Research On Decision Making And Optimization Of Intelligent Maintenance For Production Systems

Since the past decade, the impact of information technologies has added"velocity'to the design, manufacturing, and aftermarket service of a product. Velocity has been an issue in today's competition in manufacturing industry. As one of the key elements for the realization of velocity, the maintenance of equipments is becoming more and more important. The failure of equipments not only will increase the cost for maintenance, but also will decrease the productivity of production systems. For this sake, intelligent maintenance is emerged as the times require. By using predictive maintenance techniques, including intelligent machine degradation assessment methodologies, e-prognostics, and e-diagnostics, intelligent maintenance can enable manufacturers to have production systems with near-zero-breakdown conditions. In 2002, intelligent maintenance has been proposed by fortune magazine as one of the three hot technologies in the future manufacturing industry.This research is based on the long cooperation between Shanghai Jiao Tong University in China and Center for Intelligent Maintenance System in USA. The purpose of this research is to do some contribution on decision making and optimization for intelligent maintenance, which should be helpful for the development of intelligent maintenance in China.First, based on the analysis of the differences between the information flow for intelligent maintenance and for the traditional maintenance methods (i.e. periodical preventive maintenance), a decision process of intelligent maintenance for production systems was proposed. Then a three-layer decision structure was developed with consideration of the complexity of the decision process, which including component level, system level and business level. In order to realize the decision process, a decision support system for intelligent maintenance was developed.Then, the maintenance model for component level and the one for system level were proposed based on the three-layer structure, which is oriented to production systems. After analyzing the existed maintenance policies, a hybrid recursion rule for the hazard rate of equipment was developed by integrating age reduction factor and hazard rate increase factor, in which the imperfect effect of maintenance actions was considered. Based on the proposed short term maintenance model for component level, a dynamic opportunistic maintenance model was built for multi-unit series systems (system level). Then, an optimization example for a three equipment series production system was given to illustrate the implementation of the proposed decision model and the decision process.Finally, considering the fact that intelligent maintenance is not always necessary for each equipment in multi-unit production systems, a maintenance decision theory considering multi-maintenance-method was proposed. After giving the concept and the features of the new maintenance theory, a maintenance decision and optimization process was proposed, and the methodologies used in the process were discussed detailedly.This research is an important part in the intelligent maintenance field, and the research result can provide decision support for job-shop maintenance scheduling of production systems. At the same time, the proposed maintenance decision theory with consideration of multi-maintenance method, which is an extension of intelligent maintenance, can be a useful guidance for the manufacturing industry to organize the maintenance activities of production systems.