Dual Data and Mechanism Based Prediction Model for Photovoltaic Modules

Accurate photovoltaic power prediction is of great significance for the stability and safety of the power grid, but present methods lack explainability or accuracy. This paper develops a dual data and mechanism based method for photovoltaic power prediction. Its model combines physical a-priori knowledge of the photovoltaic modules for explainability and data-driven fitting for accuracy. This method is based on the single diode model and its model parameters are adjusted using a hybrid model based on convolutional neural networks and long short-term memory (CNN-LSTM). We evaluate this model by conducting experiments on open-source photovoltaic dataset. The results demonstrate the effectiveness of this dual data and mechanism driven model.