The use of cluster computer systems for NASA/JPL applications
In this paper we outline the results of using three generations of cluster machines at JPL for a select set of applications drawn from a larger development effort.
cluster↗
SEARCH · Engineering Papers
Results for “adaptive meshing”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Detailed Simulations of Shock-Bifurcation and Ignition of an Argon-diluted Hydrogen/Oxygen Mixture in a Shock Tube
Detailed simulations of the bifurcation and ignition of an Argon-diluted Hydrogen/Oxygen mixture in the two-stage weak ignition regime are performed. An adaptive mesh-refinement (AMR) technique is employed to resolve all relevant physical scales that are associated with the viscous boundary-layer, the reaction front, and the shock-wave. A high-order hybrid WENO/central-differencing method is used as spatial discretization scheme, and a detailed chemical mechanism is employed to describe the combustion of the H2/O2 mixture. The operating conditions considered in this study are p = 5 bar and T = 1100 K, and fall in the third explosion limit. The computations show that the mixing of the thermally stratified fluid, carrying different momentum and enthalpy, introduces inhomogeneities in the core-region behind the reflected shock. These inhomogeneities act as localized ignition kernels. During the induction period, these kernels slowly expand and eventually transition to a detonation wave that rapidly consumes the unburned mixture.In competition with this detonation wave are the presence of secondary ignition kernels that appear in the unreacted core-region between reflected shock and detonation wave.
Ihme, Matthias↗
Final Report for "Extending Chombo with PETSc"
Explore the source record for details and available documents.
adaptive mesh refinement, matrix assembly, mesh re↗
Adaptive Grid-based Hex Meshing for High Performance Computing.
Abstract not provided.
Owen, Steven J.↗
NEMOSYS: A PLATFORM FOR AUTONOMOUS MESH REFINEMENT AND SOLUTION VERIFICATION
Explore the source record for details and available documents.
22 GENERAL STUDIES OF NUCLEAR REACTORS↗
Characterizing Biomass Feedstock Transport Properties Using State of the Art Imaging and Computational Techniques
The microstructure of lignocellulosic biomass determines heat and mass transfer during conversion processes. We present a novel method for characterizing the transport properties of biomass using advanced imaging and computational techniques. The microstructure of two woody feedstocks, red oak and Douglas fir, before and after pyrolysis, is revealed using X-ray computed tomography (XCT). Transport properties are calculated from the XCT images, and principal permeability tensors are calculated using an immersed boundary-based finite volume solver to model gas flow through the geometries. We observe that the permeabilities of native biomass are distinctly anisotropic, however, this anisotropy is greatly reduced after pyrolysis.
adaptive mesh refinement↗
Direct Numerical Simulation of Flame-Wall Interaction for Low-Carbon Gas Turbine Combustion
Green hydrogen (H2) and ammonia (NH3) are emerging as carbon free alternatives to hydrocarbon fuels. In gas turbines, flames interact with the combustor liner (wall), which affects the pollutant emissions, the burning efficiency, and the thermal load on the liner. We lack understanding of this important flame-wall interaction (FWI) for alternative fuels. FWI occurs at the scale of the flame thickness, and such scales are only resolved by direct numerical simulation. Here, we resolve the FWI for NH3 and H2 flames. Preliminary results of two distinct projects are presented: 2D laminar anchored V-flame, and 3D turbulent swirling flame. For the laminar flame, we show flame quenching at the wall and resulting leakage of NH3 and increased N2O, but decreased NO emissions. For the swirling flame, we show CO emissions, and comparison of flame shapes and location to experimental measurements. Ultimately we show that quenching and pollutant emissions strongly affected by fuel/air ratio, and that an inclined wall can stabilize very lean turbulent CH4/H2 flames, but prone to large CO emissions.
adaptive mesh refinement↗
Cartesian based grid generation/adaptive mesh refinement
Grid adaptation has recently received attention in the computational fluid dynamics (CFD) community as a means to capture the salient features of a flowfield by either moving grid points of a structured or by adding cells in an unstructured manner. An approach based on a background cartesian mesh is investigated from which the geometry is 'cut' out of the mesh. Once the mesh is obtained, a solution on this coarse grid is found, that indicates which cells need to be refined. This process of refining/solving continues until the flow is grid refined in terms of a user specified global parameter (such as drag coefficient etc.). The advantages of this approach are twofold: the generation of the base grid is independent of the topology of the bodies or surfaces around/through which the flow is to be computed, and the resulting grid (in uncut regions) is highly isotropic, so that the truncation error is low. The flow solver (which, along with the grid generation is still under development) uses a completely unstructured data base, and is a finite volume, upwinding scheme. Current and future work will address generating Navier-Stokes suitable grids by using locally aligned and normal face/cell refining. The attached plot shows a simple grid about two turbine blades.
