Title: Geometric Programming with Voronoi Diagrams - Efficiency, Robustness and Applications
Biography: Bruno Levy is a senior researcher with Inria Nancy-Grand Est, and a member of the LORIA lab. He defended his Ph.D. thesis in 1999 and did a post-doc in Stanford. He was hired in 2000 by Inria. He is currently the head of the ALICE research team, that he created in 2004. He received the Inria young researcher award in 2011.
His main research topic is numerical geometry, that is to say mathematical algorithms for acquiring, transforming and optimizing the representation of 3D shapes. He developed several algorithms for geometry processing and mesh generation, such as Least Squares Conformal Maps, used to generate (u,v) texture mapping coordinates in several softwares, Manifold Harmonics, a Fourier-like spectral mesh analysis algorithm, and more recently, Voronoi Parallel Linear Enumeration, an algorithm for re-meshing surfaces and volumes.
Abstract: In this talk, I will focus on Voronoi diagrams, present efficient algorithms to compute them and demonstrate some applications. For instance, some mesh generation algorithms require to compute the intersection between a Voronoi diagram and another mesh. To obtain a both efficient and robust implementation of this operation, one needs to carefully consider the geometrical, combinatorial and numerical aspects of the problem. For the geometrical/combinatorial part of the problem, I will show how a basic consideration on the Voronoi cells leads to a simple yet efficient algorithm. As far as the numerical part of the problem is concerned, for the possibly degenerate cases, I will explain how to "push" all the difficulties towards the geometric predicates, and how to implement these predicates with a combination of arithmetic filters, exact arithmetics and symbolic perturbation.
The resulting algorithm computes the intersection between a nD voronoi diagram and a volumetric or a surfacic mesh, and can be used in various settings, comprising surface re-meshing, anisotropic mesh generation and computation of optimal transport maps in 3d.
The presentation will be illustrated by live demos. Some source-code is available in the Graphite software. The automatic code generator for predicates PCK (Predicate Construction Kit) is available in the (upcoming) Geogram programming library.
Title: Fifteen years of progress on anisotropic mesh adaptation for CFD and remaining challenges
Biography: Frederic Alauzet is a senior researcher with INRIA Paris-Rocquencourt and member of the Gamma3 Team. He received is PhD thesis in 2003, did a post-doc at Rensselaer Polytechnic Institute, NY. He was hired by INRIA in 2004 and was visiting Professor at Mississippi State University, MS, between 2012 and 2013. His main research topic is advanced meshing methods for enhanced numerical simulations. His major focus is anisotropic mesh adaptation for CFD with contributions to extreme anisotropy, operators on metric fields, parallelization, application to sonic boom problems, time-dependent problems, interface flows problems, ... These progresses are mainly based on the continuous mesh framework. More recently, he proposed new contributions to moving mesh methods to handle robustly and efficiently large displacement of complex geometry.
Abstract: After introducing the scientific context, the mathematical framework for anisotropic mesh adaptation will be recalled and I will introduce the continuous mesh framework and its application. Issues and progresses of anisotropic mesh adaptation will be discussed in the context of steady flows applications. I will particularly focus on feature-based and goal-oriented methods, extreme anisotropy and parallelization of the mesh adaptation loop. Then, I will give the problematics and solutions to apply anisotropic mesh adaptation to time-depend problems. In this context, space-time analysis is of main concern in order to control temporal errors. I will also point out that mesh adaptation is a way to certify numerical simulations. Finally, the remaining challenges for the next decade will be exposed. They concern the use of adapted and anisotropic meshes for moving mesh problems, high-order (curved) meshes, pseudo-structured mesh adaptation for viscous flows.
Title: Advances in high order spectral/hp element methods for high Reynolds number complex geometry flows
Prof. Spencer Sherwin Biography: Spencer Sherwin is the McLaren Racing/Royal Academy of Engineering Research Chair in the Department of Aeronautics at Imperial College London. He received his MSE and PhD from the Department of Mechanical and Aerospace Engineering Department at Princeton University. During his time at Imperial he has maintained a successful research program into the development and application of the high order spectral/hp element techniques with particular application to biomedical flow, separated unsteady aerodynamics and understanding flow physics through instability analysis. Professor Sherwin’s research group (www.sherwinlab.info) also develops and distributes the openware spectral/hp element package nektar++ which has been applied to direct numerical simulation and stability analysis to a range of applications including Biomedical Flows and separated Bluff Bodies and Vortex Flows of relevance to offshore engineering and vehicle aerodynamics. He has published over 120 peer-reviewed papers in International Journals covering topics from numerical analysis to fundamental fluid-mechanics and biomedical flow modeling and co-authored a highly cited book on the underlying spectral/hp element methods. Currently he is an associate director EPSRC/EADS funded Laminar Flow Control Centre and is the chair of the EPSRC Platform for Research in Simulation Methods (PRISM) at Imperial College London (www.prism.ac.uk).
Dr. Joaquim Peiro Biography: Dr. Joaquim Peiro is a Senior Lecturer in the Department of Aeronautics. Since 1985 he has worked on research and development of CAD geometry modeling, automatic unstructured mesh generators and CFD solvers using linear and high-order elements. His research is focused on developing geometrical methods to interpret fluid flow features and behaviour, to determine how geometry influences fluid flow development, and the development of automatic procedures for modelling fluid-structure interaction. This includes automatic mesh generation for in vivo geometries reconstructed from medical images and high-order algorithms for compressible steady and transient flows in aeronautics and haemodynamics. He is one of the developers of the FELISA system for the numerical simulation of inviscid compressible flows using unstructured meshes funded by NASA. A current interest is the development and implementation of high-order techniques for mesh generation within the p-mesh framework in the open-ware spectral/hp element package Nektar++. He has more than 50 journal articles with citation data and they have been cited more than 1000 times. He is a member of the Editorial Board of International Journal of Numerical Methods in Biomedical Engineering.
