PSI - Issue 80

Riccardo Giacometti et al. / Procedia Structural Integrity 80 (2026) 219–231 R. Giacometti, N. Grillanda, V. Mallardo / Structural Integrity Procedia 00 (2023) 000–000

224 6

(a)

(b)

Fig. 3: Analysis of dry-joint masonry panel under horizontal load proportional to self-weight: collapse mechanisms obtained via (a) rigid block limit analysis (load factor 0.3190) and (b) homogenized rigid plastic limit analysis (load factor 0.2587). In (b), normalized principal plastic strain rates are depicted via color map.

• The list of elements requiring refinement can be broadened by including the neighbouring elements. For each element refined according to the previous point, a neighbourhood region proportional to the element area is defined. All the elements located within such a region are refined through subdivision into four new elements with equal area. This step can be useful to avoid too localized mesh refinement. • Non-refined elements adjacent to the refined ones are detected. To avoid hanging nodes and maintain a con forming mesh, these elements must be subdivided accordingly. • On the new mesh, another LP problem (2) is defined and solved. A new load factor and set of plastic strain rates are then obtained. • Restart from the first step until convergence of the load factors is found. As a numerical example, a dry-joint masonry panel subjected to horizontal load proportional to self-weight is analysed here. The masonry panel was previously investigated via experimental tests during the student competition held at the Politecnico di Milano in 2018, as described in Grillanda et al. (2021). The test was conducted via tilting table: the scope was the determination of the horizontal load bearing capacity starting from the collapse angle observed during the tilting tests. It is quite easy to demonstrate that the tangent of the collapse angle correspond to the load factor with reference to a horizontal load distribution equal to the self-weight. Here, one of the panels is analysed via homogenized rigid plastic limit analysis and the outcome is compared with the one obtained via rigid blocks analysis. In Fig. 3 the obtained results are shown. Note that the homogenized model comprises some rigid blocks repre senting the lintel and the bricks above the openings: such blocks are needed to avoid structural instabilities above the openings, since the adopted homogenized material is implicitly equivalent to a texture composed of infinitely small bricks (a detailed discussion on this aspect is reported in Grillanda and Mallardo (2025)). Frictional interfaces have been applied between such rigid blocks and the homogenized material, allowing for velocity jumps along blocks’ edges only. The obtained load factors are equal to 0.3190 and 0.2587 for rigid blocks and homogenized rigid-plastic limit analysis respectiely. The lower load factor from the rigid-plastic approach is justified since the RVE is supposed to have negligible size with respect to the overall structure. A good agreement is found in the shape of the mechanisms, with a collapse given by the overturning of the three masonry pillars. Note that the map of the principal plastic strain rates provides an accurate representation of the cracks distribution during the collapse. 2.4. Numerical example