基于SOS-DBN的电能质量智能稽查识别系统的研究

Research on power quality intelligent audit identification based on SOS-DBN

为提高电能质量智能稽查识别的精度,本文针对深度置信网络模型性能受其参数影响,运用经验模态分解提取不同电能质量信号的IMF分量,在此基础上计算不同IMF分量的样本熵,将样本熵作为SOS-DBN的输入,不同电能质量信号类别作为SOS-DBN的输出,建立SOS-DBN的电能质量智能稽查识别模型。与PSO-DBN、GA-DBN和DE-DBN相比,SOS-DBN可以有效提高电能质量信号稽查识别的准确率,为电能质量信号稽查识别提供新的方法。

In order to improve the accuracy of power quality intelligent audit identification,a power quality intelligent audit identification model based on SOS-DBN was proposed to improve the performance of depth confidence network model affected by its parameters.Firstly,IMF components of power quality signals are extracted by empirical mode decomposition.Then,the sample entropy of different IMF components is calculated,and the sample entropy is taken as the input of SOSDBN,and the different power quality signal categories are taken as the output of SOS-DBN,so as to establish the power quality intelligent audit identification model of SOS-DBN.Compared with PSO-DBN,GA-DBN and DE-DBN,SOS-DBN can effectively improve the accuracy of power quality signal inspection and identification,providing a new method for power quality signal inspection and identification.