ПРОГНОЗИРОВАНИЕ РЫНКА ЭЛЕКТРОЭНЕРГИИ «НА СУТКИ ВПЕРЕД»: ПРИМЕР ВТОРОЙ ЦЕНОВОЙ ЗОНЫ

Abstract

The paper proposes techniques for medium-term and long-term price and quantities forecasting at electricity «day-ahead» market on the base of regression models. We took into account non-linear effects, time lags, and distributed in time influence. Using data for 2007-2015 years in second price zone «Siberia» we revealed the most significant factors, and obtained their quantitative estimates. The errors distribution analysis has been carried out. The hypothesis of normal distribution that doesn’t take into account the reality of thick tales was rejected using the Pearson criterion. On the base of obtained results the agents’ strategy of free spot and forward contracts signing at the wholesale market was working out. The short-term forecast of hourly prices at DAM is carried out on the base of maximum likeness extrapolation maethod.