论文部分内容阅读
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。
When the circuit composed of non-linear power electronic devices breaks down, the fault feature information can not be easily extracted and identified. This paper presents a method based on wavelet packet analysis and Elman neural network for fault diagnosis of power electronic devices. Firstly, wavelet packet analysis is used to extract the characteristic information of voltage and current signals under different fault conditions, and then the data are normalized Processing and as the input of Elman neural network, the neuron fault classifier with intelligent learning function can finish the fault identification and location. Taking 12-pulse rectifier circuit as an example, the circuit model is established under Matlab software to simulate the experiment. The results show that this method can quickly and accurately complete the fault diagnosis.