Tetrahedron VI Abstracts
Tahar Amari
Ecole Polytechnique, Palaiseau, France
Sun-Earth environment modelling from Solar Eruptions to Magnetosphere
Large scale eruptive events such as Coronal Mass Ejections represent an important source of energy release occurring in the solar atmosphere, which can make their journey to the Earth magnetosphere, representing thus a major component of Space Weather, with economic impact.
We will discuss some problems and their associated numerical aspects arising in the simulation of the physics of such phenomena which are governed by the magnetic field. We will show how the stiffness of those problems imply particular constraints among which mesh generation and adaptation can plays an important role.
Large scale eruptive events such as Coronal Mass Ejections represent an important source of energy release occurring in the solar atmosphere, which can make their journey to the Earth magnetosphere, representing thus a major component of Space Weather, with economic impact.
We will discuss some problems and their associated numerical aspects arising in the simulation of the physics of such phenomena which are governed by the magnetic field. We will show how the stiffness of those problems imply particular constraints among which mesh generation and adaptation can plays an important role.
Romain Aubry
E. Mestreau, M. Williamschen, D. Williams, W. Szymczak and S. Dey
KeyW Corp. and US Naval Research Laboratory, Washington DC, USA
Boundary layer mesh generation for arbitrary geometry and sizing
Download Presentation
Boundary layer mesh generation is a challenging area for numerous reasons. A very anisotropic smooth sizing field is required to capture the anisotropy of the physics, and provide a low interpolation error for the numerical scheme. Historically, anisotropy has been achieved by lifting the surface mesh into the volume mesh. Lately, we have shown the draw- backs of that approach at different levels, such as a lack of smoothness in the sizing field, and therefore in the mesh, as well as artificial abrupt transitions between the semistructured and unstructured mesh, and we have proposed solutions to these issues. We will discuss these aspects, and show some numerical examples to illustrate our approach.
Download Presentation
Boundary layer mesh generation is a challenging area for numerous reasons. A very anisotropic smooth sizing field is required to capture the anisotropy of the physics, and provide a low interpolation error for the numerical scheme. Historically, anisotropy has been achieved by lifting the surface mesh into the volume mesh. Lately, we have shown the draw- backs of that approach at different levels, such as a lack of smoothness in the sizing field, and therefore in the mesh, as well as artificial abrupt transitions between the semistructured and unstructured mesh, and we have proposed solutions to these issues. We will discuss these aspects, and show some numerical examples to illustrate our approach.
Mikael Berton and Olivier Allain
Lemma, Sophia-Antipolis, France
Implementation of Feflo.a, Metrix and Interpol for rotating machines
Download Presentation
Different approaches can be chosen to model rotating machines like sliding mesh, chimera, immersed boundary, edges swap, etc. The immersed boundary method allows to manage easily contact problems or to facilitate mesh motion. However, applications may be complex especialy when modeled details are very small. The use of the adaptation tools allow to reduce drastically the mesh size, to reduce the total number of processors for the simulation, and to reduce the CPU time. We will present a workflow process using mesher, interpolation, optimization, CFD tools apply to rotating machines.
Download Presentation
Different approaches can be chosen to model rotating machines like sliding mesh, chimera, immersed boundary, edges swap, etc. The immersed boundary method allows to manage easily contact problems or to facilitate mesh motion. However, applications may be complex especialy when modeled details are very small. The use of the adaptation tools allow to reduce drastically the mesh size, to reduce the total number of processors for the simulation, and to reduce the CPU time. We will present a workflow process using mesher, interpolation, optimization, CFD tools apply to rotating machines.
Maximilien Dramont
ArianeGroup, Les Mureaux, France
CASCADES: a tool to automatically generate and mesh geometries for fast CFD computations of Aerodynamic Data base (AEDB)
Download Presentation
The CASCADES tool has been designed to work with the ArianeGroup FLUSEPA CFD solver. The goal was to drastically reduce the overall duration of a typical AEDB study for almost any kind of launchers, reentry bodies and space vehicles. The CASCADES strategy is based on the decomposition of complex vehicles into a set of independent parts, thanks to a library of preexisting parametric elements. An in-house routine processes the elementary components and builds both the geometry and the hexahedral mesh grids associated with the boundary layer area. The multi-overlapping technique of FLUSEPA merges the individual pieces into a single composite grid, considering a priority level for each one. The Adaptive Mesh Refinement (AMR) technique completes the work by refining the mesh grid according to the characteristics of the local flow. The final result is a fully parametric CFD model (for both geometry and mesh) of a complete vehicle that allows the user to take into account any kind of modifications very quickly. A typical use as well as other kinds of applications will be presented during the conference.
