Exploring structural dynamics in exceptionally low ambient noise: modal identification of a deep-underground steel frame

Operational Modal Analysis (OMA) is a well-established procedure used to identify the main modal parameters of a structure —natural frequencies, mode shapes, and damping ratios— by processing experimental data acquired under operational conditions, typically from acceleration measurements at low vibration amplitudes. This type of experimental activity is also referred to ambient vibration testing (AVT), since civil structures are generally excited by broadband environmental inputs such as wind or low-amplitude ground motion. Numerous methodologies have been developed to identify dynamic properties from such measurements, generally assuming an unknown excitation source; these are commonly referred to as output-only methods. Consequently, when designing these types of tests, it is essential to ensure a low noise-to-signal ratio. The present research investigates the feasibility and effectiveness of AVT in environments characterized by exceptionally low ambient noise, such as the Gran Sasso National Laboratory (INFN-LNGS). The laboratory, located within underground galleries approximately 1,400 meters beneath the rock, hosts experimental control rooms composed of three-story bolted steel frames. To properly design the type and placement of sensors for a vibrational monitoring system, ambient vibration tests were performed under different experimental configurations. Subsequently, the main modal parameters were identified by comparing results obtained through various approaches. Finally, based on the experimental outcome, a sensitivity-based method was applied to update the numerical model for prediction analyses. This work represents a preliminary step toward the development of a Digital Twin dedicated to the laboratory’s facilities, enabling enhanced structural assessment and continuous performance monitoring in deep-underground environments.