PSI - Issue 37

Sonia Djenad et al. / Procedia Structural Integrity 37 (2022) 321–329 Djenad et al / Structural Integrity Procedia 00 (2019) 000 – 000

325 5

(2015), Ait Taleb et al, (2016)). The failure stresses measured by tensile and shear test system are given in Table 3. Composites were modelled using the Hill- Tsai failure criterion in the case of an anisotropic material.

Table 3. Composite materials properties Composite materials Thickness (mm)

Tensile stress (Mpa)

Ultimate tensile strain (‰)

Young modulus (Mpa)

GFRP fabric E-glass sheet

1.00 1.12 0.97

82 000 62 000

1400

5.6 4.9 2.2

140

Epoxy resin STR

2 975

35

3.2. FEM Approach According to the testing conditions, the geometrical model of the confined cylinders under axially compressive was established. Using the dimensions of actual specimens, a mesh sensitivity study was first performed. Indeed, a tetrahedral linear finite element model (3D), with Lagrangian formulation and four nodes with 12 degrees of freedom are used to mesh concrete. The solid elements have a dimension of approximately 1 cm, making thinly meshed elements. GFRP are meshed by quadratic finite elements model (2D), with an eight nodes solid element, with 0.5 cm dimension in the both directions. The quadratic FEM were coupled with the tetrahedral ones with no additional slip conditions i.e. no friction between the various components. With this function the concrete nodes elements are cinematically constrained to the those of solid element which it is located in. 4. Results and confrontations The finite element models proposed for the axially behaviour modelling of confined concrete are able of representing cracks initiation and cracks propagation in considered elements for different loading levels. Simulation results in terms of stress-strain response are presented and discussed. Gains and enhancement in terms of compressive strengths are also quantified. 4.1. Modeling validation In order to confirm the reliability of the simulation procedure and validate the obtained results, quantitative confrontation is focused. Indeed, FEM results were compared to those ones available within literature review for the following confinement configurations: unconfined concrete, moderate confined concrete and high rate confined concrete, as shown in Fig. 2. A good correlation between numerical and experimental results in terms of overall behavior of cylinders under compression loading is observed for all considered cases. An improvement in terms of peak and failure strength due to the confinement effect is also observed. In case of unconfined cylinder, the predicted ultimate stress is 22.69 MPa, however (Ali Ahmed and Ait tahar, 2015) experimentally obtained a stress value of around of 24 MPa for specimens with similar geometrical characteristics. On the other hand, for tested partially confined cylinders similar to 3CS1 and 5CS3B3 carried out by (Saadatmanesh et al, (1994) and Wu et al, (2006)) gives stresses around of 30 and 60 MPa respectively. Numerically these stresses are 28.38 and 62.74 MPa. The difference between the numerical values and the experimental ones is less than (<5%), which is much less than acceptable experimental dispersion. 4.2. Stress-Strain response Performance and axial behavior at ultimate limit state was numerically modelled by means of compressive loading out on confined and unconfined specimens. The imposed monotonic stresses, as a function of axial deformations until failure according to all designated cylinders is shown in Fig. 3, 4 and 5.