PSI - Issue 37

Khalil Naciri et al. / Procedia Structural Integrity 37 (2022) 469–476 Khalil Naciri et al. / Structural Integrity Procedia 00 (2019) 000 – 000

474

6

Table 5. Homogenized masonry inelastic behavior. Compressive behavior

Tensile behavior Yield stress (MPa)

Yield stress (MPa)

Inelastic strain

Inelastic displacement (mm)

2.89 2.70 2.41 2.28

0.00E+00 3.66E-04 6.00E-04 1.93E-03

0.313 0.248 0.144 0.002

0.00 0.43 0.58 1.33

5. Bending test simulation Two-dimensional simulation of the already presented experimental bending test was conducted on ABAQUS software. Two masonry arches were modeled and tested under the same experimental bending loads, the first using DMM and the second using MM. Equivalent characteristics derived in the previous subsection were assigned to the homogenized masonry arch. Experimental and numerical curves of the total vertical load are depicted in Fig. 5 against the vertical displacement applied at the load position 2. As can be seen, the DMM and the MSM reproduce correctly the experimental curve. The DMM curve is more close to the experimental curve in the first stage of the plastic deformation (between 10 and 20 mm). Numerical strengths are also in agreement with the experimental strength with slight errors of +1.27% for the DMM and -2.20% for MSM compared to the experimental value. The arch final deformed shape obtained by the DMM, amplified 2 times, is shown in Fig. 6. Compression and tensile cracks are described by the variables PEEQ and PEEQT, which are scalar measures computed from plastic strain components, where a value greater than 0 indicates cracks formation. As can be seen, compression and tension cracks are clearly visible at the load position 2. Tensile cracks are also located near the left support, which is in agreement with the experimental results. However, the other two hinges necessary for the arch collapse were observed in the external surface near the load position 1 and 3, which was not reported in the experimental work. This may be caused by the way the loads at positions 1, 3 and 4 were applied. Indeed, during the experimental test, it was intended that these loads remain equal. However, due to the arch displacements, these forces increased. Numerically, it was not possible to simulate this arbitrary variation and, for simplicity, a progressive linear loading was adopted at these position until reaching the experimental values.

Fig. 5. Experimental, DMM, and MM load displacement curves.