Data fusion for system identification of the humber bridge

Abstract

In Operational Modal Analysis (OMA) of large structures, ambient vibration data from multiple non-simultaneously recorded measurement setups is often needed to be processed. These setups share some sensors in common, while the others are moved from one setup to the next. Like this detailed mode shapes of the structure can be obtained, mimicking lots of sensors, while in fact only a few sensors are used for the measurements. Recently, the "Pre Global Estimation Re-Scaling" (PreGER) for the Stochastic Subspace Identification (SSI) was proposed to obtain global modal parameters of the structure. It is a fully automated method that takes differences in the unmeasured background excitation levels between the setups into account, merges the data and does the global system identification. Like this, the different measurement setups can be processed in one step and do not have to be analyzed separately. In this paper, system identification results of the Humber bridge are presented, which is a challenging example as a big number of setups is available and special measures need to be taken to avoid numerical explosion of the computation. The results are compared to the PoSER approach (Post Separate Estimation Re-Scaling) with data-driven SSI.