Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Helmholtz equation”

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.

At least 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Reducing Frequency Bias of Fourier Neural Operators in 3D Seismic Wavefield Simulations Through Multistage Training

The recent development of neural operator (NeurOp) learning for solutions to the elastic wave equation shows promising results and provides the basis for fast large-scale simulations for different seismological applications. In this article, we use the Fourier neural operator (FNO) model to directly solve the 3D Helmholtz wave equation for fast seismic ground-motion simulations on different frequencies and show the frequency bias of the FNO model, that is, it learns the lower frequencies better comparing to the higher frequencies. To reduce the frequency bias, we adopt the multistage FNO training, that is, after training a stage 1 FNO model for estimating the ground motion, we use a second FNO model as the stage 2 to learn from the residual, which greatly reduced the errors on the higher frequencies. By adopting this multistage training, the FNO models show reduced biases on higher frequencies, which enhanced the overall results of the ground-motion simulations. Thus the multistage training FNO improves the accuracy and realism of the ground-motion simulations.

earthquakes↗
DOE OSTI2021

Sparse Approximate Multifrontal Factorization with Butterfly Compression for High-Frequency Wave Equations

In this work, we present a fast and approximate multifrontal solver for large-scale sparse linear systems arising from finite-difference, finite-volume or finite-element discretization of high-frequency wave equations. The proposed solver leverages the butterfly algorithm and its hierarchical matrix extension for compressing and factorizing large frontal matrices via graph-distance guided entry evaluation or randomized matrix-vector multiplication-based schemes. Complexity analysis and numerical experiments demonstrate $\mathcal{O}(N\log^2 N)$ computation and $\mathcal{O}(N)$ memory complexity when applied to an $N\times N$ sparse system arising from 3D high-frequency Helmholtz and Maxwell problems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

A least-squares finite element method based on the Helmholtz decomposition for hyperbolic balance laws

In this paper, a least-squares finite element method for scalar nonlinear hyperbolic balance laws is proposed and studied. The approach is based on a formulation that utilizes an appropriate Helmholtz decomposition of the flux vector and is related to the standard notion of a weak solution. This relationship, together with a corresponding connection to negative-norm least-squares, is described in detail. As a consequence, an important numerical conservation theorem is obtained, similar to the famous Lax–Wendroff theorem. The numerical conservation properties of the method in this paper do not fall precisely in the framework introduced by Lax and Wendroff, but they are similar in spirit as they guarantee that when L 2 convergence holds, the resulting approximations approach a weak solution to the hyperbolic problem. The least-squares functional is continuous and coercive in an H -1 -type norm, but not L 2 -coercive. Nevertheless, the L 2 convergence properties of the method are discussed. Convergence can be obtained either by an explicit regularization of the functional, that provides control of the L 2 norm, or by properly choosing the finite element spaces, providing implicit control of the L 2 norm. Numerical results for the inviscid Burgers equation with discontinuous source terms are shown, demonstrating the L 2 convergence of the obtained approximations to the physically admissible solution. The numerical method utilizes a least-squares functional, minimized on finite element spaces, and a Gauss–Newton technique with nested iteration. Finally, we believe that the linear systems encountered with this formulation are amenable to multigrid techniques and combining the method with adaptive mesh refinement would make this approach an efficient tool for solving balance laws (this is the focus of a future study).

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Solving Seismic Wave Equations on Variable Velocity Models With Fourier Neural Operator

Here, in the study of subsurface seismic imaging, solving the acoustic wave equation is a pivotal component in existing models. The advancement of deep learning (DL) enables solving partial differential equations (PDEs), including wave equations, by applying neural networks to identify the mapping between the inputs and the solution. This approach can be faster than traditional numerical methods when numerous instances are to be solved. Previous works that concentrate on solving the wave equation by neural networks consider either a single velocity model or multiple simple velocity models, which is restricted in practice. Instead, inspired by the idea of operator learning, this work leverages the Fourier neural operator (FNO) to effectively learn the frequency domain seismic wavefields under the context of variable velocity models. We also propose a new framework paralleled FNO (PFNO) for efficiently training the FNO-based solver given multiple source locations and frequencies. Numerical experiments demonstrate the high accuracy of both FNO and PFNO with complicated velocity models in the OpenFWI datasets. Furthermore, the cross-dataset generalization test verifies that PFNO adapts to out-of-distribution velocity models. Finally, PFNO admits higher computational efficiency on large-scale testing datasets than the traditional finite-difference method. The aforementioned advantages endow the FNO-based solver with the potential to build powerful models for research on seismic waves.

