The efficient integration of renewable energy resources(RES)in future power systems as directed by the recent worldwide energy policy drive has given rise to challenges related to the security,sustainability and affordability of the power system("The Energy Trilemma").In particular,the effect of high RES penetration to the modern electricity markets has been the subject of extensive research.In that context,energy storage is expected to play a key role in the future power system and energy markets.There are many different services that can be provided from storage devices based on the connection point with the grid,ranging from energy arbitrage and reserves to distribution deferral and increased self-consumption.In this study we will focus on the energy arbitrage in the wholesale market and in particular in the intra-day market.Wholesale transactions are expected to move closer to real time for two reasons,namely in order to reward flexibility sources and to exploit more accurate forecasts of RES production and load [1].In this paper,the problem faced by a storage device operator participating in the Continuous Intraday(CID)market is considered.A modeling framework is implemented in order to simulate the operation of the storage device and its interaction with the market.The goal of the trading agent is the maximization of the cumulative revenues received over the entire trading horizon.Approximate Dynamic Programming(ADP)is used to develop a trading strategy that is compared to the benchmark called"rolling intrinsic".The results indicate that the trading agent is able to outperform the benchmark after a period of training.Real data from the CID market are used for training and validation purposes.
展开▼
