A Multi-Fault Recognition Method Based on Excitation Voltage Information Fingerprint for the Generator Excitation System

In order to address the challenge of lacking diverse and accurate information sources for generator excitation system fault recognition, we have proposed a multi-fault recognition method solely based on the excitation voltage information fingerprint for the generator excitation system. Firstly, this method makes the recognition approach where the fingerprint code matches with the fault fingerprint database suitable for periodic excitation voltage signals, which enhances the robustness of fingerprint encoding against time shifts in signals. Meanwhile, this method effectively achieves the rapid and accurate identification of various fault types in the three-phase bridge rectifier circuit of excitation system. Through simulation validation, the generated fingerprint data in this method is minimal, enabling extensive storage and rapid matching for fault scenarios under different parameters. As a result, it demonstrates strong discriminative and noise-resistant capabilities for asymmetric fault states at different triggering angles.