An edge resource perception gate and double threshold hysteresis control-based low-voltage power distribution network time series anomaly detection method

Abstract

The application discloses a low-voltage power distribution network time sequence anomaly detection method based on edge resource perception gating and double-threshold hysteresis control, which comprises the following steps: acquiring low-voltage power distribution network multivariate time sequence data and completing preprocessing and time slice construction; constructing a multi-scale collaborative attention model based on time slice sequences, extracting local time features, scale-reduced features and condition-triggered cross-variable collaborative features; constructing a resource state vector and calculating resource fitness; determining the current model structure state by using a double-threshold hysteresis rule, and generating a corresponding feature branch gating vector; performing forward inference on the model to obtain a reconstructed time slice sequence; calculating an anomaly score according to the reconstruction result, and performing anomaly judgment based on the double-threshold hysteresis rule to obtain a detection result. The application can more effectively distinguish typical abnormal working conditions, and can improve the recognition accuracy of the anomaly detection result and the positioning ability of the abnormal occurrence time by combining the time slice level anomaly score and the time point aggregation output mode.

Images