EVALUATION OF DEEP BEAM BEHAVIOR USING NUMERICAL MODEL WITH FINITE ELEMENT METHOD

• Main Author: Tjahyadi Tjandra Lukas, Ferdy
• Format: Final Project
• Language: Indonesia
• Subjects: Teknik sipil
• Online Access: https://digilib.itb.ac.id/gdl/view/48058
• Institution: Institut Teknologi Bandung

Reinforced conrete is commonly used for building construction. The structure is designed to withstand axial forces, shear forces, and flexure. Probability of structural failure occurs due to those external loads. In general, structural failure is divided into two types, that are shear failure and flexural failure. Shear failure is brittle and happened abruptly. Meanwhile flexural failure can be identified by the deflection and cracks that are visible before the failure. Structure component with brittle failure is dangerous due to its failure that happens without any warning, so that further study about it is required. Deep beam is an example of structure component that is vulnerable to shear failure. Finite element method is used in the research and modelling is done using ABAQUS 2017 software. Finite element analysis is used because it requires much lower cost and finite element analysis is able to identify the behavior of an object in every point, so that it is also able to identify the redistribution of stress in the object. Deep beam is modeled as two dimensional object because shear deformation is considered and existence of disturbed region, but its out of plane behavior can be neglected and can be assumed that its behavior along the thickness is uniform. Research begins with mesh sensitivity analysis, then continued with parametric study of material property, verification of numerical model, dan parametric study of deep beam. Parametric study of material property were carried on dilation angle and viscosity parameter. Parametric study of deep beam were carried on usage of bearing plate, column stub, and diagonal reinforcement in the model. Meanwhile verification of numerical model were carried on analytical calculation and experimental result. Parametric study of material properties is required before the model is verified. Mesh sensitivity analysis is done by variating the mesh size. Mesh size of 50 mm gives ultimate load that is close enough to the experimental result and shortest computational time among the mesh size tried in the research. In terms of stiffness, the numerical model is 6 times higher than the experimental sample due to variability of material in the experiment. Mesh size of 50 mm is used in parametric study of material properties. Due to significant difference in term of object’s stiffness, only ultimate load is considered in the parametric study. The chosen parameter is 28° for dilation angle and 0.0003 for viscosity parameter. Ratio of analytical ultimate load to numerical model are more than 1 for every model, showing that it is conservative. Model, which analytical calulation is done with geometric discontinuity, gives a closer result to numerical model than ones using force discontinuity. Experimental results shows higher ultimate load than numerical model, except for sample with diagonal reinforcement. Model with column stub shows higher capacity than ones with bearing plate, with ultimate load 4.24% higher and dissipated energy 25.17% higher. Model with diagonal reinforcement shows higher capacity than ones without, with ultimate load 35.95% higher and dissipated energy 54.84% higher. Those results shows that the usage of diagonal reinforcement gives deep beams a significantly higher ductility. All samples shows the occurence of stress redistribution after the top strut reached its peak stress. Stress redistribution from the top to bottom strut in the models without diagonal reinforcement occured by the formation of bottle shaped diagonal strut with stress spreading in the bottle shape. Failure in these models occurs when the bottom strut’s stress reached as high as the top strut. Meanwhile, stress redistribution in the model with diagonal reinforcement occurs through diagonal reinforcement. Concrete contributed by forming diagonal strut, but with stress higher only in area passed by the diagonal reinforcement. The occurence of stress redistribution shows that ductility of samples can be considered in analytical calculation, such as strut and tie method.