Estimating optimal replacement time for systems with two-state degradation and technological obsolescence

Abstract

This paper presents a method to estimate the optimal replacement time of systems subject to two-state degradation processes while accounting for technological obsolescence. We apply this method to cable joints, where the two states represent a functional state and a degraded state characterized by partial discharge activity. Using more than 10 years of data from online partial discharge monitoring and historical failure records, we estimate the parameters for our statistical survival models. These models are then used to simulate degradation processes and associated costs over a finite time horizon, enabling the determination of the optimal replacement time. We illustrate the methods with two use cases: one in which immediate replacement is optimal and another in which delaying replacement yields the best outcome. To quantify the uncertainty in the optimal replacement time, we employed bootstrapping and the delta method during the simulation process.