Showing 1–12 of 12 results for author: Lambrechts, J

Solving the Porous Medium Equation with the eXtreme Mesh deformation approach (X-Mesh)

Authors: Alexandre Chemin, Jonathan Lambrechts, Nicolas Moës, Jean-François Remacle

Abstract: We introduce a new scheme for solving the non-regularized Porous Medium Equation. It is mass conserving and uses only positive unknown values. To address these typically conflicting features, we employ the eXtreme Mesh deformation approach (X-Mesh), specifically designed for problems involving sharp interfaces. The method ensures that the interface is always meshed, even in the face of complex top… ▽ More We introduce a new scheme for solving the non-regularized Porous Medium Equation. It is mass conserving and uses only positive unknown values. To address these typically conflicting features, we employ the eXtreme Mesh deformation approach (X-Mesh), specifically designed for problems involving sharp interfaces. The method ensures that the interface is always meshed, even in the face of complex topological changes, without the need for remeshing or altering the mesh topology. We illustrate the effectiveness of the approach through various numerical experiments. △ Less

Submitted 6 January, 2025; originally announced January 2025.

A Coupled PFEM-DEM Model for Fluid-Granular Flows with Free-Surface Dynamics Applied to Landslides

Authors: Thomas Leyssens, Michel Henry, Jonathan Lambrechts, Vincent Legat, Jean-François Remacle

Abstract: Free surface and granular fluid mechanics problems combine the challenges of fluid dynamics with aspects of granular behaviour. This type of problem is particularly relevant in contexts such as the flow of sediments in rivers, the movement of granular soils in reservoirs, or the interactions between a fluid and granular materials in industrial processes such as silos. The numerical simulation of t… ▽ More Free surface and granular fluid mechanics problems combine the challenges of fluid dynamics with aspects of granular behaviour. This type of problem is particularly relevant in contexts such as the flow of sediments in rivers, the movement of granular soils in reservoirs, or the interactions between a fluid and granular materials in industrial processes such as silos. The numerical simulation of these phenomena is challenging because the solution depends not only on the multiple phases that strongly interact with each other, but also on the need to describe the geometric evolution of the different interfaces. This paper presents an approach to the simulation of fluid-granular phenomena involving strongly deforming free surfaces. The Discrete Element Method (DEM) is combined with the Particle Finite Element Method (PFEM) and the fluid-grain interface is treated by a two-way coupling between the two phases. The fluid-air interface is solved by a free surface model. The geometric and topological variations are therefore naturally provided by the full Lagrangian description of all phases. The approach is validated on benchmark test cases such as two-phase dam failures and then applied to a real landslide problem. △ Less

Submitted 6 January, 2025; originally announced January 2025.

The Tempered Finite Element Method

Authors: Antoine Quiriny, Václav Kučera, Jonathan Lambrechts, Nicolas Moës, Jean-François Remacle

Abstract: In this paper, we propose a new approach -- the Tempered Finite Element Method (TFEM) -- that extends the Finite Element Method (FEM) to classes of meshes that include zero-measure or nearly degenerate elements for which standard FEM approaches do not allow convergence. First, we review why the maximum angle condition [2] is not necessary for FEM convergence and what are the real limitations in te… ▽ More In this paper, we propose a new approach -- the Tempered Finite Element Method (TFEM) -- that extends the Finite Element Method (FEM) to classes of meshes that include zero-measure or nearly degenerate elements for which standard FEM approaches do not allow convergence. First, we review why the maximum angle condition [2] is not necessary for FEM convergence and what are the real limitations in terms of meshes. Next, we propose a simple modification of the classical FEM for elliptic problems that provably allows convergence for a wider class of meshes including bands of caps that cause locking of the solution in standard FEM formulations. The proposed method is trivial to implement in an existing FEM code and can be theoretically analyzed. We prove that in the case of exactly zero-measure elements it corresponds to mortaring. We show numerically and theoretically that what we propose is functional and sound. The remainder of the paper is devoted to extensions of the TFEM method to linear elasticity, mortaring of non-conforming meshes, high-order elements, and advection. △ Less

Submitted 26 November, 2024; originally announced November 2024.

