EXPERIMENTAL VS NUMERICAL CORRELATION FOR A SHAKING TABLE TEST OF A CONCRETE PRECAST MODEL STRUCTURE EQUIPPED WITH SLB DEVICES

Pantoja M.; Gonzales H.; Bozzo L.M.; Bozzo G.; Ramirez J. · World Conference on Earthquake Engineering proceedings (2021)

Resumen

This work presents the experimental and numerical correlation for a two degrees of freedom precast concrete structure equipped with SLB devices. The model was tested at the PUCP unidirectional shaking table in Peru. The shaking table limitations conditioned the whole module. The effectiveness of SLB dissipators increases as the bare structure is more flexible. However, the maximum weight on the shaking table is 150kN and the maximum height of the model 6m with 4.4 m width. The flexible structure is achieved by 5.3m height 22cm × 22cm section columns and a concentrated mass on top. Stiffness and ductility of the structure is achieved by uncoupled 10cm thick walls on both sides equipped with a single SLB device at the top middle section. If these two uncoupled walls had been installed from the shaking table base, the overall 150kN weight condition would have been exceeded. Therefore, the walls were placed 170cm from the base in order to verify the weight condition; however, that leads to a strong discontinuity. This discontinuity is not a problem for structures equipped with SLBs, since the maximum force transferred by the devices is limited and verified before manufacturing. A particular advantage of the SLB devices is that they are manufactured after material testing of the base steel plate so their final dimensions are adjusted. The paper describes the unidirectional tensile test in order to obtain the complete stress – strain curve for the steel plate. The “engineering curve” is converted into a “real stress – strain curved” for a FEM of the SLB device. According to nonlinear geometric and material properties and using ANSYS program the simplified hysteretic devices properties are generated for the specific material used for their manufacturing. The model set up on the shaking table consists of LVDTs as well as accelerometers on top and on the uncoupled walls. The test schedule consists of twelve phases. The first four were characterization of the structural period under free vibration and sinusoidal frequency sweep tests for the model with and without SLB devices. The subsequent 3 phases were for the 1974 Lima earthquake increasing its peak acceleration up to 0.927g. Similarly, the subsequent 3 phases were for the 1970 Lima earthquake increasing its peak acceleration up to 1.367g. Finally, the remaining two phases were done after uninstalling the devices. An important observation is that in all the tests, the devices were the same and their yielding, due to its initially high stiffness, started from phase 1. The experimental results are compared to numerical values using simple Wen’s hysteretic model and nonlinear time history analysis. The peak acceleration for phase 6 showed a 5% difference. For the same phase the top displacement difference is only 2%. The peak acceleration and displacement for phases 4 and 5 varies more than phase 6. In all cases, as the signal increases, its base acceleration the correlation is improved, most probably due to the fact that for smaller vibrations there are various uncertainties at the hinged connections such as friction coefficient and minimum rotation stiffness. © The 17th World Conference on Earthquake Engineering.