Short-term net feeder load forecasting of microgrid considering weather conditions

In this paper, an approach of feeder net load forecasting is proposed for mirogrid operation. Firstly, the output of intermittent renewable energy sources are took into account as a negative load and give the net feeder load definition. Then the feeder load patterns are established according to weather conditions and different solar terms that may reflect the change of season. The forecasting model of back-propagation network is developed with improved Levenberg-Maruardt (LM) training algorithm. The optimized solution of the developed model can accurately forecast the minutely net feeder loads. The validity of the proposed approach for a simplified microgrid is shown by the simulation results.