PSI - Issue 79

Mansi Gupta et al. / Procedia Structural Integrity 79 (2026) 259–265

263

duplicated with same coordinates as original. In the second stage, first the elements next to each other are identified and zero-thickness 2D cohesive (COH2D4) elements are inserted to enable discontinuities between the solid elements for crack propagation. Separate sections for mortar-mortar cohesive elements and aggregate-mortar cohesive elements are made, so that corresponding material properties can be easily assigned later on in Abaqus. Fig. 3 shows finite element models for various components of concrete mesostructure.

Fig. 3. FE of (a) aggregate, (b) mortar; cohesive elements in (c) ITZ, (d) mortar matrix After generating concrete mesostructured, the remaining homogeneous parts of three-point bend beams are generated, assembled and coupled to mesostructure using tie constraints as shown in Fig. 4. The mesh size in homogeneous region is varied from 1 mm near the assembly to 15 mm towards the end and the displacement controlled loading of 0.2 mm is provided at the top (Trawi ń ski et al. 2016).

Fig. 4. Model assembly of the simply supported beam 3. Results and discussion

The models are run in Abaqus/CAE and the simulation results provide crucial insights on the effect of aggregate shape and size on the fracture behaviour of geometrically similar concrete beams. Also, the accuracy of the simulation results was verified by extracting the load vs crack mouth opening displacement (CMOD) data for the geometrically similar beams and comparing with the experimental results of Bhowmik and Ray (2019). Fig. 5 shows the comparison of load-CMOD plots and it can be seen that the mesoscale simulations are able to capture the experimentally observed trend in concrete. The mesoscale simulations of concrete capture the intricate fracture behavior of geometrically similar beams with different aggregate shapes and sizes. It was seen from the studies on aggregate shape that models with circular or elliptical aggregates exhibit higher load carrying capacity than irregular polygonal aggregates. The differences in peaks can be attributed to the increased length of aggregate-mortar interface for the polygonal geometry, which introduces more cohesive elements and a larger ITZ length. The increase in ITZ length increases the potential crack sites, thereby leading to early failure. Also, the crack path in circular and elliptical geometries was slightly