Ambient loading and modal parameters for the Chulitna River Bridge. Issue 4 (April 2016)

• Record Type: Journal Article
• Authors: Xiao, Feng; Chen, Gang S; Hulsey, J Leroy; Dolan, J Daniel; Dong, Yongtao
• Is Part Of: Advances in structural engineering. Volume 19:Issue 4(2016)
• Page Start: 660
• Page End: 670
• Publication Date: 2016-04
• Subjects: bridge structural dynamics -- bridge structural system identification -- bridge vibration and monitoring -- structural health monitoring; Structural engineering -- Periodicals; Construction, Technique de la; Structural engineering; Periodicals; 624.1
• DOI: 10.1177/1369433216630045
• Languages: English
• ISSNs: 1369-4332

Abstract

The Chulitna River Bridge is a 790-ft five girder, five-span steel bridge on the Parks Highway between Fairbanks and Anchorage, Alaska. This bridge was built in 1970 and widened in 1993. Under the no-live load condition, five support bearings are not in contact. Heavily loaded trucks often travel across this bridge to the oil fields in Prudhoe Bay, Alaska. A virtual finite element modeling, dynamic field testing of the "ambient vibrational response, " and structural health monitoring system are used to analyze, evaluate, and monitor the structural performance. As the first stage of the research, this article presents results from the dynamic testing and evaluation of the structural responses of the bridge. In the dynamic field testing, 15 portable accelerometers were placed on centerline along the bridge length to record the structural response, and an ambient free-decay response was used to evaluate the dynamic properties of the bridge structure. Natural frequencies and modal damping ratios were identified and characterized using Hilbert–Huang transform and fast Fourier transform methods. Compared with conventional approaches, this study demonstrates that (1) the Hilbert–Huang method was found to be effective and suitable for modal parameter identification of a long steel girder bridge using ambient truck loading; (2) the nonlinear damping was, for the first time, identified based on Hilbert–Huang transform's amplitude–time slope; (3) modal frequencies are very sensitive to sensor location so their position should be optimized.