基于高斯混合模型的谐波责任估计方法

Harmonic responsibility estimation method based on gaussian mixture model

针对不完全可观系统提出一种基于高斯混合模型的谐波责任估计方法。依据谐波测量电压的概率分布特性估计各谐波负荷的谐波责任,规避因引入不可测的线路参数对量化谐波责任造成的困难。先根据测得的谐波电压样本训练高斯混合模型;然后,基于贝叶斯信息准则和Kullback‐Leibler散度比率确定混合模型中的高斯分量的数量及位置范围,并通过Z检验实现谐波电压样本的异常检测;最后,通过IEEE 14节点测试系统检验了所提方法的有效性。

A harmonic responsibility estimation method based on the Gaussian mixture model(GMM)is proposed for partially observable systems.This method estimates the harmonic responsibility of each harmonic load based on the probabilistic distribution characteristics of measured harmonic voltages,circumventing the difficulties in quantifying harmonic responsibility due to the introduction of unmeasurable line parameters.Specifically,the process begins by training a GMM using the measured harmonic voltage samples.Then,the number and range of Gaussian components in the mixture model are determined based on the Bayesian information criterion and the Kullback-Leibler divergence ratio.Additionally,anomaly detection of harmonic voltage samples is achieved through the Z-test principle.Finally,the effectiveness of the proposed method is verified using the IEEE 14-node test system.