Abstract:
Predicting the probability of dynamic component failure in aircraft is of interest to both civilian and military operators for safe and cost-effective component retirement. While aircraft OEM assumed usage spectra have been used throughout the years for fatigue forecasting and for creation of part replacement timelines, advancements in aircraft regime recognition and instrumentation have made it possible to track the portion of consumed useful life in life-limited components based on actual usage. This paper introduces a novel approach to condition based maintenance (CBM) that is designed to be used in conjunction with a probabilistic regime recognition scheme. The proposed methodology results in a probability distribution of incurred damage for specified life-limited components on an individual aircraft basis, with applications to fleet-wide probabilistic damage spectrum. The primary advantage of the proposed methodology is the rigorous treatment of regime uncertainty that is quantified in probabilistic regime recognition. A Gaussian probability density approximation is used to improve computational efficiency of the damage estimation method, and effects of the flight condition on the uncertainty in the damage estimation are investigated. Following a detailed explanation of the proposed methodology, results are provided by applying the algorithms to simulated data for the SH-60B and compared against a rule-based approach for individual aircraft damage tracking.