SCALABLE OPTIMIZATION METHOD FOR GENERATOR SCHEDULING UNDER UNCERTAINTY

Abstract

Scalable optimization methods for power system operation has been subject of research over the last 60 years. State-of-the-art methods in this research area is yet to yield the scalability desired by system operators for practical operation of electric grids. This article-based dissertation makes three significant contributions. A scalable computational method is developed to tackle a mixed-integer problem commonly referred to as Stochastic Security-Constrained Unit Commitment (SSCUC), the output of which will be beneficial to Independent System Operators to manage electric grids. Secondly, an improved model for time-progressive contingencies in security-constrained optimization problems is presented. This modeling approach is more realistic representation of contingency modeling as compared to what exists in literature. Finally, uncertainty from pulsed load transients in microgrids is tackled in the presence of energy storage units. In the first paper, a detailed SSCUC problem is considered that suffers from complexities posed by the presence of binary variables, uncertainty of renewable energy and security constraints. The second paper deals with extra challenges time-progressive contingencies such as hurricanes and wildfires pose to the Security-Constrained Optimal Power iiFlow (SCOPF) problem. The third paper deals with the MG scheduling problem in the presence of uncertainty introduced by transient load demand.