Ning Gao, Chao Wu, Ling Liu, Fei Yang, Na Liu, and Chen Shen
Support vector machine; Risk characteristics; Distribution network equipment; Operation failure risk; Digital twins; Randomized Lasso algorithm
The distribution network is a complex dynamic system comprising various types of equipment and involving extensive operational data across different time periods, which poses challenges for accurate real-time detection of operational failure risks in such equipment. To obtain full lifecycle operational data of dynamic distribution net- work equipment and accurately detect its failure risk in real time, a method based on a support vector machine (SVM) and risk char- acteristics is proposed for detecting operational failure risks in dis- tribution network equipment. Digital twin technology is employed to construct a dynamic model of the distribution network. The up- dated dynamic model, combined with load analysis and equipment condition awareness, enables accurate real-time acquisition of lifecy- cle operational data from distribution network equipment, thereby providing comprehensive and accurate data support for subsequent detection. The Randomized Lasso algorithm is used to extract key risk characteristics associated with equipment operational failures from the data, forming a key risk feature set. Using this feature set as input, a least squares support vector machine (LSSVM) model is constructed to detect and provide early warnings for operational fail- ure risk categories of distribution network equipment. Experimental results demonstrate that the proposed method can effectively detect all categories of operational failure risks in the test distribution net- work equipment. The detection results are consistent with actual conditions and allow early risk warning based on the detection out- comes. The Kappa coefficient of the detection results reaches above 0.9, the coverage rate is close to 100%, and the average bandwidth remains below 0.8, indicating high detection accuracy and coverage. State Grid Hebei Province Electric Power Co., Ltd. Zhengding County Power Supply Branch, Shijiazhuang, Hebei, 050800, China; e-mail: [email protected], [email protected], [email protected], [email protected], nao- [email protected], [email protected] Corresponding author: Ning Gao Recommended by Hasmat Malik (DOI:10.2316/J.2025.203-0615)
Important Links:
Go Back
