Research On Mechanical Product Matching Technology Based On Genetic Algorithm And Its Application

Assembly is the process of combining multiple parts into parts or products in accordance with the appropriate technical requirements.It is the last step in the production process of the entire product and has a direct impact on the quality of the product.Use computer technology to select and assemble various products according to the specified rules,and optimize and cooperate with the selected parts,and finally obtain the optimal matching scheme.This method can effectively ensure the assembly precision of the parts and reduce the accuracy.Part residual rate.In this paper,the valve product is taken as the research object.Through the research on the batch matching method of large-volume product tolerance and the matching method under the multi-quality requirements of small batch products,the valve part matching system is developed,which reduces the residual rate of the product assembly parts and improves the product assembly quality and production efficiency.The main research contents and results of this paper are as follows:(1)Aiming at the problem of selecting and assembling large-scale valve products,a grouping method based on multi-objective evolutionary algorithm is proposed.Firstly,aiming at the actual matching problem of the valve products,the coding rules of the tolerance zone grouping scheme are proposed.Then,with the matching success rate and matching precision as the optimization goal,the mathematical model of batch selection of large-volume products is constructed.Finally,the improvement is utilized.The non-dominated sorting genetic algorithm solves the grouping matching model,and obtains the optimal solution with high matching success rate and matching precision.(2)Aiming at the problem of selecting assembly under the multi-quality requirements of small batch valve products,this paper proposes a matching method based on density multi-objective evolutionary algorithm.Firstly,considering the relationship between the dimension chain and the size of the part and selecting the success rate and the matching quality as the optimization goal,construct the mathematical model of the assembly under the multi-objective and multi-quality requirements;then,using the natural number coding,with a part size value number The random sequence is used as the paired coded gene,and the fitness function is defined by the individual density.The matching problem is combined with the density-based multi-objective evolutionary algorithm(DMOEA)to construct the mathematical model.Finally,according to the actual valve product data,and other selections The feasibility and effectiveness of the method are verified by comparison with the method.(3)Completed the development of the valve part matching system.Based on the method of this paper and according to the actual production requirements of the valve products,the valve part matching system has been developed.The main functional modules include part data management,matching dimension chain definition,matching task definition,matching calculation,matching result query and Analysis and so on.The system has been put into use in a factory of China Aerospace Science and Technology Corporation,and has achieved good application results.