Simulation of multi-market trading : a case study

Abstract

This report describes a study of day-ahead (DA) and intraday (ID) trade for a hydropower producer with a single reservoir and power plant. The study simulates the trade decisions and plant operation in a rolling-horizon setup where DA decisions are reoptimized on a daily basis and ID decisions are reoptimized on an hourly basis. Market prices are given by historic German prices from 2017, while price uncertainty is modelled by sampling from historical price paths using kernel-regression. The study assesses if the producer can gain from participating in the ID market in addition to DA trade. ID trading motivated by suboptimal DA commitments due to bid curve requirements or motivated by short-term production capacity uncertainty (e.g. inflow or outages) are not captured in the study. Main findings are that the change in performance is relatively small when trading in both DA and ID market compared to only DA trade, and whether a gain or a loss is observed depends on the assessment interval. Even though ID trade is only conducted on a short horizon, within day, the end-of-horizon modelling (model horizon and water value description) can significantly affect the performance with ID trade.