Global solar energy potential forecasting through machine learning and deep learning models

Climate change is accelerating at an alarming rate, 2024 has been verified as the hottest year so far, with an average temperature of 1.55 °C warmer than upstream values set in the Paris Agreement. As such, extreme weather patterns like floods, hot weather, wild fires and glaciers melts that all pose a threat of harm to ecological systems. By investing in solar technology, nations can work towards a more sustainable energy future and addressing the pressing challenge of climate change. This study exploited the global solar photovoltaic (PV) energy potential using the Seasonal Autoregressive Integrated Moving Average with Exogenous Factors (SARIMAX) and Temporal Convolutional Network (TCN) models implemented in Python for the period 2023 to 2050 by taking input data from 2000 to 2022. The results revealed that, the solar PV capacity was 1.23 GW in the year 2000 which then increased to 1053.12 GW by 2022. SARIMAX and TCN models estimated the future of solar PV capacity which is increased from 1291.29 GW and 1094.40 GW in 2023 to 11641.41 GW and 11577.24 GW until 2050. However, the solar PV energy was 1.03 TWh in 2000 which then increased to 1323.32 TWh in 2022. SARIMAX and TCN models forecasted the future of solar PV energy which is increased from 1935.52 TWh and 1557.92 TWh in 2023 to 14967.15 TWh and 15928.52 TWh until 2050. It is observed from the results that SARIMAX model has higher accuracy as compared with the TCN model.