Search NASA⌕ Search

SEARCH · Search NASA

Results for “direct solver”

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 577 records · Page 32

Numerical simulation using vorticity-vector potential formulation

An accurate and efficient computational method is needed for three-dimensional incompressible viscous flows in engineering applications. On solving the turbulent shear flows directly or using the subgrid scale model, it is indispensable to resolve the small scale fluid motions as well as the large scale motions. From this point of view, the pseudo-spectral method is used so far as the computational method. However, the finite difference or the finite element methods are widely applied for computing the flow with practical importance since these methods are easily applied to the flows with complex geometric configurations. However, there exist several problems in applying the finite difference method to direct and large eddy simulations. Accuracy is one of most important problems. This point was already addressed by the present author on the direct simulations on the instability of the plane Poiseuille flow and also on the transition to turbulence. In order to obtain high efficiency, the multi-grid Poisson solver is combined with the higher-order, accurate finite difference method. The formulation method is also one of the most important problems in applying the finite difference method to the incompressible turbulent flows. The three-dimensional Navier-Stokes equations have been solved so far in the primitive variables formulation. One of the major difficulties of this method is the rigorous satisfaction of the equation of continuity. In general, the staggered grid is used for the satisfaction of the solenoidal condition for the velocity field at the wall boundary. However, the velocity field satisfies the equation of continuity automatically in the vorticity-vector potential formulation. From this point of view, the vorticity-vector potential method was extended to the generalized coordinate system. In the present article, we adopt the vorticity-vector potential formulation, the generalized coordinate system, and the 4th-order accurate difference method as the computational method. We present the computational method and apply the present method to computations of flows in a square cavity at large Reynolds number in order to investigate its effectiveness.

Tokunaga, Hiroshi↗

Modeling Sound Propagation Through Non-Axisymmetric Jets

A method for computing the far-field adjoint Green's function of the generalized acoustic analogy equations under a locally parallel mean flow approximation is presented. The method is based on expanding the mean-flow-dependent coefficients in the governing equation and the scalar Green's function in truncated Fourier series in the azimuthal direction and a finite difference approximation in the radial direction in circular cylindrical coordinates. The combined spectral/finite difference method yields a highly banded system of algebraic equations that can be efficiently solved using a standard sparse system solver. The method is applied to test cases, with mean flow specified by analytical functions, corresponding to two noise reduction concepts of current interest: the offset jet and the fluid shield. Sample results for the Green's function are given for these two test cases and recommendations made as to the use of the method as part of a RANS-based jet noise prediction code.

