摘要
高速泵状态监测是对高速泵运行状态评估的重要手段。首先对高速泵进行了不同状态下的试验,采集到各个状态的外特性信号以及振动信号,然后对信号进行处理,分别基于外特性信号和振动信号构建特征向量,最后利用BP神经网络对高速泵的3种不用的运行状态进行识别评估。研究表明,可以使用简单的神经网络对高速泵运行状态进行学习和识别工作;外特性信号和振动信号都能较好的反应高速泵的运行状态;对于正常工作下的高速泵运行情况有一个较好的识别效果;联合利用两种信号对高速泵运行状态进行识别,在识别率上有了显著的提高。
Absrtact:Firstly,the high-speed pump is tested in different states,the external characteristic signals and vibration signals of each state are collected,and then the signals are processed,and the feature vectors are constructed based on the external characteristic signals and vibration signals respectively.Finally,the BP neural network is used to pump the high-speed pump.The three unused operating states are evaluated for identification.
出处
《工业控制计算机》
2019年第1期11-13,共3页
Industrial Control Computer
关键词
高速泵
运行状态识别
振动信号
神经网络
high speed pump
operating state identification
vibration signal
neural network