Fatigue damage prediction based on strain field estimates using a smoothed Kalman filter and sparse measurements

In this work, we address the problem of fatigue damage prediction in a truck component excited by road induced vibrations. The damage is computed from strains estimated from sparse measurements of the dynamic response. Two different fixed-lag smoothing algorithm are employed, an augmented Kalman filter extended with a fixed-lag smoother and a smoothed joint input-state estimation algorithm. The system is described with a finite element model, and due to the complexity of reproducing the system by the model, the resulting representation may contain a number of discrepancies. Nevertheless, both smoothing algorithms succeed in capturing the dynamic behaviour of the component, although the estimated strains are affected more by the large model error than the estimated acceleration are. Furthermore, it is shown that the proposed methodology of strain estimation and fatigue damage calculations correlate well with the observed failure of a component when tested in a full-scale fatigue test of a truck chassis.