Leib, Stewart J.↗

A Geometry Based Infra-Structure for Computational Analysis and Design

The computational steps traditionally taken for most engineering analysis suites (computational fluid dynamics (CFD), structural analysis, heat transfer and etc.) are: (1) Surface Generation -- usually by employing a Computer Assisted Design (CAD) system; (2) Grid Generation -- preparing the volume for the simulation; (3) Flow Solver -- producing the results at the specified operational point; (4) Post-processing Visualization -- interactively attempting to understand the results. For structural analysis, integrated systems can be obtained from a number of commercial vendors. These vendors couple directly to a number of CAD systems and are executed from within the CAD Graphical User Interface (GUI). It should be noted that the structural analysis problem is more tractable than CFD; there are fewer mesh topologies used and the grids are not as fine (this problem space does not have the length scaling issues of fluids). For CFD, these steps have worked well in the past for simple steady-state simulations at the expense of much user interaction. The data was transmitted between phases via files. In most cases, the output from a CAD system could go to Initial Graphics Exchange Specification (IGES) or Standard Exchange Program (STEP) files. The output from Grid Generators and Solvers do not really have standards though there are a couple of file formats that can be used for a subset of the gridding (i.e. PLOT3D data formats). The user would have to patch up the data or translate from one format to another to move to the next step. Sometimes this could take days. Specifically the problems with this procedure are:(1) File based -- Information flows from one step to the next via data files with formats specified for that procedure. File standards, when they exist, are wholly inadequate. For example, geometry from CAD systems (transmitted via IGES files) is defined as disjoint surfaces and curves (as well as masses of other information of no interest for the Grid Generator). This is particularly onerous for modern CAD systems based on solid modeling. The part was a proper solid and in the translation to IGES has lost this important characteristic. STEP is another standard for CAD data that exists and supports the concept of a solid. The problem with STEP is that a solid modeling geometry kernel is required to query and manipulate the data within this type of file. (2) 'Good' Geometry. A bottleneck in getting results from a solver is the construction of proper geometry to be fed to the grid generator. With 'good' geometry a grid can be constructed in tens of minutes (even with a complex configuration) using unstructured techniques. Adroit multi-block methods are not far behind. This means that a million node steady-state solution can be computed on the order of hours (using current high performance computers) starting from this 'good' geometry. Unfortunately, the geometry usually transmitted from the CAD system is not 'good' in the grid generator sense. The grid generator needs smooth closed solid geometry. It can take a week (or more) of interaction with the CAD output (sometimes by hand) before the process can begin. One way Communication. (3) One-way Communication -- All information travels on from one phase to the next. This makes procedures like node adaptation difficult when attempting to add or move nodes that sit on bounding surfaces (when the actual surface data has been lost after the grid generation phase). Until this process can be automated, more complex problems such as multi-disciplinary analysis or using the above procedure for design becomes prohibitive. There is also no way to easily deal with this system in a modular manner. One can only replace the grid generator, for example, if the software reads and writes the same files. Instead of the serial approach to analysis as described above, CAPRI takes a geometry centric approach. This makes the actual geometry (not a discretized version) accessible to all phases of the analysis. The connection to the geometry is made through an Application Programming Interface (API) and NOT a file system. This API isolates the top-level applications (grid generators, solvers and visualization components) from the geometry engine. Also this allows the replacement of one geometry kernel with another, without effecting these top-level applications. For example, if UniGraphics is used as the CAD package then Parasolid (UG's own geometry engine) can be used for all geometric queries so that no solid geometry information is lost in a translation. This is much better than STEP because when the data is queried, the same software is executed as used in the CAD system. Therefore, one analyzes the exact part that is in the CAD system. CAPRI uses the same idea as the commercial structural analysis codes but does not specify control. Software components of the CAD system are used, but the analysis suite, not the CAD operator, specifies the control of the software session. This also means that the license issues (may be) minimized and individuals need not have to know how to operate a CAD system in order to run the suite.

Haimes, Robert↗

Scalable line and plane relaxation in a parallel structured multigrid solver

The efficient solution of sparse, linear systems that arise through the discretization of partial differential equations remains a key challenge for a range of high performance scientific simulations. One approach for reducing data movement and improving performance is by exposing and exploiting structure in a problem through the use of robust structured multilevel solvers. By choosing coarsening that preserves the structure of the problem, these methods maintain efficient structured computation and communication throughout the multigrid hierarchy. However, when coarsening is not permitted to be dependent on the operator, anisotropy must be addressed by the smoother — producing error compatible for coarse-grid correction with structured coarsening. Here, the components required in a scalable parallel structured solver are described with a focus on memory and communication efficiency of robust smoothers. While the implementation of communication and memory reduction techniques in smoothers integrated in a complete 3D solver present a significant engineering challenge, a novel approach is proposed that addresses these challenges systematically through a change to the solver’s execution model. Enabled by user-level threading paired with a set of data and communication abstractions, this approach permits seamless aggregation of communication in plane smoothers — directly reusing code for a 2D distributed multilevel cycle. Results show an effective reduction in communication costs for coarse-grid problems, and result in a speedup of 8.7x in smoothing routines shown in Fig. 12 using this approach. This produces a significant improvement to strong scalability while maintaining favorable weak scaling behavior. Finally, a parallel scaling study using a series of refined meshes is included that demonstrates the effectiveness of this approach in an application of interest.

