Research on High Voltage Cable Accessory Fault Diagnosis Method Based on Fuzzy Theory and Evidence Theory

The paper delves into the assessment of cable accessories, components integral to connecting, protecting, and securing cables in electrical systems. As a pivotal segment of these systems, the malfunctioning of cable accessories can potentially jeopardize operational stability and safety. However, traditional diagnostic approaches for cable accessory faults, often reliant on experience, may encounter limitations in accurately pinpointing faults in long-distance cables due to insufficient precision. Recognizing the health assessment of cable accessories as a multifaceted and multi-objective decision-making process, this study employs a fusion of fuzzy theory and evidence theory. Specifically, it leverages the boundary fuzziness of characteristic indicators to conduct fuzzy processing, thereby obtaining a fuzzy evaluation of these indicators. Subsequently, these fuzzy evaluations, serving as basic belief assignments, are fused using evidence theory, treating the characteristic quantities as evidence. This fusion ultimately yields an overall evaluation of the cable accessories' health status. Additionally, Euclidean distance is employed to further assess the health state of the cable accessories. The proposed framework, coupled with an auxiliary diagnostic system, is shown to effectively assist cable operators in determining the ultimate cause of failures. Specifically, the framework was able to categorize 65% of the analyzed cable accessory faults into the defined categories, while the auxiliary diagnostic system accurately predicted the root cause of failures with a 90% accuracy rate based on quantitative indicators generated from historical fault data.