Research on intelligent spectrum monitoring based on power transformer

Abstract

As the core key equipment of the power system, power transformers transmit AC power to the power grid. They are an indispensable part of long-distance power transmission and play a vital role in the safe, stable and reliable operation of the power grid. In order to perform on-line detection of partial discharge of transformers and classify faults, this paper proposes an intelligent spectrum monitoring method based on power transformers. This article first designs a power transformer electromagnetic spectrum signal acquisition system, and then performs data preprocessing on the collected power transformer spectrum data using wavelet transform. Wavelet transform can check the frequency domain characteristics of the local time domain process and the time domain characteristics of the local frequency domain process. After that, this paper use a convolutional neural network model, which is a residual network. The model makes a reference to the input of each layer and learns to form a residual function instead of learning some functions without reference. This residual function is easier to optimize and can greatly deepen the number of network layers, Input the preprocessed I/Q data into the network model for fault classification of power transformers. Based on the above, the frequency spectrum of the power transformer can be effectively monitored and fault detection can be performed. Experimental results show that our method can classify power transformer faults with an accuracy rate of 98.4%, and has a higher accuracy rate than traditional transformer fault detection techniques.