Since it is impossible to deny the possibility that buildings are attacked by earthquakes beyond the current design level, the estimation of the response of buildings against such extremely strong earthquake is demanded. In this paper, the behavior of middle-rise RC buildings which are not designed by dynamic analysis are studied against the earthquake up to 2 times for the design level using detailed 3-dimensional nonlinear FEM considering soil-foundation-structure interaction.
3 type of soil models (Type1, Type2 and Type3) and 6story building models are used for the analysis. These models are considered with the nonlinear effect of both soil and buildings and the soil structure interaction effect, and analyzed by large scale FEM. The input earthquake level is from 1.0 to 2.0 times of the Level 2 design earthquake. Since the calculation load of these analyses is very heavy even by the current computer system, the method to reduce the load is necessary. In this paper, equivalent nonlinear model of soil was studied.
Based on these study, following results are obtained.
1) In the model of the hard soil, as the input increased, the plasticity of the building advanced, but the plasticity of the pile did not proceed. In the soft soil model, as the input increases, the plasticity of the pile progresses but the progress of the plasticity of the building was small. Also, in the model of the second type of general ground, as the input increased, both the plasticity of the building and the pile advanced. These qualitative properties are almost the same as the previous studies. On the other hand, quantitative evaluation is necessary to apply it to real problems. Regarding Level 2 and higher input levels, it is not clear how much accuracy the existing model has. Therefore, it is necessary to compare the accuracy of these models and study to make more suitable model.
2) We compared the models A and B which changed the behaviors of the piles after the ultimate stress. As a result, model A, which loses stiffness after ultimate stress, does not necessarily have the larger response than model B. In the case where the response of model B is large, the subsequent seismic motion is inputted, and the maximum response value is generated there. It is considered that the difference of the vibration characteristics of the coupled system in that state is the cause of the difference of the maximum response value.
3) We investigated the equivalent linearization of the ground for the purpose of reducing the analysis load. In the general equivalent linear model based on SHAKE, the response accuracy decreased when the level exceeding the application limit of shear strain, especially in the soft soil. We proposed a method that uses the equivalent linear soil model based on the response of columnar model considering the nonlinearity of the ground, and explained the efficiency of the model.
