基于小波包神经网络的开关电源电路故障诊断研究

Study on switching power supply circuit fault diagnosis based on wavelet packet neural network

针对开关电源电路常见故障,提出一种基于小波包神经网络的开关电源电路故障诊断方法。利用小波分析对开关电源输出电压进行分析处理,依据小波多分辨分析的特点,获得信号各频段的细节系数及其能量,再利用小波包分析对小波分析中没有细分的高频信号进行分解,提高频率分解率,将各频段能量进行归一化处理后,构造故障特征向量作为神经网络的输入进行分类。将Multisim13与Matlab可以实现开关电源电路的故障诊断。

A switching power supply circuit fault diagnosis method based on wavelet packet neural network is proposed to diagnose the common faults of switching power supply.The wavelet analysis is used to analyze the output voltage of switching power supply.According to the characteristics of wavelet multi-resolution analysis,the detail coefficients and energy of the signal in each frequency band are obtained.The wavelet packet analysis is used to decompose the high.frequency signal without subdivision to improve the frequency resolution.The energy of each frequency band is normalized as the fault eigenvectors,which are taken as the inputs of neural network for classification.The Multisim13 and Matlab are combined to realize the fault diagnosis of switching power supply circuit accurately.The simulation result shows that the method can realize the fault diagnosis of switching power supply circuit.