Journal of Computational Physics: X最新文献

Essentially nonoscillatory (ENO) and weighted ENO (WENO) methods on equidistant Cartesian grids are widely used to solve partial differential equations with discontinuous solutions. The RBF-ENO method is highly flexible in terms of geometry, but its stencil selection algorithm is computational expensive. In this work, we combine the computationally efficient WENO method and the geometrically flexible RBF-ENO method in a hybrid high-resolution essentially nonoscillatory method to solve hyperbolic conservation laws. The scheme is based on overlapping patches with ghost cells, the RBF-ENO method for unstructured patches and a standard WENO method on structured patches. Furthermore, we introduce a positivity preserving limiter for non-polynomial reconstruction methods to stabilize the hybrid RBF-ENO method for problems with low density or pressure. We show its robustness and flexibility on benchmarks and complex test cases such as the scramjet inflow problem and a conical aerospike nozzle jet simulation.

A straightforward and computationally efficient Consecutive Cubic Spline (CCS) iterative algorithm is proposed for positioning the planar interface of the unstructured geometrical Volume-of-Fluid method in arbitrarily-shaped cells. The CCS algorithm is a two-point root-finding algorithm [1, chap. 2], designed for the VOF interface positioning problem, where the volume fraction function has diminishing derivatives at the ends of the search interval. As a two-point iterative algorithm, CCS re-uses function values and derivatives from previous iterations and does not rely on interval bracketing. The CCS algorithm requires only two iterations on average to position the interface with a tolerance of ${10}^{-12}$, even with numerically very challenging volume fraction values, e.g., near ${10}^{-9}$ or $1-{10}^{-9}$.

The proposed CCS algorithm is very straightforward to implement because its input is already calculated by every geometrical VOF method. It builds upon and significantly improves the predictive Newton method [2] and is independent of the cell's geometrical model and related intersection algorithm. Geometrical parameterizations of truncated volumes used by other contemporary methods [3], [4], [5], [6] are completely avoided. The computational efficiency is comparable in terms of the number of iterations to the fastest methods reported so far. References are provided in the results section to the open-source implementation of the CCS algorithm and the performance measurement data.

In ice sheet and glacier modelling, the Finite Element Method is rapidly gaining popularity. However, constructing and updating meshes for ice sheets and glaciers is a non-trivial and computationally demanding task due to their thin, irregular, and time dependent geometry. In this paper we introduce a novel approach to ice dynamics computations based on the unfitted Finite Element Method CutFEM, which lets the domain boundary cut through elements. By employing CutFEM, complex meshing and remeshing is avoided as the glacier can be immersed in a simple background mesh without loss of accuracy. The ice is modelled as a non-Newtonian, shear-thinning fluid obeying the p-Stokes (full Stokes) equations with the ice atmosphere interface as a moving free surface. A Navier slip boundary condition applies at the glacier base allowing both bedrock and subglacial lakes to be represented. Within the CutFEM framework we develop a strategy for handling non-linear viscosities and thin domains and show how glacier deformation can be modelled using a level set function. In numerical experiments we show that the expected order of accuracy is achieved and that the method is robust with respect to penalty parameters. As an application we compute the velocity field of the Swiss mountain glacier Haut Glacier d'Arolla in 2D with and without an underlying subglacial lake, and simulate the glacier deformation from year 1930 to 1932, with and without surface accumulation and basal melt.

