Guillaume Oger

SPH accuracy improvement through the combination of a quasi-Lagrangian shifting transport velocity and consistent ALE formalisms

This paper addresses the accuracy of the weakly-compressible SPH method. Interpolation defects due to the presence of anisotropic particle structures inherent to the Lagrangian character of the Smoothed Particle Hydrodynamics (SPH) method are highlighted. To avoid the appearance of these structures which are detrimental to the quality of the simulations, a specific transport velocity is introduced and its inclusion within an Arbitrary Lagrangian Eulerian (ALE) formalism is described. Unlike most of existing particle disordering/shifting methods, this formalism avoids the formation of these anisotropic structures while a full consistency with the original Euler or Navier-Stokes equations is maintained. The gain in accuracy, convergence and numerical diffusion of this formalism is shown and discussed through its application to various challenging test cases.

On distributed memory MPI-based parallelization of SPH codes in massive HPC context

Abstract Most of particle methods share the problem of high computational cost and in order to satisfy the demands of solvers, currently available hardware technologies must be fully exploited. Two complementary technologies are now accessible. On the one hand, CPUs which can be structured into a multi-node framework, allowing massive data exchanges through a high speed network. In this case, each node is usually comprised of several cores available to perform multithreaded computations. On the other hand, GPUs which are derived from the graphics computing technologies, able to perform highly multi-threaded calculations with hundreds of independent threads connected together through a common shared memory. This paper is primarily dedicated to the distributed memory parallelization of particle methods, targeting several thousands of CPU cores. The experience gained clearly shows that parallelizing a particle-based code on moderate numbers of cores can easily lead to an acceptable scalability, whilst a scalable speedup on thousands of cores is much more difficult to obtain. The discussion revolves around speeding up particle methods as a whole, in a massive HPC context by making use of the MPI library. We focus on one particular particle method which is Smoothed Particle Hydrodynamics (SPH), one of the most widespread today in the literature as well as in engineering.

Analysis and improvements of Adaptive Particle Refinement (APR) through CPU time, accuracy and robustness considerations

Abstract While smoothed-particle hydrodynamics (SPH) simulations are usually performed using uniform particle distributions, local particle refinement techniques have been developed to concentrate fine spatial resolutions in identified areas of interest. Although the formalism of this method is relatively easy to implement, its robustness at coarse/fine interfaces can be problematic. Analysis performed in [16] shows that the radius of refined particles should be greater than half the radius of unrefined particles to ensure robustness. In this article, the basics of an Adaptive Particle Refinement (APR) technique, inspired by AMR in mesh-based methods, are presented. This approach ensures robustness with alleviated constraints. Simulations applying the new formalism proposed achieve accuracy comparable to fully refined spatial resolutions, together with robustness, low CPU times and maintained parallel efficiency.

An efficient FSI coupling strategy between Smoothed Particle Hydrodynamics and Finite Element methods

Abstract An efficient coupling between Smoothed Particle Hydrodynamics (SPH) and Finite Element (FE) methods dedicated to violent fluid–structure interaction (FSI) modeling is proposed in this study. The use of a Lagrangian meshless method for the fluid reduces the complexity of fluid–structure interface handling, especially in presence of complex free surface flows. The paper details the discrete SPH equations and the FSI coupling strategy adopted. Both convergence and robustness of the SPH-FE coupling are performed and discussed. More particularly, the loss and gain in stability is studied according to various coupling parameters, and different coupling algorithms are considered. Investigations are performed on 2D academic and experimental test cases in the order of increasing complexity.

A weakly-compressible Cartesian grid approach for hydrodynamic flows

Abstract The present article aims at proposing an original strategy to solve hydrodynamic flows. In introduction, the motivations for this strategy are developed. It aims at modeling viscous and turbulent flows including complex moving geometries, while avoiding meshing constraints. The proposed approach relies on a weakly-compressible formulation of the Navier–Stokes equations. Unlike most hydrodynamic CFD (Computational Fluid Dynamics) solvers usually based on implicit incompressible formulations, a fully-explicit temporal scheme is used. A purely Cartesian grid is adopted for numerical accuracy and algorithmic simplicity purposes. This characteristic allows an easy use of Adaptive Mesh Refinement (AMR) methods embedded within a massively parallel framework. Geometries are automatically immersed within the Cartesian grid with an AMR compatible treatment. The method proposed uses an Immersed Boundary Method (IBM) adapted to the weakly-compressible formalism and imposed smoothly through a regularization function, which stands as another originality of this work. All these features have been implemented within an in-house solver based on this WCCH (Weakly-Compressible Cartesian Hydrodynamic) method which meets the above requirements whilst allowing the use of high-order ( > 3 ) spatial schemes rarely used in existing hydrodynamic solvers. The details of this WCCH method are presented and validated in this article.

