Cracks emerging from geometrically discontinuous locations under cyclic environmental loadings are critical concerns for the safety of the existing structural components. The crack-based fatigue assessment is essential for the evolving digital twin of sustainable infrastructures, including bridges, ships, and offshore platforms, to optimize the lifetime cost of these structures. This study presents an enhanced neural network-bootstrap particle filtering algorithm to construct the complex relationship between the normalized strain relaxation indicators and the crack profiles based on the numerical simulation and experimental validation. The high-cycle fatigue bending test of the welded plate connections confirms the robustness of the proposed approach in estimating the fatigue crack initiation and propagation through both strain measurement and nondestructive testing data. To overcome the uncertainties caused by the limited strain measurement, crack measurement, and different non-destructive techniques, this study combines a bootstrap particle filtering approach with an interpolation method to update the crack prediction algorithm. As validated by the experimental results, the intelligent crack sizing approach demonstrates a potential solution for crack size forecasting through affordable strain gauges in the broad framework of digitally twinning the next-generation infrastructure.
EXTENDED ABSTRACT