Direct load control using fuzzy dynamic programming

The paper develops a fuzzy dynamic programming (FDP) approach to solving the direct load control (DLC) problem of large air conditioner loads (ACLs). In the formulation, the cycling capacities of the ACLs and the system load demands are all regarded as fuzzy variables. The scheduling of the DLC and the unit commitment (UC) is integrated into an FDP structure to reduce the system peak loads as well as the system operation costs. The solution approach has been tested on the data for the Taiwan Power (Taipower) 38-unit system with 15 ACL groups. In the tests, the Monte-Carlo simulation technique is employed to assess the robustness of the solution obtained from the proposed FDP approach for various load uncertainties. The results are presented and compared to those without consideration of the load uncertainties