58 GEOSCIENCES↗
DOE OSTI2026

Enhancing photoionization rate calculations in low-temperature plasmas using spectral methods

Photoionization plays a central role in the development of streamer discharges and other non-equilibrium plasma phenomena. It creates seed electrons, which are essential for positive streamer propagation, allowing the ionization front to move forward. Because of this, accurate modeling of photoionization is very important for predicting streamer behavior and plasma evolution. The photoionization process in air (N 2 – O 2 mixture) is often described by the Zheleznyak model (1982). This model is usually solved through Helmholtz-type equations that approximate the Zheleznyak photoionization model (Zheleznyak et al. 1982) as Partial Differential Equations (PDEs). Conventional numerical methods, such as the Finite Difference Method (FDM) or Finite Volume Method (FVM), are widely used to solve these equations. Although they are prevalent, the computational cost of these methods due to their need for matrix operations and iterative solver is demanding. To address this challenge, this work develops a spectral solver based on the Fast Fourier Transform (FFT) combined with Discrete Cosine Transform (DCT) and Discrete Sine Transform (DST) to calculate the photoionization rate efficiently in an axisymmetric cylindrical domain. This method naturally satisfies the boundary conditions used in the model and converts the PDE into algebraic ones in spectral space. Thus, avoids the need for iterative matrix solvers. When compared with FDM results, it is demonstrated that the new solver not only maintains accuracy, but also reduces the computational cost, showing a performance increase of approximately 100 compared to FDM over a wide range of problem sizes. The method is parallelized using Message Passing Interface (MPI) and has been integrated into a fluid plasma model for streamer simulation. Here, this FFT-based approach provides a fast and reliable alternative for calculating photoionization in fluid models, helping large-scale plasma simulations run faster and efficiently, and allows higher-resolution simulation without extra computational cost.

Axisymmetric system↗
DOE OSTI2025

An FFT-based micromechanical model for gradient enhanced brittle fracture

Damage models incorporated within FFT-based micromechanical methods have received much attention recently because of the need to better understand and predict brittle and ductile fracture. An important aspect of a damage model is non-local regularization, which removes the mesh dependence of the predictions that otherwise become physically unacceptable upon grid refinement. In this work, the Helmholtz-type equation for non-local gradient regularization of a damage model on a distorted grid is solved using an FFT-based approach. Further, the resulting system of equations is solved using the Jacobi iterative method. The model is applied to simulate brittle fracture of an intermetallic. The influence of the time and space discretization, the length-scale parameter, and intermetallic crystallographic orientation on crack evolution is studied.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Model Development and Analysis of a High-Fidelity Neutron Transport Sensor: The Quadrupole Detector Concept for Measurement of the Neutron Flux Gradient

Accurate reconstruction of the neutron flux distribution within a reactor core is essential for safe and efficient reactor operation. Traditional power shape synthesis in Light Water Reactors relies on hundreds of in-core detectors. However, this approach becomes impractical for Advanced Reactors and Microreactors due to limited space and harsh environments. To address this challenge, we propose a data-driven methodology that combines high-fidelity modeling with real-time ex-core sensor measurements, enabling the reconstruction of core power distribution while minimizing the reliance on intrusive in-core instrumentation. This project began in FY24 and achieved two initial milestones: (1) the definition of a three-year development plan for a Digital Twin framework and (2) the development of high-fidelity neutronics models of the Purdue University Reactor One (PUR-1) using both MCNP6 and OpenMC. The PUR-1 reactor, a zero-power facility, was selected due to its suitability for neutronics-focused modeling and the availability of experimental data for validation. Both models were benchmarked using neutron flux measurements obtained from irradiated gold foils, which were strategically placed within the core during a dedicated campaign in July 2024. This report marks the continuation and completion of those foundational tasks. The OpenMC model has been refined (improved geometric accuracy, expanded cross-section libraries, and refined sampling) and validated using additional experimental data. An updated sensor design—based on quadrupole configuration—was designed to measure both ex-core flux and its spatial gradient. These measurements will serve as inputs to a neural network-based reconstruction algorithm. Finally, the methodology was demonstrated on a two-dimensional test case representative of the heterogeneous material composition of the PUR-1 reactor core. A neural network implementation of the Kirchhoff-Helmholtz integral equation was employed to solve the boundary value problem using peripheral sensor measurements. The preliminary results confirm the strong potential of the proposed approach for accurate and minimally invasive neutron flux reconstruction.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2024