Download Presentation
The CASCADES tool has been designed to work with the ArianeGroup FLUSEPA CFD solver. The goal was to drastically reduce the overall duration of a typical AEDB study for almost any kind of launchers, reentry bodies and space vehicles. The CASCADES strategy is based on the decomposition of complex vehicles into a set of independent parts, thanks to a library of preexisting parametric elements. An in-house routine processes the elementary components and builds both the geometry and the hexahedral mesh grids associated with the boundary layer area. The multi-overlapping technique of FLUSEPA merges the individual pieces into a single composite grid, considering a priority level for each one. The Adaptive Mesh Refinement (AMR) technique completes the work by refining the mesh grid according to the characteristics of the local flow. The final result is a fully parametric CFD model (for both geometry and mesh) of a complete vehicle that allows the user to take into account any kind of modifications very quickly. A typical use as well as other kinds of applications will be presented during the conference.
Mark Gammon
ITI, Cambridge, UK
Why CAD Geometry is Not CAE Geometry
Download Presentation
Over the last 30 years commercial CAD tools have matured at an impressive rate. Today they provide the de facto industry solution for the creation of sophisticated, highly realistic, parametric geometry, capturing ever increasing levels of detail and complexity. However, for engineers and analysts performing advanced simulation, CAD geometry is frequently quoted as the single biggest bottleneck in their workflow, accounting for hours and days of time lost performing manual clean-up and simplification. The end result of this work is a new CAE geometry model suitable for the selected analysis. In this talk I will attempt to give an overview of some of the most pressing geometry issues affecting the simulation industry based on customer experiences and feedback from the ITI CADfix™ tool. I will explain their source and illustrate the nature of their negative impact on producing an acceptable CAE geometry. I will also share some highlights of the progress ITI is making in deploying novel techniques, such as the 3D medial axis transform, to unlock some of the key CAE geometry challenges.
Download Presentation
Over the last 30 years commercial CAD tools have matured at an impressive rate. Today they provide the de facto industry solution for the creation of sophisticated, highly realistic, parametric geometry, capturing ever increasing levels of detail and complexity. However, for engineers and analysts performing advanced simulation, CAD geometry is frequently quoted as the single biggest bottleneck in their workflow, accounting for hours and days of time lost performing manual clean-up and simplification. The end result of this work is a new CAE geometry model suitable for the selected analysis. In this talk I will attempt to give an overview of some of the most pressing geometry issues affecting the simulation industry based on customer experiences and feedback from the ITI CADfix™ tool. I will explain their source and illustrate the nature of their negative impact on producing an acceptable CAE geometry. I will also share some highlights of the progress ITI is making in deploying novel techniques, such as the 3D medial axis transform, to unlock some of the key CAE geometry challenges.
Vladimir Garanzha and Liudmila Kudryavtseva
Dorodnicyn Computing Center, Moscow, Russia
Hybrid Voronoi meshing
Download Presentation
Approximation of the computational domain boundary by the Voronoi edges and faces is challenging problem of mesh generation. For 2d multimaterial domains one can build hybrid Voronoi meshes with structured anisotropic orthogonal Voronoi layers near boundaries. In 3d one can recover surfaces features, such as sharp edges ad conical vertices via careful choice of Voronoi seeds, but the problem of elimination of Voronoi faults, or small unaligned boundary Voronoi edges/faces cannot be easily solved which leads us either to the concept of approximate curved Voronoi-like faces for near-boundary polyhedral cells, or to the generalizations of the surface circular patterns concept - covering empty ball patterns.
