The continuous growth of global energy demand and the increasing awareness of environmental protection have made load forecasting in distribution networks a crucial part in operation and planning of power systems.The present work made a preliminary attempt to establish an attention-mechanism-optimized GRU hybrid neural network model to improve accuracy and robustness of distribution network load forecasting.The proposed model was verified by a test on public data-base,compared with conventional GRU model,superior in forecasting accuracy and stability,and thereby potentially con-ducive to efficient operation of smart distribution networks.
distribution networkload forecastingGRU neural networkAttention mechanismdeep learningtime se-ries forecasting
维普
万方数据
