Time series damage state estimation using empirical transfer function estimation approach and broadband chirp active sensing

The present paper proposes a state estimation approach that estimates the fatigue induced damage state of a metallic structure under biaxial fatigue load. Ultrasound based broadband active sensing technique is used. A novel dual sensor signal based damage index estimation technique is developed which removes the contribution from high frequency input dependant noise. Nonparametric system identification based empirical transfer function estimation approach is used to evaluate the scalar damage index. The time series estimation of the damage indices (or damage states) shows good correlation with visual measurements such as crack lengths. The time series 2s error bound was also evaluated to study the effect of measurement noise on damage state estimation.