Speed of Sound Measurements of R-1130(E) and an Azeotropic Blend of R-1336mzz(Z)/1130(E)

Sound speed data measured using a dual-path pulse-echo instrument are reported for pure trans-1,2-dichloroethene (R-1130(E)) and an azeotropic blend of cis-1,1,1,4,4,4-hexafluorobutene (R-1336mzz(Z)) and R-1130(E) with a composition of 74.8 mass % R-1336mzz(Z) with the balance being R-1130(E). The azeotropic blend of R-1336mzz(Z)/1130(E) is classified as R-514A in ANSI/ASHRAE standard 34. Liquid phase speed of sound data are reported from just above the saturation pressure of pure R-1130(E) or the bubble point pressure of R-514A to a maximum pressure of 26.7 MPa. The relative combined expanded uncertainty in the speed of sound varies from 0.032 % to 0.148 % with the greatest deviations occurring at the lowest sound speeds. At present, no reference Helmholtz-energy-explicit equation of state (EOS) is available for R-1130(E). Therefore, the reported data for pure R-1130(E) are compared to an extended corresponding states (ECS) model. Deviations between the pure R-1130(E) sound speed data and the ECS model were found to be consistently negative ranging between − 4.1 % and − 3.5 %. The R-514A data are compared to a multifluid model inclusive of the established reference Helmholtz-energy-explicit EOS for R-1336mzz(Z) and ECS model for R-1130(E) with estimated binary interaction parameters. Deviations between the experimental speed of sound data and the multifluid model were also found to be consistently negative. However, deviations from the multifluid model were found to be as great as − 17.1 %. The large deviations from the ECS model and multifluid model underscore the need for a robust Helmholtz-energy-explicit EOS for R-1130(E).

42 ENGINEERING↗
DOE OSTI2022

Memory-efficient emulation of physical tabular data using quadtree decomposition

Computationally expensive functions are sometimes replaced in simulations with an emulator that approx-imates the true function (e.g., equations of state, wavelength-dependent opacity, or composition-dependent materials properties). For functions that have a constrained domain of interest, this can be done by discretizing the domain and performing a local interpolation on the tabulated function values of each local domain. For these so-called tabular data methods, the method of discretizing the domain and mapping the input space to each subdomain can drastically influence the memory and computational costs of the emulator. This is especially true for functions that vary drastically in different regions. We present a method for domain discretization and mapping that utilizes quadtrees, which results in significant reductions in the size of the emulator with minimal increases to computational costs or loss of global accuracy. We apply our method to the electron-positron Helmholtz free energy equation of state and show over an order of magnitude reduction in memory costs for reasonable levels of numerical accuracy.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Speed of Sound Measurements of Binary Mixtures of trans -1,2-Difluoroethylene (R-1132(E)) with Difluoromethane (R-32) or 2,3,3,3-Tetrafluoropropene (R-1234yf)

The speed of sound of two binary mixtures containing 0.336 mole fraction trans-1,2- difluoroethylene (R-1132(E)) with difluoromethane (R-32) and 0.435 mole fraction R-1132(E) with 2,3,3,3-tetrafluoropropene (R-1234yf) were measured with a dual-path pulse-echo instrument. The speed of sound was measured along several pseudo-isochores for each blend at temperatures ranging from 230 K to 325 K for blends with R-32 and to 342 K for blends with R- 1234yf. Measurements started at pressures just above each mixtures bubble point pressure and were limited to 8 MPa to avoid potential disproportionation reactions of the R-1132(E). The data were compared to multi-fluid models incorporating Helmholtz-energy-explicit equations of state (EOS) for each pure fluid.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Bubble Point Measurements of cis-1,1,1,4,4,4-Hexafluorobutene [R-1336mzz(Z)] + trans-1,2-Dichloroethene [R-1130(E)] mixtures

