PSI - Issue 64

Nicola Nisticò et al. / Procedia Structural Integrity 64 (2024) 2238–2245 Author name / Structural Integrity Procedia 00 (2019) 000–000

2239

2

1. Introduction In the field of cultural heritage, the acquisition and reconstruction of three-dimensional (3D) data are crucial. Specific technologies for augmented reality perception are essential for disseminating Digital Twins to museums worldwide, including Metaverse museums managed in the cloud. Several methods and technologies for 3D digitization have been discussed by Nisticò (2024). These methods are utilized for implementing geometric models to support various purposes, including structural analyses typically based on finite or discrete numerical models, whose historical perspective is outlined by Liu et al. (2022). Within the scope of this work, three approaches will be discussed: 1) Microplane and Lattice methods, for which a comprehensive reviews have been respectively provided in (Gambarelli et al., 2016a; Fascetti et al., 2016a) and 2) Discrete element Method. Regarding Microplanes , Bažant (1984), referring to Taylor (1938), Batdorf and Budianski (1949) as pioneers, proposed to describe inelastic phenomena through a set of strains acting on planes, differently oriented, denoted as microplanes . Various strategies have been proposed in the past, as documented in Bažant and Gambarova (1984), Bažant and Prat (1988), Bažant and Ožbolt (1990), Ožbolt and Bažant (1992), Carol and Bažant (1995), Bažant et al. (1996a, b), Ožbolt and Kožar (2001). Ožbolt et al. (2011) extended microplanes to model Fiber Reinforced Polymers. The Lattice model, historically rooted in Hrennikoff's work (1941), conceptualizes a continuous material through a system of appropriately oriented bars. Subsequent advancements include the proposal by Schlangen and Van Mier (1992) of a lattice comprised of beam elements to simulate the fracture of concrete materials. Here, brittle-breaking elements were removed from the mesh upon surpassing specific stress or displacement thresholds. Zubelewicz and Bažant (1987) proposed a 2D random particle model for fracture in aggregate composites, and Bazant et al. (1990) extended it, proposing a simple procedure to randomly generate a given particle distribution. The methodology, implemented can be framed in the Distinct Element Method proposed by Cundall (1971), Cundall and Strack (1979). The software BALL (Cundall, 1978) also needs to be mentioned. Bolander and Saito (1998) proposed a random geometry to define an equivalent rigid-body-spring networks. Cusatis et al (2003a,b) proposed a 3D lattice to predict the mechanical behaviour of the mesostructured concrete. Yip et al. (2005) adopted an irregular 3D lattice to model combination of a bulk material, curvilinear reinforcement, and their interfaces. Bolander and Sukumar (2005) proposed an irregular lattice model, showing its performance when short fibers are embedded in a matrix material. Geometrical meshing is essential for describing the mechanical model. The domain can be discretized as outlined in Bolander and Sukumar (2005). Voronoi (1908) tessellation offers a method for discretizing the material based on a set of points. These points are utilized for the Delaunay (1934) triangulation, which defines the element connectivity of the lattice network. Voronoi tessellation (Aurenhammer, 1991; Amenta and Bern, 1999, Martinez et al., 2007) serves different disciplines, among which image processing. Here it is proposed for the digitalization process and consequently the mechanical analysis of the digitized object. Multiscale modelling and analysis play an important role in comprehending materials and structures and are widely applicable across different disciplines, including image recognition (Drăgut and Eisank, 2011). The integration of object digitization, geometrical and mechanical modelling, and the performance evaluation of objects can be significantly enhanced through the application of Artificial Intelligence (AI) technologies proposed for both fields (Liu et al. 2022; Janga et al, 2023). 2. Geometrical Modelling Multiscale modelling, spanning from nano- to micro- and meso-scales, find widespread in image recognition (Curtis and Woodcock, 1987; Drăgut and Eisank, 2011), contribut ing to both coarse-grained and fine-grained approaches (Fig. 1).