Research On Equipment Maintenance Scheduling For Complex Manufacturing Systems

With the rapid development of modern industrial and informationtechnology, complex manufacturing systems have been expanding andbecome highly automated. Especially in recent decades, many digitalmanufacturing systems have been proposed, such as agile manufacturing,intelligent manufacturing and scale manufacturing with characteristics ofinformation. These complex and automated manufacturing systems require ahigher demand on the stability and utilization of production equipment. Howto quickly and accurately use the available maintenance resources to develop areasonable equipment maintenance program is a major issue in academic andengineering research. Reasonable equipment maintenance program not onlyreduce losses due to equipment failure and yield, but also improve the valueand market competitiveness of manufacturing enterprises. To solve theequipment maintenance scheduling problem of semiconductor manufacturingprocess, this paper proposes reasonable equipment maintenance strategies interms of equipment maintenance modeling and maintenance strategies:1Monte Carlo algorithm is applied to simulation equipment continuousdegradation process considering the characteristics of the continuous state deterioration model, a maintenance strategy is implemented based on a set ofmaintenance and safety thresholds.2On the basic of equipment economic relevance and systemconfiguration, a combination of break-down maintenance and condition-basedmaintenance is used for multi-device system according to the level ofequipment deterioration, failure type, etc.3Considering the equipment failure rate, yield and other performanceindicators of semiconductor manufacturing systems, a two-level hierarchicalmaintenance modeling for semiconductor production equipment is proposed.The device layer is formulated as a Markov decision process(MDP), whichseeks to develop a sequential testing strategy for a single equipment, multipleproduct system. Short-term maintenance scheduling is developed in systemlayer based on a variety of information provided by device layer and limitedmaintenance resources.The maintenance models and methods to deal with the problem ofcomplex manufacturing system maintenance scheduling are complicated andneed to consider many factors, such as production planning and WIP. Thereare still many unresolved issues. The research work of this dissertation onlyfocus on the maintenance scheduling of complex manufacturing equipmentand can provide some reference for the future work.