An application of the GWO-ELM hybrid model to the accurate 2 prediction of solar radiation for the purpose of sustainable energy 3 integration

Despite their drawbacks, fossil fuels like gas, oil, and coal are widely used globally for energy purposes. On the contrary, nascent stages of the energy market are occupied by renewable energy sources such as solar, wind, and hydro, even though they are depletable and secure. Solar energy, which can be utilized on both small and large scales, is the purest form of renewable energy currently accessible. Accurate solar irradiance prediction is vital for sustainable solar energy use. Direct normal irradiance significantly influences solar power generation and can predict solar energy output. This study aims to develop a new hybrid model combining the grey wolf optimizer and extreme learning machine algorithm for accurate direct normal irradiance prediction despite forecasting complexities. Ten input characteristics were chosen for the suggested model from a set of ten features that were gathered in Qinghai from June 30, 2023, to July 1, 2022. Following this, a feature selection process was conducted on these attributes, and the algorithms proceeded to process only the most significant ones. Upon comparing the alternative model utilized in this study to the suggested model its capability and effectiveness were unequivocally established. Consequently, employing the suggested model enables one to reliably estimate direct normal irradiance for solar energy generation. Performance indicators for the proposed GWO-ELM model include RMSE of 63.17, MAE of 46.68, MSE of 3990.80, and RSE of 89.39. These findings show that the model can accurately and reliably estimate direct normal irradiance, despite the difficulties of solar energy forecasting.