LAUSR

Abstract Solar irradiance fluctuates within a very short period of time that creates a lot of hindrances to estimate the injection of output power into the grid. During the operation of solar power plant, short-term PV power forecasting supports load dispatching, planning, and also the regulatory actions. But this short term PV power forecasting is a very complicated problem in order to solve it. This paper represents short-term PV power forecasting by constructing a 3-stage approach which is formed by combining empirical mode decomposition (EMD) technique, sine cosine algorithm (SCA), and extreme learning machine (ELM) technique. At the initial phase of the proposed technique, a de-noised series is obtained by adopting a signal filtering strategy based on EMD decomposition technique. Next three different time interval data series are opted for the training and forecasting stage. The selected sets of data are quarterly, half-hourly and hourly PV data observations. The simulation results signify that the recommended technique performs in an out-standing manner than the conventional ones while addressing short term PV power forecasting.