Generator fault diagnosis system and method based on current waveform similarity

Abstract

The invention provides a generator fault diagnosis system and method based on current waveform similarity. The method comprises the following steps of 1, obtaining the waveform data of the three-phase current of a generator stator, carrying out normalization processing of the waveform data of the three-phase current, and obtaining the waveform data after normalization processing; 2, respectively acquiring waveform data of A-phase current and waveform data of B-phase current in a set time from the normalized waveform data obtained in the step 1, and respectively recording the waveform data of the A-phase current and the waveform data of the B-phase current as a time sequence A and a time sequence B; 3, calculating similarity between the time sequence B and the time sequence A; 4, judging a state of the generator according to the similarity obtained in the step 3; the method aims at the characteristics that electrical monitoring parameters of the wind turbine generator are complex and rapid in change, and feature quantity extraction is not easy; and the wind turbine generator equipment state in real time is reflected by using the collected current waveform similarity, and the equipment fault probability is analyzed.

Description

technical field [0001] The invention is a diagnosis method applied to electrical / mechanical faults of wind turbines, in particular to a generator fault diagnosis system and method based on current waveform similarity. Background technique [0002] The operating conditions of wind turbines are relatively complex, and they often face more complex environmental conditions such as high temperature, high humidity, salt spray corrosion, lightning, etc., and the wind power fluctuates violently. Its fault manifestations are closely related to the operating state, and it is difficult to realize fault identification. Therefore, it is necessary to master an effective fault feature extraction method for wind turbine components, so as to improve the accuracy of fault diagnosis and state prediction of the unit, and predict the health decline trend of the unit in advance. , reasonably adjust the operation and arrange maintenance, which will improve the safety and reliability of the wind tu...

AI Technical Summary

Problems solved by technology

However, there are still problems in frequency domain analysis such as spectrum leakage, limited resolution, spectrum ambiguity, and the selection of wavelet basis functions have a great influence on the analysis results.

Images