A Summer Load Forecasting Method for Power Systems Considering the Characteristics of Dog Days Data

To improve the accuracy of the long-term load prediction in the summer, this paper proposes a summer load forecasting method for power systems considering the characteristics of historical dog days data. Firstly, the historical load data information of the power system is obtained, and the training data set and testing data set are generated. The load forecasting model based on the Light Gradient Boosting Machine (LightGBM) is established, and the model parameters are trained and tested based on the data set. Secondly, the historical meteorological data information of the dog days is excavated, and the future load forecasting data set of the power system in different scenarios is constructed by combining the weather forecast data of the next 7 days. Finally, the probability prediction of the summer load in multiple scenarios is carried out to obtain the final summer forecast load. The simulation results prove that the proposed method can not only realize the summer load forecasting of power systems for a long period of time but also has high accuracy and reliability.