Assessment of Electric Power Equipments Aging Trend Based on Robust Locally Weighted Regression Algorithm

As a usual phenomenon, equipment aging has gradually been paid more attention by a large number of power companies. Until now, the deterministic method based on engineering judgment is used in most of power companies to judge electric power equipment out of service for aging. Aiming at the disadvantages brought by engineering judgment, a method based on robust locally weighted regression algorithm is proposed to evaluate the aging trend of electric power equipment in this paper. Combined with the actual operation states and characters of electric power equipment, the method uses robust locally weighted regression algorithm to establish the aging model of electric power equipments, and analyzes the aging status of the electric power equipment to approach the aging process of electric power equipment, so that make out very lively estimation to the electric power equipment aging trend. Thereby, the results can provide an early forecasting for equipment fault and scientific evidences for equipment maintenance and out of service, which are caused by equipment aging. Finally, the method is applied to analyze an actual transformer. The result in accord with the actual situation verified the availability of the method. For power system planning, analysis of equipment standby, and the course of electric power marketization, the application of the method have a great significance.