Short‐Term Multi‐Horizon Line Loss Rate Forecasting of a Distribution Network Using Attention‐GCN‐LSTM

ABSTRACT Accurately predicting line loss rates is crucial for effective management in distribution networks, particularly for short‐term multihorizon forecasts ranging from 1 hour to 1 week. In this study, we propose attention‐GCN–LSTM, a novel method that integrates graph convolutional networks (GCN), long short‐term memory (LSTM) and a three‐level attention mechanism to address this challenge. By effectively capturing spatial and temporal dependencies, our model enables precise forecasting of line loss rates across multiple horizons. Comprehensive evaluations using real‐world data from 10 kV feeders demonstrate that the attention‐GCN–LSTM model consistently outperforms existing algorithms, achieving superior prediction accuracy and excelling in multihorizon forecasting. This model holds significant potential for improving line loss management in distribution networks.