Price Forecasting Techniques in Power Systems

Abstract

This tutorial paper provides a brief overview of forecasting techniques for hourly electricity price prediction in both the short and the long term, with an emphasis on analytical, nonheuristic procedures. Appropriate background material on time‐series analysis is reviewed first. Short‐term hourly price forecasting (from 12 to 168 hours in advance) is then addressed considering mostly time‐series tools. Illustrative examples based on both European and North American electricity markets are provided to clarify the functioning of the tools described. Unobserved component models are then introduced to address the long‐term forecasting (from 1 week to 1 year in advance) of hourly electricity prices. An illustrative example based on data from a North‐American electricity market is used to clarify the working of an unobserved component model. Appropriate conclusions are finally drawn.

Automated Variable Selection and Shrinkage for Day-Ahead Electricity Price Forecasting

“…Short-term EPF generally involves predicting 24 hourly (or 48 half-hourly) prices in the day-ahead market, cleared typically at noon on the day before delivery, i.e., 12-36 h before delivery, the adjustment markets, cleared a few hours before delivery, and the balancing or real-time markets, cleared minutes before delivery [32]. The spot market, especially in the literature on European electricity markets, is often used as a synonym of the day-ahead market.…”

On the importance of the long-term seasonal component in day-ahead electricity price forecasting

“…Short-term EPF generally involves predicting prices in the day-ahead market -cleared typically at noon on the day before delivery, i.e. 12 to 36 hours before delivery, the adjustment markets -cleared a few hours before delivery, and the balancing or real-time markets -cleared minutes before delivery (Garcia-Martos and Conejo, 2013). The spot market, especially in the literature on European electricity markets, is often used as a synonym of the day-ahead market.…”

On the importance of the long-term seasonal component in day-ahead electricity price forecasting

“…In fact, most mid-term EPF models work with average daily prices and focus on the annual or long-term seasonal component (LTSC; also called the trend-seasonal component). However, in short-term EPF the daily and weekly seasonalities are always taken into account, but the LTSC is believed to add unnecessary complexity to the already parameter-rich models and is generally ignored (for recent reviews see Garcia-Martos and Conejo, 2013;Weron, 2014). But is this the right approach?…”