Research On Flexible-Interval Preventive Maintenance Policy In Finite Time Horizon For Production System

Maintenance on equipments in a production system has an important influence on the manufacturing cost and the production operation, and it is one of the key factors determining the economic benefits and market competition of a manufacturing enterprise. Compared with traditional maintenance, preventive maintenance banishes the maintenance mode mainly depending on experiences and the qualitative judgment. It uses different kinds of maintenance modeling methods and optimization techniques, emphasizes quantitative description and analysis of the maintenance work to improve the operate performance of the enterprise. Under the market competition environment, to discover the potential rules of preventive maintenance and to construct effective preventive maintenance policies and models are very important for a modern manufacturing enterprise.Since preventive maintenance has an important influence on production cost and production task, the flexible-interval preventive maintenance policy is proposed in this dissertation. The time interval between equipment maintenance actions has a selective range for the maintenance policy. The problem of flexible-interval preventive maintenance policy modeling and optimizing is studied, which includes the decision of flexible-interval preventive maintenance planning in the finite time horizon and integration of flexible-interval preventive maintenance planning with production scheduling. The main contents of this dissertation are outlined as follows:(1) For a single unit system, multi-objective optimization problem of flexible-interval preventive maintenance planning in the finite time horizon is presented. The objectives of the problem are to minimize maintenance cost and maintenance time under the premise of satisfying the requirement of the system reliability and availability. Based on Pareto rank and penalty function, the multi-objective genetic algorithm is used to solve the flexible-interval preventive maintenance optimization problem. The numerical experiment shows that flexible-interval preventive maintenance policy can save more maintenance cost and maintenance time than periodic preventive maintenance.(2) Optimization problem of flexible-interval preventive maintenance planning based on reliability in the finite time horizon is presented for a multi-unit series system. The objective of the problem is to minimize the maintenance cost. In order to reduce the shut down loss caused by the frequently preventive maintenance of the units, several preventive maintenance actions of different units are performed together to make the intervals between preventive maintenance actions performed on the system always within a certain range. The genetic algorithm with the function of repairing infeasible chromosome is used to obtain the optimal preventive maintenance planning. The numerical experiment shows that flexible-interval preventive maintenance policy can effectively save the maintenance cost.(3) Considering the impact of the stochastic failure on production task, the joint optimization problem of flexible-interval preventive maintenance and production scheduling is presented for a single machine. The purpose of the scheduling problem is to choose an optimal preventive planning as well as an optimal sequence for the jobs which minimize the total expected weighted completion time of the jobs. Genetic algorithm is used to search the optimal flexible-interval preventive maintenance planning and job scheduling. The performance of the proposed policy was compared to other policies which address the same problem, and the result shows that flexible-interval preventive maintenance policy saves the completion time of the jobs. Besides, the numerical experiments also show that the optimal preventive maintenance planning for the joint optimization problem is often performed before the jobs.(4) In order to optimize the overall performance of the production system, multiple objectives, including minimization of makespan, total weighted completion time of jobs, total weighted tardiness, maintenance cost, maintenance time and maximization of machine availability are simultaneously considered to optimize the integrated problem of before- job preventive maintenance and production scheduling. Multi-objective genetic algorithm with a random weighted fitness function is used to obtain the near Pareto optimal solutions. The total weighted percent deviation provides the basis for decision-makers to select the best solution among the Pareto optimal solutions in practice.(5) Considering the practice of equipment maintenance management in manufacturing enterprises of the country, an equipment maintenance management information system is constructed. Maintenance management and preventive maintenance scheduling for the injection workshop of a manufacturing enterprise as an example that illustrates the application process and operating result of the equipment maintenance management information system. The result verifies that the policy of flexible-interval preventive maintenance can meet the actual requirement of equipment maintenance management in manufacturing enterprises.