A Study on Deep Learning-Based Photovoltaic Power Generation Prediction Model Using Meteorological Data

Due to climate change and environmental concerns, interest in renewable energy is growing. Photovoltaic power is recognized as a clean, sustainable energy source, generating electricity through various systems including ground-mounted and building-integrated photovoltaic(BIPV) installations. Photovoltaic power generation is significantly influenced by meteorological factors such as temperature, humidity, and solar radiation. Therefore, accurate photovoltaic power prediction is essential to ensure power system stability and reliability. This paper proposes a deep learning model for photovoltaic power prediction using meteorological data. The model integrates Graph Attention Networks and Bi-LSTM to extract features from meteorological data and sequentially predict solar irradiation and power generation. The evaluation of the proposed model was conducted on both ground-mounted and building-integrated photovoltaic systems. Results demonstrate that the proposed model can be effectively applied to photovoltaic systems installed in various orientations, including east-facing and west-facing installations.