1 Optimisation and Economic Feasibility of Battery Energy Storage Systems in Electricity Markets Inês Gaspar 1) , Rui Castro 2)*) 1) Instituto Superior Técnico, University of Lisbon, Portugal 2) INESC-ID/IST, University of Lisbon, Portugal *) Corresponding author: rcastro@tecnico.ulisboa.pt ABSTRACT This study identifies the optimal operating strategy of storage systems on the electricity markets, from the perspective of a market participant with a renewables’ portfolio. The energy storage system provides a balancing service for re- newable sources, while also performing energy arbitrage at the considered three short-term markets. Spot price and renewable generation predictions guide the bidding decision-making process to maximise agents’ economic profit. Thus, a Long Short-Term Memory (LSTM) model was developed to forecast these variables. The designed decision support models considered market rules and the technical constraints in the operation of the storage system. The influ- ence of storage systems in the optimal operation, regulation costs and revenues is analysed on a daily and yearly basis. The economic viability of sodium-sulphur, lithium-ion, zinc hybrid and vanadium redox flow batteries in 2018 and 2025 is studied. The results suggest that the use of forecasting techniques and battery implementation sharply reduce regulation costs. The analysis shows that acting in the balancing market can be a motivation for the storage system viability. The economic evaluation proves that only the lithium-ion battery is a profitable investment in the considered years. Keywords: Renewables; Energy Storage Systems; Forecasting; Optimal Operation Strategy; Energy Markets. Nomenclature   day-ahead market bid at hour h   intraday market bid at hour h  charge assigned to day-ahead strategy at hour h  charge assigned to intraday strategy at hour h  discharge assigned to day-ahead strategy at hour h  discharged assigned to intraday strategy at hour h   energy storage system maximum capacity  predicted renewable production assigned to the day-ahead strategy at hour h  predicted renewable production assigned to the intraday strategy at hour h   deviation from the day-ahead bid at hour h