97 MATHEMATICS AND COMPUTING↗

Direct inference of nuclear equation-of-state parameters from gravitational-wave observations

The observation of neutron star mergers with gravitational waves (GWs) has provided a new method to constrain the dense-matter equation of state (EOS) and to better understand its nuclear physics. However, inferring nuclear microphysics from GW observations necessitates the sampling of EOS model parameters that serve as input for each EOS used during the GW data analysis. The sampling of the EOS parameters requires solving the Tolman–Oppenheimer–Volkoff (TOV) equations a large number of times—a process that slows down each likelihood evaluation in the analysis on the order of a few seconds. Here, we employ emulators for the TOV equations built using multilayer perceptron neural networks to enable direct inference of nuclear EOS parameters from GW strain data. Our emulators allow us to rapidly solve the TOV equations, taking in EOS parameters and outputting the associated tidal deformability of a neutron star in only a few tens of milliseconds. We implement these emulators in PyCBC to directly infer the EOS parameters using the event GW170817, providing posteriors on these parameters informed solely by GWs. We benchmark these runs against analyses performed using the full TOV solver and find that the emulators achieve speed ups of nearly two orders of magnitude, with negligible differences in the recovered posteriors. Additionally, we constrain the slope and curvature of the symmetry energy at the 90% upper credible interval to be $L$ sym ≲ 106 MeV and $K$ sym ≲ 26 MeV.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Convex relaxation for Fokker–Planck equation

We propose an approach to directly estimate the moments or marginals for a high-dimensional equilibrium distribution in statistical mechanics by solving the high-dimensional Fokker–Planck equation in terms of low-order cluster moments or marginals. With this approach, we bypass the exponential complexity of estimating the full high-dimensional distribution and directly solve the simplified partial differential equations for low-order moments/marginals. Moreover, the proposed moment/marginal relaxation is fully convex and can be solved via off-the-shelf solvers. We further propose a time-dependent version of the convex programs to study non-equilibrium dynamics. In a specific setting, we show the proposed method can recover a mean-field-type equilibrium density. Numerical results are provided to demonstrate the performance of the proposed algorithm for high-dimensional systems.

Chen, Yian↗

Three Dimensional Solution of Pneumatic Active Control of Forebody Vortex Asymmetry

Pneumatic active control of asymmetric vortical flows around a slender pointed forebody is investigated using the three dimensional solution for the compressible thin-layer Navier-Stokes equation. The computational applications cover the normal and tangential injection control of asymmetric flows around a 5 degree semi-apex angle cone at a 40 degree angle of attack, 1.4 freestream Mach number and 6 x 10(exp 6) freestream Reynolds number (based on the cone length). The effective tangential angle range of 67.5 approaches minus 67.5 degrees is used for both normal and tangential ports of injection. The effective axial length of injection is varied from 0.03 to 0.05. The computational solver uses the implicit, upwind, flux difference splitting finite volume scheme, and the grid consists of 161 x 55 x 65 points in the wrap around, normal and axial directions, respectively. The results show that tangential injection is more effective than normal injection.

Kandil, Osama A.↗

Transport Equation Based Wall Distance Computations Aimed at Flows With Time-Dependent Geometry

Eikonal, Hamilton-Jacobi and Poisson equations can be used for economical nearest wall distance computation and modification. Economical computations may be especially useful for aeroelastic and adaptive grid problems for which the grid deforms, and the nearest wall distance needs to be repeatedly computed. Modifications are directed at remedying turbulence model defects. For complex grid structures, implementation of the Eikonal and Hamilton-Jacobi approaches is not straightforward. This prohibits their use in industrial CFD solvers. However, both the Eikonal and Hamilton-Jacobi equations can be written in advection and advection-diffusion forms, respectively. These, like the Poisson s Laplacian, are commonly occurring industrial CFD solver elements. Use of the NASA CFL3D code to solve the Eikonal and Hamilton-Jacobi equations in advective-based forms is explored. The advection-based distance equations are found to have robust convergence. Geometries studied include single and two element airfoils, wing body and double delta configurations along with a complex electronics system. It is shown that for Eikonal accuracy, upwind metric differences are required. The Poisson approach is found effective and, since it does not require offset metric evaluations, easiest to implement. The sensitivity of flow solutions to wall distance assumptions is explored. Generally, results are not greatly affected by wall distance traits.

