Abstract: A power quality disturbance classification method based on markov transition field (MTF), convolutional neural network (CNN), and particle swarm optimization (PSO) algorithms is proposed to address the trend of complex and diverse power quality disturbances (PQDs) in new power systems. Firstly, MTF is used to convert one-dimensional time-series power quality disturbance signals into two-dimensional modal images, which creat favorable conditions for subsequent data feature extraction. Secondly, a power quality disturbance classification model based on image modal data is constructed using CNN, and the parameters in the model are optimized using PSO algorithm to achieve optimal learning rate and avoid the phenomenon of underfitting or overfitting. Finally, the optimized classification model is used to obtain the classification results of power quality disturbances, and its performance is evaluated using evaluation indicators. The simulation results show that the constructed classification model can effectively classify power quality disturbance signals, with higher classification accuracy and noise resistance.

Key words: power quality disturbance(PQDs), markov transition field(MTF), convolutional neural networks(CNN), particle swarm optimization algorithm(PSO)