Essential Tools and Strategies for the Participation of Battery Energy Storage Systems in Ancillary Services

Abstract

This thesis aims to develop the necessary tools to optimize the participation of both standalone Battery Energy Storage Systems (BESSs) and hybrid Photovoltaic (PV) installations with BESS in energy and frequency regulation markets. Firstly, the thesis reviews the current and future state of the European power systems and electricity markets, focusing on Spain and Ireland, to identify profitable markets for these installations. The thesis then analyzes and predicts market prices and energy variables using different time series analyses and forecasting techniques, including those based on Machine Learning (ML) and Deep Learning. Additionally, it examines the impact of battery degradation in grid applications, assessing ML algorithms for diagnosing battery health and the impact of market structures on battery life. Finally, the thesis proposes Reinforcement Learning strategies to optimize hybrid PV-BESS operations in different dynamic environments, including cost minimization for residential prosumers and energy volume bidding in both energy and ancillary services markets.