Field testing of the Wolf Creek Curved Girder Bridge: Part I: Vibration Tests

About the report:

The Wolf Creek Bridge is a curved, multi-girder three span steel composite bridge located south of Narrows, Virginia, that was completed in 2006.  A finite element model of the bridge revealed that pier flexibility may be important in modeling the bridge.  In addition, questions have been raised as to the effectiveness of the C15x33 diaphragms in providing lateral transfer of loads between members.

This study was conducted as Phase I of a project for which the overall goal was to use field testing to obtain a better understanding of the behavior of multi-span curved girder bridges. An array of vertically oriented accelerometers was located along the inner and outer edges of the bridge, along with radially oriented accelerometers along the outer edge, a tangentially oriented accelerometer on the outer edge, and an additional vertical accelerometer placed in the middle of the center span. Dynamic response data were collected under a variety of excitations, including sinusoidal forcing induced by an electro-dynamic shaker, impulse loadings at various locations, and several different vehicular loads.

The dynamic data were transformed into the frequency domain and analyzed using a simple frequency domain algorithm to extract vibration frequencies and mode shapes. The resulting frequencies and mode shapes were compared with the existing finite element model. The findings indicated that not only is pier flexibility important, as had been hypothesized, but also that end constraints imposed by highway guardrails change both the natural frequencies and the mode shapes in ways that had not been anticipated. Frequencies of modes with strong pier participation and modes with strong transverse (hogging) components were lower than predicted by the computer model, suggesting that pier stiffness may be less than the model predicted and that transverse stiffness, to which the diaphragms contribute, may also be estimated.

Implications of this study could have a significant effect on future health monitoring applications as they pertain to both curved and straight girder bridges.  It is essential that finite element models in such long-term applications be able to reproduce the “as-built” response characteristics of a bridge. The current study raised significant issues about the ability to model the behavior of curved girder bridges correctly. Thus, it will be important to perform subsequent numerical research studies to develop models that will result in more precise predictions and to use these and other methods being developed in any health monitoring applications.

Authors