Prediction of solar irradiance using convolutional neural network and attention mechanism-based long short-term memory network based on similar day analysis and an attention mechanism

Attention-Based Models for Multivariate Time Series Forecasting: Multi-step Solar Irradiation Prediction

“…Li et al [ 46 ] suggest a two-channel method employing LSTM, WGAN, and CEEMDAN, splitting solar output into frequency-based subsequences for prediction, and integrating their values for final output. Hou et al [ 47 ] introduce CNN-A-LSTM, employing comparable day analysis and attention processes, surpassing various models on the NSRDB dataset for accurate solar irradiance prediction, particularly excelling in unclouded and partly cloudy conditions. Munsif et al [ 48 ] explore the CT-NET model, a transformer variation combining CNN and multi-head attention for both local and global information utilization, outperforming CNN-RNN, CNN-GRU, and CNN-LSTM across seasons using the Alice Springs dataset.…”

Unlocking renewable energy potential: Harnessing machine learning and intelligent algorithms

“…The comprehension of the intricate and dynamic character of solar PV generation is greatly enhanced by these techniques, which also raise the precision and effectiveness of solar power forecasts (R. Luo et al, 2021). A lot of researchers have concentrated on encouraging the application of ML in the domain of solar energy including photovoltaic (PV) array layout optimization (Subhashini et al, 2023), PV energy predictions (Mellit and Pavan, 2010), solar irradiance predictions (Hameed et al, 2019;Hou et al, 2023) and enhancing the efficiency of solar chimney facilities (Mandal et al, 2024;Taki et al, 2021). An ANN architecture employed for the prediction of global solar irradiance (GSR) depicts the way it is employed in Figure 9 (Aljanad et al, 2021).…”