Optimization of opportunistic replacement activities: A case study in the aircraft industry
Preprint, 2007

In the aircraft industry maximizing availability is essential. Maintenance schedules must therefore be opportunistic, incorporating preventive maintenance activities within the scheduled as well as the unplanned ones. At the same time, the maintenance contractor should utilize opportunistic maintenance to enable the minimization of the total expected cost to have a functional aircraft engine and thus to provide attractive service contracts. This paper provides an opportunistic maintenance optimization model which has been constructed and tested together with Volvo Aero Corporation in Trollhättan, Sweden for the maintenance of the RM12 engine. The model incorporates components with deterministic as well as with stochastic lives. The replacement model is shown to have favourable properties; in particular, when the maintenance occasions are fixed the remaining problem has the integrality property, the replacement polytope corresponding to the convex hull of feasible solutions is full-dimensional, and all the necessary constraints for its definition are facet-inducing. We present an empirical crack growth model that estimates the remaining life and also a case study that indicates that a non-stationary renewal process with Weibull distributed lives is a good model for the recurring maintenance occasions. Using one point of support for the distribution yields a deterministic replacement model; it is evaluated against classic maintenance policies from the literature through stochastic simulations. The deterministic model provides maintenance schedules over a finite time period that induce fewer maintenance occasions as well as fewer components replaced.


Michael Patriksson

Chalmers, Mathematical Sciences, Mathematics

University of Gothenburg

Torgny Almgren

Niclas Andreasson

Dragi Anevski

University of Gothenburg

Chalmers, Mathematical Sciences

Ann-Brith Strömberg

Chalmers, Mathematical Sciences, Mathematics

University of Gothenburg

Johan Svensson

Subject Categories

Computational Mathematics

Probability Theory and Statistics

Preprint - Department of Mathematical Sciences, Chalmers University of Technology and Göteborg University: 2008:45

More information