基于快速S变换的电能质量主成分分析方法

针对电能质量分析中S变换计算量大，及支持向量机分类识别需要设定特征参数的问题，文章将一种新的s变换的快速算法（FST）与主成分分析（PCA）相结合，并应用于电能质量分类识别。首先从FST域提取几种电压扰动的模系数；进而利用PCA进行降维处理并提取主要特征成分；然后获得投影矩阵；最后待识别电压信号投影后，根据最近邻分类器进行分类识别。仿真结果表明，电压扰动FST域模系数的特征成分主要在低频段，因此识别结果准确率高、计算时间短，并且算法本身有一定的抗噪声能力，能较好地实时处理电能质量扰动。

An Analysis Method of Principal Components of Power Quality Based on Fast S-transform

To handle the problems in power quality analysis such as a large amount of calculation for S-transform, requiring setting feature parameters to the support vector machine classification, etc, it proposed a new approach which combined the fast S-transform （FST） algorithm with the principal component analysis （PCA） to classify the power quality disturbances. Firstly, the module coefficients of several voltage disturbances were extracted from FST domain. Then, PCA was applied to descend dimension and extract the main feature components and the projection matrix was obtained accordingly. Finally, the voltage signals were projected and classified by the nearest neighbor classifier. The simulation results show that the feature components of the FST module coefficients for voltage disturbances are focused on low-frequency band, so that the algorithm has the advantages of high accuracy, short runtime and somewhat anti-interference ability. Furthermore, It is a good choice to deal with PQDs via real-time identification.