The Method of Short-term Forecast Electricity Load with Combined a Sinusoidal Function and an Artificial Neural Network

摘要

Currently, in Russian Federation operates the Electricity Market. Market entities - retail energy companies provide for company "Administrator of the trading system", data about purchases the required volumes of electricity load on the "Day-ahead market". All actual deviations from the filed applications are supported by the Balancing Market, where the cost of electricity differs significantly from the "One-day ahead" market. This specific process pushes the retail energy companies to develop short-term forecast electricity load methods. Otherwise, the energy retail companies may suffer significant financial losses. The proposed hybrid method of short-term forecast electricity load based with combined a sinusoidal function and an artificial neural network for forecast electricity load, the approximation coefficients are calculated using an artificial neural network.