Behavior of Riveted Connections in Steel Truss Bridges
Jost, Saura Michael
MetadataShow full item record
Current recommendations provided by the Federal Highway Administration (FHWA) for load rating of steel truss bridges were shown to be overly conservative in estimating the rivet capacity in shear by researchers from the University of Washington. Many steel truss bridges in place were constructed during the mid 20th century and require such load rating to determine if retrofitting, tear down, or a live loading limit is necessary. Past research on riveted connections has not included tests on as-built riveted connections, but rather shop fabricated specimens. Therefore, an enhanced understanding of ultimate rivet shear strength in gusset plates and the relation to connection length could improve the current procedure. Experiments on secondary riveted joints with various connection lengths were tested in a Universal Testing Machine (UTM) to determine the average ultimate shear strength of the rivets. The riveted connection was assembled with supporting members to improve capacity and stability and allow for testing of longer connections through shear failure. The test setup used for this investigation placed each rivet under a uniform load. This differed from previous research performed by Fisher and Rumpf (1967) on lap splices. Their loading mechanism caused the average ultimate strength of the rivets to decrease as connection length increases, and therefore a reduction factor was recommended by AASHTO for connections greater than 50 inches. The results of this test program in combination of past research will serve as an upper and lower bound, respectively, for connection behavior. Results showed that the average ultimate rivet shear strength was not a function of connection length under the designed load mechanism. The average ultimate rivet shear strength was larger than the values used in the load rating procedure and past research. The connecting elements deformed prior to failure of the rivets, which is a result of the rivets being stronger than the nominal values from the time of construction. This suggest that for gusset plate connections loaded in a distributed manner, that a reduction factor for ultimate rivet strength may not be needed, and there is potential that current reduction factors for longer connections could be reduced.
- Civil engineering