An experimental investigation into soil–structure interactions of integral bridge abutments due to cyclic translations

Abstract

Integral bridges are built without expansion joints to avoid many of the persistent issues with joint degradation and maintenance problems regularly experienced by traditional bridges. In the absence of expansion joints, the cyclic thermal displacements of the integral bridge would be transferred unmitigated by the joints to the abutment resulting in accentuated soil–structure interactions. Previous studies have mainly focussed on the rotational mode of integral abutment movement while the translational mode of movement and backfill deformations have received much less attention. This paper presents the results of laboratory tests conducted on a semiscale (1:2) physical model, which was designed using dimensional analysis to achieve field stress and strain similarity with a full-scale 1.92 m high-prototype integral abutment. The comparative effects from five tests covering low to high cyclic translation amplitude on backfill surface settlement and heave, and lateral earth pressures behind the abutment were investigated over assumed 120 seasonal cycles. Asymptotic settlement depth of the backfill was obtained in each test. Full passive failure surface was observed when the cyclic translation amplitude of the test exceeded the threshold amplitude. The findings also suggest that the settlement depth seemed to be linked to the development of passive failure surface in the following sense—the asymptotic settlement depth in those tests which developed full passive failure surface would attain the same (ultimate) asymptotic depth; conversely, when the passive failure surface had not or only partially developed in that test, the asymptotic settlement depth was dependent on the cyclic translation amplitude, but it was less than the ultimate observed for full passive failure surface. Lateral earth pressure only stabilized to an asymptotic state for the test with the lowest cyclic translation amplitude during the 120 cycles. Shakedown limit state was observed for this case. At larger cyclic translation amplitudes, the lateral earth pressure did not stabilize but continued to evolve with the elapsed cycles, and shakedown limit state was not observed for these cases.