基于小波包分析和Elman神经网的故障诊断方法研究

由非线性电力电子装置组成的电路发生故障时,故障特征信息不易提取和识别。对此提出一种基于小波包分析和Elman神经网的电力电子装置故障诊断的方法,先运用小波包分析法提取电路在不同故障状态下电压及电流信号的特征信息,然后对数据进行归一化处理并作为Elman神经网的输入,由具有智能学习功能的神经元故障分类器完成故障识别和定位。以12脉冲整流电路为例,在Matlab软件下建立电路模型进行仿真实验,结果表明该方法能快速、准确的完成故障诊断。

Research on a fault diagnosis method based on wavelet packet analyzing and Elman neural network

It＇s hard to extract and identify effective fault feature,when the fault of circuit composed of nonlinear power electronic device happened.A method based on wavelet packet analyzing and Elman neural network was proposed to fault diagnose of power electronic device.Firstly,the signal feature of different states about voltage and current of power electronic device was extracted by using the wavelet packet analysis.Then it was normalized and being inputted the Elman neural network,the neural network classifier will identifying and diagnosing different faults.Taking the 12-pulse circuit of electronic power electronic rectifier device as an example,completed the circuit model under Matlab and put up emulation experiment.The result of emulation shows that the method can deal with the fault diagnosis rapidly and accurately.