A novel procedure for the optimal scheduling of operating reserve based on stochastic optimization

Operating generation reserve plays a key role for ensuring a stable balance between production and load, enabling the system to deal with sudden events like faults or unexpected variations of loads and renewable generation. The proper size of the operating reserve has been traditionally tuned by the System Operators using deterministic approaches or, more recently, using probabilistic methods like Montecarlo simulations. This paper proposes to use a double-stage stochastic approach to optimize the amount of stand-by and spinning reserve to be set and ensured in the afternoon of day D-1. The idea is to model the day-ahead ancillary service market as the typical 'first stage' of Stochastic Optimization, while quasi real-time sessions of balancing market perfectly fit with 'second stage' decisions to be taken by the Operator. A case study is proposed discussing technical, economic and convergence aspects of different operational approaches (deterministic, average scenario, stochastic), also proposing interesting variants to the classical models of load uncertainty.