Multi-Stage Scenario Reduction for Optimal Energy Storage Operation in Real-Time Markets

The growing share of renewable energy sources in the electricity grid has led to increased volatility in real-time balancing prices. While this volatility creates valuable opportunities for energy storage systems (ESS), the inherent uncertainty of future prices significantly complicates the optimization process. In this paper, we propose a multi-stage scenario-based stochastic model predictive control (SMPC) strategy to optimize the operation of ESS in real-time balancing markets. Central to our approach is a stagewise refinement algorithm, inspired by optimal transport theory. This algorithm constructs non-symmetric multi-stage scenario trees through clustering methods and is benchmarked in a simulation on the Belgian real-time balancing market with a battery energy storage system. Results demonstrate that our approach, particularly when combined with k-means clustering, yields higher profits compared to traditional multi-stage reduction techniques.