| Summary: | Malaysia has been strongly dependent on non-renewable energy such as coal and natural gas to power up the country. As the country’s natural resources are now depleting, solar energy is seen as the most suitable future energy specifically due to Malaysia’s strategic location at the equator of the Earth. In Malaysia, many Large-Scale Solar Photovoltaic (LSSPV) plants have been developed as a result of effective policy by the government. However, one of the challenges faced by the independent power producers is the uncertainty of the output power from the LSSPVs due to fluctuation of weather conditions. This paper presents a forecasting power generation model of LSSPV farm using Artificial Neural Network (ANN) and Particle Swarm Optimization (PSO) technique. The UiTM 50MW LSSPV in Gambang, Pahang has been used as a case study. The PSO technique is utilized to optimize the weight of ANN for determining the best Mean Square Error (MSE) and regression performance. The forecasting model uses total global horizontal irradiance, global irradiation on the module plan and PV module temperature as input variables while Alternating Current (AC) output power as output variable. The input variables were chosen from a filtration process of the historical data. The historical data used in the training and testing process are from the month of May 2019 until August 2019. The data is forecasted at every 30 minutes’ basis and compared with the actual AC output power. The result shows that the ANN-PSO method outperforms the traditional ANN with a better MSE and regression performance. © 2022. All Rights Reserved.
|
|---|