Solar Irradiance Estimation at Certain Location Using Artificial Neural Network and ASHRAE Clear-Sky Model

Solar energy is one of the renewable energies with great potential in tropical countries including Indonesia. Harvesting solar energy through a photovoltaic system has a major challenge due to its intermittency and uncertainty as well as the unavailability of the measured data at every location. The purpose of this study is to estimate the solar irradiance at certain location by employing the ASHRAE Clear-Sky model and the local weather information through an Artificial Neural Network (ANN) algorithm. The ASHRAE Clear-Sky model is used as the basis of maximum irradiance which will be calibrated by considering the local weather information. The estimation model is developed by using a backpropagation algorithm of ANN. The proposed method is simulated at the department of electrical engineering Universitas Indonesia. The result shows that the proposed method can provide an accurate estimation of solar irradiance with the mean absolute error in the three different test days are 71.44 in day 1, 43.24 in day 2, and 60.23 in day 3,also the average from mean absolute error of these days is 58.30.