Diode Envelope Extraction and Fault Diagnosis of On-Load Tap-Changer Vibration Signals in Converter Transformers

This paper proposes a fault diagnosis method for the on-load tap changer (OLTC) of a converter transformer based on diode envelope detection, empirical mode decomposition (EMD), and support vector machine (SVM). First, a diode envelope detection algorithm for envelope extraction was designed and implemented to process the current and vi-bration signals of the OLTC drive motor, and the influence of RC parameters was studied. Experiments show that selecting RC values of 400 and 5000, respectively, achieves optimal envelope extraction effects for different signals. Then, multi-frequency features were extracted using EMD decomposition, and fault types were classified using the SVM model. The experiments covered three typical faults: spring loosening, fatigue, and fracture. Results show that the method achieved an accuracy rate of 95.5%, effectively identifying different fault types. The study demonstrates that diode envelope detection offers strong peak preservation, and EMD feature extraction improves classification performance.