Download Presentation
Approximation of the computational domain boundary by the Voronoi edges and faces is challenging problem of mesh generation. For 2d multimaterial domains one can build hybrid Voronoi meshes with structured anisotropic orthogonal Voronoi layers near boundaries. In 3d one can recover surfaces features, such as sharp edges ad conical vertices via careful choice of Voronoi seeds, but the problem of elimination of Voronoi faults, or small unaligned boundary Voronoi edges/faces cannot be easily solved which leads us either to the concept of approximate curved Voronoi-like faces for near-boundary polyhedral cells, or to the generalizations of the surface circular patterns concept - covering empty ball patterns.
Bob Haimes
MIT, Cambridge, MA, USA
The Use of Geometry from within the Engineering Sketch Pad
The EGADS API
Download Presentation
The Engineering Sketch Pad (ESP) is a collection of integrated APIs built to solve the problem of providing a consistent underpinning for 'Design through Analysis'. ESP supports the ability to have a single set of parameters drive the design in a true multi-fidelity and multidisciplinary setting. The API that communicates with grid generators and solvers is CAPS (Computational Aircraft Prototype Syntheses), where the API that contains a differentiated parametric build engine is OpenCSM (the Open Constructive Solid Modeler). EGADS (the Electronic Geometry Aircraft Design System) is the API that performs both 'Bottom-up' and 'Top-down' geometric construction and is foundational for all APIs and plug-ins found within ESP.
This presentation discusses the use of EGADS and EGADSlite (a read-only, no construction, version of EGADS used in HPC environments) for mesh generation and adaptation. The sections of the talk include:
The EGADS API
Download Presentation
The Engineering Sketch Pad (ESP) is a collection of integrated APIs built to solve the problem of providing a consistent underpinning for 'Design through Analysis'. ESP supports the ability to have a single set of parameters drive the design in a true multi-fidelity and multidisciplinary setting. The API that communicates with grid generators and solvers is CAPS (Computational Aircraft Prototype Syntheses), where the API that contains a differentiated parametric build engine is OpenCSM (the Open Constructive Solid Modeler). EGADS (the Electronic Geometry Aircraft Design System) is the API that performs both 'Bottom-up' and 'Top-down' geometric construction and is foundational for all APIs and plug-ins found within ESP.
This presentation discusses the use of EGADS and EGADSlite (a read-only, no construction, version of EGADS used in HPC environments) for mesh generation and adaptation. The sections of the talk include:
- The ESP collection of APIs
- The Boundary Representation (BRep)
- EGADS/EGADSlite Objects
- Geometry objects
- Topology objects
- Tessellation objects - The EGADS/EGADSlite API
- Base & Utility Functions
- Attribution
- Geometry
- Topology
- Tessellation
- Top-Down Construction (not in EGADSlite)
- Meshing Considerations
Yasushi Ito
JAXA, Tokyo, Japan
Automatic Local Remeshing Method for High-Fidelity Computational Fluid Dynamics Simulations
Download Presentation
This presentation describes high-fidelity computational fluid dynamics simulations for aircraft in various configurations. Reynolds-averaged Navier-Stokes simulations were conducted with an unstructured flow solver, the TAS code, and a hybrid mesh generator, MEGG3D. To evaluate the performance of an aircraft with aerodynamic add-on devices, such as vortex generators, we sometimes need to predict the aerodynamic difference of a baseline geometry with and without the add-on devices. Unstructured hybrid meshes are effective to represent complex geometries and to perform high Reynolds number viscous flow simulations. However, it is often difficult to update an existing volume mesh by only remeshing around the add-on devices in order to avoid unnecessary mesh dependency and to ease the evaluation process by reusing the existing mesh. We will present the application of an automatic local remeshing method to a few practical aircraft geometries to discuss its usefulness.
Download Presentation
This presentation describes high-fidelity computational fluid dynamics simulations for aircraft in various configurations. Reynolds-averaged Navier-Stokes simulations were conducted with an unstructured flow solver, the TAS code, and a hybrid mesh generator, MEGG3D. To evaluate the performance of an aircraft with aerodynamic add-on devices, such as vortex generators, we sometimes need to predict the aerodynamic difference of a baseline geometry with and without the add-on devices. Unstructured hybrid meshes are effective to represent complex geometries and to perform high Reynolds number viscous flow simulations. However, it is often difficult to update an existing volume mesh by only remeshing around the add-on devices in order to avoid unnecessary mesh dependency and to ease the evaluation process by reusing the existing mesh. We will present the application of an automatic local remeshing method to a few practical aircraft geometries to discuss its usefulness.
