Neural network representation of fatigue damage dynamics

Abstract

Recently, a model of fatigue damage dynamics has been reported (Ray et al., 1994), which allows the damage information on critical plant components to be integrated with the plant dynamics for both online life prediction and offline control synthesis. This paper proposes a neural network implementation of the fatigue damage model in order to alleviate the problem of slow computation via conventional numerical methods. The results of simulation experiments reveal that a neural network algorithm could be used as an intelligent instrument for online monitoring of fatigue damage and also as a tool for failure prognostics and service life prediction.