Solar Irradiance Forecasting Model for Pulau Pinang using Artificial Neural Network

Abstract

One of the important parameters for solar photovoltaic (PV) optimization is solar irradiance. Solar irradiance can be defined as the rate of solar energy when it falls onto the surface. Solar irradiance can make an importance decision, efficiency, performance, and maintenance on energy yield in the future. However, solar irradiance has high degree of uncertainty due to environmental and meteorological condition such as cloud cover, haze, fog and rapid change in ambient temperature. Thus, by forecasting solar irradiance, it would improve the efficiency, performance, and operation of the PV system in order to generate a maximum power. This research aims to develop a forecasting model of solar irradiance based on meteorological data from year 2018 in Pulau Pinang. The forecasting model is developed based on Artificial Neural Network (ANN) Multilayer Perceptron (MLP) method. The accuracy of the forecasting model is compared with the data from The National Aeronautics and Space Administration (NASA) and Sustainable Energy Development Authority (SEDA). The result shows that the forecasting model is able to deliver a good result with the correlation coefficient result of r = 0.82 (forecasted vs NASA) and r = 0.78 (forecasted vs SEDA). Thus, this solar irradiance forecasting model is able to predict almost the same value as the NASA and SEDA and can potentially be used to assist in evaluating and predicting the power output efficiency of the solar PV plant. Hence, this model can be regarded as an important tool in planning and managing the operation of PV system. others’ blogs. Moreover, interview results showed that reflective journals contributed their personal and professional development.