The capability of representing fracture processes in non-homogeneous media is of great interest among the scientific community for at least two reasons: the first one stems from the fact that the use of composite materials is ubiquitous within structural applications, since the advantages of the constituents can be exploited to improve material performance; the second consists of the need to assess the non-linear post-peak behavior of such structures to properly determine margins of safety with respect to strong excitations (e.g. earthquakes, blast or impact loadings). Different kinds of theories and methodologies have been developed in the last century in order to model such phenomena, starting from linear elastic equivalent methods, then moving to plastic theories and fracture mechanics. Among the different modeling techniques available, in recent years lattice models have established themselves as a powerful tool for simulating failure modes and crack paths in heterogeneous materials. The basic idea dates back to the pioneeristic work of Hrennikoff: a continuum medium can be modeled through the interaction of unidimensional elements (e.g. springs or beams) spatially arranged in different ways. The set of nodes that interconnect the elements can be regularly or irregularly placed inside the domain, leading to regular or random lattices. It has been shown that lattices with regular geometry can strongly bias the direction of cracking, leading to incorrect results. A variety of lattice models have been developed. Such models have seen a wide field of applications, ranging from aerodynamics (using Lattice-Boltzman models) to heat transfer, crystallography and many others. Every material used in civil and infrastructure engineering is constituted of different phases. This is due to the fact that the different features of different elements are usually coupled in order to obtain greater advantages with respect to the original constituents. Even structural steel, which is usually thought of as a homogeneous continuum-type medium, includes carbon particles that can be seen as inhomogeneities at the microscopic level. The mechanical behavior of concrete, which is the main object of the present work, is strongly affected not only by the presence of inclusions (i.e. the aggregates pieces) but also by their arrangement. For this reason, the explicit, statistical representation of their presence is of great interest in the simulations of concrete behavior. Lattice models can directly account for the presence of different phases, and so are advantageous from this perspective. The definition of such models, their implementation in a computer program, together with validation on laboratory tests will be presented. The present work will briefly review the state of the art and the basic principles of these models, starting from the geometrical and computing tools needed to build the simulations. The implementation of this technique in the Matlab environment will be presented, highlighting the theoretical background. The numerical results will be validated based on two complementary experimental campaigns,which focused on the meso- and macro-scales of concrete. Whereas the aim of this work is the representation of the quasi-brittle fracture processes in cementitious materials such as concrete, the discussed approach is general, and therefore valid for the representation of damage and crack growth in a variety of different materials.
Random lattice particle modeling of fracture processes in cementitious materials / Fascetti, Alessandro. - (2016 Feb 26).
|Titolo:||Random lattice particle modeling of fracture processes in cementitious materials|
|Data di discussione:||26-feb-2016|
|Appartiene alla tipologia:||07a Tesi di Dottorato|