Jean-François Lague
Distene, Bruyeres-le-Chatel, France
Parallel mesh generation with mesh sizing control
Download Presentation
This talk will focus on the automatic meshing technologies being developed at Distene with an emphasis on our parallel approach for frontal Delaunay tetrahedral meshing on shared or distributed memory systems. An overview of this approach will be presented and some recent performance and robustness improvements will be discussed. We will also show how this approach can be applied to the parallel isotropic or anisotropic mesh generation and boundary layers meshing.
Download Presentation
This talk will focus on the automatic meshing technologies being developed at Distene with an emphasis on our parallel approach for frontal Delaunay tetrahedral meshing on shared or distributed memory systems. An overview of this approach will be presented and some recent performance and robustness improvements will be discussed. We will also show how this approach can be applied to the parallel isotropic or anisotropic mesh generation and boundary layers meshing.
Adrien Loseille
Inria Saclay Ile-de-France, Palaiseau, France
Some challenges in industrialising mesh adaptation : a focus on accuracy, stability, geometry and parallelism
Download Presentation
If anisotropic mesh adaptation has been an active field of research for more than a decade, it is still barely used in academia, industry or proposed in commercial packages. One major bottleneck in the industrialising process is the long list of components implied in the loop not fully related to core mesh generation: error estimates, numerical schemes, interpolation, … Each component is usually a field of research on its own and the required high level of expertise in each field seems to be an additional difficulty. The scope of this talk is to explore how the standard meshing pipeline, almost unchanged since the 90’s, can be modified to become industry-ready.
Download Presentation
If anisotropic mesh adaptation has been an active field of research for more than a decade, it is still barely used in academia, industry or proposed in commercial packages. One major bottleneck in the industrialising process is the long list of components implied in the loop not fully related to core mesh generation: error estimates, numerical schemes, interpolation, … Each component is usually a field of research on its own and the required high level of expertise in each field seems to be an additional difficulty. The scope of this talk is to explore how the standard meshing pipeline, almost unchanged since the 90’s, can be modified to become industry-ready.
Ed Luke
Mississippi State University, Starkville, MS, USA
Mesh quality evaluation for hybrid low dissipation flux schemes on unstructured meshes
Download Presentation
An economical technique for obtaining low dissipation production compressible flow solvers suitable for LES simulations can be found by hybridizing upwind schemes with skew symmetric central difference schemes. However, such hybrid schemes place more stringent requirements on mesh quality in order to control dispersion errors. We will show how to evaluate mesh quality using a linearized modified equation approach in order to better understand the mesh quality requirements of these hybrid approaches.
Download Presentation
An economical technique for obtaining low dissipation production compressible flow solvers suitable for LES simulations can be found by hybridizing upwind schemes with skew symmetric central difference schemes. However, such hybrid schemes place more stringent requirements on mesh quality in order to control dispersion errors. We will show how to evaluate mesh quality using a linearized modified equation approach in order to better understand the mesh quality requirements of these hybrid approaches.
Todd Michal
Boeing, Saint Louis, MO, USA
Development of an anisotropic solution adaptive meshing tool for production aerospace applications
Download Presentation
The Edge Primitive Insertion and Collapse (EPIC) tool was developed to provide a general solution adaptive meshing capability for Boeing Computational Fluid Dynamics (CFD) applications. EPIC utilizes a combination of edge and cavity based operators to drive a mesh into conformance with an anisotropic metric field derived from a solution error estimate. EPIC has been tightly integrated with Boeing’s unstructured grid CFD processes simplifying use on commercial, military, and space vehicle applications.
In this talk, an overview of EPIC will be presented and recent improvements in complex geometry modeling, parallel scalability, and grid convergence will be discussed. Some of the challenges of developing a production adaptive meshing tool for a wide range of flow solvers, geometry sources and applications will be reviewed. Recent adaptive meshing applications on workshop and Boeing production applications will be used to illustrate the current benefits and remaining challenges to use of adaptive meshing.
