Least square support vector machine technique for short term solar irradiance forecasting

Abstract

Application of support vector machine (SVM) has been widely used in regression and prediction. Accurate prediction of daily solar irradiance is important in photovoltaic power system since rapid changes in solar irradiance will easily affect the performance of the whole system. This paper presents the short term future prediction of the solar irradiance using least square support vector machine (LSSVM). Historical solar data set including daily solar irradiance over a period of three years (1 January 2014 to 31 December 2016) has been collected at Green Energy Research Centre, Universiti Teknologi Mara (UiTM) Shah Alam, Selangor. This related information shall be used in prediction of the future solar irradiance which useful for predicting electrical parameters of a PV system especially large scale solar (LSS) farm. The simulation was carried out using SVM Toolbox in MATLAB software. The results show good agreement between the predicted and the measured values.Application of support vector machine (SVM) has been widely used in regression and prediction. Accurate prediction of daily solar irradiance is important in photovoltaic power system since rapid changes in solar irradiance will easily affect the performance of the whole system. This paper presents the short term future prediction of the solar irradiance using least square support vector machine (LSSVM). Historical solar data set including daily solar irradiance over a period of three years (1 January 2014 to 31 December 2016) has been collected at Green Energy Research Centre, Universiti Teknologi Mara (UiTM) Shah Alam, Selangor. This related information shall be used in prediction of the future solar irradiance which useful for predicting electrical parameters of a PV system especially large scale solar (LSS) farm. The simulation was carried out using SVM Toolbox in MATLAB software. The results show good agreement between the predicted and the measured values.