Research on Power Load Characteristics of Power Users Based on Deep Confidence Network

The non-technical loss of power system represented by abnormal power consumption of power users will usually cause a significant increase in the operating cost of power supply companies. In order to reduce the operation cost of power supply companies, this paper proposes an abnormal power consumption detection method of power users based on deep neural network. According to the power load characteristics of power users, the deep confidence network (DBN) is used to extract the features of the original power load data and obtain the consistent features. The feature classification is completed based on the limit learning machine (ELM), so as to establish the basic model of abnormal power consumption detection of power users.