NSF Award
9160169
The objective of the research is to develop scheduling algorithms for electric generation that account directly for system security requirements. Virtually all algorithms in use currently do not optimize generating unit commitment decisions against security requirements from the outset and yield, therefore, suboptimal results when the commitment decisions must be adjusted to comply with these requirements. Utilities can obtain significantly better results if they use optimal scheduling algorithms that comply directly with security requirements. While algorithms of this type had severe computational requirements, five-to-ten years ago, recent advances in network programming have now made these algorithms more attractive. The PI proposes to demonstrate the practical feasibility of a new algorithm for solving the security-constrained unit scheduling problem. The innovation of our approach is to capture the system security requirements within a network flow subproblem that is solved in conjunction with the unit commitment problem. In Phase One of this effort we will develop a prototype implementation of this method by integrating a state-of-the-art network flow algorithm into an advanced unit commitment algorithm. We will demonstrate the cost savings possible with this algorithm on realistic test cases using representative utility data.
