Better understanding of the soil-structure interaction of integral bridge abutments is clearly desired when there has been an increasing number of integral bridges being built in recent years. Daily and seasonal changes in the length of the bridge deck generate cyclic motion to the integral abutments and the soil behind the abutment wall. It is very difficult to predict any increase in the earth pressure behind and accumulative movement of the abutment due to the soil densification during cyclic motion. Existing field measurements together with centrifuge tests have proven to be insufficient to fully understand the overall mechanisms involved....[ Read more ]

Better understanding of the soil-structure interaction of integral bridge abutments is clearly desired when there has been an increasing number of integral bridges being built in recent years. Daily and seasonal changes in the length of the bridge deck generate cyclic motion to the integral abutments and the soil behind the abutment wall. It is very difficult to predict any increase in the earth pressure behind and accumulative movement of the abutment due to the soil densification during cyclic motion. Existing field measurements together with centrifuge tests have proven to be insufficient to fully understand the overall mechanisms involved.

Previous results of centrifuge modelling of a spread-base integral bridge abutment were interpreted with internal consistency check. Finite difference analyses of the abutment were performed using the Mohr-Coulomb model, and the UBCSAND model (Byrne & Puebla 1996) that is able to reasonably predicting both the volumetric and stress changes of sand under the strain-controlled cyclic triaxial loading conditions. Four empirical parameters determined from laboratory tests were added to the formulations of the UBCSAND model at two different imposed cyclic strains. The effects of the height of fill on the excavated side, the flexural stiffness of the abutment and an 'L-shaped' wall were also studied parametrically.

The interpreted interaction behaviour in the centrifuge tests with a 'smooth' wall is consistent with the published results from abutments with a 'rough' wall (Ng et al, 1998a&b). The differences between the sliding and ratchetting displacement of integral abutments are distinguished. Interpretations of the effects of wall friction and soil density using the centrifuge data show that the soil density dominates the response of the abutment. Also, the numerically simulated responses are found qualitatively consistent with the centrifuge results in terms of the stress increases, surface settlement, wall movements, bending moment and deck force. The influence of the excavated height on the interaction response is the most pronounced, while the difference between an 'inverted-T' wall and 'L-shaped' wall is insignificant. Further work is recommended mainly to improve the constitutive relations of the sand model under cyclic loading and to incorporate interface elements in the numerical simulations.