Automatic Condition Monitoring and Fault Diagnosis System for Power Transformers Based on Voiceprint Recognition

Power transformer emits continuous vibration signals during working, which contain plenty of impulses and fluctuations caused by mechanical faults, and these signals are the main data source to evaluate the operational condition of a power transformer. This study researches the voiceprint features of vibration sound signals within power transformers under different operational conditions and constructs an automatic and noninvasive condition monitoring and fault diagnosis system based on acoustic characteristics. First, a normal/faulty operational condition classification module is proposed using acoustic short-time energy and zero-crossing rate to evaluate the real-time operational condition of a power transformer. Second, a multifault recognition module with finite signal samples is constructed based on the time–frequency characteristics of acoustic signals, and an automatic fault diagnosis system for power transformers is built. To improve the accuracy and robustness under dynamic noisy environments and complicated operational conditions, a multidimensional performance optimization method is proposed to handle the problems caused by environmental noises, mixed faults, and unknown faults. Experimental results demonstrate that the proposed condition monitoring and fault diagnosis system can fast distinguish the faulty operational conditions and diagnose six types of typical faults of power transformers, as well as mixed faults, under different noisy environments with an accuracy of more than 97.83%.