Download Presentation
The Edge Primitive Insertion and Collapse (EPIC) tool was developed to provide a general solution adaptive meshing capability for Boeing Computational Fluid Dynamics (CFD) applications. EPIC utilizes a combination of edge and cavity based operators to drive a mesh into conformance with an anisotropic metric field derived from a solution error estimate. EPIC has been tightly integrated with Boeing’s unstructured grid CFD processes simplifying use on commercial, military, and space vehicle applications.
In this talk, an overview of EPIC will be presented and recent improvements in complex geometry modeling, parallel scalability, and grid convergence will be discussed. Some of the challenges of developing a production adaptive meshing tool for a wide range of flow solvers, geometry sources and applications will be reviewed. Recent adaptive meshing applications on workshop and Boeing production applications will be used to illustrate the current benefits and remaining challenges to use of adaptive meshing.
Vincent Moureau
P. Bénard, G. Lartigue, R. Mercier, M. Cailler, A. Froehly, C. Dobrzynski
CORIA, Saint Etienne du Rouvray, France
Parallel and dynamic mesh adaptation of tetrahedral-based meshes for propagating fronts and interfaces: application to premixed combustion and primary atomization
Download Presentation
Thanks to the steady growth of computational resources and a large effort on solver optimization, Large-Eddy Simulation (LES) of realistic systems has become attainable. In these systems, the presence of turbulent multi-physics flows involves a large range of scales. Some of these scales need to be resolved by the mesh to capture the proper flow dynamics. Adaptive or dynamic mesh adaptation (AMR) is an appealing technique to reduce the modeling errors in LES of multi-physics flows. AMR of tetrahedral-based meshes is difficult to perform as it requires numerous mesh topology changes. It is even more challenging for LES as high-quality grids are required to resolve the turbulent scales that are close to the cut-off frequency of the mesh. A parallel AMR strategy has been developed recently (Bénard et al., IJNMF 2015) in the YALES2 flow solver (www.coria-cfd.fr). It combines adaptation and repartitioning steps to enable the AMR of massive grids counting billion cells exploiting up to tens of thousand cores. Mesh adaptation relies on the work of Dapogny et al. (JCP 2014) available in the MMG library (www.mmgtools.org). The presentation will focus on the parallel adaptation strategy for both volume and surface meshes, its optimization on modern super-computers and on various academic and industrial applications related to turbulent combustion and primary atomization.
Download Presentation
Thanks to the steady growth of computational resources and a large effort on solver optimization, Large-Eddy Simulation (LES) of realistic systems has become attainable. In these systems, the presence of turbulent multi-physics flows involves a large range of scales. Some of these scales need to be resolved by the mesh to capture the proper flow dynamics. Adaptive or dynamic mesh adaptation (AMR) is an appealing technique to reduce the modeling errors in LES of multi-physics flows. AMR of tetrahedral-based meshes is difficult to perform as it requires numerous mesh topology changes. It is even more challenging for LES as high-quality grids are required to resolve the turbulent scales that are close to the cut-off frequency of the mesh. A parallel AMR strategy has been developed recently (Bénard et al., IJNMF 2015) in the YALES2 flow solver (www.coria-cfd.fr). It combines adaptation and repartitioning steps to enable the AMR of massive grids counting billion cells exploiting up to tens of thousand cores. Mesh adaptation relies on the work of Dapogny et al. (JCP 2014) available in the MMG library (www.mmgtools.org). The presentation will focus on the parallel adaptation strategy for both volume and surface meshes, its optimization on modern super-computers and on various academic and industrial applications related to turbulent combustion and primary atomization.
Mike Park
NASA Langley, Hampton VA, USA
Democratizing Mesh Adaptation for Computational Fluid Dynamics
Download Presentation
Collaboration fostered via the informal Unstructured Grid Adaptation Working Group has enabled the development and verification of new and existing mesh adaptation processes. These integrated methods automatically control the interpolation error of Computational Fluid Dynamics (CFD) solutions with initial to final grids conforming to the geometry. The combination of restricted-release and open-source tools that interact through Application Programming Interfaces and open file formats enables verification of analytic solutions. Code-to-code comparisons of integrated process and their components to verify correctness of implementations and support the confidence of CFD practitioners.
