Preserving the past with modern techniques: Cloud2FEM and Artificial Intelligence for accurate meshing and analysis of historic structures

The cultural heritage structures (CHs) complexity represents a challenging task for the numerical model generation. Anyway, the importance of cultural heritage protection is paramount and needs to be aided by the most suitable technologies at disposal. The baseline of this process is underlined by the massive use of point clouds generation, widely implemented in the cultural heritage field for restoration and research purposes. A solution developed to directly obtain a 3D mesh for Finite Element Method (FEM) analysis, without the use of 3D CAD-based software, is provided by Cloud2FEM. Cloud2FEM is a Python-based libraries software that produces a voxalized mesh of an uploaded points cloud. Once the voxels are obtained, they are converted in mesh ready to be used in FEM analysis software. Two of the main drawbacks of this procedure are: presence of curved shapes and Python coding implementation. Geometries different from the regular ones such as curved shapes, arches, vaults, etc., are affected by a surface faceting, decreasing the reliability of the analysis. These solutions can be set up by Python encodings and analysed with a mesh generation. The aim of this thesis is to generate and analyse a mesh obtained with Cloud2FEM, proposing solutions for the main drawbacks presented before with the supporting of Artificial Intelligence (AI). This is provided by the model case-study of UNESCO world heritage Mausoleum of Theodoric. Furthermore, to speed up the process and develop familiarity with the Python coding implementation task, the strongly contribution of AI is adopted. AI will be the cornerstone for the whole process providing an innovative method for accurate meshing and analysis of historic structures.