Scipedia: Presentations and videos to 17th International Conference on Computational Plasticity (COMPLAS 2023)

Additive Manufacturing: some dreams, some nightmares

Jesús Sánchez Pinedo — Tue, 12 Dec 2023 10:39:03 +0100

Laser-based Powder Bed Fusion of Metals (PBF-LB/M) is an additive manufacturing technology suitable for producing metal components with complex geometries and remarkable mechanical properties and performances. However, a widespread adoption of this technology in many industrial context is yet hindered due to the high stochasticity of the process. In fact, the complex process-structure-property relationships occurring in PBF-LB/M are today not yet fully understood. Therefore, suitable physical and numerical models need to be developed to shed light on these complex phenomena to boost a broader adoption of AM technologies in industrial applications. It is well known for example that the elastic behavior of lattice structures is dramatically underestimated when computed on the as-designed geometry. Furthermore, due to the inherent variability of PBF-LB/M process parameters, several sources of uncertainty hinder a full understanding of the complex process-structure-property relationships. In the presentation we will highlights some of the interesting applications open by the power of AM but also some limitations due the problems highlighted above.

Phase field for compaction band formation: capture of grain crushing and permeability evolution in heterogeneous media

Jesús Sánchez Pinedo — Tue, 12 Dec 2023 10:43:45 +0100

Compaction bands form when the pore spaces between the solid grains of a rock mass collapse into a narrow zone. This deformation style has attracted much attention in the theoretical and numerical modeling community since the porosity reduction associated with pore collapse reduces the overall permeability of the rock, thus enhancing its potential to serve as a fluid flow barrier. Recent publications [1, 2] demonstrate the capability of the phase-field modeling approach for capturing the formation and propagation of compaction bands in porous rocks. In this talk, the phase-field approach is utilized to show how grain crushing and fluid flow impact the formation and propagation of compaction bands. In the context of the finite element method, a three-field variational formulation in terms of solid displacement, fluid pressure, and the phase-field variable is employed for this purpose. Using material parameters calibrated from real rocks, we show how the volume constraint imposed by fluid flow could impact the stress-strain responses of the rock as well as the ensuing geometric style of the compaction band.

Physics-based and data-driven hybrid modelling of materials and their processing

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 09:59:15 +0100

Physics aware digital twins of materials, processes, components and systems enable emulating real assets while ensuring fidelity and efficiency. They embrace physics-based and data-driven functionalities, both enriching mutually. Both should proceed in almost real-time, and the last being able to proceed in the scarce data limit. When applied to materials and processes, model order reduction technologies enable the construction of the so-called surrogate model, whereas data-driven modelling, based in advanced machine learning and artificial intelligence technologies, must be informed by the physics to encompass rapidity and accuracy, in the low data limit. This hybrid approach allows improving accuracy of existing models, as well as constructing models when the existing ones remains too poor or uncertain. Moreover, this setting allows to speed-up predictions, enabling real-time control, decision-making as well as the exploration of the whole design space, crucial in the design of materials and components

High-resolution computational plasticity at the micron scale

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:07:13 +0100

The persistent demand for green, strong and ductile advanced high strength steels, with a reduced climate footprint, calls for novel and improved multi-phase microstructures. The development of these new steels requires an in-depth understanding of the governing plasticity mechanisms at the micron scale. In order to address this challenge, novel numerical-experimental methods are called for that account for the discreteness, statistics and the intrinsic role of interfaces. This lecture sheds light on recent and innovative developments unravelling metal plasticity at the micron scale

Phase-field Fracture: Toward a General Purpose Technology

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:10:45 +0100

Fracture with Phase-Field Models: Discretization, Acceleration, Application

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:17:21 +0100

This presentation will provide an overview and an evaluation of modern discretizational techniques of phase-field methods for fracture. To this end, this talk will open with a motivating example of a geometrically very complex 3D fracture of a core rock sample (see Figure 1). It is well known that such computations are challenging. This is mainly because: • the phase-field regularization of the sharp crack is based on a length-scale parameter that requires very fine computational meshes, • domains of interest may possess very complex topologies described e.g. by CTscans, which in turn leads to very large systems of equations. • phase-field models for fracture result in an unsymmetric coupled problem of at least two fields whose staggered solution typically suffers from slow convergence. This talk will present a set of recently developed numerical tools to address these challenges. First, a type of local refinement is introduced, which is particularly well suited for transient situations and for which very efficient open-source implementations now exist. This discretization is then combined with the Finite Cell Method, which delivers the possibility to compute phase-field fracture models on complex domains in a straightforward manner. At this point, an extension of the phase-field method for the modelling of rock will be introduced

Automated model discovery – A new paradigm in computational mechanics?

