ABSTRACT
Photovoltaic (PV) energy is a highly intermittent renewable energy source. Due to this intermittent nature, when integrating the PV system with a power system, the reliability and the stability of the power system will decrease. Therefore, predicting the PV generation will help to maintain the power system in a stable way without any uncontrollable and unanticipated variations. Thus, the task of solar power forecasting becomes vital. Furthermore, during the operation of the PV system, a considerable number of outliers will occur due to array anomalies and faults, sensor failure and irregular array shutdowns. These outliers will cause the accuracy of the forecasting model to decrease. Therefore, data pre-processing should be done before modelling the forecast model. This chapter presents interpolation and exponential smoothing as data pre-processing techniques, and an artificial neural network (ANN) is trained to create a day-ahead solar forecasting model.
