Sparse and Low-rank Methods for Structural Identification and SHM

Recent advances in Statistical Learning techniques have enabled the development of new algorithms for monitoring, data cleansing, data compression, structural identification, damage identification, and structural dynamics in general. This keynote presents novel sparse and low-rank methods to address the inverse problems in structural dynamics, identification, and data-driven health monitoring. In particular, the emerging mathematical tools such as sparse representation (SR) and compressed sensing (CS), as well as the unsupervised multivariate blind source separation (BSS), are used to harness the structural dynamic features and damage information intrinsic within the structural vibration response measurement data, which is found to have sparse and low-rank structure. Data-driven approaches are developed towards rapid, unsupervised, and effective system identification, damage detection, as well as massive SHM data management.