With increased automation and mechanisation, many modern plants have installed flexible computer-controlled automatic and unmanned equipment. Along with the installation of this complex and capital-intensive equipment, maintenance work and costs have increased substantially. Wireman [1] conducted a benchmarking exercise and found that the maintenance cost for industrial firms in the USA has increased by 10–15% per year since 1979. High maintenance costs highlight the need to define clearly the maintenance objectives, to develop and
Correspondence and offprint requests to: Dr R. C. M. Yam, Department of Manufacturing Engineering and Engineering Management, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong
enhance modern maintenance management methods continously, to integrate maintenance and production activities effectively, and to use intelligent computer-based maintenance systems. In industry, failure-driven and time-based maintenance are two major maintenance management approaches. Failure-driven maintenance (FDM) is also called run-to-failure maintenance [2]. It is a reactive management approach, where corrective maintenance is often dominated by unplanned events, and is carried out only after the occurrence of an obvious functional failure, malfunction, or breakdown of equipment. Corrective maintenance action can restore an item of failed equipment by either repairing or replacing the failed component. If the equipment is non-critical or is easily repaired, unplanned stoppages of the equipment will cause minimal disruption to production, and FDM could be an acceptable maintenance approach. However, in the case of purely random breakdown of equipment that would have a serious impact on production, an emergency corrective maintenance action is necessary to avoid the serious consequences of failure [3]. The practical implication of emergency corrective maintenance often results in the unpredictable performance in a plant, i.e. high equipment downtime, high cost of restoring equipment, extensive repair time, high penalties associated with the loss of production, and a high spare parts inventory level [4]. Time-based maintenance (TBM) is also called periodic preventive maintenance. In order to slow down the deterioration processes leading to faults, primary preventive maintenance is carried out by periodically lubricating, calibrating, refurbishing, inspecting, and checking of equipment on a regularly scheduled basis. TBM assumes that the estimated failure behaviour of the equipment, i.e. the mean time between functional failures (MTBF) is statistically or experientially known for equipment and machinery degrading in normal usage [5]. TBM also involves minor or major planned shutdowns of systems for overhaul or predetermined repair activities on still functioning equipment. This can prevent functional failures by replacing critical components at regular intervals just shorter than their expected useful lifetime. System overhaul and critical item replacement at fixed intervals are widely adopted by many modern automated plants. Although TBM can reduce the probability of system failure or the frequency of unplanned
384 R. C. M. Yam et al.
emergency repairs, it cannot eliminate the occurrence of random catastrophic failure. Some TBM practices may be out of date and do not cope with the actual operating requirement of the modern automated plant. Most of the maintenance decisions are made by experienced planners according to the original equipment manufacturer’s recommendations, the reported breakdown history or failure data and the operating experience and judgement of the maintenance staff and technicians. Under a situation of uncertainty, it is very difficult to plan the maintenance activities properly in advance, and so, very often, the maintenance staff are required to work under “fire fighting” conditions [6].