Dynamic Optimal Design of Bevoloid Gear RV Reducer Based on Neural Network

Yu Guangbin,Li Guixian,Wu Junfei,Wang Yanfei
DOI: https://doi.org/10.3321/j.issn:1004-132x.2001.09.003
2001-01-01
Abstract:In order to improve the performance of optimal problem based on BP neural-network, a improved wavelet neural network is used in the optimal problem instead of the BP neural-network. The mapping relation between the design variables and the dynamic parameters of a bevoloid gear RV reducer is established by using the nonlinear mapping ability of improved wavelet neural network based on improved particle swarm optimization (IPSO), This method combines the global optimization searching performance of the improved particle swarm optimization (PSO) algorithm and the time-frequency localization of the wavelet neural network which solves the difficult problem of establishing target function in dynamic optimal design So a complicated dynamic optimal problem converts into a simple optimal problem. This provides a new way to obtain a design scheme with good dynamic behavior in design stage.
What problem does this paper attempt to address?