Voltage Prediction Method Based on Data-driven and Graph Neural Networks

In recent years, with the large-scale integration of new energy sources such as photovoltaic and wind power and nonlinear loads into distribution transformer areas, the problem of voltage fluctuations in distribution transformers has become increasingly prominent, and the power quality issue in distribution transformer areas has become increasingly prominent. Therefore, in order to better predict and analyze the voltage trend changes of regional distribution transformers, this paper proposes a voltage prediction method based on data-driven and graph neural networks, and conducts experiments based on the operation data of 32 distribution transformers in a certain area of Shanghai. The results show that the ATGCN-TCN graph neural network model proposed in this paper has a very good fitting ability for the simultaneous prediction of multiple distribution transformer voltages.