SUMMARY

Condition based maintenance (CBM) programs are of interest to both the U.S. military and commercial aviation sectors for use in rotary wing and fixed wing aircraft. Cost savings in fleet maintenance and safety improvements for personnel are two of the most immediate benefits. Current programs rely heavily on an assumed mission usage spectrum, and post processing recorded flight data from the aircraft health and usage monitoring system (HUMS) using regime recognition (RR) coupled with known regime-based fatigue rates. This dissertation presents a novel approach to regime recognition by casting the problem in a probabilistic framework that captures the uncertainty inherent in classifying aircraft flight regimes. RR algorithms utilizing an interacting multiple model (IMM) estimator are presented. A developed framework and associated algorithms are then presented for generating probabilistic cumulative damage estimates of life-limited components that captures uncertainty in the damage, which is capable of using probabilistic RR results as an input. Results for the proposed regime recognition and damage estimation algorithms using simulated data from the SH-60B are presented, along with comparisons to RR results produced from traditional rule-based methodologies. The regime recognition and damage estimation algorithm viability are verified using real world HUMS data gathered from a generic utility-scale helicopter. Finally, the RR algorithms are adapted to work on fixed wing aircraft, and usage spectrum results are shown using simulated data for the F-16.