ANN Based Power Management Strategy For Standalone Microgrids

Abstract

This paper presents a solar power generation prediction technique using artificial neural network. The predicted data is then applied to the adaptive power management strategy for Photovoltaic (PV) generation units in a standalone microgrid. The intermittent nature of solar power generation leads to major challenges in power system planning and load sharing. Prediction of solar power generation based on weather conditions and the proper use of this data in power management strategies improve the performance of existing standalone systems. This paper proposes an adaptive control strategy, which uses the predicted value of solar generation to determine the mode of operation. An ANN model is developed and trained using the dependency of solar power generation on weather parameters. The trained model is used to predict the expected solar power generation at any time. The applicability of the proposed adaptive control method is analyzed using Matlab/Simulink based simulation studies.