Study on load forecasting of an actual micro-grid system in Taiwan

In microgrid operation control, load forecasting plays an essential role in enhancing the security of the microgrid operation and in optimizing the energy dispatch. In this paper, two models of load forecasting very short term and short term models are studied. The very short term models forecast the load from several minutes ahead until 1 hour ahead. It can be used for security control in the prevention and emergency condition. Short term models that forecast the load from several hours ahead until several days ahead can be used to plan the energy dispatch. This study provides online load forecasting models for both very short and short terms. The models are based on Back-Propagation Neural Networks (BPNN). The prediction results are verified with the actual time series load.