Photovoltaic Power Prediction Based on GA-Elman Model

Abstract Aiming at the time-varying and fluctuating characteristics of photovoltaic power output, a GA-Elman model is proposed for short-term prediction active power. This can use historical value daily meteorological information to directly forecast power, which convenient dispatching department supply bureau obtain output data in advance enhance reliability supply. Firstly, input variables Elman are determined by analyzing factors that affect Then, genetic algorithm adopted optimize parameters network establish model. Finally, usability proved an example simulation.