Load forecasting is an important part of power management, and the forecasting results directly affect the security, reliability and economy of system operation. Aiming at the problem that the traditional load forecasting method is not highly accurate, as well as the problem that it cannot combine the historical change rule and influencing factors of the load itself, we proposes the load forecasting method of distribution network based on LSTM algorithm. Considering the historical load and various types of load influencing factors, LSTM with time-series memory function is utilized to construct short-term and medium-term load forecasting models. The experimental results show that the average accuracy of the short-term prediction results of our proposed model reaches 0.92, and the medium-term prediction accuracy reaches 0.84, which can accurately carry out the prediction of distribution network loads, and has the practicality and good application value.
Research on LSTM-based load forecasting method for distribution networks
11.10.2023
2749250 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Taylor & Francis Verlag | 2024
|