Saturation pressures of pure R-1336mzz(Z) and R-1130(E) and bubble point pressures of three R-1336mzz(Z)/1130(E) blends were measured from 265 K to 360 K. For each pure refrigerant or refrigerant blend, a total of twenty unique saturation pressures or bubble points were measured. In total 100 unique state points were obtained. Presently, no Helmholtz-energy-explicit type equation of state (EoS) is available for R-1130(E). While an extended corresponding states EoS for R-1130(E) is available to estimate the properties of the R-1336mzz(Z)/1130(E) blend, this model does not resolve the azeotropic behavior of the mixture. Therefore, the perturbed-chain statistical-associating fluid theory (PC-SAFT) EoS is used to model the vapor–liquid equilibria of the R-1336mzz(Z)/1130(E) blend. PC-SAFT model parameters are reported, and the overall performance of the model is characterized by deviations from the experimental data.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

H2OI95: A Stand-Alone Fortran Code for Evaluating the IAPWS-95 Equation-of-State Model for Water (Rev. 1)

H2OI95 is a stand-alone Fortran code for evaluating the IAPWS-95 equation-of-state model for the thermodynamic properties of water. It also evaluates the corresponding thermochemical properties of water consistent with the CODATA recommendations. The IAPWS-95 model is based on a model equation for the dimensionless Helmholtz energy for which the primary variables are the inverse reduced temperature (τ = T cr /T ) and reduced density (δ = ρ/ρ cr ). Here T is the absolute temperature (K), ρ is density (kg/m³), and the subscript “ cr ” refers to the critical point of water (647.096 K and 22.064 MPa pressure in this model, for which ρ cr is 322 kg/m³). The code solves four basic types of problems, distinguished by the specified inputs: 1. Temperature (K) and density (ρ kg/m³) or reduced density (δ) 2. Temperature (K) and pressure (MPa). 3. Temperature (K) on the saturation (liquid-vapor equilibrium) curve 4. Pressure (MPa) on the saturation curve. Each type of problem is run using a corresponding input (text) file. All but the first type of problem require iteration. For example, to solve for desired temperature and pressure, the reduced density must be adjusted to give the desired pressure. Iteration is accomplished using the Newton-Raphson method, though the secant method is also used in solving the fourth type of problem. For the last three types of problems, H2OI95 has been used to conduct numerical studies of convergence and the problem of multiple numerical solutions, some of which are not physically valid. Obtaining valid results depends mainly on appropriate choice of starting values for the density. The default values used in H2OI95 appear to consistently lead to generally desired results. With modification (not addressed here), H2OI95 can be used to support SUPCRT92 and similar codes that compute chemical thermodynamic properties of species and reactions over a wide range of temperature and pressure (273.16-1273K and 0-1000 MPa). Version 1.1 of this software calculates an extended set of water properties, including some derivable solely from IAPWS-95, such as the isobaric thermal expansivity and the isothermal compressibility, and still other properties based on supplementary models sanctioned by IAPWS, including the dynamic viscosity, thermal conductivity, surface tension, and static dielectric constant, and various other properties (notably Debye-Hückel parameters) derivable from the dielectric constant and its partial derivatives with respect to temperature and pressure. A coding error in Version 1.0 was identified while testing the implementation of the viscosity model. This error is described in Appendix C. It mainly affects calculation of “higher order” properties such as compressibility and heat capacity near the critical point.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Equation of State for the Thermodynamic Properties of Trans-1,2-dichloroethene [R-1130(E)]

We present an empirical equation of state in terms of the Helmholtz energy for trans-1,2-dichloroethene [R-1130(E)]. The range of validity is from the triple-point temperature, 223.31 K to 525 K with pressures up to 30 MPa. It may be used to calculate all thermodynamic properties in the fluid phase, including liquid, gas, and supercritical regions. Comparisons are given with existing literature data and estimated uncertainties are provided. In addition, checks were made for correct extrapolation behavior so that the equation behaves in a physically realistic manner when used outside of its range of validity, enabling its use in mixture models. The estimated uncertainties (at a k = 2 or 95 % level of confidence) are based on comparisons with critically assessed data and are 0.25 % for vapor pressure for temperatures in the range 300 K < T < 454 K, rising to 1.5 % as the temperature decreases from 300 K to 265 K. For density in the liquid phase the estimated uncertainty is 0.14 % for temperatures 270 K < T < 410 K and for pressures up to 30 MPa. For the vapor phase the estimated uncertainty in density is 3 %. The uncertainty for liquid-phase heat capacity is 1 % at atmospheric pressure over the temperature range 268 K < T < 309 K, and the uncertainty for the speed of sound in the liquid phase is 0.25 % for temperatures 230 K < T < 420 K and for pressures up to 30 MPa. The uncertainties are larger outside of these specified ranges and in the critical region.

