Abstract

Lebesgue-sampling-based prognostic algorithms have demonstrated computational advantages and execution time savings of several orders of magnitude over Riemman-based prognostic methods. In this regard, some remarkable achievements have been obtained when using Lebesgue-sampling (LS) methods to build empirical degradation models in terms of statistics of the Time-of-Failure distribution, such as the Just-in-Time Point (JITP). This article studies some theoretical issues related to the implementation of LS-based prognostic methods, and analyzes the impact associated with the utilization of alternative procedures for the computation of Time-of-Failure probability density function statistics.