Battery energy storage systems for energy arbitrage: market selection and price forecasting optimization

The rapid integration of variable renewable energy (VRE) sources into electricity grids has disrupted traditional power market dynamics and created new opportunities for grid-scale battery energy storage system (BESS) ventures. This thesis presents an integrated, data-driven framework for optimizing the investment and operational decision-making processes of battery energy storage system-enabled energy arbitrage (BESS-EA) systems. The first component develops a novel energy arbitrage feasibility (EAF) scoring methodology to evaluate the locational suitability of 29 European electricity markets. Using eight years of hourly wholesale electricity price data (2015–2023) along with regulatory, volatility, and macroeconomic indicators—including electricity supply-origin risk (ESOR), GDP per capita (purchasing power parity), and index of economic freedom (IEF)—the EAF score facilitates standardized, cross-country investment assessment across five investor preference scenarios. Results identify Estonia and Luxembourg as optimal BESS-EA markets across five scenarios. The second component addresses the critical challenge of electricity price forecasting for bid optimization by introducing a hybrid long short-term memory (LSTM) and exponentially weighted moving average (EWMA) architecture. Trained on Austria's market data, the model achieves a root mean square error (RMSE) of €1.2207/MWhe and incorporates 95% bootstrap confidence intervals for risk quantification. By simultaneously capturing long-term temporal trends and short-term volatility patterns, the model supports robust intraday BESS bid decisions in wholesale and ancillary service markets.