Machine learning algorithms are progressively used in structural health monitoring (SHM) applications. However, damage identification in a supervised learning context is challenging due to insufficient training data for various damage states of the structure. It may be feasible to acquire unhealthy sensor data for low-cost structures, but not for high-value complex structures, such as aircraft. In this work, numerical modelling is employed to simulate hard to attain damage scenarios that are in turn used to create training databases for damage identification through machine learning. In order to take into account modelling uncertainty, a comprehensive set of models representing different damaged states with varying parameters is created. Clustering techniques are then used to explore robust features between the model and the physical structure which are finally used to train artificial neural networks. The framework is validated by an experimental case study of damage detection, in the form of loosened bolts on a laboratory tested structure.

The Grand LAgrangian Deployment (GLAD) used multiscale sampling and GPS technology to observe time series of drifter positions with initial drifter separation of O(100 m) to O(10 km), and nominal 5 min sampling, during the summer and fall of 2012 in the northern Gulf of Mexico. Histograms of the velocity field and its statistical parameters are non-Gaussian; most are multimodal. The dominant periods for the surface velocity field are 1–2 days due to inertial oscillations, tides, and the sea breeze; 5–6 days due to wind forcing and submesoscale eddies; 9–10 days and two weeks or longer periods due to wind forcing and mesoscale variability, including the period of eddy rotation. The temporal e-folding scales of a fitted drifter velocity autocorrelation function are bimodal with time scales, 0.25–0.50 days and 0.9–1.4 days, and are the same order as the temporal e-folding scales of observed winds from nearby moored National Data Buoy Center stations. The Lagrangian integral time scales increase from coastal values of 8 h to offshore values of approximately 2 days with peak values of 3–4 days. The velocity variance is large, O(1)m2/s2, the surface velocity statistics are more anisotropic, and increased dispersion is observed at flow bifurcations. Horizontal diffusivity estimates are O(103)m2/s in coastal regions with weaker flow to O(105)m2/s in flow bifurcations, a strong jet, and during the passage of Hurricane Isaac. The Gulf of Mexico surface velocity statistics sampled by the GLAD drifters are a strong function of the feature sampled, topography, and wind forcing