Setting sail towards predictive maintenance : developing tools to conquer difficulties in the implementation of maintenance analytics

Abstract

Unexpected downtime of equipment is disruptive in complex manufacturing supply chains and imposes high costs due to forgone productivity. Executives in asset-intensive industries therefore regard such unexpected failures of their physical assets as a primary operational risks to their business. Predictive maintenance (PdM) (including condition-based maintenance) can aid practitioners in preventing these unexpected failures and getting insight into current and future behaviour of their assets. However, the use of PdM in practice seems to lag behind recent technological advancements and our theoretical understanding. The current study therefore aims to further develop our understanding on the use and adoption of predictive maintenance and, based on these observations, develop tools to better support the practical application of predictive maintenance. This research is guided by the following research question: How can the practical application of predictive maintenance better be supported? To be able to answer this question, an explorative multiple-case study is conducted including fourteen cases from various industries in the Netherlands to study successful applications of predictive maintenance. The focus in this multiple-case study lays on both the technical and the organizational aspects of PdM, because the organizational application process of PdM seems overlooked by the academic literature. The multiple case study reveals that almost all organizations who applied PdM successfully have followed a costly trial and error process. This appears to be the result of the technical and organizational complexity of the application of PdM and the absence of effective theoretical guidance in: (i) selecting the most suitable techniques for PdM; (ii) identifying the most suitable candidates for PdM; and (iii) evaluating the added value of PdM. To conquer the three main identified problems and to assist practitioners in the implementation of PdM, three corresponding decision support tools – which can be used together – have been designed in the remainder of this dissertation. The three solutions are designed using a structured design science process. Therefore, after studying the problems in-depth to define design criteria and select design principles, the developed solutions are demonstrated in practice using case studies in various industries. Future research should be guided towards the refinement and testing of the provided methods.