Abstract

For certain important structural systems such as nuclear reactor facilities, dynamic response analysis would be required in the future considering multi-component effects of ground motion input. In a case of simulation of three-dimensional ground motions by use of stochastic processes, a question immediately arises "Should the components of motion be cross correlated in a statistical sense?". If correlated, in addition to properties which establish the respective components, one must specify cross-correlations for the components. An orthogonal set of principal axes is defined for three-dimensional ground motion processes. These principal axes are determined to have maximum, minimum and intermediate values of variance and have zero values of covariance. This property reveals that the corresponding components of motion along the principal axes are fully uncorrelated with one another. Since real earthquake accelerograms are well represented by Gaussian random processes, the three components of motion along a set of the principal axes are statistically independent of one another. Using the concept of principal axes and applying a moving-window technique to the accelerograms recorded during the San Fernando earthquake of February 9, 1971, time-dependent characteristics of three-dimensional ground motion along a set of principal axes are determined. Results of the analysis indicate a significant correlation between horizontal directions of the major and intermediate principal axes and directions from the recording stations to the fault slip zone. In view of the existance of principal axes, it is concluded that the three components of motion can be generated in a stochastic manner being independent of one another, provided that a set of axes along which they are directed is treated as a set of principal axes.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call