Evaluation of Dynamic Programming Based Methods and Multiple area Representation for Thermal Unit Commitments

TLDR

The study compares four unit commitment methods, three using dynamic programming, and introduces a linear flow network model to incorporate inter‑area transmission constraints. Realistic evaluations were performed on two large systems—EPRI Scenario System C and a midwestern utility—each with up to 96 units. The results demonstrate that multi‑area representation reduces schedule costs and guarantees realizable unit commitment schedules.

Abstract

This paper compares the performance of four unit commitment methods, three of which are based on the dynamic programming approach. The paper also presents the modeling of inter-area flow constraints by linear flow network, so that such multi-area representation recognizes any existing transmission limitations which are normally ignored in most unit commitment methods. To assure realistic results, data for two large systems (up to 96 units) were used: EPRI Scenario System C and a midwestern utility. Unit commitment results for these two systems are presented in the paper and they demonstrate the importance of multiple area representation of a system in unit commitment. Such representation affects the schedule costs and assures the determination of realizable schedules.