Seismic behavior of post-tensioned gravity dams:: Shake table experiments and numerical simulations

Displacement controlled shear tests on small concrete lift joint specimens with different surface roughnesses reinforced with an unbonded post-tension cable are conducted to study a related joint constitutive model, and to determine the failure mechanisms associated to it. To investigate the applicability of the joint model, shake table tests are conducted on a 3.4-m high plain concrete post-tensioned gravity dam that has a cold lift joint about midheight. Numerical simulations using the joint constitutive model are compared with the experimental results. From the displacement controlled shear tests, it is shown that the post-tension cable provides additional strength to the joint by applying an additional normal load on the joint, and by the shear resistance that can be mobilized (cable dowel action). The tests performed on rough surface joints have demonstrated that the dilatancy phenomenon due to asperities increases the post-tension force. Shake table tests on the 3.4-m high dam model have shown that post-tensioning largely reduces residual sliding displacements of the joint. This reduction is a function of the frequency content of the base excitation. However, a single post-tension cable placed near the upstream face increases the upstream rocking response of the upper block, which is also a function of the frequency content of the excitation. Strong seismic excitations can induce cable failure at a lower value than its uniaxial tensile strength when significant shear displacements take place. To describe this brittle failure mechanism, a cable failure criterion that considers the mobilized shear and tensile strengths was proposed and was found to be adequate. Moreover, the post-tensioned joint model proposed showed good agreement with the static and dynamic tests performed to evaluate the joint shear force.