摘要
针对磁力泵滑动轴承磨损在线监测,设计制作了传感器数据采集试验台,大量采集各种工况下的数据,综合使用EXCEL,MATLAB,NEUROSHELL等数据处理软件,用神经网络法和回归分析法进行了数据分析.在此基础上,提出了一种多传感器数据采集系统和多神经网络模型迭代求解的方法,用该方法实现的磁力泵滑动轴承磨损检测,可消除磁力泵轴向窜动变量的影响.试验结果表明测量值与实际值很接近,建立的检测系统能检测滑动轴承的间隙,并达到一定的测量精度.
An experimental device with data collecting sensors to on-line monitor sliding bearings of magnetic drive pump is developed and massive data under different work condition are gathered. The data are processed with data processing softwares such as EXCEL, MATLAB, NEUROSHELL, neural network and regression analysis. A method of multi-sensors data acquisition system and multi-neural network model iteration is proposed. This testing method can eliminate the effect caused by axial move of sliding bearings of magnetic drive pump. The experimental results indicate that the observed value and the actual value agree well.
出处
《排灌机械》
2006年第5期36-39,共4页
Drainage and Irrigation Machinery
基金
江苏省高新技术研究资助项目(BG2002014)
江苏省科技建设计划资助项目(MB2002805)
关键词
磁力泵
滑动轴承
神经网络
磨损监测
magnetic drive pump
sliding bearing
neural network
wear monitoring