Design and validation of a shaking-table test model on a long-span cable-stayed bridge with inverted-Y-shaped towers

Abstract

Existing studies on the seismic responses of long-span cable-stayed bridges under strong earthquake excitations have mostly been performed through numerical methods. In this work, a shaking-table experiment was conducted based on a long-span cable-stayed bridge with a main span of 1088 m and typical inverted-Y-shaped towers. The test model had a total length of 59.65 m and a height of 9.1 m. Micro-concrete was used for the model to meet the payload limit of the shaking tables. Pretest numerical analyses were carried out and the results showed that both the flexural and compressive/tensile properties of the column elements had effects on the seismic responses of the inverted-Y-shaped towers. Thus, the frequently used scaling strategy for conventional columns, with only strictly scaled sectional flexural stiffness and bending moment capacity, is not suitable for structures with stable triangle frames. The steel girder was designed to have properly scaled flexural stiffness, and the decreased strain responses in turn allowed the simplification of the cable system. A total of 28 pairs of cables were used for the bridge model compared to 136 pairs of cables for the prototype bridge, and 42 and 50 additional mass blocks were utilized for the girder and for each of the model towers, respectively. The errors induced by the simplification of the cable system and the additional lumped mass were no greater than 8% when compared with the theoretical mass distribution. Based on the white-noise tests, the identified natural modes of the scaled bridge model are consistent with those extracted from the numerical simulation. The observed failure mode matched well with the numerical prediction, and comparative responses between the numerical simulation and experimental test were shown. These results validated the design of the scaled bridge model.