A Non-invasive Load Identification Method Based on V-I Trajectory Mixing Feature

Abstract

Non-invasive load decomposition technology is an important part of the smart grid technology system. Aiming at the problem that traditional methods cannot accurately identify household loads containing high-order harmonics, a load identification method based on V-I trajectory matrix-power and high-order harmonics multi-feature fusion was proposed. First, the V-I trajectories, power characteristics, and harmonic characteristics of 11 typical household loads are analyzed, and a hybrid feature matrix construction method based on pixel image transformation is proposed. Through binary coding transformation, the power and high-order harmonic characteristics of the load are fused with the basic V-I pixel trajectory, enriching the characteristic information of the sample. Second, a recognition algorithm based on Gaussian Mixture Model (GMM) and Support Vector Machine (SVM) is established, and the probability distribution of the two algorithms is fused to generate the final recognition result. Finally, the Plug Level Appliance Identification Dataset (PLAID) was used to verify the proposed method, and the accurate identification rate of various household loads was over 93%. Experimental results show that the proposed method can accurately identify household load, including high harmonic load.