NLFEA of one-way slabs in transition between shear and punching: Recommendations for modeling

Abstract

Several studies investigated the precision of non-linear finite element analyses (NLFEA) to predict the shear capacity of beams or the punching capacity of slab-column connections. However, in the literature, there is little discussion regarding the results of NLFEA to predict the ultimate capacity and failure mechanism of slabs susceptible to different failure mechanisms. In this study, a set of tests from literature that developed different shear failure mechanisms was evaluated by NLFEA. A total of thirteen slabs were modeled, where slab’s width and shear span varied. A coupled damaged-plasticity model was employed to simulate the concrete behavior. The proposed model was calibrated to simulate specimens that failed as wide beams in one-way shear and specimens that failed by punching shear. Besides, the effect of different modeling choices was investigated: (i) the assumed stress–strain behavior in compression; (ii) tensile stress–strain behavior; (iii) the inclusion or not of damage parameters and (iv) the viscosity parameter. The results indicated that, on average, the proposed modeling strategy represented the failure mechanism and ultimate loads well. Also, the same calibrated model was found capable of representing one-way shear failure and punching shear failure or mixed modes. The sensitivity study demonstrated that the tensile stress–strain behavior and viscosity parameter influences the results of the numerical models more significantly than the assumed stress–strain behavior in compression or including/not including the damage parameters. This paper concludes that modeling strategies during the calibration process shall be checked carefully before performing any parametric analyses to identify the accuracy of the numerical models to represent the different failure mechanisms. Moreover, the validation step of the modeling strategy shall identify possible limits of the numerical model to be considered in the parametric analyses.