Root Cause Analysis Model of Abnormal Line Loss in Substations Based on Multi-source Data Fusion

This paper proposes a root cause analysis model of abnormal line loss in substations based on multi-source data fusion. Using multi-dimensional data such as power equipment operation status, user load characteristics, and meteorological conditions, a fusion algorithm is designed to improve the accuracy of anomaly detection and the scientific nature of root cause analysis. Through data preprocessing and feature extraction methods, a multi-source data fusion framework is constructed to ensure the spatiotemporal consistency between different data sources. Then, this paper proposes an adaptive weighted data fusion algorithm (AWDF) to dynamically adjust the weights of different data sources in anomaly detection, so that the model can more accurately capture the main influencing factors of line loss anomalies. Subsequently, combined with graph neural network (GNN) and anomaly detection algorithm, intelligent recognition of abnormal data is realized, and the possible root causes of abnormal line loss in substations are derived through Bayesian reasoning. This paper uses the operation data of power substations in a certain area, including more than 10,000 equipment monitoring records, load data, and meteorological data. Through simulation experiments and comparative analysis, the effectiveness of this method is verified. Experimental results show that compared with traditional methods (such as anomaly detection based on a single data source), this method improves the anomaly detection accuracy by 17.3%, the recall rate by 20.8%, and significantly enhances the accuracy of root cause analysis.