Modular Subspace-based System Identification and Damage Detection on Large Structures

In Operational Modal Analysis (OMA) of large structures it is often needed to process sensor data from multiple non- simultaneously recorded measurement setups, especially in the case of large structures. In order to obtain global modal parameters (natural frequencies, damping ratios, mode shapes) with covariance-driven Stochastic Subspace Identification (SSI), the data from all the setups is normalized, merged and processed together. With this “Pre Global Estimation Re-Scaling” (PreGER) approach the global modal parameters are obtained automatically, but lots of data is processed at the same time, which can easily lead to memory problems when dealing with a big number of measurement setups and sensors. In this paper, a new efficient variant of the PreGER algorithm is presented that avoids the numerical explosion of the calculation by using a modular approach, where the data from the measurement setups is processed setup by setup and not at the same time. Furthermore, a new efficient variant of the subspace-based stochastic damage detection for multiple measurement setups is presented and we show the efficiency of the identification and damage detection algorithms on a relevant industrial example.