Prediction of Regional PV Power Generation Based on LSTM-CNN

The increasing installed capacity of PV plants across a country has transformed the power grid into a decentralized system, resulting in heightened grid complexity. Predicting regional PV power is essential to ensure grid stability and effective planning. In this work, we propose an LSTM-CNN model to predict regional PV power by utilizing actual weather data from the investigated region and considering PV installed capacity in the investigated region over time. Experimental results indicate that the proposed model can achieve lower RMSE and MAE scores, approximately 13.12 MW and 6.29 MW, respectively, with R2 coefficients around 0.94. Despite its good performance, this model requires a longer learning process duration.