Abstract
The main objective of a maintenance policy consists of conducting maintenance actions at lower costs. This paper proposes an approach for comparing numerically three maintenance strategies, involving minimal repairs at failure, replacement with complete renewal only at the first failure, and replacement with complete renewal at each failure. These strategies are integrated into a modified block replacement policy that includes corrective and preventive maintenances. The approach proceeds by presenting the mathematical models at the component level and at the system level. As the renewal function for generalised Weibull distributions is impossible to obtain, a novel asymptotic algorithm is introduced for estimating the replacements number. However, a multi-component industrial example is proposed for selecting the strategy that minimises the maintenance costs. A sensitivity analysis is performed for comparing an opportunistic maintenance policy with the proposed replacement policy to check if substantial cost reduction still possible. The experiment results show clearly that the third strategy is the most efficient and reduces maintenance costs to a very low level. Finally, we think that the developed study provides a flexible and less costly solution to deal with maintenance decision-making for systems that do not have modern technological equipment to collect data from system breakdowns.
Disclosure statement
No potential conflict of interest was reported by the authors.
Additional information
Funding
Notes on contributors
![](/cms/asset/e6918d87-07aa-4995-b81b-67beabfdf1fc/tprs_a_1832275_ilg0001.gif)
Mohamed Larbi Rebaiaia
Mohamed-Larbi Rebaiaia holds a doctorate degree (PhD) in Industrial Engineering from Laval University (Canada), a doctorate degree in Computer Science from Batna University, and a Master by Science in Operational Research from Annaba university (Algeria). He is currently a researcher in Production & operations management, and reliability & maintenance engineering, at the Science and Engineering faculty (Laval University), since 2008. Before 2007, he was an assistant professor in Computer Science and Operational Research at Annaba and Batna universities. His research interests include Networks Reliability Evaluation & Optimization, Maintenance Engineering, Production Management & Planning, and Software Engineering, Machines Learning & AI.
![](/cms/asset/f70ce195-03e4-4b94-bacb-e84793836fe0/tprs_a_1832275_ilg0002.gif)
Daoud Ait-kadi
Daoud Ait-Kadi holds a doctorate degree (PhD) in Industrial Engineering and a master’s degree in applied sciences from Polytechnique school at Montreal University (Canada). He is a Full professor and researcher in the fields of Modeling, Optimization and Validation of the Reliability, Maintainability and Availability of systems subject to one or more degradation modes, Integrated Logistical Support, Design and Management of value creation networks, Management of end-of-life Products and Reverse Logistics, Management of spare parts, Optimization of Systems’ Performance with a view to Sustainable Development.