Fuzzy C-Means Clustering Algorithm for Non-Intrusive Load Identification Considering the Difference in Information Expression of the Features

Non-intrusive load monitoring is an important means to achieve grid-load friendly interaction and efficient demand management, which has attracted much attention in the construction of new power system. In order to improve the accuracy and generalization performance of the non-intrusive load identification (NILI) algorithm, a fuzzy C-Means (FCM) algorithm modified by considering the difference in information expression of the feature is proposed in the paper. Based on the entropy weight method, the weight coefficients of the features which contain different information for load identification are configured, and the distance measurement matrix modified with feature weighted correction is constructed. Therefore, the modification of the membership matrix and the cluster centers in the identification algorithm is closely combined with the information expression ability of the feature. To improve the solving ability of the algorithm, a method of determining the optimal cluster numbers is built based on contour coefficient, and a partition method of transient-state and steady-state interval of the measurement data is also presented based on the principle of minimum spanning tree. These effectively solve the problem of identification accuracy caused by improper setting of cluster numbers and promote the accurate capture of the feature. The simulation results show that the proposed method can effectively improve the load identification accuracy and generalization performance in application.