Utilizing Fiber-Reinforced Polymers to Retrofit Steel Bridge Girders Damaged by Fatigue Loading.

Abstract

Tens of thousands of steel bridges constructed prior to the mid-1980s are susceptible to distortion-induced fatigue, which caused nearly 90% of cracking observed in steel bridge girders (Connor, R. and Fisher, J. 2006). Distortion-induced fatigue is caused by secondary stresses induced in cross-frames or diaphragms as differential deflection occurs between bridge girders. Girders are interconnected at discrete locations with cross bracing; as one girder deflects due to traffic loads, the cross bracing induces a force on the adjacent girder, causing high stresses at connections which can eventually lead to the formation of fatigue cracks. Retrofits which can be easily installed without disrupting bridge traffic are needed in order to retard the growth of these fatigue cracks and prevent bridge failure. This research utilized computer models as well as experimental testing to develop and test multiple retrofit techniques to prevent the growth of fatigue cracks. A retrofit measure referred to herein as the "composite block retrofit" was developed and successfully tested in a laboratory setting utilizing a 2.82-meter (9.25-ft.) model of a steel girder. This retrofit uses a block of carbon fiber reinforced polymer (CFRP) material applied at a cracked location to distribute stresses away from the crack. This retrofit measure was successful in preventing crack growth under severe fatigue loading. The magnitude of stress reduction calculated with the computer models was compared with the magnitude of stress reduction recorded from a strain gage attached to the girder subassembly. The computer analyses indicated a 96% stress reduction and the strain gage on the girder subassembly recorded a 92% stress reduction. Thus, this process extended the fatigue life of the retrofitted girder considerably, possibly as much as 4 million cycles or more. This comparison showed that if properly used, computer models could provide an accurate representation of the magnitude of stress reduction that may be expected during physical test trials.