摘要
对液压机械无级变速器机械故障的振动和噪声信号进行了分析:采用双谱分析法识别齿轮故障,希尔伯特信号包络法识别滚动轴承故障,小波变换信号分离法识别传动箱故障。对液压机械无级变速器液压故障的试验数据进行了研究:采用BP神经网络法识别电液比例伺服机构故障,频段分布法识别变量泵故障,核方法识别湿式离合器故障。研究表明:六种不同的方法对变速器的故障都有独特的识别作用,应根据变速器零部件的特性选择恰当的识别模式,以提高故障识别水平。
A nalyzed the vibration and noise signals caused by the mechanical faults of hydro-mechanical continuously variable transmission： bispectra analysis method for gear fault, Hilbert signal envelope method for rolling bearing fault, wavelet transform signal separation method for transmission box fault. Test data of hydraulic faults of hydro-mechanical continuously variable transmission were studied：BP neural network method for electro hydraulic proportional servomechanism fault, frequency distribution method for variable pump fault, kernel method for wet clutches fault. The results show that the six different methods above play respective roles in the fault identifications of hydro-mechanical continuously variable transmission. So the appropriate fault identification methods should be chosen according to the characteristics of transmission components, in order to improve the level of fault identifications.
出处
《机械设计与制造》
北大核心
2017年第10期44-49,共6页
Machinery Design & Manufacture
基金
国家自然科学基金项目(51575001)
安徽省自然科学基金项目(1508085ME70)
安徽工程大学科研启动基金(2015YQQ002
2015YQQ003)
关键词
液压机械无级变速器
故障识别
机械故障
液压故障
Hydro-Mechanical Continuously Variable Transmission
Fault Identifications
Mechanical Faults
Hydraulic Faults