Coirier, William J.↗
Self-Avoiding Walks over Adaptive Triangular Grids
In this paper, we present a new approach to constructing a "self-avoiding" walk through a triangular mesh. Unlike the popular approach of visiting mesh elements using space-filling curves which is based on a geometric embedding, our approach is combinatorial in the sense that it uses the mesh connectivity only. We present an algorithm for constructing a self-avoiding walk which can be applied to any unstructured triangular mesh. The complexity of the algorithm is O(n x log(n)), where n is the number of triangles in the mesh. We show that for hierarchical adaptive meshes, the algorithm can be easily parallelized by taking advantage of the regularity of the refinement rules. The proposed approach should be very useful in the run-time partitioning and load balancing of adaptive unstructured grids.
Heber, Gerd↗
Self-Avoiding Walks Over Adaptive Triangular Grids
Space-filling curves is a popular approach based on a geometric embedding for linearizing computational meshes. We present a new O(n log n) combinatorial algorithm for constructing a self avoiding walk through a two dimensional mesh containing n triangles. We show that for hierarchical adaptive meshes, the algorithm can be locally adapted and easily parallelized by taking advantage of the regularity of the refinement rules. The proposed approach should be very useful in the runtime partitioning and load balancing of adaptive unstructured grids.
Heber, Gerd↗
LDRD Abbreviated report: High-Order General-Discrete-Ordinates Method Enabling Efficient Deterministic Transport in Hydrodynamic Simulations
Deterministic transport simulations for national-security and energy applications often operate in high-dimensional phase-space, where accuracy and cost both become major challenges. A common numerical artifact in such problems is the “ray-effect,” which appears as unphysical streaks. Beyond misinterpretation, these artifacts can contaminate tightly coupled physics, such as fluid dynamics, radiation-hydrodynamics, and laser-plasma interactions, eroding the predictive capability of entire multiphysics workflows. Our objective was to make high-dimension studies practical on modern hardware while mitigating the ray-effect without relying on prohibitively expensive sampling approaches such as Monte Carlo methods. We developed the Generic Discretization Library (GenDiL), a Graphics Processing Unit (GPU)-first framework that uses high-order Discontinuous Galerkin (DG) methods and matrix-free algorithms to reduce memory usage and improve computational efficiency, critical for phase-space simulations. GenDiL supports phase-space adaptivity in both mesh size and polynomial order (hp-adaptivity) to place resolution only where it is needed. A central capability is Local Dimensional Refinement (LDR), which couples lower-dimension continuum models to higher-dimension kinetic models through stable and conservative interfaces, so that high-fidelity physics is applied only in regions where it is essential. Building on the GenDiL framework, we developed the General SN (GSN) family of algorithms as a true generalization of the polar SN approach (discrete ordinates, often denoted SN). Rather than tying discrete ordinates to a specific polar change of coordinates, GSN formulates transport on an arbitrary change of coordinates chosen to reduce ray-effect. We studied two complementary variants: an analytic variant, where the coordinate map is prescribed in advance by a closed-form function; and a data-driven variant, where a quantity of interest, such as the net flux, guides the coordinate system. GenDiL provides the library infrastructure for efficient GPU execution, but the GSN concept is algorithmic and independent of any one library. Across representative high-dimension tests, including non-symmetric solutions, both variants delivered strong ray-effect mitigation at practical cost, moving four- to six-dimensional analysis toward repeatable, routine studies.
73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
A new procedure for dynamic adaption of three-dimensional unstructured grids
A new procedure is presented for the simultaneous coarsening and refinement of three-dimensional unstructured tetrahedral meshes. This algorithm allows for localized grid adaption that is used to capture aerodynamic flow features such as vortices and shock waves in helicopter flowfield simulations. The mesh-adaption algorithm is implemented in the C programming language and uses a data structure consisting of a series of dynamically-allocated linked lists. These lists allow the mesh connectivity to be rapidly reconstructed when individual mesh points are added and/or deleted. The algorithm allows the mesh to change in an anisotropic manner in order to efficiently resolve directional flow features. The procedure has been successfully implemented on a single processor of a Cray Y-MP computer. Two sample cases are presented involving three-dimensional transonic flow. Computed results show good agreement with conventional structured-grid solutions for the Euler equations.
Biswas, Rupak↗
A structured and unstructured remeshing method for high speed flows
An adaptive remeshing method using both triangular and quadrilateral elements suitable for high speed flows is presented. For inviscid flows the method generates completely unstructured meshes. For viscous flows the boundary layer edge is identified adaptively and a structured mesh is generated in the boundary layer, and an unstructured mesh is generated in the inviscid region. Examples of inviscid and viscous mesh adaptations for high speed flows are presented. A comparison is made between first order and higher order finite element algorithms when used in association with the remeshing method.
