Prognosis of Component Degradation Under Uncertainty: A Method for Early Stage Design of a Complex Engineering System

Abstract

This paper proposes a method that dynamically improves a statistical model of system degradation by incorporating uncertainty. The method is illustrated by a case example of fouling, or degradation, in a heat exchanger in a cogeneration desalination plant. The goal of the proposed method is to select the best model from several representative condenser fouling models including linear, falling rate, and asymptotic fouling, and to validate and improve model parameters over the duration of operation. Maximum likelihood estimation (MLE) was applied to obtain a stochastic distribution of condenser fouling. Akaike's Information Criterion (AIC) and the Bayesian Information Criterion (BIC) were then computed at time intervals to assess the accuracy of the MLE results. The degradation model was further evaluated by estimating future prognoses and then crossvalidating with real world fouling data. The results show the accuracy of a prognosis can be improved substantially by continuously updating fouling model parameters. The proposed method is a step toward facilitating prognosis of engineering systems in the early design stages by improving the prediction of future com- * Address all correspondence to this author, tomonori@mit.edu. ponent degradation.

A maintenance-focused approach to complex system design

“…In CBM, system components are continuously monitored by various sensors to measure the state of the component, and computer algorithms estimate the "health" of the component, and predict the remaining useful life (RUL) using either physics-based models or data-driven methods (Wang & Vachtsevanos, 2001;Wang & Wang, 2015). PHM is the research area that focuses on modeling RUL distribution and reliability in a number of different areas, including thermal and electronic components (Xing et al, 2011;Yu et al, 2012), civil infrastructure systems (Peng et al, 2012;Loyola et al, 2013), airplane maintenance (Bodden et al, 2005), and battery technologies (He et al, 2011). Information provided by PHM technologies is used in CBM to schedule maintenance accordingly.…”