A Delaunay Refinement Algorithm for the Particle Finite Element Method applied to Free Surface Flows

Authors: Thomas Leyssens, Michel Henry, Jonathan Lambrechts, Jean-Francois Remacle

Abstract: This paper proposes two contributions to the calculation of free surface flows using the particle finite element method (PFEM). The PFEM is based on a Lagrangian approach: a set of particles defines the fluid. Then, unlike a pure Lagrangian method, all the particles are connected by a triangular mesh. The difficulty lies in locating the free surface from this mesh. It is a matter of deciding which… ▽ More This paper proposes two contributions to the calculation of free surface flows using the particle finite element method (PFEM). The PFEM is based on a Lagrangian approach: a set of particles defines the fluid. Then, unlike a pure Lagrangian method, all the particles are connected by a triangular mesh. The difficulty lies in locating the free surface from this mesh. It is a matter of deciding which of the elements in the mesh are part of the fluid domain, and to define a boundary - the free surface. Then, the incompressible Navier-Stokes equations are solved on the fluid domain and the particles' position is updated using the resulting velocity vector. Our first contribution is to propose an approach to adapt the mesh with theoretical guarantees of quality: the mesh generation community has acquired a lot of experience and understanding about mesh adaptation approaches with guarantees of quality on the final mesh. We use here a Delaunay refinement strategy, allowing to insert and remove nodes while gradually improving mesh quality. We show that this allows to create stable and smooth free surface geometries. Our PFEM approach models the topological evolution of one fluid. It is nevertheless necessary to apply conditions on the domain boundaries. When a boundary is a free surface, the flow on the other side is not modelled, it is represented by an external pressure. On the external free surface boundary, atmospheric pressure can be imposed. Nevertheless, there may be internal free surfaces: the fluid can fully encapsulate cavities to form bubbles. The pressure required to maintain the volume of those bubbles is a priori unknown. We propose a multi-point constraint approach to enforce global incompressibility of those empty bubbles. This approach allows to accurately model bubbly flows that involve two fluids with large density differences, while only modelling the heavier fluid. △ Less

Submitted 27 March, 2024; originally announced March 2024.

X-Mesh: A new approach for the simulation of two-phase flow with sharp interface

Authors: Antoine Quiriny, Jonathan Lambrechts, Nicolas Moës, Jean-François Remacle

Abstract: Accurate modeling of moving boundaries and interfaces is a difficulty present in many situations of computational mechanics. We use the eXtreme Mesh deformation approach (X-Mesh) to simulate the interaction between two immiscible flows using the finite element method, while maintaining an accurate and sharp description of the interface without remeshing. In this new approach, the mesh is locally d… ▽ More Accurate modeling of moving boundaries and interfaces is a difficulty present in many situations of computational mechanics. We use the eXtreme Mesh deformation approach (X-Mesh) to simulate the interaction between two immiscible flows using the finite element method, while maintaining an accurate and sharp description of the interface without remeshing. In this new approach, the mesh is locally deformed to conform to the interface at all times, which can result in degenerated elements. The surface tension between the two fluids is added by imposing the pressure jump condition at the interface, which, when combined with the X-Mesh framework, allows us to have an exactly sharp interface. If a numerical scheme fails to properly balance surface tension and pressure gradients, it leads to numerical artefacts called spurious or parasitic currents. The method presented here is well balanced and reduces such currents down to the level of machine precision. △ Less

Submitted 1 February, 2024; v1 submitted 8 February, 2023; originally announced February 2023.

