基于SA-BP算法的主减速器品质诊断研究  

Main Reduction Gear Quality Diagnosis Research Based on SA-BP Algorithm

作　　者：潘昊[1] 张华伟[1] 高美铃[1] 

机构地区：[1]武汉理工大学计算机科学与技术学院,武汉430070

出　　处：《武汉理工大学学报》2011年第1期161-164,共4页Journal of Wuhan University of Technology

基　　金：湖北省自然科学基金(2008CDB288)

摘　　要：汽车噪声是影响汽车品质的重要因素,而后桥主减速器齿轮噪声是微车噪声的主要来源。齿轮故障诊断是一个复杂的非线性问题,目前只能靠人工经验来诊断故障。提出了改进的模拟退火——BP神经网络融合算法,该融合算法克服了BP算法易陷入局部极小的缺点,将其应用于主减速器品质诊断,研究结果表明优于传统算法,效果突出。The auto noise is the important factor affecting autos quality,and the rear axle main reduction gear noise is the important source of micro vehicle noise.At present,the gear failure diagosis can only be diagnosed by artificial experience for it is a complex non-linear problem.This paper proposed an improved simulation annealing BP neural network fusion algorithm,which overcame the outcoming of the BP algorithm easily falling into the local minimum.Applied in the main gear box quality diagnosis,the findings indicated that it surpassed the traditional algorithm,and the effect was splendid.

关 键 词：齿轮故障 BP网络 模拟退火算法 融合算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]