NONLINEAR ANALYSIS OF SHEAR CRITICAL RC MEMBERS USING CURRENT FE SOFTWARE

The shear behaviour of concrete has been widely studied in the field of structural engineering, yet it is still considered an unresolved issue. In one hand, the design codes provide diverse methods to estimate the shear capacity of concrete with different results. On the other hand, analytical models to assess the performance of shear critical concrete members have been proposed by several authors. These models have been developed using different assumptions. Some of the theories have been implemented in finite element software packages that are used by engineers in academia and in the industry. This study presents an investigation of the state-of-the-art software packages for the analysis of shear critical concrete elements. The theory implemented on each software is detailed and analysed. To evaluate the capacities of the packages, some benchmark tests have been used for both monotonic and cyclic loading cases. The presented results show the capabilities and limitations of the programs from the experience obtained during the duration of this work. Just as well, the results among the programs are compared among themselves as well as the results from other authors. 4897 Available online at www.eccomasproceedia.org Eccomas Proceedia COMPDYN (2017) 4897-4913 © 2017 The Authors. Published by Eccomas Proceedia. Peer-review under responsibility of the organizing committee of COMPDYN 2017. doi: 10.7712/120117.5770.18864 P. Ceresa and A. González Víquez 1 BACKGROUND AND MOTIVATION Even though the mechanics of shear behaviour in reinforced concrete (RC) has been studied for a long time, it is still an unsolved problem. On one hand, many design codes provide different methods to define and estimate the shear capacity of concrete. Among these codes there are the National Building Code of Canada by the CSA [1], AASHTO LRFD Bridge Specifications by AASHTO [2], and the Fib Model Code by the International Federation for Structural Concrete [3] and ACI 318-11 by the ACI Committee 318 [4]. These codes and their methods of design generally lead to very different shear capacity estimations for RC elements (and therefore, different amount of required transverse reinforcement). On the other hand, several analytical models have been proposed in order to assess the performance of shear critical members. Authors such as Litton [5], Bazant and Oh [6], Cervenka [7], Vecchio and Collins [8], De Borst [9], Hsu [10] and Maekawa et al. [11], have developed methods to determine the shear behaviour of RC, using diverse types of assumptions in the process. As defined in Ceresa et al. [12] [13], in the framework of smeared crack theory for RC the following main models can be identified: • Compression Field Theory (see Collins [14], Vecchio and Collins [15]) • Modified Compression Field Theory, MCFT (see Vecchio and Collins [16]) • Rotating-Angle Softened Truss Model, RA-STM (see Belardi and Hsu [17], Belardi and Hsu [18], Pang and Hsu [19]) • Fixed-Angle Softened Truss Model, FA-STM (see Pang and Hsu [20], Hsu and Zhang [21], Zhang and Hsu [22]) • Cracked membrane model (see Kaufmann and Marti [23]) • Softened membrane model (see Zhu [24], Hsu and Zhu [25]) • Disturbed Stress Field Model, DSFM (see Vecchio [26], Vecchio and Lai [27]). Some of these analytical models have been used in the development of commercial finite element (FE) software packages used by researchers and engineers in both industry and academia. Yet the complete range and capabilities of these programs is not completely investigated. Therefore, the aim of this work is to provide a thorough review of the theoretical models implemented in state-of-the-art FE packages for the nonlinear analysis of shear critical RC members. Similarly, verification examples are provided for the assessment of the accuracy, performance, advantages and limitations for the studied programs. This is performed for both the monotonic and the cyclic behaviour of reinforced concrete. 2 BRIEF OVERVIEW OF THE ANALYZED SOFTWARE Three are the software packages taken into account in this study: VecTor2 [28], DIANA [29] and ATENA[30]. The detailed review of the theory implemented in each software can be found elsewhere [31]. VecTor2 is a nonlinear finite element program for the analysis of two-dimensional RC member structures subjected to in-plane normal and shear stresses. It models cracked concrete as an orthotropic material with smeared, rotating cracks. It considers the element to act as in the plane stress idealization. The program uses an incremental total load, iterative secant stiffness algorithm to produce the nonlinear solution. The two main bases of the modelling of RC are the MCFT and the DSFM. The latter two are then complemented in VecTor2 with other models for compressive and tensile behaviour of concrete, tension softening and tension stiffening. The main aspects that greatly affect the response of RC elements are the crack width check, the crack slip calculation and the models for hysteretic response of concrete and reinforcement steel, using models that consider the damage due to load reversal, repetitions and plastic behaviour.