1997 BMW, Audi, Toyota, Mecedes, Nissan and Jaguar have expanded Computer Clusters for Stochastic Car Crash Simulation taking 10s of pounds from car model designs. Aerospace companies applying to improve aerospace designs. Alenia reduced weight of new commercial airliner tail by 6%.Courtesy, Alenia AeronauticaProcess for Decision Support Model a multi-disciplinary design-analysis process Randomize the process model Run Monte Carlo simulation of the model Process Results Correlation Maps showing Cause and Effect Outlier identification showing anomalies Direction for Design Improvement Identify what influences functionality Address Uncertainty and Variation Provides credibility in modeling & simulation Results clouds represent what is possible Easy to use No methods or algorithms to learn Reduces risk through better engineering Takes all inputs into account vice using initial assumptions Changing the general engineering processCorrelation Maps - Filter Complexity while Modeling Reality