Noise feature extraction and judgment method for transformer fault diagnosis

Abstract

The invention discloses a noise feature extraction and judgment method for transformer fault diagnosis, and solves the problem that an existing human-based sensory experience analysis and transformerfault diagnosis technology is not perfect. The method comprises the following steps of collecting a normal operation state of a transformer and characteristic frequency spectrums of various faults, and establishing a database; receiving a sound wave signal of the transformer, and judging whether the sound wave signal is sent by the transformer or not; extracting an eigenvector of the noise of thetransformer, and building a membership function; calculating a distance between the eigenvector of the noise and a noise template stored in the database to obtain a close membership degree; searchingfor the same or similar templates according to the close membership degree; and outputting a recognition result. According to the method, the eigenvector of the noise is extracted; according to a fuzzy recognition algorithm, the distance between the eigenvector of the noise and the noise template stored in the database is calculated; the same or similar templates are found; and the recognition result is output, so that potential faults of the transformer can be discovered as early as possible, and safe operation of the transformer can be guaranteed.

Description

technical field [0001] The invention belongs to the technical field of transformer monitoring, in particular to a noise feature extraction and judgment method for transformer fault diagnosis. Background technique [0002] The power transformer is an important substation equipment in the power system, and its operating status directly affects the safety and stability of the system. Early detection of latent faults of transformers to ensure the safe operation of transformers, thereby improving the reliability of power supply is an important issue concerned by the power sector. Transformer faults are the result of the comprehensive effect and long-term accumulation of the transformer itself and its application environment. Therefore, the symptoms of transformer faults are various, and the relationship between fault symptoms and fault mechanisms is also intricate. This has caused great difficulties in establishing a general transformer fault control method. big difficulty. So ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a noise feature extraction and judgment method for transformer fault diagnosis. By determining the characteristic spectrum of the normal operation state and different faults of the monitored transformer, these characteristic values ​​are used to construct the membership function and obtain the closest membership. The operating state of the equipment can be judged by using the fuzzy recognition method, which solves the problems that the existing analysis of the operating state of power equipment based on human sensory experience fails to achieve quantitative analysis and the fault diagnosis technology of transformers is not perfect.

Images