Clonal selection algorithm for power generators maintenance scheduling

• Decoupling the generators maintenance scheduling problem into two interrelated sub-problems.
• Introduction of a new technique based on clonal selection algorithms (CSA).
• The CSA is used to search for the optimal maintenance schedule of each unit.
• The iterative lambda technique is used to find the optimal output power form each generating unit.
• Local search technique is used to repair the feasibility of the new solutions.

This paper introduces a technique based on the one of the artificial immune system (AIS) technique known as the clonal selection algorithm (CSA) to obtain the optimal maintenance schedule outage of generating units. Based on a weekly load profile, the proposed technique provides the optimal maintenance window and calculates the optimal output power from each generator over a one year horizon. The maintenance scheduling problem is decoupled into two interrelated sub-problems namely, the maintenance scheduling and the power system sub-problems. The CSA is used to solve the maintenance scheduling subproblem to obtain the optimal maintenance outage of each unit. Based on the schedule generated by the CSA, the economic dispatch iterative lambda technique is used to find the optimal output power from each unit. Due to the search nature of the CSA, infeasible solutions may be introduced during the solution process. Therefore, a local search technique is used to watch the feasibility of the new solutions. The paper reports test results of the proposed algorithm to find the optimal maintenance schedule of the IEEE 30 bus system with 6 generating units and the IEEE 118 bus system with 33 generating units. Results are compared against the results obtained by complementary decision variables structure (CDV) and the evolutionary programming based techniques. Results are encouraging and indicate the viability of the proposed CSA technique.