Research on Capacity Estimation of Photovoltaic System Carrying Capacity Improvement in Low-Voltage Distribution Network Based on Data Mining

This paper proposes a capacity estimation method for photovoltaic system carrying capacity improvement in low-voltage distribution network, which combines LSTM and GRU joint algorithms. First, the historical operation data of low-voltage distribution network is collected by big data technology, including current, voltage, load and other parameters, and a time series data set containing multiple variables is constructed. Then, a deep learning algorithm based on LSTM and GRU joint optimization is designed, which can effectively capture the operation law and potential risks of photovoltaic system in distribution network and accurately predict the carrying capacity. This model can not only accurately predict the carrying capacity of photovoltaic system under different operating conditions, but also realize efficient capacity estimation. The influence of PV capacity on load distribution and stability of grid is investigated by numerical simulation. Experiments prove the validity of the model. Simulation results show that the proposed algorithm can increase the computing precision by about 15%, and satisfy the demand of capacity upgrading in different load situations.