无石棉制动带摩擦磨损性能预测系统的研究  

Research on prediction system of asbestos-free brake band's friction and wear properties

作　　者：李屹[1] 周元康[1] 吴兵[1] 

机构地区：[1]贵州大学机械工程学院,贵州贵阳550003

出　　处：《机械》2010年第4期5-7,共3页Machinery

基　　金：国家自然科学基金(50563001)

摘　　要：以实验为基础,利用Matlab软件平台上的BP神经网络工具箱,建立并训练了无石棉制动带复合纤维优化设计中所需的各纤维组分体积百分含量与制动带摩擦磨损性能指标之间的人工神经网络映射模型。研究结果表明,该BP网络映射模型的网络误差小,映射结果与实际实验结果的吻合度高,能较为准确地预测复合纤维组成对应制动带的摩擦磨损性能指标,解决了二者关系无法用显函数表征的问题,为制动带无石棉复合纤维组成的优化设计奠定了基础。BP network mapping model between the content of each constituent in composite fiber and the friction and wear properties of brake band was established and trained by using BP network toolbox of Matlab based on experiments. The results indicate that this BP work has lower network error; the mapping result has higher goodness of fit with experimental result. The model can be used to precisely predict the friction and wear properties of brake band corresponding to the content of each constituent in composite fiber. Thus, the difficulty in constructing the relational function between them has been solved. This work lays a foundation for further optimization design of composite fiber constituent of brake band.

关 键 词：BP网络 制动带 复合纤维 摩擦磨损性能 

分 类 号：TP389.1[自动化与计算机技术—计算机系统结构]