Lean load forecasting method for distribution substation area
Hongjun Wen, Xiaohshuang Li, Bowei Chen, Jiqiao Chen, Songkun Yang, Qian Wu · 2024
In order to improve the practicality level of forecasting methods, refine the granularity of load forecasting results, and achieve the integration of multiple elements of "source network load storage", this paper adopts a "bottom-up" forecasting approach to conduct research on lean load forecasting methods based on distribution transformer areas. The research task is divided into four steps. The first step is to classify the differentiated electricity consumption scenarios in the distribution substation area. The second step is to conduct flexible resource research and characteristic analysis. The third step is to conduct research on the load characteristics of typical substation areas. The fourth step is to quantitatively analyze the sensitivity level of various influencing factors on the load of the substation area. The fifth step is to carry out precise load forecasting in the distribution substation area. In the process of load forecasting, a combination of multiple methods such as big data clustering, neural network, and Monte Carlo were used. To verify the rationality of the research results, detailed calculation cases were provided. The method proposed in this article is simple and easy to implement, with a clear physical meaning. Based on this method, it can gradually extend from bottom to top to 10kV lines, substations, power supply units/grids, and ultimately form load forecasting results suitable for engineering practice.