Abstract: In this presentation we will discuss the challenges of performing high Reynolds number simulations using spectral/hp element discretisations. The spectral/hp element methods combines the good phase and dispersion properties of spectral methods with the geometric flexibility of finite element methods and can be considered as a high order finite element technique. Whilst these properties are attractive from a numerical discretisation a number of challenges exist in apply these methods to high Reynolds number complex geometry problems.
Mesh generation is a clearly an enabling technology for complex geometries, in particular those of interest to the aeronautical and automotive industry. Whilst generation of arbitrary high-order meshes is a viable goal the lack of robust mesh generators is still a significant bottleneck. In this presentation we will review our work on the generation of boundary conforming meshes of high-order elements that incorporates curvature driven discretisation to account for higher derivatives of surface and preserve hp-convergence. We will also discuss how to incorporate surface and volume mesh optimisation to reduce element deformation and improves solution accuracy. We will also present a novel transfinite interpolation technique for generating the highly stretched boundary-layer meshes required for the simulation of high Reynolds number flows near solid surfaces as shown in sub figure (a).
Finally we will further discuss the role of polynomial de-aliasing and spectral vanishing viscosity to stabilise high Reynolds number transient flows as shown in figure 1(b).
Figure: (a) high-order near wall meshing using curved prismatic elements that maintain positive of the mapping, (b) flow over a Naca 0012 wing time at R=1.2M using a six order accurate spectral/hp element discretisation.
Title: CFD Meshing Challenges in Aerospace: It's About More Than Just Lift and Drag
Biography: Deryl Snyder is the Director of Aerospace and Defense at CD-adapco, globally the largest privately-owned CFD/CAE software and service provider. Deryl has over 15 years of computational fluid dynamics expertise related to the aerospace and defense industry. He received a Ph.D. jointly from Utah State University in the U.S. and the von Karman Institute for Fluid Dynamics in Belgium, specializing in numerical algorithms for CFD. He has worked as an engineering support contractor for the Munitions Directorate of the US Air Force Research Laboratory at Eglin Air Force Base solving aerodynamic technical issues for various missile systems and tactical UAVs. Deryl has also been a faculty member in the Mechanical Engineering Department at Brigham Young University, specializing in computational methods in the thermal/fluid sciences as well as small- and micro-UAV development. He led the CFD efforts in the Aerodynamics Center of Excellence at Lockheed Martin Missiles and Fire control from 2007 to 2011, where he was responsible for overseeing numerical analysis methods, procedures, practices, and tools. Since 2011, Deryl has overseen the aerospace and defense business at CD-adapco, working with customers, the product-delivery team, and developers to address simulation-related challenges in the industry.
Abstract: Computational fluid dynamics (CFD) meshing challenges in aerospace come in all sizes and shapes. Aerodynamics has long been the bread and butter of aerospace CFD, and has posed meshing challenges since the first time the Navier-Stokes equations were numerically solved on a 2d airfoil. Yet, other applications push the capability envelope even more in terms of geometric complexity and cell count. From block-structured to unstructured tetrahedron and now general polyhedron, meshing technologies have evolved, allowing great improvements in geometric fidelity and productivity. And still, ask any practitioner and you'll find key challenges still remain.
This presentation will introduce examples, challenges, and even a few solutions related to meshing for CFD in aerospace. Application areas discussed include aerodynamics, thermal management systems, and aero acoustics. Through application examples, I will highlight meshing techniques and technologies that practitioners utilize in a production setting, with a focus on modern unstructured meshing topologies. Practical challenges, strengths and weaknesses of these meshing approaches will be discussed based on industry use.
Title: Industrial Perspectives on Geometry Handling for Aerodynamics
Biography: Nigel graduated from the University of Southampton with a first-class honours degree in Aeronautics & Astronautics in 1987. Following a period of integrated experimental and computational research leading to the award of a PhD, he joined the Aerodynamics Department at DRA Farnborough in 1993. Here, his principal areas of work were in the design and assessment of high-lift/manoeuvre devices for combat aircraft and wind tunnel technique development. In 1999, he joined the Aerodynamics Group at MBDA Filton and became a Technical Expert for Missile Aerodynamics in 2004. He assumed his current role as Capability Leader, Aerodynamic Tools & Methods in 2006. Since 2003, Nigel has represented MBDA and the UK weapon aerodynamics community in various national fora, including the Aerodynamics National Advisory Committee, the Aerodynamics National Technical Committee and the Council of the UK Centre for Aerodynamics. He is also currently a member of the AIAA Meshing, Visualisation & Computational Environments Technical Committee.
Abstract: One of the most fundamental properties affecting the aerodynamic performance of a body is its shape. With progressively increasing demands for performance, the need to explore and optimise the performance of novel airframe shapes rapidly and with robust, efficient processes is becoming increasingly important. This poses significant challenges for the ways in which the associated geometry is generated and manipulated (in support of design) both on its wetted surfaces and in the adjacent air flow (i.e. the computational mesh). This presentation will review the processes associated with handling geometry to support industrial aerodynamic analyses – covering its receipt and preparation for use in analysis, though to its onward delivery - and will explain how they are influenced by the stage in the product lifecycle in which the analysis is being undertaken. Examples from the aircraft, turbo-machinery and missile sectors will be provided to illustrate key technology aspects. Insight will also be provided as to how current practice may develop in the future.