This dissertation demonstrates how flexibility in hourly electricity operations can impact long-term planning and analysis for future power systems, particularly those with substantial variable renewables (e.g., wind) or strict carbon policies. Operational flexibility describes a power system’s ability to respond to predictable and unexpected changes in generation or demand. Planning and policy models have traditionally not directly captured the technical operating constraints that determine operational flexibility. However, as demonstrated in this dissertation, this capability becomes increasingly important with the greater flexibility required by significant renewables (>=20%) and the decreased flexibility inherent in some low-carbon generation technologies. Incorporating flexibility can significantly change optimal generation and energy mixes, lower system costs, improve policy impact estimates, and enable system designs capable of meeting strict regulatory targets.

Methodologically, this work presents a new clustered formulation that tractably combines a range of normally distinct power system models, from hourly unit-commitment operations to long-term generation planning. This formulation groups similar generators into clusters to reduce problem size, while still retaining the individual unit constraints required to accurately capture operating reserves and other flexibility drivers. In comparisons against traditional unit commitment 3 formulations, errors were generally less than 1% while run times decreased by several orders of magnitude (e.g., 5000x). Extensive numeric simulations, using a realistic Texas-based power system show that ignoring flexibility can underestimate carbon emissions by 50% or result in significant load and wind shedding to meet environmental regulations.

Contributions of this dissertation include: (1) Demonstrating that operational flexibility can have an important impact on power system planning, and describing when and how these impacts occur; (2) Demonstrating that a failure to account for operational flexibility can result in undesirable outcomes for both utility planners and policy analysts; and (3) Extending the state of the art for electric power system models by introducing a tractable method for incorporating unit commitment based operational flexibility at full 8760 hourly resolution directly into planning optimization. Together these results encourage and offer a new flexibility-aware approach for capacity planning and accompanying policy design that can enable cleaner, less expensive electric power systems for the future.

[Thesis submitted October 2012; Doctoral Degree date February 2013]