Disturbance Identification of Power Quality Based on Markov Transition Field and Deep Residual Network

Electric energy is an indispensable secondary energy in today's society and an important foundation for the development of the national economy. With the rapid development of renewable energy and the application of a large number of power electronic devices, power quality disturbances(PQDs) are gradually becoming more complicated and diversified. Since unqualified power quality will cause a series of adverse effects on the production user side and electronic equipment, this requires real-time monitoring of power quality and rapid and effective identification and classification of various disturbance problems, thereby effectively improving power quality. Due to the slow computational speed and low identification rate of traditional power quality disturbance recognition algorithms when dealing with large amounts of power quality data, this paper uses the powerful representation learning ability of the deep residual network (Resnet) to classify power quality disturbances. First, the Markov transition field(MTF) is used to transform the one-dimensional power quality disturbance time series signal into a two-dimensional visualization graph, and then the deep residual network is used to extract the disturbance features in the two-dimensional image and perform classification and recognition. The experimental results show that the recognition algorithm based on Markov transition field and deep residual network has good accuracy in the classification of power quality disturbance signals, and the effective recognition rate reaches 99.875%.