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:20:21 +0100

Convolution Hierarchical Deep Learning Neural Network (C-HiDeNN)-AI: From Topological Optimization to Additive Manufactured Materials

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:23:41 +0100

Interactions of dislocations and twins with grain boundaries: unraveling the mechanisms of plastic deformation in polycrystals

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:26:44 +0100

Grain boundaries play a critical role in the plastic deformation of metallic materials. They may hinder slip transfer, leading to the formation of dislocation pile-ups and to size effects (HallPetch) or allow slip transfer, allowing the localization of deformation in suitably oriented grains clusters. They can also absorb and/or emit dislocations, nucleate twins, induce fracture, etc. As a result, interaction of dislocations and twins with grain boundaries as well as the role of grain boundaries in the nucleation of twins has received a lot of attention from the scientific community. However, most of experimental results are limited to surface observations in which the 3D nature of grains boundaries is not accounted for while atomistic simulations are mostly focused on coincidence lattice site boundaries that are different from the disordered grain boundaries of polycrystals. Thus, reliable criteria to quantify the role of grain boundaries in polycrystal deformation are still lacking. In this talk, large experimental data sets of the interaction of dislocations with grain boundaries were obtained using state-of-the-art characterization techniques to assess the influence of grain boundary orientation in 3D on the possibility of slip transfer/blocking in Ti and Mg. Similarly, in situ mechanical tests within the scanning electron microscope were carried to study the nucleation of tensile twins in Mg near grain boundaries. This information was analyzed using statistical methods and machine learning tools to determine the critical microstructural parameters that dominate slip transfer/blocking and twin nucleation. This information was then implemented in crystal plasticity finite element models to predict the mechanical behavior of polycrystals.

Towards Multiscale Computational Design of Shock-absorbing Metamaterials: (II) From the low-scale to the upper-scale

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:30:37 +0100

To explore the computational design of shock-absorbing metamaterials, this work is a continuation of the one that was presented in COMPLAS 2021 " Towards the Multiscale Computational Design of Shock-absorbing Metamaterials: (I) From the Upper-Scale to the Low-Scale" (see [1]). Once explored there the mechanisms for mechanical dissipation that arise from propagating shocks on the high scale via “theoretical" nonconvex hyperelastic materials, the concept of multiscale metamaterial design is retrieved by defining a mesoscale constituted by a beams lattice, which buckles due to the interaction with the macro-scale (Hill-Mandel energetic equivalence principle), thus giving rise to a homogenized constitutive behavior exhibiting, on the macro-scale, the nonconvexity requested to exhibit “extrinsic” dissipation features. The goal now is exploring the computational challenges associated to this computational modeling i.e. 1) The homogenization of a representative volume element (RVE), made of 1D buckling beams at the mesoscale, into a 2D constitutive model at the macro-scale, 2) the controversial issue of the dependence of resulting homogenized macro-scale behavior on the RVE size, 3) the efficiency in generating mechanical dissipation at the upper scale, this qualifying the proposed setting as amenable for shock absorbing metamaterial design purposes. Representative examples show the degree of achievement of solutions to the aforementioned challenges.

Interaction of particulate fluids and structures. Modelling and computational challenges

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:33:00 +0100

Advances In Particle Finite Element Method For The Simulation Of Phase Change Problems And Fluid-Structure Interactions

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:37:12 +0100

Particle Finite Element Method (PFEM) is a still rather young method that tries to combine the advantages of classical methods such as FEM and more recent methods known as particle methods (SPH…) The method is quite versatile and can be applied to both solid and fluids material behavior. It is a Lagrangian method that combines computations over one time step using FEM with a fast remeshing algorithm trying to avoid mesh distortions consequent to very large deformations such as the ones encountered for fluid flow with free surfaces. New developments will be presented here, such as the use of a level set function, instead of the traditional alpha-shape algorithm to determine the new boundaries of a body after remeshing, as well as the implementation of phase-change algorithm, including vaporization. The lecture will cover several applications including the simulation of the fluid behavior in a melt pool during LPBF (laser Powder Bed Fusion) where the initial powder is melted by the laser, and then solidifies again when the laser goes away. During this process, due to the high power density of the laser, some part of the material is not only melted but also vaporized. Other applications of the PFEM will illustrate fluidstructure interactions simulations including contact between different solid parts and plastic deformation of some components of the system.

Enhancing Material Point Method for Disaster Simulation

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:40:52 +0100

Several recently developed enhancements to material point methods (MPMs) are presented to increase the reliability and predictability of disaster simulations. The details of the enhanced techniques are as diverse as those listed below.

Application of virtual element methods for numerical simulation of inelastic response

Jesús Sánchez Pinedo — Wed, 13 Dec 2023 10:44:25 +0100

The Virtual Element Method (VEM) is a novel technology for the approximate solution of partial differential equations that shares the variational background of the finite element method. VEM has the flexibility to deal with general polygonal/polyhedral meshes, including “hanging vertices” and non-convex element shape, while retaining the conformity of the method. This allows different applications in the area of inelastic materials which include homogenization of materials with polycrystalline microstructure, thermo-mechanical responses at finite strains and impact problems. In this presentation we will discuss different aspects of the formulation of low order three-dimensional virtual elements for the class of problems mentioned above