Application of stochastic filtering for lifetime prediction

This paper introduces a stochastic filtering modeling approach for predicting the remaining lifetime of a component based on information on the stochastic degradation process and uncertain condition monitoring measurements. The model is illustrated by a case study, where the degradation is assumed to be a simplified fatigue crack growth process. The model accounts for uncertainties in both degradation process and condition measurements in a sound way. If completed with information on costs of monitoring, failure and replacement, such model could be used in optimizing both the condition monitoring intervals and, e.g. the replacement time for the component. (C) 2005 Elsevier Ltd. All rights reserved.