BP神经网络理论对磨损自补偿过程的预测  

Research of BP Neural Network Theory on Predicting the Process of Wear-self-compensation

作　　者：莫易敏[1] 史玉升[1] 邹岚[2] 赵源[2] 

机构地区：[1]石油大学,北京昌平102200 [2]武汉材料保护研究所,430030

出　　处：《润滑与密封》1998年第6期6-9,共4页Lubrication Engineering

基　　金：国家自然科学基金资助项目

摘　　要：本文通过钢—钢摩擦副的摩擦学性能的对比研究,发现在磨损自补偿添加剂SW4作用下,钢—钢摩擦副也象钢—铜摩擦副一样具有磨损自补偿性能,其摩擦系数和磨损量相对下降了很多;同时,利用BP神经网络理论对钢—钢摩擦副的摩擦学过程,特别是其磨损自补偿过程进行了预测,该方法与传统的建模方法相比,具有运算时间短,使用数据少,不需要严格的数学模型等优点,其预测值与试验值非常接近。Through the contrast research of the tribological characteristic of steel - steel tubbing pair, it was found that the steel -steel rubbing pair of using additive SW4 is of wear-setf-compensation's characteristic as well as steel- copper rubbing pair and the frictional coefficient and wear drop down largely in this paper. The method of BP neural network theory that was used to predict the tribological pro cess of steel-steel nibbing μair and specially the wear-self-compensation's process need not to build up mathematical model in contrast with traditional modeling methods, and reduce computer time and application data. The predicting results are close to experiment data.

关 键 词：磨损自补偿 BP神经网络 钢-钢摩擦副 添加剂SW4 

分 类 号：TH117.1[机械工程—机械设计及理论]