Remaining useful lifetime estimation and noisy gamma deterioration process

In many industrial issues where safety, reliability, and availability are considered of first importance, the lifetime prediction is a basic requirement. In this paper, by developing a prognostic probabilistic approach, a remaining lifetime distribution is associated to the system or component under consideration. More particularly, the system's deterioration is modelled by a non-homogeneous gamma process. The model considers a noisy observed degradation data and by using the Gibbs sampling technique, the hidden degradation states are approximated and afterwards the system's remaining useful lifetime distribution is estimated. Our proposed prognosis method is applied to the Prognostic and Health Management (PHM) 2008 conference challenge data and the interest of our probabilistic model is highlighted. To point out the interest of the prognostic, a maintenance decision rule based on the remaining lifetime estimation results is proposed. (C) 2016 Elsevier Ltd. All rights reserved.