Dynamic Evaluation Method of Running State of Charging Pile Based on Multi-Source Data of Fault Operation and Maintenance

Abstract This paper proposes a dynamic health status assessment method for charging station faults and operation based on multi-source data, aiming to achieve real-time monitoring and health assessment of charging station operation status through subjective and objective weight optimization and dynamic health scoring. Firstly, by collecting real-time data on the operation status of charging stations, multi-modal abnormal data is filtered using the 3-method and interquartile range method; secondly, missing data is filled using regression analysis, and data is normalized using the minimum-maximum value normalization method; thirdly, a charging station health status assessment system is constructed, combining the entropy weight method and health index as the comprehensive health weight of the analytic hierarchy process to determine the corresponding health level; finally, an engineering example verification is conducted based on the actual charging order data of a power company in Shandong Province in 2024. The practice proves that the proposed method reduces subjective bias through weight fusion and achieves an assessment accuracy of 95%. The entropy weight method compensates for the deficiency of the analytic hierarchy process in being insensitive to data variation, addressing the subjective limitations; the health index introduces the time variable t, enabling the weights to adaptively adjust with equipment aging, which is superior to the static weight model. This method significantly improves the fault diagnosis ability of charging stations and the refinement level of operation and maintenance management.