The eXtreme Mesh deformation approach (X-MESH) for the Stefan phase-change model

Authors: Nicolas Moes, Jean-Francois Remacle, Jonathan Lambrechts, Benoit Le, Nicolas Chevaugeon

Abstract: The eXtreme Mesh deformation approach (X-MESH) is a new paradigm to follow sharp interfaces without remeshing and without changing the mesh topology. Even though the mesh does not change its topology, it can follow interfaces that do change their topology (nucleation, coalescence, splitting). To make this possible, the key X-MESH idea is to allow elements to reach zero measure. This permits interf… ▽ More The eXtreme Mesh deformation approach (X-MESH) is a new paradigm to follow sharp interfaces without remeshing and without changing the mesh topology. Even though the mesh does not change its topology, it can follow interfaces that do change their topology (nucleation, coalescence, splitting). To make this possible, the key X-MESH idea is to allow elements to reach zero measure. This permits interface relaying between nodes as well as interface annihilation and seeding in a time continuous manner. The paper targets the Stefan phase change model in which the interface (front) is at a given temperature. Several examples demonstrate the capability of the approach. △ Less

Submitted 24 January, 2023; v1 submitted 7 November, 2021; originally announced November 2021.

Comments: 18 pages

Quaternionic octahedral fields: SU(2) parameterization of 3D frames

Authors: Pierre-Alexandre Beaufort, Jonathan Lambrechts, Christophe Geuzaine, Jean-Francois Remacle

Abstract: 3D frame fields are auxiliary for hexahedral mesh generation. There mainly exist three ways to represent 3D frames: combination of rotations, spherical harmonics and fourth order tensor. We propose here a representation carried out by the special unitary group. The article strongly relies on \cite{du1964homographies}. We first describe the rotations with quaternions, \cite[§13-15]{du1964ho… ▽ More 3D frame fields are auxiliary for hexahedral mesh generation. There mainly exist three ways to represent 3D frames: combination of rotations, spherical harmonics and fourth order tensor. We propose here a representation carried out by the special unitary group. The article strongly relies on \cite{du1964homographies}. We first describe the rotations with quaternions, \cite[§13-15]{du1964homographies}. We define and show the isomorphism between unit quaternions and the special unitary group, \cite[§16]{du1964homographies}. The frame field space is identified as the quotient group of rotations by the octahedral group, \cite[§20]{du1964homographies}. The invariant forms of the vierer, tetrahedral and octahedral groups are successively built, without using homographies \cite[§39]{du1964homographies}. Modifying the definition of the isomorphism between unit quaternions and the special unitary group allows to use the invariant forms of the octahedral group as a unique parameterization of the orientation of 3D frames. The parameterization consists in three complex values, corresponding to a coordinate of a variety which is embedded in a three complex valued dimensional space. The underlined variety is the model surface of the octahedral group, which can be expressed with an implicit equation. We prove that from a coordinate of the surface, we may identify all the quaternions giving the corresponding 3D frames. We show that the euclidean distance between two coordinates does not correspond to the actual distance of the corresponding 3D frames. We derive the expression of three components of a coordinate in the case of frames sharing an even direction. We then derive a way to ensure that a coordinate corresponds to the special unitary group. Finally, the attempted numerical schemes to compute frame fields are given. △ Less

Submitted 6 January, 2020; v1 submitted 14 October, 2019; originally announced October 2019.

Comments: NOT submitted. Main reference: Patrick Du Val, 1964. Homographies, quaternions and rotations

Adversarial Robustness vs Model Compression, or Both?

