Development and application of automated OMA algorithms

Within the contest of vibration-based monitoring, the paper summarizes the conceptual development of automated procedures of modal parameter estimation (MPE) and modal tracking (MT). The performance of the proposed algorithms is demonstrated using data collected both in single data-sets and over a period of continuous monitoring. The MPE process, based on the automated interpretation of the stabilization diagram associated to any parametric identification methods, is exemplified by using the SSI-Cov technique and consists of three key steps: (1) filtering a high number of spurious poles in the stabilization diagram (2) clustering process and (3) improving the accuracy of the estimates. Eventually, the mean modal parameters are computed for each clustered mode. The MT task is based on the definition of: (a) a pre-selected list of baseline modes with adaptive thresholds and (b) a dynamic reference list of modes associated to fixed thresholds.