基于振动信号和小波神经网络的变压器故障诊断

Fault Diagnosis of Power Transformer Based on Vibration and Wavelet Neural Network

提出一种基于振动信号和小波神经网络的电力变压器故障诊断方法。采用变压器油箱表面的振动信号作为采样信号进行频谱分析提取特征频率信号,并以此特征频率信号乘以电流标么值的平方作为训练样本进行小波神经网络训练,小波神经网络输出量能够反映出频谱特征向量和变压器故障类型之间的映射关系,从而实现变压器的故障诊断。实验结果表明,使用该方法能够有效地对变压器进行故障分类及其诊断,并且小波神经网络具有很好的泛化能力。

A fault diagnostic method of power transformer based on vibration and wavelet neural network is presented, which gets the characteristics of the vibration in frequency domain from the vibration sampled from the tank of transformers to train for the wavelet neural network （WNN）. With the output of the wavelet neural network, we can get the relationship between the faults and the frequency characteristics can be obtained. The experiment results show that the proposed method can be used for diagnosis of power transformer and output the type of the fault, and the wavelet neural has a good generalized performance.