Real-time pricing related short-term load forecasting

Abstract

The production cost of electricity is not constant over time. It is dependent on the instantaneous load being supplied, the available generation and the state of the network. Real-time pricing (RTP), which sets the electricity selling price approximately equal to marginal cost, is proposed as a potential method for ensuring overall economic rationality, and for limiting the demands required by all consumers at times of limited supply or emergency conditions. Any tariff change will influence customer's electricity consumption behavior. Some customers will respond to the real-time pricing by modifying or rescheduling electricity usage. This makes the short-term load forecasting problem more complicated than before. By combining the power of supervised and unsupervised neural networks, this paper presents a new solution for RTP related short-term load forecasting problem. The simulation result on realistic load and weather data confirms the good performance of this load forecaster.