讲义参考分析star文章类优化mf01.pptxVIP

INVERSE DESIGN OF SHAPE IN HEAT TRANSFER PROBLEMS USING modeFRONTIER AND FEMLAB;OUTLINE Introduction Software and methods Accomplishments Two dimensional conductive problem Three dimensional conductive problem Two dimensional conjugate problem Concluding remarks;INTRODUCTION - 1 Common approach is the “analysis problems” solution: Definition of geometry, equations and boundary condition; Evaluation of the fitness; Cut and try method; Weak improvements with a large number of design variables. Optimization with conjugate gradient method (CGM): Lack of generality, robustness and flexibility; Very fast convergence. ;INTRODUCTION - 2 Use of proprietary and/or problem-specific methods (solvers and optimisers): High accuracy and computational efficiency; Lack of generality, robustness and flexibility – e.g. hard to apply in real-world applications; Industrial heat transfer problems are frequently compute-intensive, with several conflicting objectives. Use of a general optimization tool, modeFRONTIER?, in pair with an unstructured FEM solver, FEMLAB? : Powerful, flexible and general optimization method; Multi-objective search allowed; Useful in industrial applications.;;INVERSE DESIGN OF SHAPE IN HEAT TRANSFER PROBLEMS;INVERSE DESIGN OF SHAPE IN HEAT TRANSFER PROBLEMS;INVERSE DESIGN OF SHAPE IN HEAT TRANSFER PROBLEMS;INVERSE DESIGN OF SHAPE IN HEAT TRANSFER PROBLEMS;INVERSE DESIGN OF SHAPE IN  HEAT TRANSFER PROBLEMS ;Two dimensional conductive problem – optimization evolution ;Two dimensional conductive problem ;INVERSE DESIGN OF SHAPE IN HEAT TRANSFER PROBLEMS; Surface parameterization: Two joined bi-cubic Bézier patches; Smooth curvature on joining edges; 22 free parameters.;INVERSE DESIGN OF SHAPE IN  HEAT TRANSFER PROBLEMS ;INVERSE DESIGN OF SHAPE IN  HEAT TRANSFER PROBLEMS ;INVERSE DESIGN OF SHAPE IN  HEAT TRANSFER PROBLEMS ;INVERSE DESIGN OF SHAPE IN  HEAT TRANSFER PROBLEMS ;Two dimensional conjugate problem ;Two dimensional conjugate problem ;;Two dim