Discrete unified gas-kinetic scheme (DUGKS) has been developed recently as a general method for simulating flows at all Knudsen numbers. In this study, we extend DUGKS to simulate fully compressible thermal flows. We introduce a source term to the Boltzmann equation with the Bhatnagar-Gross-Krook (BGK) collision model [1] to adjust heat flux and thus the Prandtl number. The fully compressible Navier-Stokes equations can be recovered by the current model. As a mesoscopic CFD approach, it requires an accurate mesoscopic implementation of the boundary conditions. Using the Chapman-Enskog approximation, we derive the “bounce-back” expressions for both temperature and velocity distribution functions, which reveal the need to consider coupling terms between the velocity and thermal fields. To validate our scheme, we first reproduce the Boussinesq flow results by simulating natural convection in a square cavity with a small temperature difference $(ϵ=0.01)$ and a low Mach number. Then we perform simulations of steady natural convection $(Ra=1.0\times {10}^6)$ in a square cavity with differentially heated side walls and a large temperature difference $(ϵ=0.6)$, where the Boussinesq approximation becomes invalid. Temperature, velocity profiles, and Nusselt number distribution are obtained and compared with the benchmark results from the literature. Finally, the unsteady compressible natural convection with $Ra=5.0\times {10}^9,ϵ=0.6$ is studied and the turbulent fluctuation statistics are computed and analyzed.

We propose a sparse grid-based adaptive noise reduction strategy for electrostatic particle-in-cell (PIC) simulations. By projecting the charge density onto sparse grids we reduce the high-frequency particle noise. Thus, we exploit the ability of sparse grids to act as a multidimensional low-pass filter in our approach. Thanks to the truncated combination technique [1], [2], [3], we can reduce the larger grid-based error of the standard sparse grid approach for non-aligned and non-smooth functions. The truncated approach also provides a natural framework for minimizing the sum of grid-based and particle-based errors in the charge density. We show that our approach is, in fact, a filtering perspective for the noise reduction obtained with the sparse PIC schemes first introduced in [4]. This enables us to propose a heuristic based on the formal error analysis in [4] for selecting the optimal truncation parameter that minimizes the total error in charge density at each time step. Hence, unlike the physical and Fourier domain filters typically used in PIC codes for noise reduction, our approach automatically adapts to the mesh size, number of particles per cell, smoothness of the density profile and the initial sampling technique. It can also be easily integrated into high performance large-scale PIC code bases, because we only use sparse grids for filtering the charge density. All other operations remain on the regular grid, as in typical PIC codes. We demonstrate the efficiency and performance of our approach with two test cases: the diocotron instability in two dimensions and the three-dimensional electron dynamics in a Penning trap. Our run-time performance studies indicate that our approach can provide significant speedup and memory reduction to PIC simulations for achieving comparable accuracy in the charge density.

A high-order in space spectral-element methodology for the solution of a strongly coupled fluid-structure interaction (FSI) problem is developed. A methodology is based on a partitioned solution of incompressible fluid equations on body-fitted grids, and nonlinearly-elastic solid deformation equations coupled via a fixed-point iteration approach with Aitken relaxation. A comprehensive verification strategy of the developed methodology is presented, including h-, p- and temporal refinement studies. An expected order of convergence is demonstrated first separately for the corresponding fluid and solid solvers, followed by a self-convergence study on a coupled FSI problem (self-convergence refers to a convergence to a reference solution obtained with the same solver at higher resolution). To this end, a new three-dimensional fluid-structure interaction benchmark is proposed for a verification of the FSI codes, which consists of a fluid flow in a channel with one rigid and one flexible wall. It is shown that, due to a consistent problem formulation, including initial and boundary conditions, a high-order spatial convergence on a fully coupled FSI problem can be demonstrated. Finally, a developed framework is applied successfully to a Direct Numerical Simulation of a turbulent flow in a channel interacting with a compliant wall, where the fluid-structure interface is fully resolved.

While fast multipole methods (FMMs) are in widespread use for the rapid evaluation of potential fields governed by the Laplace, Helmholtz, Maxwell or Stokes equations, their coupling to high-order quadratures for evaluating layer potentials is still an area of active research. In three dimensions, a number of issues need to be addressed, including the specification of the surface as the union of high-order patches, the incorporation of accurate quadrature rules for integrating singular or weakly singular Green's functions on such patches, and their coupling to the oct-tree data structures on which the FMM separates near and far field interactions. Although the latter is straightforward for point distributions, the near field for a patch is determined by its physical dimensions, not the distribution of discretization points on the surface.

