Title: Application of gradient boosting decision tree algorithm in operation quality evaluation of electric energy metering device in electric power company

Authors: Zhixia Bai; Xinhui Liu; Jun Zhang; Yang Li; Muqing Wang

Addresses: Asset Operation Office, State Grid Shanxi Power Company Marketing Service Centre, No. 10, Wuluo Street, Tanghuayuan District, Taiyuan, Shanxi Transformation and Comprehensive Reform Demonstration Zone, Taiyuan, 030032, China ' Asset Operation Office, State Grid Shanxi Power Company Marketing Service Centre, No. 10, Wuluo Street, Tanghuayuan District, Taiyuan, Shanxi Transformation and Comprehensive Reform Demonstration Zone, Taiyuan, 030032, China ' Asset Operation Office, State Grid Shanxi Power Company Marketing Service Centre, No. 10, Wuluo Street, Tanghuayuan District, Taiyuan, Shanxi Transformation and Comprehensive Reform Demonstration Zone, Taiyuan, 030032, China ' Asset Operation Office, State Grid Shanxi Power Company Marketing Service Centre, No. 10, Wuluo Street, Tanghuayuan District, Taiyuan, Shanxi Transformation and Comprehensive Reform Demonstration Zone, Taiyuan, 030032, China ' Asset Operation Office, State Grid Shanxi Power Company Marketing Service Centre, No. 10, Wuluo Street, Tanghuayuan District, Taiyuan, Shanxi Transformation and Comprehensive Reform Demonstration Zone, Taiyuan, 030032, China

Abstract: This study uses the gradient boosting decision tree (GBDT) algorithm to help the power company evaluate and predict the operational quality of electric energy metering devices. The results show that when testing the algorithm performance, the P value, R value and F1 value of GBDT algorithm model were better than the other two algorithm models. By using the established random forest algorithm (RF), classification and regression tree (CRAT), GBDT metering device operating quality identification models to predict and calculate the fault conditions of 7,000 electric metres, it is found that the prediction accuracy rates of the three algorithms are 74.56%, 71.41%, and 85.31%, respectively. In conclusion, the feasibility of the algorithm is verified, and the accuracy of the algorithm for predicting severe faults in the power system is verified, thereby improving the security of the entire power system.

Keywords: metering device; power company; decision tree algorithm; GBDT; quality assessment.

DOI: 10.1504/IJPT.2023.136145

International Journal of Powertrains, 2023 Vol.12 No.4, pp.323 - 338

Received: 26 Sep 2022
Accepted: 05 Jan 2023
Published online: 18 Jan 2024 *

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