Maintenance scheduling (MS) of power-system generating units and transmission lines is one of the typical power-system long-term operation and planning problems. MS is an important tool for increasing both generating equipment availability and system reliability while minimizing both system operational and maintenance costs.; Maintenance scheduling is an optimization problem subjected to maintenance and system constraints. The MS problem can be decomposed into two interrelated subproblems namely, the maintenance subproblem and the power-system subproblem. The classical decomposition methods impose additional constraints on both subproblems to allow a unique solution. This dissertation introduces a new solution technique for the decomposed problem. The proposed technique is based on evolutionary programming (EP) to search for a near-optimal solution and the feasibility watch of the hill-climbing technique. The search nature of the EP allows the sequential solution of the two subproblems without additional constraints.; The MS problem has many uncertainties. These include load uncertainties, market price changes for maintenance and fuel. Ignoring these uncertainties may yield to misleading results or a local optimal solution. To achieve an improved solution, the MS problem formulation is extended to include the above mentioned uncertainties using fuzzy techniques. The integrated fuzzy model for the MS problem is then solved by the evolutionary programming-based technique. The results of the proposed solution approach are the fuzzy overall production and maintenance cost that reflect the uncertainties associated with the MS problem. A related contribution of this dissertation is the development of a new fuzzy comparison technique suitable for the MS problem.; To test the proposed approach, a modified IEEE 30-bus test system with 41 lines and 6-generating units and the IEEE 118-bus test system with 179 lines and 33 generating units were used. Test results on the IEEE 30-bus system using the proposed approach against those obtained using the complementary decision variables approach indicates improvement of results by the proposed approach and established its viability to solve the MS problem.
展开▼
