Reserve Requirements for Wind Power Integration: A Scenario-Based Stochastic Programming Framework

We present a two-stage stochastic programming model for committing reserves in systems with large amounts of wind power. We describe wind power generation in terms of a representative set of appropriately weighted scenarios, and we present a dual decomposition algorithm for solving the resulting stochastic program. We test our scenario generation methodology on a model of California consisting of 122 generators, and we show that the stochastic programming unit commitment policy outperforms common reserve rules.