Download Presentation
Collaboration fostered via the informal Unstructured Grid Adaptation Working Group has enabled the development and verification of new and existing mesh adaptation processes. These integrated methods automatically control the interpolation error of Computational Fluid Dynamics (CFD) solutions with initial to final grids conforming to the geometry. The combination of restricted-release and open-source tools that interact through Application Programming Interfaces and open file formats enables verification of analytic solutions. Code-to-code comparisons of integrated process and their components to verify correctness of implementations and support the confidence of CFD practitioners.
Jeanne Pellerin
Total, La Défense, France
Mesh generation: a transverse challenge
Download Presentation
Today, most numerical simulations rely on meshes and the generation of these meshes still remain an important bottleneck for our engineers. Indeed, mesh generation needs completely depends on the application. In this talk, I will present a transverse overview of the meshing needs in Total and will detail the challenges more specifically encountered in subsurface modeling.
Download Presentation
Today, most numerical simulations rely on meshes and the generation of these meshes still remain an important bottleneck for our engineers. Indeed, mesh generation needs completely depends on the application. In this talk, I will present a transverse overview of the meshing needs in Total and will detail the challenges more specifically encountered in subsurface modeling.
Per-Olof Persson
University of California, Berkeley, CA, USA
Optimization-based PDE-constrained discontinuity tracking with high-order curved unstructured meshes
Download Presentation
We present a framework for generating curved moving meshes which align the element faces with discontinuities in an evolving solution. The solution comes from solving a conservation law discretized using the discontinuous Galerkin (DG) method and an arbitrary Lagrangian-Eulerian formulation. The meshes are evolved by solving an optimization problem, where a carefully chosen indicator penalizes misaligned faces. Our discontinuity indicator monotonically approaches a minimum as element faces approach the discontinuity surface, which allows for efficient gradient-based optimizers. We also include a mesh skewness measure to ensure the meshes are well-shaped. For problems with large deformations, we use local element topology changes such as edge flips on the curved elements to improve the mesh qualities. We demonstrate our methods on a number of problems with moving discontinuities, such as convection problems and flow problems with shocks.
Download Presentation
We present a framework for generating curved moving meshes which align the element faces with discontinuities in an evolving solution. The solution comes from solving a conservation law discretized using the discontinuous Galerkin (DG) method and an arbitrary Lagrangian-Eulerian formulation. The meshes are evolved by solving an optimization problem, where a carefully chosen indicator penalizes misaligned faces. Our discontinuity indicator monotonically approaches a minimum as element faces approach the discontinuity surface, which allows for efficient gradient-based optimizers. We also include a mesh skewness measure to ensure the meshes are well-shaped. For problems with large deformations, we use local element topology changes such as edge flips on the curved elements to improve the mesh qualities. We demonstrate our methods on a number of problems with moving discontinuities, such as convection problems and flow problems with shocks.
Jean-François Remacle
Université Catholique de Louvain, Louvain, Belgium
Finding Hexahedrizations/Tetrahedrizations for Small Quadrangulations/Triangulations of the Sphere
This talk tackles the challenging problem of constrained hexahedral and tetrahedral meshing. Starting with a quadrangulation/triangulation of the 2-sphere, our goal is to build combinatorial hexahedral/tetrahedral meshes whose boundary facets exactly match the surface mesh [1].
At first, we show that our approach is able to compute small hexahedral meshes of quadrangulations for which the previously known best solutions could only be built by hand or contained thousands of hexahedra. These challenging quadrangulations include the boundaries of transition templates that are critical for the success of general hexahedral meshing algorithms.
Then, we will show that the same kind of combinatorial approach, combined with geometry-based filtering, can be used to find all tetrahedrizations of a given ball-shaped cavity. This algorithm will be used for both (i) recovering missing edges and faces in a tetrahedrization (boundary recovery) and (ii) optimizing tetrahedral meshes.
