In general notice, the horizontal component of ground motions will normally always play the dominant role for response of buildings compared to the response from the vertical component. Therefore, the vast literature is well written for modeling of buildings to simulate the effect of horizontal components of ground motions. However, with the relatively recent recognition that the vertical component of ground motion can exceed its horizontal counterpart, there is a renewed interest in vertical ground motions and their impact on buildings.
In practical earthquake engineering, modeling of buildings to simulate effects of ground motions are based on simplified methods such as the lumped mass approach and Bernoulli beam elements. The objective in this thesis has been to maintain the simplicity in these assumptions and generalize them to determine whether a simplified model is suitable to simulate the effect of vertical motion. To evaluate the accuracy of simplified models, they are compared to an exact model which includes extremely refined element mesh.
The eigenvalue analysis has to a great extent been dominant for the investigation. Comparing both natural horizontal and vertical mode shapes and periods for different models, has been essential to determine which simplified model with least amount of computational effort can simulate realistic vertical motion. It is the Author's belief that studies in this thesis show that simplified models can be used to simulate vertical motion. Nevertheless, the common modeling assumption in earthquake pratice cannot be used to simulate realistic vertical motion. This applies especially to the rigid diaphragm assumption when modeling the slab.
Furthermore, response parameters from time history analysis of a suite of ground motions, shows that simplified models can simulate the effect of vertical ground motion with reliable accuracy compared to the exact solution.