Structure Optimization Design Of RV Reducer Based On Genetic Algorithm And BP Neural Network

RV reducer has advantages such as high transmission,high efficiency,high load capacity,compact structure and wide application in the field of robotics.At present,although RV reducer can meet the usage requirements,the parameters chosen in the traditional design method are not always optimal and have a long design cycle and high cost,which affects the development and application of RVs.due to the above,this paper fully takes into account many factors affecting the design of RV reducer based on the advanced BP neural network and genetic algorithms,including:(1)Based on the analysis of the basic principles and characteristics of the RV reducer drive and using 3D modeling,virtual assembly and secondary development software CATIA,parametric modeling of key RV reducer components was completed,reducing the workload of designers and increasing design efficiency.using ADAMS dynamic analysis software,ADAMS simulates a virtual model,compares theoretical values with model results to check the justification.(2)Based on the finite element method of the contact problem,contact stress analysis was performed on the cycloid needle wheel transmission part using ANSYS Workbench software,and the structural parameters of the cycloid needle wheel transmission part were used as input samples and the performance parameters obtained from the finite element analysis were used as output samples.The design of the multielement,multi-level test plan was completed using the parameterized design platform built with CATIA secondary development technology,models with different combinations of parameters were built,and contact stress analysis was performed on each group of models to complete the acquisition of neural network training samples.(3)Optimized the standard BP neural network parameters using genetic algorithm,and combined with the above acquired sample collection,completed the training of the cycloid needle wheel transmission partial model of the standard and neural network;through analyzing the results of the two types of models,selected the model of BP neural network fusion genetic algorithm,and acquired the complex nonlinear relationship between the parameters.(4)With the volume of the RV reducer as the objective function,the structural parameters as the design variables,the size requirements of each component and the performance requirements of the cycloidal needle wheel transmission obtained from the BP neural network mapping as the constraints,an optimized mathematical model of the RV reducer mechanism was constructed and obtained using the MATLAB genetic algorithm kit.The optimized volume of the RV reducer is reduced by 16.9%,and the design variable is obtained.The diameter of the center circle of the needle tooth is changed to 77 mm,which is 5mm smaller than the original data.The radius of the needle teeth is changed to 4mm,which is 1mm smaller than the original data.The cycloid width is changed to 8mm,which is reduced by 1mm compared with the original data,and the pin diameter is changed to 5mm,which is reduced by 1.32 mm compared with the original data,realizing the optimization goal of reducing the volume of the RV reducer.