Efficient market clearing for manual Frequency Restoration Reserve capacity with stochastic activation costs and nested decomposition

Key Points

• This research aims to optimize the market clearing process for manual Frequency Restoration Reserve (mFRR) capacity by incorporating stochastic activation costs.
• Developed a solution algorithm combining Benders Decomposition with Stochastic Activation Value Linear Relaxation (SAVLR).
• Utilized Bayesian Neural Networks for scenario generation to improve cost predictions.
• Analyzed potential cost reductions in a market setting in Sweden.
• Identified potential daily cost reductions ranging from 600 to 14,500 €.
• Showed that the implementation of the new market design can significantly lower operational costs.
• Provided a framework for efficient clearing in capacity markets with defined activation costs.