Abstract:
In this paper, we examine the optimal solution to the thermal generation scheduling problem using three versions of Dynamic Programming (DP), Conventional DP (CDP), Sequential Combination DP (SC-DP), and Truncated Combination DP (TC-DP) to solve the Unit Commitment (UC) problem. The Unit Commitment (UC), one of the essential problems in power system applications, consists of producing a minimum cost output schedule for a setting horizon while satisfying the units constraints on the system generation as a whole. The lambda-iteration technique is applied to solve the embedded economic dispatch sub-problem. The presented techniques are implemented for 5-unit and 10-unit thermal systems for 24-hour. In comparative simulation results, it is shown that the CDP can find accurate solutions for smaller systems, but it is computationally expensive for large systems. In order to strike a better trade-off between solution quality and computational efficiency, TC-DP inevitably prunes non-promising decision paths, making it more efficient than classical DP by reducing the state-space explosion. The Results indicate that TC-DP is an efficient and scalable optimization framework for modern complex power systems.
Keywords: