论文部分内容阅读
为了改变传统的基于软件的故障诊断模式,发挥神经网络超大规模集成电路(VLSI)的优势,提出了一种用于故障诊断的改进脉冲频率调制(PFM)模拟神经网络脉冲流VLSI电路。利用单层感知器网络、场效应管电路实现了一种新的数字模拟混合突触乘法/加法器电路。以此电路为基础,设计了进行主轴承磨损故障诊断的神经网络故障识别系统。用含有故障信息的噪声信号代替振动信号进行特征值提取,经过前置信号处理分析、故障特征值提取和神经网络运算,最后得出代表待诊断测试信号与标准故障模板之间“欧氏距离”的VLSI电路输出端电容的电压值。根据各个电压值,可以判断出故障类别。该电路具有较高的识别精度,可以实现实时在线的故障诊断。
In order to change the traditional fault diagnosis mode based on software and take advantage of neural network VLSI (VLSI), an improved pulse frequency modulation (PFM) neural network pulse stream VLSI circuit is proposed for fault diagnosis. A new digital-analog hybrid synaptic multiplication / adder circuit is implemented using single-layer sensor networks and field-effect transistor circuits. Based on this circuit, a neural network fault identification system for fault diagnosis of main bearing wear is designed. The noise signal which contains the fault information is used to replace the vibration signal to extract the eigenvalue. After the pre-signal processing and analysis, the fault eigenvalue extraction and the neural network operation, the Euclidean distance between the test signal to be diagnosed and the standard fault template "VLSI circuit output capacitor voltage value. According to each voltage value, you can determine the fault type. The circuit has high recognition accuracy, real-time online fault diagnosis.