Design of Power Transformer Fault Diagnosis System Based on Wavelet Transform

Partial discharge of large-scale power equipment has always been one of the important research issues in the field of insulation. The research results of this problem are of great significance to insulation diagnosis, life prediction, online monitoring, safety and economic operation of power system, etc. Transformer accidents will lead to local or even large area power outages, which will seriously affect the normal production and life. Therefore, its running state is related to the reliability of electric power production. Judging the current situation of power transformers in this paper can improve the reliability and safety of transformers, while reducing blind maintenance. However, the traditional diagnosis methods are not accurate enough and timely, which brings great security risks. According to the strong time-frequency local processing ability of wavelet, the theory of wavelet transform is analyzed firstly, and its application in signal detection is analyzed; On this basis, a new wavelet energy ratio criterion based on signal energy is proposed. The quantitative diagnosis method of entropy distance is used to diagnose transformer partial discharge fault more accurately. The above diagnosis process is simulated by MATLAB signal processing toolbox, which proves that the method is effective.