基于PSO-SVM的风电整流器故障诊断

Fault Diagnosis of Wind Power Rectifier Based on PSO-SVM

摘要: 针对风力发电系统中整流器故障诊断的问题，以三相全控整流电路为例，提出粒子群优化支持向量机(particle swarm optimization-support vector machine, PSO-SVM)的分类算法。首先用MATLAB进行仿真得到故障信号，再使用快速傅立叶变换(FFT)预处理故障信号，主成分分析(PCA)提取其中主要频域特征，以及采集一周期内的时域特征，最后使用SVM和PSO-SVM分别对提取的时域、频域以及结合时频域的特征进行训练和测试。实验结果表明，采用PSO-SVM的方法对时域、频域、时频域特征的故障诊断率都比SVM要高，诊断时间也得到了大大提升，并且选用时频域特征进行故障诊断要比单独用时域或者频域特征的效果要好。

Abstract: Aiming at the problem of rectifier fault diagnosis in wind power generation system, a classification algorithm based on particle swarm optimization-support vector machine (PSO-SVM) is proposed by taking a three-phase fully controlled rectifier circuit as an example. Firstly, the fault signal is obtained by MATLAB simulation, and then the fast fourier transform (FFT) is used to preprocess the fault signal. The principal component analysis (PCA) is used to extract the main frequency domain features, and the time domain features in a cycle are collected. Finally, SVM and PSO-SVM are used to train and test the extracted time domain, frequency domain and combined time-frequency domain features. The experimental results show that the fault diagnosis rate of time-domain, frequency-domain and time-frequency-domain features using the PSO-SVM method is higher than that of SVM, and the diagnosis time is also greatly improved. The effect of fault diagnosis using time-frequency-domain features is better than that using time-domain or frequency-domain features alone.