A fuzzy approach of online reliability modeling and estimation.

International audienceIn maintenance field, traditional concepts like preventive and corrective strategies are progressively completed by new ones like predictive and proactive maintenance. For that purpose, a fundamental task is the estimation of the provisional reliability of equipment as well as its remaining useful life. However, traditional approach of reliability based on statistical analysis can be not suitable as very few knowledge can be available. Within this frame, the general purpose of the work is to explore the way of developing a fuzzy approach of on-line reliability modeling and estimation in order to take into account the uncertainty as welle as possible. A federative point of view of the reliability modeling process and of the prognostic of degradation activity is proposed. From that, two ways of considering uncertainty in reliability modeling are discussed (probabilistic, fuzzy/possibility approaches), and the inherent limits of both methods are pointed out