Search

Prognostics is essential in decision making to prevent cost and downtime. Prognostics is especially useful in variable loading conditions, such as loads in wind turbine gearboxes. Literature suggests that accurate prediction of remaining useful life will improve reliability and reduce maintenance cost and that there is a need for a comprehensive framework for quantifying uncertainties in prognostics. The research question of this study is “can probabilistic analysis quantify uncertainties in prognostics?” The specific aims of the research are to construct a generic probabilistic prognostics framework, to estimate the remaining useful life of structures under variable amplitude loading, and to estimate the remaining useful life of cracked gears under variable loading. A computational framework for probabilistic prognostics was developed to perform rainflow counting, dynamic analysis of a gearbox, finite element analysis of a gear tooth, and estimation of remaining useful life. The framework analyzes uncertainties related to state, loading, and material properties. The remaining useful life of structures under variable loading was analyzed by quantifying material and loading uncertainties and employing cumulative fatigue damage. The complete remaining useful life trajectories of structures under variable loading were determined. The remaining useful life estimation of a gear to predict crack growth in the gear tooth was also analyzed. The probability of failure at any loading cycle was quantified. The proposed probabilistic prognostics framework could quantify uncertainties in remaining useful life prediction for fault mitigation. Hence, it can be used as a helpful tool to prepare maintenance schedules in advance to reduce cost and avoid downtime.