Detection of Abnormal Data of Gateway Energy Meter Based on User Dynamic Behavior Mining

Since the quality of measuring data of gateway energy meter has significant impacts on the fairness and impartiality of trade settlement, detection of abnormal data is of great importance. However, recent related researches of meter measuring data focus on the abnormal causes and the design of an on-line monitoring system, the value of data mining itself is ignored. For this reason, this paper proposes a method of detecting abnormal data in the gateway energy meter based on user dynamic behavior mining. According to the regularity of power load data, a detection method for random abnormal points based on time series decomposition and autocorrelation analysis is proposed. At the same time, for the continuous abnormal situation of the data caused by the user’s power theft behavior that the power grid company is concerned about, the line loss rate and the continuous decline trend of the power consumption are extracted. Then, a detection model is established to realize the detection of abnormal data in the gateway energy meter. Finally, the validity of this method is verified based on the measured data of the power grid. The result proves that the proposed method can accurately detect the abnormal data of the gateway energy meter and meet the needs of engineering practice.