Multi-Objective Optimization Design Of High-Speed And High-Precision Feed System Based On Neural Network Agent Model And Genetic Algorithm

Under the background of the rapid development of engineering fields such as aviation industry,automobile industry and shipbuilding and Marine industry,CNC machine tools are also increasingly developing towards the direction of high speed and precision.By virtue of its excellent static,dynamic and thermal performance,the double screw drive feed system is gradually used in the field of high-end CNC machine tools.Because the structural parameters of the double screw drive feed system will have an important impact on the overall performance of the machine tool,this paper takes the multi-objective optimization of the feed system structure as the purpose to establish its statics,dynamics and thermodynamic finite element models.Combined with the neural network agent model and the non-inferior sequencing genetic algorithm(NSGA-Ⅱ),the optimal structural parameters of the double screw drive feed system were obtained,and finally the static,dynamic and thermal performance of the machine tool was improved.Aiming at the multi-objective optimization of the structure of the machine tool double screw drive feed system,based on the finite element analysis,this paper selected the span of ball screw pair,the span of rolling guide pair,the diameter of the ball screw center hole,the outer diameter of ball screw and the length of the ball screw thread as the structural optimization design variables.Taking the static deformation,thermodynamic coupling deformation,mass,natural frequency and other performance parameters of the feed system as optimization objectives,the test sample points corresponding to the design variables and optimization objectives were obtained by changing the structural optimization design variables and conducting static,dynamic and thermodynamic finite element analysis.After that,the black box function of neural network was constructed using the test sample points,and on this basis,the optimization mathematical model of the double screw drive feed system was established.The optimization termination conditions were set,and the optimization was solved by the non-inferior sorting genetic algorithm and the optimal results were output.Finally,after structure optimization,the static deformation,thermal coupling deformation and mass of the double screw drive feed system are reduced by 7.11%,11.93% and 1.27%,respectively,and the first order natural frequency is increased by 11.66%.Thus,on the basis of satisfying the lightweight of the double screw drive feed system,the static dynamic thermal comprehensive performance is improved.It also provides ideas and methods for the optimization design of the structure of the feed system in the future...