Post-test Data Analysis

Prof. Liel will be using test data from hybrid simulations to explore how well existing models capture three limit states, (i) column shear failure, (ii) column axial failure and (iii) progressive collapse resulting from column axial failure, in nonlinear analysis of real buildings. We are particularly interested in identifying characteristics of reinforced concrete frames for which vertical collapse modes may lead to progressive collapse. Emphasis will be placed on improved methods of detection and simulation of these failure modes, given variation in axial loading, transverse reinforcement, RC frame geometry/configuration and retrofit (if any). An improved understanding of these vertical collapse modes will be essential in quantifying the expected collapse performance of non-ductile reinforced concrete frame structures and in comparing that performance with retrofitted structures. A set of typical non-ductile reinforced concrete frame structures, representative of construction before significant improvements in building-codes for reinforced concrete in the U.S. and elsewhere (1975), will serve as the illustration of this assessment. Structures retrofitted with FRP will also be evaluated.

Performance-based earthquake engineering methods rely on detailed nonlinear simulation models to obtained probabilistic predictions of structural response, damage and collapse. In the case of non-ductile reinforced concrete frame structures, given recent damage and deadly collapses in China (2008), Pakistan (2005), Indonesia (2004), and other earthquake events world-wide, it is particularly important to quantify the collapse resistance of these structures as a measure of seismic safety. Performance-based assessments of collapse safety can be used to identify particularly vulnerable structures and evaluate the effectiveness of different retrofit strategies, such as FRP.

The accuracy of simulation models for predicting the seismic collapse performance of non-ductile reinforced concrete frame structures depend upon their ability to faithfully represent critical failure modes in the structure. Simulation of column shear and axial failure leading to, possibly, the subsequent progressive collapse of the structure has been a central challenge in developing seismic performance assessments. Of particular importance in understanding the behavior of non-ductile reinforced concrete frame structures is ability to predict when shear and axial failure occur in columns and, critically, whether the collapse of one column leads to progressive collapse of the structure (a life safety concern) or if the structure is able to redistribute these loads. The hybrid simulations conducted in this study will provide much needed data for investigation of when and how vertical column collapse occurs in order to develop better seismic performance predictions.