Forecasting spot electricity market prices using time series models

Abstract

The worldwide electricity industry is in an era where an overwhelming transition towards deregulation is taking place. Since its start in the early 1980s, the industry has been in a continuous change to a different atmosphere; the ultimate benefit being providing the enduse customer with a reliable but yet cheaper electricity supply. In the old monopolistic system, utilities were the only authoritative body to set the tariff based on their aggregate cost. In the contrary, as a newly emerging structure, deregulation has come-up with a new way of functioning; leading generation, transmission and distribution to be independent activities. This market is generally a customer driven market and since the early days of deregulation, price forecasting has become an important task to all the market players engaged. Unless each body is fairly aware of the market, it might result in losses to a generating company or inadequate supply to the system; resulting in a huge crisis. Hence, understanding the market behavior is of vital importance for the well-functioning of the industry and the benefit of each party.This thesis addresses the importance of electricity price forecasting in the deregulated market. After a comprehensive review of previous works in the subject, it is concluded that forecast accuracy varies depending on the forecast method used and the electricity market under study. In this thesis multiple linear regression approach are proposed to predict next day's electricity prices. The developed models are tested in the Nord Pool and the Ontario electricity markets and satisfactory results are achieved. Comparing the forecast results from the two markets, results in the Nord Pool market are significantly more accurate than the Ontario market. This arises from the fact that Ontario market is very volatile and its market prices are hardly predictable. The forecast results in this thesis work show that the proposed models perform very well with the absence of volatility. For instance, the developed models for the first week of December 2007 generated results with weekly Mean Absolute Percentage Error (MAPE) of about 5.04%. In early spring even better results are achieved where the models generated results with a daily MAPE of up to as low as 1.83% and weekly MAPE of about 2.96%. On the other hand, during summer where prices vary considerably even between hours of the same day, the models generated a relatively poor forecast. A day with a price of as low as 2.96 EURO/MWh to as high as 23.52 EURO/MWh is observed and this price fluctuation contributed for poor forecast. For instance, the models generated forecasts with a weekly MAPE of about 13% for the period from August 29 to Sept. 4, 2007. For the Ontario market, these models generated forecast results with a weekly MAPE of about 17%.

Development оf combined method for predicting the process of the occurrence of emergencies of natural character

“…The statistical models include the following groups: regression models, autoregressive models, exponential smoothing models. Structural models set the functional relationship between future and actual values of a time series, as well as external factors, structurally [17,18]. Structural models include the following groups: neural network models, models based on Markov chains, models based on classification -regression trees.…”

“…Works [18,19] show that it is possible to use models based on Markov chains in case of insufficient information on a process of emergency occurrence. However, this model takes into consideration only the current state of a process at forecasting of a future state of a process.…”

Double layer interval weighted graphs in assessing the market risks

“…Следующим этапом апробации предложенной модели является сравнение точности ДИВГпрогноза с точностью авторегрессионного прогноза цен на стальные биллеты методами ARIMA(1,1,1) и GARCH(1,1), определенных на всем объеме выборки, представленном ценами на стальные биллеты с 24.07.2008 г. по 09.02.2011 г. Выбор этих методов прогнозирования не является случайным: из всего многообразия математических методов прогнозирования временных рядов ARIMA и GARCH являются наиболее используемыми [Бокс, Дженкинс, 1974;Marengia, 2008;Alfares, Nazeeruddin, 2002].…”

A hybrid model for short term real-time electricity price forecasting in smart grid

“…The participants in electricity market can regulate the operating time of electrical devices automatically or manually during high-price periods and gain the benefits from low-price periods via DRM, thus achieving the aims of reducing energy usage and saving electric bills for themselves [5,[9][10][11]. Therefore, the research on RTP tariff is of interest to researchers, production companies, investors, independent market operators and large industrial consumers in recent years [12,13].…”