Vemaganti, Gururaja R.↗
Analysis of Hydrocyclone Geometry via Rapid Optimization Based on Computational Fluid Dynamics
Abstract Hydrocyclones exploit density gradients for the centrifugal separation of dispersions in a continuous liquid. Selection of the geometrics for optimal separation is case specific, like the media characteristics. The existing optimization method based on computational fluid dynamics (CFD) provides a powerful analytical tool but requires long computational times. The most common praxis for CFD optimization is via the single‐factor optimization method (SFOM). In this study, a novel approach is presented as an improved rapid optimization method that implements a dynamic‐mesh and user‐defined function optimization method (DUOM). The DUOM adapts the dynamic‐mesh approach from other applications to the optimization analysis of hydrocyclones. The DUOM reduced the computational time by 31.1 %, compared to the SFOM.
Liu, Lin↗
An improved random walk algorithm for the implicit Monte Carlo method
In this work, we introduce a modified Implicit Monte Carlo (IMC) Random Walk (RW) algorithm, which increases simulation efficiency for multigroup radiative transfer problems with strongly frequency-dependent opacities. To date, the RW method has only been implemented in “fully-gray” form; that is, the multigroup IMC opacities are group-collapsed over the full frequency domain of the problem to obtain a gray diffusion problem for RW. This formulation works well for problems with large spatial cells and/or opacities that are weakly dependent on frequency; however, the efficiency of the RW method degrades when the spatial cells are thin or the opacities are a strong function of frequency. To address this inefficiency, we introduce a RW frequency group cutoff in each spatial cell, which divides the frequency domain into optically thick and optically thin components. In the modified algorithm, opacities for the RW diffusion problem are obtained by group-collapsing IMC opacities below the frequency group cutoff. Particles with frequencies above the cutoff are transported via standard IMC, while particles below the cutoff are eligible for RW. This greatly increases the total number of RW steps taken per IMC time-step, which in turn improves the efficiency of the simulation. We refer to this new method as Partially-Gray Random Walk (PGRW). We present numerical results for several multigroup radiative transfer problems, which show that the PGRW method is significantly more efficient than standard RW for several problems of interest. In general, PGRW decreases runtimes by a factor of ∼2–4 compared to standard RW, and a factor of ∼3–6 compared to standard IMC. While PGRW is slower than frequency-dependent Discrete Diffusion Monte Carlo (DDMC), it is also easier to adapt to unstructured meshes and can be used in spatial cells where DDMC is not applicable. This suggests that it may be optimal to employ both DDMC and PGRW in a single simulation.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
Temporal adaptive Euler/Navier-Stokes algorithm involving unstructured dynamic meshes
Explore the source record for details and available documents.
Kleb, W. L.↗
Progress in the Simulation of Steady and Time-Dependent Flows with 3D Parallel Unstructured Cartesian Methods
The proposed paper will present recent extensions in the development of an efficient Euler solver for adaptively-refined Cartesian meshes with embedded boundaries. The paper will focus on extensions of the basic method to include solution adaptation, time-dependent flow simulation, and arbitrary rigid domain motion. The parallel multilevel method makes use of on-the-fly parallel domain decomposition to achieve extremely good scalability on large numbers of processors, and is coupled with an automatic coarse mesh generation algorithm for efficient processing by a multigrid smoother. Numerical results are presented demonstrating parallel speed-ups of up to 435 on 512 processors. Solution-based adaptation may be keyed off truncation error estimates using tau-extrapolation or a variety of feature detection based refinement parameters. The multigrid method is extended to for time-dependent flows through the use of a dual-time approach. The extension to rigid domain motion uses an Arbitrary Lagrangian-Eulerlarian (ALE) formulation, and results will be presented for a variety of two- and three-dimensional example problems with both simple and complex geometry.
Aftosmis, M. J.↗
Target Formulation and Construction in Mesh Quality Improvement
In the Target-matrix Optimization Paradigm (TMOP), it’s long been understood that one must create a set of target matrices before the mesh can be optimized. But there is still no general method to create correct, effective targets in response to a specific mesh quality improvement goal. The TMOP literature describes how certain sets of target matrices can be used to control the shape or size of mesh elements, but those examples address only a fraction of the problems that can occur in mesh quality improvement and were not derived from a general framework for target matrix construction. In this work, a general method of target construction is introduced based on an independent set of geometric parameters which are intrinsic to the Jacobian matrices upon which TMOP is based. The parameters enable a systematic approach to target definition and construction. The approach entails two parts. The first part defines correspondences between available raw data (stuff about the mesh and/or the physical solution) and intermediate data (e.g., a field of error estimates). Once the correspondences are established, the raw data is processed into intermediate field data existing on mesh sample points. The second part is to create a model that represents the values of the geometric target parameters as functions of the intermediate data. The model is then tested numerically to establish model constants and effectiveness. This systematic approach to target construction is illustrated in a set of examples to show how it can be applied to common problems in mesh optimization such as equidistribution of geometric properties, preservation of existing good quality, and adaptation of the mesh to the solution. The result is a systematic method of target construction for TMOP which can be applied to a wide variety of planar and volume mesh quality improvement tasks.
97 MATHEMATICS AND COMPUTING↗