基于改进神经网络的少齿差行星齿轮参数优化设计  

Optimal Design of Parameter for Few-tooth-difference Planetary Gear Transmission Based on Improved Neural Network

作　　者：吕俊峰[1] 陈小安[1] 赵孟娜[2] 

机构地区：[1]重庆大学机械传动国家重点实验室,重庆400044 [2]重庆理工大学重庆汽车学院,重庆400054

出　　处：《重庆理工大学学报（自然科学）》2012年第2期55-59,共5页Journal of Chongqing University of Technology：Natural Science

基　　金：国家自然科学基金资助项目(51074507)

摘　　要：在对少齿差行星齿轮进行结构分析的基础上,根据其传动特点和设计要求,运用模糊数学的原理进行了模糊可靠性分析,建立了可靠性数学模型,将模糊设计优化模型转化为了常规的优化模型。通过所建立的模型可以实现最优参数选取,同时针对传统BP神经网络的不足,将模拟退火和BP网络相结合,设计了一种新型的改进神经网络。实验结果表明,此种算法得出的绝对误差和相对误差都较小。This paper, on the basis of analyzing few - tooth - difference planetary gear structure and according to its transmission characteristics and design requirements, builds a reliability mathematical model based on ambiguity reliability analysis through ambiguity reliability principle, which transforms the ambiguity design optimization model into the common optimization model. Through this model, the optimal parameters can be selected. And considering the deficiency of the traditional BP neural net- work, this paper designs a new type of improved neural network by combining stimulated annealing and BP neural network. The results showed that the absolute error and relative error through this algo-rithm are smaller. This is a comparatively accurate method for optimization design.

关 键 词：少齿差行星齿轮 BP神经函数 模拟退火 模糊可靠性 

分 类 号：TH132.41[机械工程—机械制造及自动化]