Authors: Shaokai Ye, Kaidi Xu, Sijia Liu, Jan-Henrik Lambrechts, Huan Zhang, Aojun Zhou, Kaisheng Ma, Yanzhi Wang, Xue Lin

Abstract: It is well known that deep neural networks (DNNs) are vulnerable to adversarial attacks, which are implemented by adding crafted perturbations onto benign examples. Min-max robust optimization based adversarial training can provide a notion of security against adversarial attacks. However, adversarial robustness requires a significantly larger capacity of the network than that for the natural trai… ▽ More It is well known that deep neural networks (DNNs) are vulnerable to adversarial attacks, which are implemented by adding crafted perturbations onto benign examples. Min-max robust optimization based adversarial training can provide a notion of security against adversarial attacks. However, adversarial robustness requires a significantly larger capacity of the network than that for the natural training with only benign examples. This paper proposes a framework of concurrent adversarial training and weight pruning that enables model compression while still preserving the adversarial robustness and essentially tackles the dilemma of adversarial training. Furthermore, this work studies two hypotheses about weight pruning in the conventional setting and finds that weight pruning is essential for reducing the network model size in the adversarial setting, training a small model from scratch even with inherited initialization from the large model cannot achieve both adversarial robustness and high standard accuracy. Code is available at https://github.com/yeshaokai/Robustness-Aware-Pruning-ADMM. △ Less

Submitted 22 June, 2021; v1 submitted 29 March, 2019; originally announced March 2019.

Comments: Accepted by ICCV 2019

High quality mesh generation using cross and asterisk fields: Application on coastal domains

Authors: Christos Georgiadis, Pierre-Alexandre Beaufort, Jonathan Lambrechts, Jean-François Remacle

Abstract: This paper presents a method to generate high quality triangular or quadrilateral meshes that uses direction fields and a frontal point insertion strategy. Two types of direction fields are considered: asterisk fields and cross fields. With asterisk fields we generate high quality triangulations, while with cross fields we generate right-angled triangulations that are optimal for transformation to… ▽ More This paper presents a method to generate high quality triangular or quadrilateral meshes that uses direction fields and a frontal point insertion strategy. Two types of direction fields are considered: asterisk fields and cross fields. With asterisk fields we generate high quality triangulations, while with cross fields we generate right-angled triangulations that are optimal for transformation to quadrilateral meshes. The input of our algorithm is an initial triangular mesh and a direction field calculated on it. The goal is to compute the vertices of the final mesh by an advancing front strategy along the direction field. We present an algorithm that enables to efficiently generate the points using solely information from the base mesh. A multi-threaded implementation of our algorithm is presented, allowing us to achieve significant speedup of the point generation. Regarding the quadrangulation process, we develop a quality criterion for right-angled triangles with respect to the local cross field and an optimization process based on it. Thus we are able to further improve the quality of the output quadrilaterals. The algorithm is demonstrated on the sphere and examples of high quality triangular and quadrilateral meshes of coastal domains are presented. △ Less

Submitted 7 June, 2017; originally announced June 2017.

Computing cross fields -- A PDE approach based on the Ginzburg-Landau theory

Authors: Pierre-Alexandre Beaufort, Christos Georgiadis Jonathan Lambrechts, François Henrotte, Christophe Geuzaine, Jean-François Remacle

Abstract: This paper proposes a method to compute crossfields based on the Ginzburg-Landau theory. The Ginzburg-Landau functional has two terms: the Dirichlet energy of the distribution and a term penalizing the mismatch between the fixed and actual norm of the distribution. Directional fields on surfaces are known to have a number of critical points, which are properly identified with the Ginzburg-Landau a… ▽ More This paper proposes a method to compute crossfields based on the Ginzburg-Landau theory. The Ginzburg-Landau functional has two terms: the Dirichlet energy of the distribution and a term penalizing the mismatch between the fixed and actual norm of the distribution. Directional fields on surfaces are known to have a number of critical points, which are properly identified with the Ginzburg-Landau approach: the asymptotic behavior of Ginzburg-Landau problem provides well-distributed critical points over the 2-manifold, whose indices are as low as possible. The central idea in this paper is to exploit this theoretical background for crossfield computation on arbitrary surfaces. Such crossfields are instrumental in the generation of meshes with quadrangular elements. The relation between the topological properties of quadrangular meshes and crossfields are hence first recalled. It is then shown that a crossfield on a surface can be represented by a complex function of unit norm with a number of critical points, i.e., a nearly everywhere smooth function taking its values in the unit circle of the complex plane. As maximal smoothness of the crossfield is equivalent with minimal energy, the crossfield problem is equivalent to an optimization problem based on Ginzburg-Landau functional. A discretization scheme with Crouzeix-Raviart elements is applied and the correctness of the resulting finite element formulation is validated on the unit disk by comparison with an analytical solution. The method is also applied to the 2-sphere where, surprisingly but rightly, the computed critical points are not located at the vertices of a cube, but at those of an anticube. △ Less