[1] Verhetsel, K., Pellerin, J., & Remacle, J. F. (2019). Finding Hexahedrizations for Small Quadrangulations of the Sphere. arXiv preprint arXiv:1904.11229, accepted for SIGGRAPH 2019.
This talk tackles the challenging problem of constrained hexahedral and tetrahedral meshing. Starting with a quadrangulation/triangulation of the 2-sphere, our goal is to build combinatorial hexahedral/tetrahedral meshes whose boundary facets exactly match the surface mesh [1].
At first, we show that our approach is able to compute small hexahedral meshes of quadrangulations for which the previously known best solutions could only be built by hand or contained thousands of hexahedra. These challenging quadrangulations include the boundaries of transition templates that are critical for the success of general hexahedral meshing algorithms.
Then, we will show that the same kind of combinatorial approach, combined with geometry-based filtering, can be used to find all tetrahedrizations of a given ball-shaped cavity. This algorithm will be used for both (i) recovering missing edges and faces in a tetrahedrization (boundary recovery) and (ii) optimizing tetrahedral meshes.
[1] Verhetsel, K., Pellerin, J., & Remacle, J. F. (2019). Finding Hexahedrizations for Small Quadrangulations of the Sphere. arXiv preprint arXiv:1904.11229, accepted for SIGGRAPH 2019.
Xevi Roca
Barcelona Supercomputing Center, Barcelona, Spain
Curved meshing: approximation and accommodation of curved features
Download Presentation
To exploit the promising features of unstructured high-order methods, practitioners generate meshes composed by curved elements that interpolate the domain boundaries, and that accommodate the boundary curvature to the interior of the mesh while untangling the initially invalid elements. This approach is the standard solution to the problem referred to as mesh curving, and it has been the main driver in the research on generating curved meshes. This talk summarizes our research to solve the mesh curving problem, mainly applied to inviscid and viscous flow simulation, but also the evolution towards to solve a more general curved meshing problem. Specifically, the talk details our efforts in the problem of approximating, not only interpolating, those salient curved features, determined by the domain boundary but also by the solution, while iteratively accommodating the curvature of the features to the rest of the mesh without introducing invalid elements. The talk also highlights our past, current, and future vision of the challenges in curved meshing.
Download Presentation
To exploit the promising features of unstructured high-order methods, practitioners generate meshes composed by curved elements that interpolate the domain boundaries, and that accommodate the boundary curvature to the interior of the mesh while untangling the initially invalid elements. This approach is the standard solution to the problem referred to as mesh curving, and it has been the main driver in the research on generating curved meshes. This talk summarizes our research to solve the mesh curving problem, mainly applied to inviscid and viscous flow simulation, but also the evolution towards to solve a more general curved meshing problem. Specifically, the talk details our efforts in the problem of approximating, not only interpolating, those salient curved features, determined by the domain boundary but also by the solution, while iteratively accommodating the curvature of the features to the rest of the mesh without introducing invalid elements. The talk also highlights our past, current, and future vision of the challenges in curved meshing.
Rubén Sevilla
Swansea University, Swansea, Wales
A feature-independent mesh generation and integrated solution framework
Download Presentation
Although computational simulation is extensively employed in industry, its wider use is limited by the complexity of the geometric models involved. This limitation is due to the excessive number of human hours, ranging from days to months, required to transfer information from a computer aided design (CAD) model to a computer aided engineering (CAE) model suitable for simulation. CAD models frequently involve a level of detail much greater than that required to perform a computational simulation with a CAE system.
This talk will introduce a new computational environment that includes a feature-independent mesh generation paradigm and the numerical techniques required to integrate such meshes in a finite element type solver. The main advantage of the proposed mesh generation technique is that the new meshes will completely remove the uncertainty introduced by defeaturing CAD models. Instead of relying on the opinion of experts, to decide which features might not be relevant in a simulation, the CAD model will not be altered, leading to higher fidelity simulations and more confidence in the results.
Download Presentation
Although computational simulation is extensively employed in industry, its wider use is limited by the complexity of the geometric models involved. This limitation is due to the excessive number of human hours, ranging from days to months, required to transfer information from a computer aided design (CAD) model to a computer aided engineering (CAE) model suitable for simulation. CAD models frequently involve a level of detail much greater than that required to perform a computational simulation with a CAE system.
