21.11: Asymmetrical cross sections in historic riveted bridges

Abstract

ABSTRACTThe assessment of the load‐carrying capacity of two steel bridges dating from 1907 and 1912, showed that most of the heavily loaded cross‐sections were asymmetrical. The bridges are of the two‐hinged portal frame type. Compared to simple beam bridges, portal frames with clamped or hinged supports allow redistribution of bending moments from the span centre towards the frame nodes. Consequently, the beam height at the span centre can be reduced, which is exactly what is needed for obtaining maximum vertical clearance of the subjacent waterway. Both bridges show heavy corrosion, some parts of the flanges and angle profiles of the end portals being seriously weakened. Full inventory has been made of the corrosion damage, as it appeared mostly in the edge frames of both structures. Reduction or suppressing of flanges has been introduced in the FE‐models of both bridges to assess the remaining load‐carrying capacity for real traffic loading.A particular characteristic are the asymmetrical cross‐sections of the main portal frames. Because of this, the compression and tensile stresses in both flanges are of unequal magnitude, favouring the latter. The reason for this became evident rather quickly. In tensile parts of the riveted structure, weakening of cross‐sections by holes has been taken into account, whereas it is not considered in compressed areas. Older standards and literature confirm the fact that this was allowed at the time riveted structures were still being built.As today little is known about riveting, it was felt that this needed further underpinning. Apart from the longitudinal effects, lateral shrinkage of rivets during placing and plastic deformation also exist within the holes. The plastic deformation in the hole introduces a high compression between the rivet and the hole edge. As the heated rivet is cooling down and before further shrinking of the rivet diameter, this pressure must be reversed. Hence, cooling does not reduce the rivet diameter entirely, a portion being used to decompress the connection. Approximations of the process have been worked out and showed that the final reduction of the rivet section is less than the tolerance for fitting bolts. Hence, the assumptions made in the past may be underpinned to a certain extent.This surely was resourceful during assessment of historic structures, requiring considerably more detailed numerical modelling than new structures, thus having better chances to preserve these valuable bridges for continuing use.