Submitted 6 January, 2020; v1 submitted 2 June, 2017; originally announced June 2017.

Comments: Promoted version of: Proceeding for the 26th International Meshing Roundtable, IMR26, 18-21 September 2017, Barcelona, Spain

Fast and robust mesh generation on the sphere Application to coastal domains

Authors: Jean-François Remacle, Jonathan Lambrechts

Abstract: This paper presents a fast an robust mesh generation procedure that is able to generate meshes of the earth system (ocean and continent) in matters of seconds. Our algorithm takes as input a standard shape-file i.e. geospatial vector data format for geographic information system (GIS) software. The input is initially coarsened in order to automatically remove unwanted channels that are under a des… ▽ More This paper presents a fast an robust mesh generation procedure that is able to generate meshes of the earth system (ocean and continent) in matters of seconds. Our algorithm takes as input a standard shape-file i.e. geospatial vector data format for geographic information system (GIS) software. The input is initially coarsened in order to automatically remove unwanted channels that are under a desired resolution. A valid non-overlapping 1D mesh is then created on the sphere using the Euclidian coordinates system $x,y,z$. A modified Delaunay kernel is then proposed that enables to generate meshes on the sphere in a straightforward manner without parametrization. One of the main difficulty in dealing with geographical data is the over-sampled nature of coastline representations. We propose here an algorithm that automatically unrefines coastline data. Small features are automatically removed while always keeping a valid (non-overlapping) geometrical representation of the domain. A Delaunay refinement procedure is subsequently applied to the domain. The refinement scheme is also multi-threaded at a fine grain level, allowing to generate about a million points per second on 8 threads. Examples of meshes of the Baltic sea as well as of the global ocean are presented. △ Less

Submitted 7 July, 2016; originally announced July 2016.

Comments: Submitted to the 25th international meshing round table

Optimizing the geometrical accuracy of curvilinear meshes

Authors: Thomas Toulorge, Jonathan Lambrechts, Jean-François Remacle

Abstract: This paper presents a method to generate valid high order meshes with optimized geometrical accuracy. The high order meshing procedure starts with a linear mesh, that is subsequently curved without taking care of the validity of the high order elements. An optimization procedure is then used to both untangle invalid elements and optimize the geometrical accuracy of the mesh. Standard measures of t… ▽ More This paper presents a method to generate valid high order meshes with optimized geometrical accuracy. The high order meshing procedure starts with a linear mesh, that is subsequently curved without taking care of the validity of the high order elements. An optimization procedure is then used to both untangle invalid elements and optimize the geometrical accuracy of the mesh. Standard measures of the distance between curves are considered to evaluate the geometrical accuracy in planar two-dimensional meshes, but they prove computationally too costly for optimization purposes. A fast estimate of the geometrical accuracy, based on Taylor expansions of the curves, is introduced. An unconstrained optimization procedure based on this estimate is shown to yield significant improvements in the geometrical accuracy of high order meshes, as measured by the standard Haudorff distance between the geometrical model and the mesh. Several examples illustrate the beneficial impact of this method on CFD solutions, with a particular role of the enhanced mesh boundary smoothness. △ Less

Submitted 14 July, 2015; originally announced July 2015.

Comments: Submitted to JCP

MSC Class: 65D99; 76M10