Extraction of bridge mode shapes from the response of a two-axle passing vehicle using a two-peak spectrum idealized filter approach

Abstract

Bridge mode shape is an essential indicator in the assessment of structural health of bridges. Previous studies have demonstrated the feasibility of extracting bridge mode shapes from the response of a passing vehicle. However, the majority of these studies modeled the vehicle as a single degree-of-freedom (DOF) spring mass, and the accuracy of the recovered bridge mode shapes was not desirable. Furthermore, limited experimental validations exist for bridge mode shape extraction. In this study, a practicalmulti-DOF two-axle vehicle for scanning bridge mode shapes was investigated. First, a closed-form solution for the responses of a passing symmetric two-axle vehicle interacting with a simply supported bridge was obtained, in which the bridge mode shapes were extracted from the vehicle responses using a novel two-peak spectrum idealized filter approach. Then, a numerical finite element simulation was performed to verify the proposed approach for synchronously identifying bridge frequencies and mode shapes, and the impacts of vehicle speed, vehicle parameters, damping, road roughness, and environmental noise were parametrically analyzed. Finally, laboratory vehicle-bridge interaction tests were conducted to extract bridge mode shapes from the measured accelerations of a passing two-axle vehicle. The proposed approach only considered the two-peak spectra of the left-shifted and right-shifted bridge frequencies in the frequency spectrum of the vehicle response, which was theoretically derived to be capable of accurately extracting bridge mode shapes. Numerical investigations revealed that the proposed approach can synchronously identify bridge frequencies and mode shapes for high vehicle speeds (e.g., 20 m/s), various vehicle parameters, and the existence of vehicle and bridge damping. Road roughness and environmental noise contaminated the two-peak spectra of bridge frequencies, which negatively impacted the effectiveness of the proposed approach. However, provided that the bridge frequency is visible in the frequency spectrum, the proposed approach is feasible. Results of the laboratory tests demonstrated that bridge frequencies and mode shapes were successfully identified from the accelerations of a passing two-axle vehicle. This study verified that a two-axle passing vehicle can be used to scan bridge mode shapes, which could facilitate wide engineering applications of the vehicle scanning method for indirect monitoring of bridge infrastructures.