ESTIMATION OF BRIDGE FREQUENCIES FROM THE VIBRATION RESPONSE OF A MOVING VEHICLE USING AN INTEGRATED VEHICLE-BRIDGE INTERACTION ANALYSIS

Abstract eng:
Traditionally, frequencies and other modal characteristics of existing bridges are measured with vibration sensors installed at multiple locations along the bridge. However, this approach is costly, time-consuming and in many cases not applicable; e.g. for common short and medium spans bridges, which represent the majority of bridges in service. During the last decade, there is a growing interest in an alternative approach for extracting the modal properties of bridges which relies on the response of passing vehicles. The idea is that a vehicle crossing a bridge with a specific speed excites the bridge. At the same time, the vibrating bridge affects the response of the passing vehicle. Thus, by processing the data obtained from the vehicle response, one can extract the dynamic characteristics of the bridge. However, this approach is not straightforward. The frequencies of the bridge cannot be easily identified directly from the vehicle response due to the vehicle frequency which usually appears as the dominant frequency of the coupled vehicle- bridge system and the effect of the road surface roughness. The present study assesses the accuracy and the effectiveness of two methods that are used to identify the frequencies of a straight reinforced concrete bridge, from the vibration response of a two-axle truck. The response of the vehicle is estimated from an integrated vehicle-bridge interaction analysis that considers the effect of the road roughness condition of the deck. The coupled system consists of the vehicle subsystem and the bridge subsystem. The truck vehicle is modeled as an assembly of rigid bodies, connected with linear springs and viscous dashpots representing the properties of the suspension system. The bridge frequencies are estimated using the Stochastic Subspace Identification (SSI) method. The equations of motion of the coupled vehicle-bridge system are expressed in a state space format, including the effects of road surface roughness. In this way, we separate the “known” parameters (i.e. response of the vehicle, and the road roughness excitation) from the “unknown” parameters (i.e. dynamic characteristics of the bridge). Using algorithms from numerical linear algebra (i.e. Hankel matrix, orthogonal projection theorem, and singular value decomposition), we derive the observability matrix that contains the bridge dynamic properties. To validate the results the study uses an independent method to identify the frequencies of the specific bridge that relies on signal processing of the moving truck vibration. The vertical accelerations of the moving truck, as obtained from the VBI analysis, are transformed into a Fourier spectrum, from which the dominant bridge frequencies are identified. In this context, the singular spectrum analysis with band-bass filter is applied to identify the bridge natural frequencies.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket