Modelling the Evolution of Wind and Solar Power Infeed Forecasts

Abstract

With the increasing integration of wind and photovoltaic power in the whole European power system, there is a longing for detecting how to trade energy in the ever-changing intraday market from electric power industries. Intraday trading becomes even more relevant in the wake of the European Cross-Border Intraday (XBID) project, which aims at integrating electricity trading across Europe. Therefore, optimal trading strategies to address forecast fluctuations in renewables output are growingly required to be designed. In this study, we model, simulate and predict the evolution of wind and PV infeed forecasting errors over eight days preceding the start of a given quarter-hourly delivery period and updated in 15-minute steps. We test comparatively the performance of several stochastic and probabilistic models, and recommend their complementary use, depending on the frequency in which forecast values are adjusted. Since ex-ante updated forecasting errors of renewables infeed are usually not available to researchers, simulations based on our proposed models break the ground for further applications to intraday pricing and optimization.

Keywords: Wind/Photovoltaic forecasting errors, intraday market, GMM, stochastic models

JEL Classification: C22; C53; C55; G10; Q20; Q40; Q41; Q42