1,2-Dichloroethene↗
DOE OSTI2022

Quasi-Helmholtz decomposition, Gauss' laws and charge conservation for finite element particle-in-cell

Development of particle-in-cell (PIC) methods using finite element based methods (FEMs) have been a topic of renewed interest; this has largely been driven by (a) the ability of finite element methods to better model geometry, (b) better understanding of function spaces that are necessary to represent all Maxwell quantities, and (c) more recently, the fundamental rubrics that should be obeyed in space and time so as to satisfy Gauss' laws and the equation of continuity. In that vein, methods have been developed recently that satisfy these equations and are agnostic to time stepping methods. While this development is indeed a significant advance, it should be noted that implicit FEM transient solvers support an underlying null space that corresponds to a gradient of a scalar potential ∇Φ(r) (or t∇Φ(r) in the case of wave equation solvers). While explicit schemes do not suffer from this drawback, they are only conditionally stable with time step sizes that are mesh dependent and very small. Furthermore, the null space produces spurious charge that can corrupt the desired physics of a PIC simulation. The way to overcome this bottleneck, and indeed, satisfy all four Maxwell's equation is to use a quasi-Helmholtz formulation on a tessellation. In the re-formulation presented, we strictly satisfy the equation of continuity and Gauss' laws for both the electric and magnetic flux densities. Results illustrating the efficacy of this scheme will be demonstrated by analyzing non-neutral beams with space-charge and the adiabatic expansion of a neutral plasma with realistic parameters (Debye length and real mass ratios).

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2021

Self-similar Reynolds-averaged mechanical–scalar turbulence models for Rayleigh–Taylor, Richtmyer–Meshkov, and Kelvin–Helmholtz instability-induced mixing in the small Atwood number limit

Analytical self-similar solutions to two-, three-, and four-equation Reynolds-averaged mechanical–scalar turbulence models describing incompressible turbulent Rayleigh–Taylor, Richtmyer–Meshkov, and Kelvin–Helmholtz instability-induced mixing in planar geometry are derived in the small Atwood number (Boussinesq) limit. The models are based on the turbulent kinetic energy K and its dissipation rate ε, together with the scalar (heavy-fluid mass fraction) variance S and its dissipation rate Χ modeled either differentially or algebraically. The models allow for a simultaneous description of mechanical and scalar mixing, i.e., mixing layer growth and molecular mixing, respectively. Mixing layer growth parameters and other physical observables relevant to each instability are obtained explicitly as functions of the model coefficients. The turbulent fields are also expressed in terms of the model coefficients, with their temporal power-law scalings obtained by requiring that the self-similar equations are explicitly time-independent. The model calibration methodology is described and discussed. Expressions for a subset of the various physical observables are used to calibrate each of the two-, three-, and four-equation models, such that the self-similar solutions are consistent with experimental and numerical simulation data corresponding to these values of the observables and to specific canonical Rayleigh–Taylor, Richtmyer–Meshkov, and Kelvin–Helmholtz turbulent flows. A calibrated four-equation model is then used to reconstruct the mean and turbulent fields, and late-time turbulent equation budgets for each instability-induced flow across the mixing layer. The reference solutions derived here can provide systematic calibrations and better understanding of mechanical–scalar turbulence models and their predictions for instability-induced turbulent mixing in the very large Reynolds number limit.

