Abstract
It is important to ensure that deep underground structures can withstand the potential strongest earthquakes throughout their design life. To achieve this goal, in the seismic design of such structures, it is necessary to select the real ground motions that put the underground structures in the most perilous state. That is, to identify the severest design ground motions. Based on the author’s previous research, some of the severest design ground motions were obtained from the Pacific Earthquake Engineering Research Center (PEER) strong ground motion database through the method of constructing composite intensity measure. However, the accuracy of the severest design ground motions has only been validated through numerical analysis up to now. As the existing limitations of the numerical method, for example, the idealized constitutive model cannot accurately describe the complex nonlinear dynamic mechanism of soil. Therefore, a large-scale shaking table test was conducted in this paper. The selected severest design ground motions, along with El Centro and Taft ground motions, were used as inputs for the shaking table test to validate. The validation results indicate that the severest design ground motions identified through the previous proposed method are significantly more dangerous than El Centro and Taft, according to the indicators of story drift, shear, bending moment, and axial force of key components of underground structures. Therefore, this study has provided the first validation of the reasonableness of selecting the severest design ground motions for underground structures from the perspective of shaking table tests.