Load Forecasting for Station Areas Based on User Portrait

• Published in: 2024 3rd International Conference on Energy and Electrical Power Systems (ICEEPS) pp. 1270 - 1280
• Main Authors: Yu, Zongchao, Wen, Ming, Chen, Zizi, Luo, Shuchen, Li, Peiqiang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.07.2024
• Subjects: Accuracy; Distribution Transformer Load; Electricity; Industries; Industry Load Forecasting; Libraries; Load forecasting; Load modeling; Power grids; Prediction algorithms; Predictive models; Transformers; User Portrait
• DOI: 10.1109/ICEEPS62542.2024.10693270

Distribution transformer load forecasting is an important supporting technology for safe and efficient operation of distribution network. In this paper, a method of load forecasting for station area based on user portrait is proposed. First, the industry labels of station areas are obtained, the load portrait of the industry of station area is carried out in the way of "upward aggregation and downward classification", and the load characteristics of the historical load of each industry are analyzed. Secondly, the prediction model base is constructed, considering the characteristic labels of various industries, and the optimal prediction model is matched. Finally, combining the user portrait results of the industry of the station area and considering PV stripping when forecasting, the power demand of the station area load plus the distributed power supply output is predicted first, and the predicted value of distributed PV output is deducted from the forecast result to obtain a more accurate station area load forecast value, and the validity of the model is verified by real examples.