42 ENGINEERING↗
DOE OSTI2021

Non-intrusive data-driven model reduction for differential–algebraic equations derived from lifting transformations

In this paper we present a non-intrusive data-driven approach for model reduction of nonlinear systems. The approach considers the particular case of nonlinear partial differential equations (PDEs) that form systems of partial differential–algebraic equations (PDAEs) when lifted to polynomial form. Such systems arise, for example, when the governing equations include Arrhenius reaction terms (e.g., in reacting flow models) and thermodynamic terms (e.g., the Helmholtz free energy terms in a phase-field solidification model). Using the known structured form of the lifted algebraic equations, the approach computes the reduced operators for the algebraic equations explicitly, using straightforward linear algebra operations on the basis matrices. The reduced operators for the differential equations are inferred from lifted snapshot data using operator inference, which solves a linear least squares regression problem. The approach is illustrated for the nonlinear model of solidification of a pure material. The lifting transformations reformulate the solidification PDEs as a system of PDAEs that have cubic structure. The operators of the lifted system for this solidification example have affine dependence on key process parameters, permitting us to learn a parametric reduced model with operator inference. Numerical experiments show the effectiveness of the resulting reduced models in capturing key aspects of the solidification dynamics.

42 ENGINEERING↗
DOE OSTI2026

A buoyancy–shear–drag–scalar-based turbulence model for power-law acceleration-driven Rayleigh–Taylor, reshocked Richtmyer–Meshkov, and Kelvin–Helmholtz mixing

A previously developed phenomenological turbulence model for Rayleigh–Taylor, reshocked Richtmyer–Meshkov, and Kelvin–Helmholtz instability-induced mixing based on a general buoyancy–shear–drag model [O. Schilling, “A buoyancy–shear–drag-based turbulence model for Rayleigh–Taylor, reshocked Richtmyer–Meshkov, and Kelvin–Helmholtz mixing,” Physica D 402, 132238 (2020)] is extended to include active or passive scalar mixing and power-law acceleration-driven Rayleigh–Taylor mixing. The buoyancy–shear–drag equations are coupled to a scalar variance equation that is used to define the molecular mixing parameter θ m , and when the scalar is active, modifies the Rayleigh–Taylor and Kelvin–Helmholtz mixing layer growth parameters to depend on the asymptotic value of this parameter, θ mol . Here, the scalar variance equation is closed by algebraically or differentially modeling the scalar variance dissipation rate. Nonlinear analytical solutions of the model are obtained in the total and separate bubble and spike mixing layer width formulations with the algebraic scalar variance dissipation rate for each instability, which are then used to calibrate the mechanical and scalar equation coefficients to predict specific values of physical observables and molecular mixing parameters. Surrogate mechanical and scalar turbulent fields can be constructed by multiplying a presumed self-similar spatial profile by appropriate functions of the width and its time derivative, and of the scalar obtained by solving the ordinary differential model equations. The explicit modeling and solution of turbulent transport equations are not required. The bubble and spike mixing layer width and scalar variance equations are then solved numerically for constant-acceleration Rayleigh–Taylor, impulsively reshocked Richtmyer–Meshkov, and Kelvin–Helmholtz mixing, confirming that the prescribed level of molecular mixing is correctly predicted and illustrating the spatiotemporal evolution of the scalar fields.

Buoyancy–drag↗
DOE OSTI2020

High-order matrix-free incompressible flow solvers with GPU acceleration and low-order refined preconditioners

In this work, we present a matrix-free flow solver for high-order finite element discretizations of the incompressible Navier-Stokes and Stokes equations with GPU acceleration. For high polynomial degrees, assembling the matrix for the linear systems resulting from the finite element discretization can be prohibitively expensive, both in terms of computational complexity and memory. For this reason, it is necessary to develop matrix-free operators and preconditioners, which can be used to efficiently solve these linear systems without access to the matrix entries themselves. The matrix-free operator evaluations utilize GPU-accelerated sum-factorization techniques to minimize memory movement and maximize throughput. The preconditioners developed in this work are based on a low-order refined methodology with parallel subspace corrections, as described for diffusion problems in [1]. The saddle-point Stokes system is solved using block-preconditioning techniques, which are robust in mesh size, polynomial degree, time step, and viscosity. For the incompressible Navier-Stokes equations, we make use of projection (fractional step) methods, which require Helmholtz and Poisson solves at each time step. The performance of our flow solvers is assessed on several benchmark problems in two and three spatial dimensions.

97 MATHEMATICS AND COMPUTING↗