Time constrained scenario selection for stochastic unit commitment

Uncertain elements in power systems can have an enormous impact on system dynamics. Yet, these uncertainties are seldom incorporated into industry power systems optimization models. A scenario-based stochastic unit commitment (SUC) formulation explicitly models uncertainties to improve grid reliability and cost efficiency. Unlike deterministic unit commitment, stochastic unit commitment considers multiple possible scenarios. In practical applications such as operation in anticipation of a natural disaster, the usefulness of SUC is hindered by its burdensome computation time. This limits the number of scenarios SUC can practically utilize out of the entire scenario space. In weather related preventative operation the scenario space grows to astronomical proportions due in large part to the high degree of uncertainty in meteorological forecasting. This introduces the need for an effective scenario selection method that considers the unique requirements of the problem. A new scenario selection method is developed in this paper to address these challenges. The method reduces the temporal dimension of scenarios based on the theory that this reduction will decrease SUC computation time without sacrificing solution quality. The method is tested with line failure probabilities derived from a simulation of a synthetic hurricane over a synthetic 500 bus system. The results are compared to another scenario selection method of equivalent scope and purpose.