Research on Data Anomaly Detection and Repair Methods for Smart Meters

This research suggests a deep learning model that combines CNN and LSTM for detecting and fixing anomalies in smart meter data. The model can accurately detect and correct abnormal data by extracting time series and local features from meter data effectively. First, the paper examines the technologies related to managing and detecting anomalies in smart meter data, before developing and putting into action a system for detecting and fixing anomalies in smart meter data. The primary loss function used is the mean square error (MSE), which is paired with a self-supervised learning algorithm for the automatic identification and correction of abnormal meter data. In the application part of the model, this paper verifies the effectiveness and practicality of the proposed method through actual case analysis in the power system. The experimental results show that the designed model has achieved high repair accuracy in multiple indicators, and effectively solves the anomaly problem in meter data, providing a guarantee for the data integrity and accuracy of the power system.