Research On Multi-objective Optimization Design Method Based On Surrogate Model And Its Application

With the increasing complexity and non-linearity of practical engineering design problems,it is difficult to express the relationship between design variables and performance indicators,and it takes a lot of time to solve the performance index values corresponding to design parameters in the whole design space.Therefore,the surrogate model is introduced into the multi-objective optimization design.In the process of the construction of the surrogate model,there is a conflict between the fitting accuracy of the surrogate model and the construction time of the surrogate model.Therefore,it is worth studying to improve the construction efficiency of the surrogate model on the premise of ensuring the accuracy of the surrogate model.The drilling support structure is the key structure of glass edge grinding and drilling machine.Optimal design with minimum deformation and minimum mass will help to improve the precision of drilling and reduce cost.In this paper,a multi-objective optimization design method based on surrogate model is studied and applied to the optimization design of drilling support structure.The main research contents are as follows:(1)First of all,the performance indicators under the initial size of the drilling support structure are obtained through static analysis and modal analysis.Secondly,the parameters of the drilling support structure are extracted according to the actual engineering requirements.Then,the dimension reduction of the parameters of the drilling support structure is carried out based on the correlation analysis theory to reduce the complexity of the optimization model.(2)The theory of multi-objective optimization design based on surrogate model is introduced.In the aspect of experimental design,the typical central composite experimental design and Latin hypercube sampling design are introduced.In terms of surrogate model,the polynomial response surface model and Kriging model are studied.In the aspect of surrogate model updating strategy,we summarize the common strategies of adding points and reducing space.In the aspect of multi-objective optimization,the multi-objective genetic algorithm and the second generation nondominated genetic algorithm are introduced.(3)A multi-objective optimization design method based on Kriging intelligent layout points is proposed.In order to improve the accuracy of the surrogate model,a surrogate model updating strategy combining the addition of sampling points in the Pareto region with the addition of verification points in the sparse region was adopted.The drilling support structure was optimized by using this method,and the Pareto solution meeting the accuracy requirements was obtained.Compared with Kriging automatic refinement method and Kriging no-refinement point method in ANSYS Workbench,the proposed multi-objective optimization design method based on Kriging intelligent point layout is proved to have advantages in solving efficiency and solving precision.(4)A multi-objective optimization design method based on a two-layer surrogate model is studied.In order to combine the advantages of different surrogate models,the polynomial response surface model is used as the first layer of the two-layer surrogate model,and the Kriging surrogate model is used as the second layer of the two-layer surrogate model.The design space reconstruction method based on preference is adopted to reduce the design space,thereby improving the efficiency of surrogate model update.The method is applied to the optimization design of surrogate support structure,and the performance indexes of surrogate support structure are balanced by matching each dimension parameter.