Abstract:In order to achieve the purpose of remotely monitoring the status of power meters and accurately discovering abnormalities of power meters in time, this paper proposes a VAR-based power meter dimensionality reduction error estimation model, which is based on acquiring, analyzing and screening the power meter data, adopting the principal component analysis (PCA) method to reduce the dimensionality of the original data, and extracting the features in the time series through the vector autoregressive model (VAR), so as to accurately predict the power meter consumption and compare the abnormal power meters. The energy meter uses electricity and compares the abnormal energy meters. The PCA dimensionality reduction algorithm deals with the unsolvability of the actual model, and the VAR autoregressive algorithm improves the stability and accuracy of the estimation, which is characterized by high prediction accuracy and small amount of required data compared with the traditional method. In order to verify the effectiveness and practicability of the method, the method is applied to the actual station area for testing, and eight abnormal meters are accurately located by analyzing the electricity consumption data of 127 meters in the station area within half a year. The results show that the method does not need to calculate the network loss independently in advance, and can estimate the smart meter error and network loss rate in real time.