Shear Failure of Steel Fiber and Bar Reinforced Concrete Beams Without Stirrups : Predictions based on Nonlinear Finite Element Analyses

Sammanfattning: Shear failure in concrete beams are often brittle in nature and potentially dangerous without adequatereinforcing measures. In design of concrete, it is commonly recommended to install transversalreinforcement along the shear span to induce a more ductile structural response, improving the shearcapacity all together and providing sufficient warning prior to collapse. However, it is more frequentlybeing assessed whether analogous performance can be achieved in fiber reinforced concrete beamswithout stirrups, and multiple attempts in literature confirm that it is possible. This alternative technologyintroduces need for better understanding of the modeling aspects of FRC in numerical simulations, as it isbecoming more common for engineers to resort to the finite element method in quality assurance ofstructures.In this thesis, the possibility of predicting shear failure numerically in simply supported fiber reinforcedconcrete beams with flexural bar reinforcement but without stirrups was investigated by means ofnonlinear finite element analysis, using the software package ATENA 2D Engineering. The ultimate aimwas to, as accurately as possible by means of numerical analyses on representative FE-models, replicatethe results from physical three-point-bending tests on simply supported FRC beams of various sizesperformed by Minelli et al. (2014). These beams were merely equipped with flexural reinforcement andexhibited shear failure.This thesis revolved around development and comparative assessment of material models for FRC basedon the smeared crack approach, adopting two different strategies: (1) The first strategy was to calibratematerial parameters based on results from 3PBT on notched FRC beams that were carried out prior totesting of the reinforced FRC beams, as reported by Minelli et al. (2014). Nonlinear finite element analysiswas used on representative FE-models for the notched 3PBT specimens, from which material parameterswere obtained iteratively by employing inverse analysis methods proposed by Červenka Consulting s.r.o.(2). The second strategy comprised of utilizing recommended constitutive relations from designrecommendations in SS812310 and RILEM TC 162-TDF. All of the constructed material models werefinally coupled with the FE-models that represented the beams with flexural reinforcement for evaluationof their performance based on their consistency with experiment data.It was found that the material models that were generated from inverse analysis in general would haveyielded successful predictions for the occurrence of shear failure in the reinforced FRC beams, providedthat the governing post-cracking residual tensile parameters were processed with respect to relevantassumptions as to describe uniaxial tensile behavior. However, although it was possible to utilize theproposed calibration method to replicate the load-displacement data for the notched 3PBT specimens withsufficient conformity, it was not possible to arrive at only one unique solution. Instead, multiple outcomescould be obtained based on the initial choice for the input value of the uniaxial tensile strength, leading tothe conclusion that experience and the engineering judgment of the user is of high importance whenadopting this method.Regarding the material models that were derived from constitutive relations in design recommendations,satisfactory estimates for the shear capacity could be obtained from the FE-models that were based onrecommendations by RILEM. The models that were based on SS812310, on the other hand, demonstratedover-stiff behavior and they were unable to provide accurate graphical visualizations of characteristicshear cracking, although the obtained load bearing capacity overall matched the experiment data in caseswhen size effects seemingly had a minor influence. An important observation from the comparison ofthese material models was that the initial drop in tensile strength during crack initiation within an elementis crucial in modeling of FRC, as it accounts for a more realistic behavior through a gradual transitionfrom aggregate bridging mechanisms of PC to the added fiber bridging mechanisms of FRC. Forsituations with high residual tensile strengths in relation to tensile strength at crack initiation, theguidelines in SS812310 become less practical for predicting shear failure by means of NLFEA.