A Method to Overcome Power Grid Voltage Fluctuation in the Application of Non-Invasive Load Identification Technology

Non-invasive load identification technology is one of the important technologies in smart grid construction. Accurate identification of the operation status of electrical equipment is of great value for demand side dispatching. For event-based NILM, the background voltage (power entry voltage) of household appliances is generally different every time they are turned on and off, This leads to the deviation of feature quantities extracted by some electrical appliances under different voltage levels, which affects the accuracy of load identification. In view of this situation, a feature extraction method to overcome the change of background voltage is proposed. The core idea is the measured feature values of different electrical appliances under different background voltages, The load characteristic can be compensated to the standard voltage (220V), and a feature extraction scheme to overcome the change of background voltage is realized. The test process and results show that this scheme has low computational complexity. Compared with the uncompensated extraction scheme, it can effectively improve the accuracy of feature extraction and has certain application value.