Investigation Of Macro And Micro Trans-scale Parallel Topology Optimization Design With Variable Density Approach

Topology optimization is a subject with mechanics and mathematics as the core and optimization theory as the basis.It is widely used in aerospace,automotive industry,biomedical and other fields by finding the optimal distribution of materials in a given design domain to maximize the structural performance.Traditional topology optimization mainly focuses on the macro scale of structure design.With the increasing requirements of modern industrial products for structural performance,single scale structural optimization can't meet such requirements.Microstructure is more and more popular among researchers because of its light weight,high specific stiffness and strength,good vibration isolation and heat insulation effect.Multi-scale topology optimization considering micro structure can give full play to the performance of the structure.In multi-scale topology optimization,in order to make the iteration converge stably,macro optimization and micro optimization are carried out separately,and the change of microstructure configuration has no effect on the macro material distribution.In order to further improve the structure performance,a multi-scale parallel topology optimization model based on variable density method is established to optimize the macro structure and micro structure simultaneously.Based on the model,the single objective optimization of the structure with force load and thermal load and the multi-objective optimization of the structure with thermal load are studied.Firstly,a multi-scale parallel topology optimization model and a non parallel topology optimization model are proposed.The results of parallel optimization and non parallel optimization are compared and analyzed.It is verified that the parallel optimization can greatly improve the structure performance.Aiming at the problem of large variables in multi-scale optimization design,uniform partition method and cluster partition method are used to divide the macro structure design domain respectively.The optimization results of different partition methods are analyzed and compared.It is found that the cluster partition method is more effective and greatly reduces the cost of computing time.The improved model of clustering method is designed to prevent the jumping phenomenon of clustering boundary value in the later stage of optimization and improve the convergence speed of optimization.Secondly,the multi-scale parallel topology optimization design of force loaded structure is carried out,and the multi-scale parallel topology optimization design model of force loaded structure based on clustering method is proposed.The model considers the influence of microstructure on the performance of the overall macro structure.According to the results of macro optimization,the micro structure obtained each time is filled into the macro structure to re evaluate the performance of the macro structure,guide the next macro optimization,and obtain the volume and location distribution of the micro structure.Then the topology optimization design of the micro structure is carried out,and the optimal micro structure is filled In the macro structure,the performance of the macro structure is evaluated,and the cycle is repeated in turn until the optimal microstructure configuration and location distribution are obtained,thus the optimal performance of the macro structure is also obtained.Taking cantilever beam structure,MBB beam structure and Michelle structure as optimization objects,the correctness and effectiveness of the proposed method are further verified by analyzing the results of macro meso multi-scale parallel optimization.Then,the multi-scale parallel topology optimization design of thermal load structure is carried out,and the multi-scale parallel topology optimization design model of thermal load structure based on clustering method is proposed.The model is used to optimize the heat dissipation performance of four boundary constant temperature structure,four vertex constant temperature structure and central constant temperature structure under steady-state heat conduction conditions.From the optimization results,it can be seen that the multi-scale parallel topology optimization model is universal.Finally,the multi-scale and multi-objective parallel topology optimization design is carried out for the structure under the joint action of thermal load and force load,and the mechanical and thermal properties of the structure under different weight coefficients are analyzed.It can be clearly seen that the multi-objective optimization greatly improves the comprehensive performance of the structure.