Bridge Structural Analysis Using Computational Structural Mechanics

Background

Bridge failure due to wind has been observed as far back as 1823. The latest concept for an efficient and cost-effective bridge design is the cable-stayed bridge. Bridge stay cables, however, have exhibited large-amplitude vibrations as a result of wind loadings, sometimes in combination with rain. In recent years, attempts have been made to model this problem both in the laboratory and on the computer. Several wind tunnel tests have been conducted in North America in an effort to better understand the mechanisms involved in this problem so that better or more efficient mitigation methods can be established.

While some tests have focused on simulation of rain or water rivulets on the surface of the cable and its influence on stability, others have concentrated on dry cable behavior, looking at response to vortex shedding and galloping, and measurement of pressure fluctuations and distribution of the cable surface. While some experimental work has looked at the effectiveness of aerodynamic surface treatments, cable suppliers have performed this research, and the results are proprietary. Computational and simulation methods are needed to efficiently evaluate the direct effects of wind loads on cables under a range of expected wind conditions (turbulence intensity, speed, and direction) and to assess the efficacy of various mitigation methods, including surface treatment.

It is also possible for undesirable cable vibrations to result from wind-induced motion of the cable anchorages. In this case, the anchorage motion can be the result of either traffic on the bridge, wind loading on the bridge deck or towers, or combinations of the two. Here, cable vibration is not primarily the result of wind load directly on the cable, but on some other part of the structure and feeding energy into the cable. Where there is a relatively close match between deck/tower and cable frequencies or multiples thereon, large cable vibrations can result. To better understand the nature and extent of this problem, computational and simulation methods are needed to efficiently model the anchorage motions resulting from buffeting and other wind effects on the main structure and assess the effect of these motions on cable stability.

TRACC Research Activities

Prior analyses performed by the Federal Highway Administration (FHWA) on the dynamics of bridge decks, specifically the Bill Emerson Memorial Bridge, were used as part of the structural analysis benchmark suite for TRACC’s high-performance computer acceptance testing. The FHWA baseline results were obtained on a dual Xeon 3.2-GHz machine and on a 16-core 2.6-GHz Thinkmate workstation. Execution of the full-size, detailed cable-stayed Bill Emerson Memorial Bridge model for one-half the span required 2,600 hours/108 days on the dual Xeon machine and 400 hours/16.7 days on the 16-core Thinkmate workstation. Performing the same analysis on the TRACC cluster with 256 CPUs only required 32.5 hours, which is a significant reduction in computing time on this problem.