Application of Grey Model in Long-Term Solar Energy Forecasting: A Case Study in Taiwan

• Published in: 2021 14th International Conference on Advanced Computer Theory and Engineering (ICACTE) pp. 65 - 71
• Main Authors: Mariano, June Raymond L., Ay, Herchang, Liao, Mingyu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2021
• Subjects: forecasting; grey model; Radiation effects; Renewable energy sources; satellite-derived data; Solar energy; Temperature measurement; Uncertainty; Weather forecasting; Wind speed
• ISSN: 2154-7505
• DOI: 10.1109/ICACTE53799.2021.00018

Promoting green energy is one of the main goals of Taiwan, and the key objective is to increase the renewable sources of energy supply by 2025. To achieve the government's plan, solar energy forecasting plays a critical role in attaining the conversion of solar irradiance to useful output power. In this study, monthly solar irradiance is used in the GM(1,1) to forecast six years of solar irradiance data at two locations in Taiwan, one at Weather Station (467440) in Kaohsiung and another at St John's University (SJU) in Tamsui. The MAPE rates of historical and predicted GHI for the weather station is 6.28%, while SJU is 10.56%. Uncertainties pose the main difficulty in solar energy forecasting, leading to an increase in prediction errors. Based on the MAPE scores, solar energy forecasting from 2020 to 2025 is highly feasible, producing satisfactory to excellent prediction accuracy. This study ascertains the high potential of achieving Taiwan's clean solar energy target, while the GM(1,1) serves as a useful forecasting method for long-term solar irradiation in Taiwan.