Title: Research on short-term load forecasting of power system based on gradient lifting tree

Authors: Taofang Xia; Yajuan Zhou; Shian Zhan; Hua Lin; Tuo Zhang; Yueqian Lan

Addresses: State Grid Fujian Marketing Service Centre, Fujian 350013, China ' Energy Internet Research Institute, Tsinghua University, Beijing 100084, China ' State Grid Fujian Marketing Service Centre, Fujian 350013, China ' State Grid Fujian Marketing Service Centre, Fujian 350013, China ' China Urban Construction Design and Research Institute Co., Ltd., Beijing 100120, China ' Energy Internet Research Institute, Tsinghua University, Beijing 100084, China

Abstract: To reduce the average absolute error of short-term load forecasting and improve the calculation speed, a short-term load forecasting method of power system based on gradient lifting tree is designed. Firstly, after determining the input of short-term load forecasting, the fuzzy probability is used to quantify the short-term load influencing factors and complete the load data pre-processing. Secondly, the short-term load series are processed by the difference decomposition method. Finally, in the direction of the negative gradient of the loss function, a strong regression gradient lifting tree is established, and the historical short-term load sequence is input into it to obtain the load forecasting results. The experimental results show that the maximum average absolute error of the prediction results of this method is only 1.41%, the minimum prediction calculation speed is 7.5 s, and the maximum prediction calculation speed is only 9.6 s.

Keywords: power system; short-term load; load forecasting; pre-treatment; differential decomposition; gradient lifting tree.

DOI: 10.1504/IJPEC.2022.130951

International Journal of Power and Energy Conversion, 2022 Vol.13 No.3/4, pp.235 - 247

Received: 25 Aug 2022
Accepted: 22 Dec 2022
Published online: 14 May 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article