Here, we present a general framework for efficiently coupling locally corrected quadratures with FMMs, relying primarily on what are called generalized Gaussian quadratures rules, supplemented by adaptive integration. The approach, however, is quite general and easily applicable to other schemes, such as Quadrature by Expansion (QBX). We also introduce a number of accelerations to reduce the cost of quadrature generation itself, and present several numerical examples of acoustic scattering that demonstrate the accuracy, robustness, and computational efficiency of the scheme. On a single core of an Intel i5 2.3 GHz processor, a Fortran implementation of the scheme can generate near field quadrature corrections for between 1000 and 10,000 points per second, depending on the order of accuracy and the desired precision. A Fortran implementation of the algorithm described in this work is available at https://gitlab.com/fastalgorithms/fmm3dbie.

Concurrent multiscale finite element analysis (FE²) is a powerful approach for high-fidelity modeling of materials for which a suitable macroscopic constitutive model is not available. However, the extreme computational effort associated with computing a nested micromodel at every macroscopic integration point makes FE² prohibitive for most practical applications. Constructing surrogate models able to efficiently compute the microscopic constitutive response is therefore a promising approach in enabling concurrent multiscale modeling. This work presents a reduction framework for adaptively constructing surrogate models for FE² based on statistical learning. The nested micromodels are replaced by a machine learning surrogate model based on Gaussian Processes (GP). The need for offline data collection is bypassed by training the GP models online based on data coming from a small set of fully-solved anchor micromodels that undergo the same strain history as their associated macroscopic integration points. The Bayesian formalism inherent to GP models provides a natural tool for online uncertainty estimation through which new observations or inclusion of new anchor micromodels are triggered. The surrogate constitutive manifold is constructed with as few micromechanical evaluations as possible by enhancing the GP models with gradient information and the solution scheme is made robust through a greedy data selection approach embedded within the conventional finite element solution loop for nonlinear analysis. The sensitivity to model parameters is studied with a tapered bar example with plasticity and the framework is further demonstrated with the elastoplastic analysis of a plate with multiple cutouts and with a crack growth example for mixed-mode bending. Although not able to handle non-monotonic strain paths in its current form, the framework is found to be a promising approach in reducing the computational cost of FE², with significant efficiency gains being obtained without resorting to offline training.

We introduce a numerical method to describe the propagation of two-dimensional nonlinear water waves over a flat bottom. The free surface is described in terms of a Lagrangian representation, i.e. by following the position and the velocity potential of a set of surface particles. The method consists in a mixed Eulerian-Lagrangian modification of the classical High-Order Spectral (HOS) method. At each time step, the Eulerian velocity potential inside the domain and the velocity of the surface particles are estimated by using a spectral decomposition along with a perturbation expansion at an arbitrary order M. The Lagrangian description of the surface makes it possible to use lower approximation orders and fewer Fourier modes to capture steep nonlinear waves, which also improves the numerical stability of the method. Its accuracy is established for steep regular waves by comparing simulations to existing Lagrangian and Eulerian solutions, as well as to traditional HOS-simulations. For irregular bichromatic waves, we show with an example that the obtained solution converges with respect to the Lagrangian conservation equations as the order M increases. Finally, the ability of the proposed method to compute the velocity field in steep irregular waves is demonstrated.

A parallel and scalable stochastic Direct Simulation Monte Carlo (DSMC) method applied to large-scale dense bubbly flows is reported in this paper. The DSMC method is applied to speed up the bubble-bubble collision handling relative to the Discrete Bubble Model proposed by Darmana et al. (2006) [1]. The DSMC algorithm has been modified and extended to account for bubble-bubble interactions arising due to uncorrelated and correlated bubble velocities. The algorithm is fully coupled with an in-house CFD code and parallelized using the MPI framework. The model is verified and validated on multiple cores with different test cases, ranging from impinging particle streams to laboratory-scale bubble columns. The parallel performance is shown using two different large scale systems: with an uniform and a non-uniform distribution of bubbles. The hydrodynamics of a pilot-scale bubble column is analyzed and the effect of the column scale is reported via the comparison of bubble columns at three different scales.