基于改进贝叶斯分类法的电能质量扰动分类方法

Power Quality Disturbance Classification Method Based on Improved Bayes Method

基于改进贝叶斯分类法提出了2种改进的暂态电能质量扰动分类方法。该分类方法保留了原贝叶斯分类法“最优分类”的性质,使原贝叶斯分类法转化为非参数分类法,扩大了分类法的适用范围,可对数量有限的交叉样本进行最优分类。采用交流暂态仿真软件对5种典型的电能质量扰动信号如电压振荡、电压中断等进行仿真和分类识别。对暂态电压扰动的分类结果表明,上述改进的暂态电能质量扰动分类方法分类特性良好、适用范围较广。

Based on improved Bayes method, two improved classification methods for transient power quality disturbances are proposed. In these classification methods the feature of optimal classification in original Bayes method is reserved, it makes the original Bayes method turning into non-parametric classification and enlarges the applicability range of classification, thus the optimal classification of intercrossed samples which amount is limited. By use of alternative transients program （ATP） five typical power quality disturbance signals, such as voltage oscillation, voltage interruption etc., are simulated, classified and recognized, the classification results of transient voltage disturbances show that the proposed two classification methods for transient power quality disturbances possess good classification characteristics and can be applied in wider scope.