Construction And Application Of Multi-objective Optimization Platform Based On CAD/CAE Integration

Multi-objective optimization problems are common in fields such as engineering applications.Compared to a single optimal solution for a single-objective optimization problem,in a multi-objective optimization problem,a set of non-inferior solution sets is searched.When optimizing actual engineering problems,it is impossible to obtain these optimal solutions by manually modifying the CAD model and/or CAE analysis conditions multiple times.Therefore,research on multi-objective optimization problems based on CAD / CAE integration has become a focus of attention.This article focuses on the following aspects around this direction:(1)Based on CAD / CAE integration,a design analysis platform for multi-objective optimization is proposed.Through secondary software development and script-driven methods,the CAD subsystem can complete the automatic updating of parameterized models and export simulation models;the CAE subsystem can complete the parameterization and automation tasks of finite element analysis,such as adding loads and boundary conditions,set process parameters,etc.,and automatically extract the corresponding response function,which solves the problem of a large number of human-computer interactions in the optimization process and realizes optimization analysis automation.(2)A Kriging agent model was used,combined with a multi-objective genetic algorithm(MOGA)and a pattern search algorithm(PS),to build a hybrid optimization algorithm MOGA＿PS based on the agent model.First,Latin hypercube sampling is used to obtain the initial sample points,and the corresponding response values are obtained by simulation analysis.Then the Kriging agent model is used to fit the nonlinear relationship between the design variables and the response values,and then the MOGA＿PS hybrid algorithm is used to optimize Kriging.The algorithm first uses the MOGA algorithm for fast global optimization to avoid falling into a local optimal solution.After finding the optimal region,it uses the pattern search algorithm PS to perform local optimization to search for a solution that meets the requirements.Use the advantages of different algorithms at different stages to balance the speed and accuracy of global and local optimization,so as to obtain more efficient optimization algorithms.(3)A method for warping optimization of thin-walled injection molded parts is proposed.For injection products with large warpage deformation,an optimization method combining reverse deformation and process optimization is proposed.First,the compensation coefficient of the anti-deformation design is optimized,and the compensation coefficient with the smallest warpage deformation under the same process conditions is selected to obtain an anti-deformation model,and then the process optimization is performed.An improved warpage calculation formula is proposed,and a warping quantification method that is convenient to calculate and fully reflects the warpage deformation of the workpiece is obtained.Based on the methods proposed above,the feasibility and effectiveness of the solution proposed in this paper are proved by example analysis.And with the aid of a graphical user interface application development framework,an optimized design interactive interface was developed,making the entire integration optimization process more intuitive and easy to use.