Systems of unattached, or freestanding, structures are highly vulnerable to sliding and/or overturning when subject to seismic excitation. This class of structures includes many rigid and heavy systems such as classical multi-drum columns, statue-pedestal systems, and various building contents. Damage to these structures can result in loss of irreplaceable heritage, limited functionality of critical facilities, and even loss of life. As a result, accurate predictions of their response are essential. However, traditional analytical methods cannot be feasibly extended to account for the multiple modes engaged during the excitation, the three dimensional nature of the body, and (if present) additional blocks in a system. All of which may be necessary to model realistic structures.

In an effort to develop robust predictive tools for freestanding structural systems, three research objectives are addressed in this dissertation. The first is to understand the general geometric and mass characteristics of freestanding structural systems. This was accomplished through a field survey targeting the geometrically complex statue-pedestal system. The geometry of these structures was captured using light detection and ranging scanning and computer vision techniques. The resulting range of geometry guided the design of two phases of shake table testing addressing the second objective, which is to generate a comprehensive database of the experimental response of freestanding structural systems. In these tests, a tower structure, which was adaptable to a variety of geometries and mass configurations, was subject to seismic excitation in unanchored single-body and dual-body arrangements. Test results emphasize the multi-modal and three-dimensional nature of the seismic response of these systems.

The final objective is to develop and validate a numerical model for predicting the seismic response of freestanding structures. This model was developed utilizing the contact-impact capabilities within the multi-physics solver, LS-DYNA. An analytical extension was developed in conjunction with the numerical model accounting for warped interfaces, an aspect that was revealed to have a pronounced effect on the response. The numerical model was validated against the experimental results, with particular attention given to capturing the multi-modal and multi-body behavior. Numerical simulations indicate that this model can be used to predict the dynamic response of freestanding structures with high fidelity.