One hundred and forty seven pairs of ground motion records for earthquakes with moment magnitude greater than 6.5 and site-to-source distance smaller than 15 km were selected from the NGA strong motion dataset to study the relationship between maximum and geometric mean demands. In the near-fault region, the geometric mean spectral demand can be much smaller than the maximum spectral demand for a pair of ground motions. The Next Generation Attenuation (NGA) models for shallow crustal earthquakes in the Western United States, like most other ground motion attenuation relationships, predict the geometric mean of horizontal spectral demand. We recommend adjustments to make design ground motions compatible with target risk levels. Because this is unlikely, design ground motions have lower probability of occurrence than intended, with significant societal costs. The use of MD ground motions effectively assumes that the azimuth of maximum ground motion coincides with the directions of principal structural response.
In order to achieve structural designs consistent with the collapse risk level given in the NEHRP documents, we argue that design spectra should be compatible with expected levels of ground motion along those principal response axes. This assumption may be true for some in-plan symmetric structures, however, the response of most structures is dominated by modes of vibration along specific axes (e.g., longitudinal and transverse axes in a building), and often the dynamic properties (especially stiffness) along those axes are distinct.
These maximum-direction (MD) ground motions operate under the assumption that the dynamic properties of the structure (e.g., stiffness, strength) are identical in all directions. The 2009 NEHRP Provisions modified the definition of horizontal ground motion from the geometric mean of spectral accelerations for two components to the peak response of a single lumped mass oscillator regardless of direction.
0 kommentar(er)
