Weather Forecasting Based Intelligent Distribution Feeder Load Prediction

Increased penetration of renewable energy based generators throughout modern distribution networks makes it crucial to seek elevated levels of accuracy in forecasting methods. This paper presents a new load forecasting method for residential distribution feeders. It uses, load time series decomposition to distinguish between all types of loads and events on feeder. Then, a generalized regression neural network (GRNN) is used to fit the weather variables for weather dependent load component. Proposed algorithm has been evaluated on residential feeder. Results show very accurate prediction for active power as well as the lighting load peaks.