Modelling industrial maintenance systems and the effects of automatic condition monitoring

摘要

In this dissertation, industrial maintenance activities are researched from a systemic point of view. Maintenance is considered from the selected viewpoint as a control system for controlling the reliability of machines in a process environment. The control takes place through communication of information. Thus, maintenance can also be considered as an information processing system. Therefore, the viewpoint also supports development of future maintenance information systems. The specific interest of the research is in modelling the effects of automatic condition monitoring systems enabled by embedded electronics and software in industrial machines.The research problem is to model maintenance systems and the effects of condition monitoring. Because the focus is on the maintenance systems, the research approach differs from the usual reliability engineering approaches, which focus on the reliability of mechanical systems. The applied methods are a literature study of the main reliability paradigms and the systems theory to derive a theory of maintenance systems. The theory is then applied by developing a UML knowledge model, an applied Gorry-Morton control activities model, a stochastic simulation model and a dynamic simulation model of the maintenance systems. Also, one additional simulation model is developed to study the effect of remote condition monitoring on the spare parts supply chain.The main results indicate that the subjective nature of failures, the available information about the machines as well as the repeatability of the maintenance actions are the most important factors in the maintenance system control. Also, correct selection of the maintenance system cost functions is critical in determining the applied maintenance policies.Condition monitoring increases observability of the states of the machines and therefore enables more effective maintenance systems with the help of condition-based maintenance. However, the effectiveness of condition-based maintenance depends on the accuracy of the monitoring and the coverage of the failure diagnosis. The failure patterns and the repeatability of the maintenance actions also contribute significantly to the effectiveness of condition-based maintenance. In spare parts supply chains, remote condition monitoring can be used to stabilise the supply chain variability and to reduce the supply chain sensitivity to random noise and sudden changes in the consumption.