This talk will introduce a new computational environment that includes a feature-independent mesh generation paradigm and the numerical techniques required to integrate such meshes in a finite element type solver. The main advantage of the proposed mesh generation technique is that the new meshes will completely remove the uncertainty introduced by defeaturing CAD models. Instead of relying on the opinion of experts, to decide which features might not be relevant in a simulation, the CAD model will not be altered, leading to higher fidelity simulations and more confidence in the results.
Alexander Skovpen and Nicolas Delsate
NUMECA International, Bruxelles, Belgium
Surface recovery inside hex-dominant mesh using Imprint method
Download Presentation
The proposed method recovers the triangulated surface inside the existing basic volume mesh. The basic mesh could be tetrahedral or hex-dominant with the presence of hexahedrons, tetrahedra, prisms and pyramids. The initial and resulting meshes are conformal. To provide smooth size propagation in final mesh the basic mesh could be created with adaptation to the local linear size of the imprinted surface. The method uses the division of the basic mesh polyhedrons into tetrahedra; optimization of the tetrahedral mesh in the proximity of the imprinted surface; Delaunay based nodes insertion and faces recovery similar to the constraining step in tetrahedral meshing.
Download Presentation
The proposed method recovers the triangulated surface inside the existing basic volume mesh. The basic mesh could be tetrahedral or hex-dominant with the presence of hexahedrons, tetrahedra, prisms and pyramids. The initial and resulting meshes are conformal. To provide smooth size propagation in final mesh the basic mesh could be created with adaptation to the local linear size of the imprinted surface. The method uses the division of the basic mesh polyhedrons into tetrahedra; optimization of the tetrahedral mesh in the proximity of the imprinted surface; Delaunay based nodes insertion and faces recovery similar to the constraining step in tetrahedral meshing.
Carolyn Woeber
Pointwise, Fort Worth, TX, USA
A Shared Vision for Mesh Generation
Download Presentation
Contributors to the CFD community have been developing mesh generation and CFD software side-by-side from the beginning. CFD solvers have been developed toward numerically solving some form of the Navier-Stokes equations. Mesh generation algorithms and techniques have evolved in reaction to CFD users’ practical needs for high-quality meshes suitable for specific aerodynamic applications and solvers. Over the years, this applied focus has contributed to the creation of roadblocks in mesh generation processes.
The NASA CFD Vision 2030 Study (hereinafter the Study) noted these meshing roadblocks were some of the primary contributors to the costliness of current CFD simulations. The Study challenged the community to reduce the required level of human interaction, increase mesh scalability (both desktop and HPC), leverage mesh adaptation, and invest in other high-risk (potentially) disruptive new meshing ideas/technologies.
A paradigm shift in how we think about, develop, and interact with mesh generation in the CFD ecosystem is necessary to move forward and achieve these goals. This talk will focus on a number of meshing technologies being developed for Pointwise and how they align with the CFD community’s vision for the future of mesh generation.
Download Presentation
Contributors to the CFD community have been developing mesh generation and CFD software side-by-side from the beginning. CFD solvers have been developed toward numerically solving some form of the Navier-Stokes equations. Mesh generation algorithms and techniques have evolved in reaction to CFD users’ practical needs for high-quality meshes suitable for specific aerodynamic applications and solvers. Over the years, this applied focus has contributed to the creation of roadblocks in mesh generation processes.
The NASA CFD Vision 2030 Study (hereinafter the Study) noted these meshing roadblocks were some of the primary contributors to the costliness of current CFD simulations. The Study challenged the community to reduce the required level of human interaction, increase mesh scalability (both desktop and HPC), leverage mesh adaptation, and invest in other high-risk (potentially) disruptive new meshing ideas/technologies.
A paradigm shift in how we think about, develop, and interact with mesh generation in the CFD ecosystem is necessary to move forward and achieve these goals. This talk will focus on a number of meshing technologies being developed for Pointwise and how they align with the CFD community’s vision for the future of mesh generation.
Sponsors
