Evaluating OMA System Identification Techniques for Drive-by Health Monitoring on Short Span Bridges

Abstract

Drive-by health monitoring (DBHM) is an indirect structural health monitoring (SHM) strategy developed to reduce costs associated with traditional SHM systems on a variety of bridge structures. Experimental studies have successfully demonstrated DBHM’s bridge system identification capabilities; however, there exists a noticeable lack of full-scale experiments validating the methodology’s feasibility on highway bridges shorter than 18.28 m. Furthermore, few studies have used existing operational modal analysis (OMA) techniques, as the DBHM methodology violates fundamental OMA assumptions. This study addresses these research gaps by experimentally investigating the feasibility of employing OMA techniques in DBHM to identify the modal properties of a 9.14 m bridge span. Multiple OMA techniques were employed to identify their strengths and weaknesses under the DBHM framework. Results demonstrated that histograms constructed of frequencies identified across multiple tests were necessary to consistently identify bridge frequencies and that a tradeoff exists between vehicle mass and the speed at which bridge frequencies can be identified; this tradeoff also has an effect on the vehicle-on-bridge occupation time and the resolution of bridge frequencies. No single OMA technique was found to yield the best system identification capabilities in all test conditions; rather a combination of techniques is recommended, where the continuous identification of frequencies across multiple methods provides a means of reliably distinguishing between real and spurious frequencies and helps with accurately labeling vehicle and bridge frequencies.