Abstract：Due to the cross-effects of various factors, the electricity sales of distribution substations have irregular fluctuations in their time series, and it is difficult to directly predict their satisfactory results. In this paper, a method for forecasting the electricity sales of distribution stations based on a combined model is proposed. First, variational mode decomposition (VMD) is used to decompose the time series of electricity sales of distribution stations into low-frequency modes and high-frequency modes to reduce time series Non-stationarity; secondly, the Prophet time series prediction model and gated recurrent unit (GRU) neural network are used to predict each mode; finally, the prediction results of high-frequency and low-frequency modes are summed and reconstructed To get the forecast result of the electricity sales amount of the substation. The simulation example selects the daily electricity sales data of a power distribution station in the east of China in recent years for verification. The results show that the proposed method has better accuracy than the traditional prediction method.
刘成, 陈光宇, 张仰飞. 基于组合模型的台区售电量预测研究[J]. 电气技术, 2020, 21(11): 25-31.
Liu Cheng, Chen Guangyu, Zhang Yangfei. Research on electricity sales forecast of substation based on combined model. Electrical Engineering, 2020, 21(11): 25-31.
2020, Vol. 21 
