Fault Diagnosis of Transformer Inter-Turn Short Circuit Based on Automatic Multiscale-Based Peak Detection Algorithm Combined with Magnetic Flux Leakage Analysis

The power transformer is crucial to the power system. Traditional transformer protection measures cannot quickly and sensitively identify coil-related faults. Magnetic flux leakage (MFL), as an intermediate product of energy conversion during the operation of a transformer, undergoes changes when there is a coil fault in the transformer. However, there is limited research on the analysis and fault diagnosis of the MFL field before and after transformer faults. In this study, a three-dimensional finite element simulation model of the transformer is established. The simulation model data are analyzed to select appropriate measurement paths. Through local analysis of measurement point data and overall analysis of measurement lines, the periodic characteristics of the transformer MFL field distribution are further elucidated. Find the significant MFL change characteristics after transformer inter-turn short circuit fault, and try to analyze the causes of the characteristics. This article proposes to use the peak point detection algorithm- Automatic Multiscale-based Peak Detection (AMPD) algorithm to find the peak point of the characteristic waveform, and realize the fault diagnosis of the transformer through the MFL data. The research supplements the correlation analysis of the MFL characteristics of inter-turn short circuit faults in transformers. The proposed method provides new insights into the detection, localization, and assessment of inter-turn short circuit faults in transformers.