A Cartesian Explicit Solver for Complex Hydrodynamic Applications

In order to efficiently address complex problems in hydrodynamics, the advances in the development of a new method are presented here. This method aims at finding a good compromise between computational efficiency, accuracy, and easy handling of complex geometries. The chosen method is an Explicit Cartesian Finite Volume method for Hydrodynamics (ECFVH) based on a compressible (hyperbolic) solver, with a ghost-cell method for geometry handling and a Level-set method for the treatment of biphase-flows. The explicit nature of the solver is obtained through a weakly-compressible approach chosen to simulate nearly-incompressible flows. The explicit cell-centered resolution allows for an efficient solving of very large simulations together with a straightforward handling of multi-physics. A characteristic flux method for solving the hyperbolic part of the Navier-Stokes equations is used. The treatment of arbitrary geometries is addressed in the hyperbolic and viscous framework. Viscous effects are computed via a finite difference computation of viscous fluxes and turbulent effects are addressed via a Large-Eddy Simulation method (LES). The Level-Set solver used to handle biphase flows is also presented. The solver is validated on 2-D test cases (flow past a cylinder, 2-D dam break) and future improvements are discussed.Copyright

Development of a New Cartesian Explicit Solver for Hydrodynamics Flows

In order to solve complex problems in hydrodynamics, a new method is developed. This method aims at finding a good compromise between computational efficiency, accuracy, and easy handling of complex geometries. The chosen method is an Explicit Cartesian Finite Volume method for Hydrodynamics (ECFVH) based on a compressible (hyperbolic) solver, with an embedded method for interfaces and geometry handling. The explicit nature of the solver is obtained through a weakly-compressible approach chosen to simulate nearly-incompressible flows. The explicit cell-centered resolution allows for an efficient solving of very large simulations together with a straightforward handling of multi-physics. The use of an embedded Cartesian grid ensures accuracy and efficiency, but also implies the need for a specific treatment of complex solid geometries, such as the cut-cell method in the fixed or moving body frame. Robustness of the cut-cell method is ensured by specific procedures to circumvent small cell volume numerical errors. A characteristic flux method for solving the hyperbolic part of the Navier-Stokes equations is used for which upwinding is necessary, also introducing numerical viscosity. This numerical viscosity is evaluated before trying to model viscous and turbulent effects. In a first approach viscous effects are computed via a finite difference Laplacian operator introduced as a source term. This solver is validated on 2-D test cases and future improvements are discussed.Copyright

Next-generation Multi-mechanics Simulation Engine in a Highly Interactive Environment

We describe the development of a highly interactive approach to simulation of engineering multi-mechanics problems, using the smoothed particle hydrodynamics mesh-free method as the computational engine, for applications including ship survival, medical devices and Pelton turbines.

Fast and accurate SPH modelling of 3D complex wall boundaries in viscous and non viscous flows

Abstract The treatment of wall boundary conditions is a difficult issue in the SPH method and still represents a challenging topic in the scientific community. After a review of state of the art wall treatment methods, an SPH method for modelling viscous and non-viscous flows in the presence of 3D complex wall boundaries is presented. New developments embedded in the proposed method include the addition of a Laplacian operator adapted to the adopted formalism, as well as a cutface process for calculating the particle/wall interactions on any type of geometry. Validations are proposed on a 2D Poiseuille flow, a flow around a 2D cylinder, a 3D hydrostatic tank, and a 3D dambreak. Comparisons are performed with results from the literature. Finally, an industrial automotive application is presented as illustration of the method ability to deal with arbitrarily complex geometries.

A Coupled SPH-Spectral Method for the Simulation of Wave Train Impacts on a FPSO

In order to help in achieving a correct design of structures subjected to wave impacts, CFD tools with a sufficient accuracy should be developed. But nowadays, modelling accurately both wave propagations and the resulting impact of a wave train on a complex geometry is still challenging. This paper deals with the introduction of a weak coupling between a Spectral method and Smoothed Particle Hydrodynamics (SPH), used as complementary techniques for modelling respectively wave propagation and high-dynamic impact of a wave train on a complex-shaped floating body. Comparisons with experiments are provided as a validation of these preliminary developments.Copyright