SEISMIC BEHAVIOR AND VULNERABILITY OF INDIAN RC FRAME BUILDINGS WITH URM INFILLS

Abstract

Reinforced Concrete (RC) framed buildings with Un-Reinforced Masonry (URM) infills are the most popular structural systems for multistory buildings in India and many other parts of the world. These buildings have shown poor performance during past earthquakes and suffered severe damage or collapse, even under moderate earthquakes. It is general practice to ignore the infills in design as the uncertainty in infill-frame interaction results in complex modes of failure, rendering the simulation of seismic behavior of infilled frames a challenging task. The behavior is further affected by the construction sequence of infilled frames, as the infills are usually added after completion of the frame, and it results in a gap between the infill and soffit of the beam above. Despite significant research effort dedicated to such buildings in the past decades, the understanding of seismic behavior of infilled frames is still not adequate and guidelines for modeling and analysis are lacking in the design codes. In this Thesis, a stock of available earthquake damage survey reports, experimental studies, analytical models and design codes, is taken to identify various failure modes of such buildings. A review of available models for estimating the strength of infills and frame members in various failure modes is also presented with an objective to identify the most reliable models. A statistical exercise is carried out to select a representative building plan, based on the analysis of 50 buildings selected from a field survey, conducted earlier (DEQ 2009). The effect of infills and their construction sequence, on the seismic performance of generic RC buildings with the selected plan and designed as per relevant Indian Standards is studied. A comparative study of the available 1-, 2-, and 3-strut models is performed to examine their capability to predict the failure modes observed in post earthquake damage surveys and experimental studies. It is found that the single eccentric diagonal strut model has higher reliability in predicting the observed failure modes. A macro-model is proposed to simulate the seismic behavior of URM infilled frames, which can be easily implemented on available software. The infills are modeled as eccentric diagonal struts connected to columns, with stiffness as defined in ASCE-41 (2007) and strength in various modes of failure is considered. Nonlinear ‘Gap’ iv elements are used to simulate the gap between the infill and the beam and to release the struts in tension. Sequential analysis is performed to take into account the construction sequence of infill panels relative to frames. It is observed that conventional simultaneous analysis ignoring the construction sequence may be highly erroneous and can almost nullify the effect of infills in some cases. The complex behavior of infilled frames under lateral loading gets further complicated when infills are placed irregularly in plan and/or elevation to maximize the usage of available space. Mixed occupancy, lack of enforcement, and inadequate guidelines have resulted in a huge stock of seismically deficient irregular buildings throughout India. A survey (DEQ 2009) of multistorey buildings in the New Okhla Industrial Development Authority (NOIDA), a model township in the National Capital Region (NCR) in India, revealed that 95% of the surveyed buildings have ground storey open (without partitions) for parking and as many as 62% buildings suffer torsion irregularity. Taking a note of the widespread failure of buildings having irregularly placed infills, the seismic behavior and vulnerability of the three most common configurations of RC frames with irregular placement of infills, viz. open ground storey, front bay open in the ground storey, and three bays open in the ground storey, are studied. Nonlinear Static Pushover analysis is performed to compare the seismic performance of Indian code designed RC frame buildings with and without URM infills. It is observed that the buildings, designed for gravity loads only, as per Indian Standards can generally survive seismic excitation up to MCE of Indian Seismic Zone IV (PGA = 0.24g), without collapse. However, inclusion of infills deteriorates the performance of RC frames significantly and the collapse of the gravity load designed infilled frame buildings is caused by brittle shear failure of column. It is observed that the open ground storey buildings designed for code (i.e. ground storey columns and beams designed for 2.5 times the normal base shear) is able to attain the stiffness and strength close to those of the corresponding uniformly infilled frame buildings and the estimated performance of such buildings is slightly better than the uniformly infilled frame buildings, indicating the adequacy of the code provisions for open ground storey buildings. To study the seismic performance of buildings with asymmetric placement of infills in the ground storey, Incremental Dynamic Analysis is carried out using bi-directional v ground motions with a wide range of source and site parameters. The static capacity spectra are convoluted with the demand spectra to compare these with the dynamic capacity spectra. The convoluted static and dynamic capacity spectra match well and the ultimate drift ratio obtained from the two curves remain mostly the same, indicating that the strength and displacement parameters derived from static pushover curves, are comparable with IDA results. Fragility curves of the studied buildings are developed using the different variabilities defined in HAZUS. Capacity of structure against “Incipient Collapse” damage state and variability in seismic demand are determined from the results of Incremental Dynamic Analysis. The comparative study of the fragility curves and Damage Probability Matrices suggest that URM infills result in significant increase in seismic vulnerability of RC frames. The vulnerability of infilled frames is further increased due to irregular placement of infills. A comparative study of risk assessment methodologies based on macroseismic intensity and response spectrum approaches is also carried out. To facilitate the comparative study, a spreadsheet-based software tool ‘SeisVARA’ is developed using the vulnerability functions for Indian RC frames buildings with and without URM infills, developed in this Thesis. In this open software tool, seismic hazard can be specified either in terms of macroseismic intensity, or Peak Ground Acceleration in combination with the spectral shapes and soil amplification models of various earthquake design codes, or in terms of inelastic response spectra using the ‘Next Generation Attenuation Relationships’. A comparison of these different approaches is conducted for a typical city in northern India. Effect of different parameters, e.g. level of PGA, spectral shape, source-site parameters, and soil amplification models, on estimated loss, is also studied. It is observed that not only the different approaches result in widely varying damage and loss estimates, but also the variation of parameters within a given approach can result in considerable differences.