Key points are not available for this paper at this time.
This paper presents a novel framework for optimizing preventive maintenance (PM) intervals under interval-dependent maintenance effectiveness in repairable systems. Traditional PM models often assume constant effectiveness, overlooking the empirical reality that restoration quality varies with the timing of intervention. Using failure and maintenance data from underground mining Load-Haul-Dump (LHD) trucks, we calibrate a virtual-age-based model where restoration effectiveness, denoted as Formula: see text, is a function of the PM interval Formula: see text. System failures are modeled via a non-homogeneous Poisson process (NHPP), and parameters are estimated through maximum likelihood techniques combined with global optimization algorithms including Genetic Algorithms (GA), Simulated Annealing (SA), and Particle Swarm Optimization (PSO). Univariate and multivariate sensitivity analyses reveal strong nonlinear and asymmetric relationships between PM intervals and availability, especially for moderate maintenance (Type II). Optimized schedules achieve significantly improved availability compared to OEM policies, and robustness checks show that small deviations from optimal intervals incur only marginal losses, providing operational flexibility. A set of three-dimensional surface plots further illustrates the interaction effects among PM types, while local perturbation analyses quantify local robustness. The proposed methodology enables maintenance planners to jointly evaluate effectiveness and timing, providing a scalable approach to real-world reliability optimization. The findings underscore the importance of interval calibration in maintenance scheduling and offer practical decision support for high-stakes industrial applications.
Uthman Said (Sat,) studied this question.