Tucker, Paul G.↗

Transport Equation Based Wall Distance Computations Aimed at Flows With Time-Dependent Geometry

Eikonal, Hamilton-Jacobi and Poisson equations can be used for economical nearest wall distance computation and modification. Economical computations may be especially useful for aeroelastic and adaptive grid problems for which the grid deforms, and the nearest wall distance needs to be repeatedly computed. Modifications are directed at remedying turbulence model defects. For complex grid structures, implementation of the Eikonal and Hamilton-Jacobi approaches is not straightforward. This prohibits their use in industrial CFD solvers. However, both the Eikonal and Hamilton-Jacobi equations can be written in advection and advection-diffusion forms, respectively. These, like the Poisson's Laplacian, are commonly occurring industrial CFD solver elements. Use of the NASA CFL3D code to solve the Eikonal and Hamilton-Jacobi equations in advective-based forms is explored. The advection-based distance equations are found to have robust convergence. Geometries studied include single and two element airfoils, wing body and double delta configurations along with a complex electronics system. It is shown that for Eikonal accuracy, upwind metric differences are required. The Poisson approach is found effective and, since it does not require offset metric evaluations, easiest to implement. The sensitivity of flow solutions to wall distance assumptions is explored. Generally, results are not greatly affected by wall distance traits.

Tucker, Paul G.↗

Adjoint-Based Minimization of X-59 Sonic Boom Noise Via Control Surfaces

A multidisciplinary design optimization methodology to directly minimize the ground-level noise generated by a supersonic aircraft's sonic boom is presented. A gradient-based optimizer is coupled with a Cartesian Euler flow solver and an atmospheric propagation tool to forge a unique design capability. Adjoint formulations for both the flow solver and propagation tool are also coupled to provide sensitivities in an efficient manner. The design method is demonstrated on the X-59 Low-Boom Flight Demonstrator by optimizing control surface deflections to improve ground-level noise while maintaining trimmed cruise flight. Two noise minimization examples and a noise maximization are presented. All optimized designs result in measurable improvement in the ground-level noise objective. Extensive surveys of the design space confirm that the optimization method is effective in finding a local optimum. Moreover, repeated application of the method with varying initial design points also demonstrates the robustness of the method.

ARMD↗

Comparison of OVERFLOW Computational and Experimental Results for a Blunt Mars Entry Vehicle Concept During Supersonic Retropropulsion

Simulations of supersonic retropropulsion (SRP) flow over a Hypersonic Inflatable Aerodynamic Decelerator (HIAD) blunt-body vehicle were performed using the Overflow Computational Fluid Dynamics (CFD) solver. Simulation conditions and geometry were designed to match specific test runs in the Descent System Study (DSS) testing campaign. The relative accuracy of simulation predictions are assessed by direct comparison to experimental data. Computational predictions of the SRP flowfield and bow shock shape are compared to experimental schlieren images. Comparisons of the model surface pressure environment are presented for unsteady and static discrete tap data as well as time-averaged Pressure-Sensitive Paint (PSP) data.

Supersonic Retropropulsion↗

CFD 2030 Grand Challenge: CFD-in-the-Loop Monte Carlo Flight Simulation for Space Vehicle Design

Flight qualification of space vehicles is markedly different from those typically employed for aircraft. The concept of an extensive flight test campaign for a space vehicle does not exist, and vehicle designers must look to alternative techniques for demonstrating robust and reliable performance of their vehicles prior to operational flight. A space vehicle may undergo only a handful of flight tests in its development cycle, with each flight representing a drastically different flight phase or flight configuration. For instance, NASA’s Space Launch System (SLS) launch vehicle and Orion spacecraft will only see a total of four flight demonstrations before flying a crew on its first operational mission, and each flight demonstrates a unique vehicle configuration and/or set of flight conditions. The SLS will be flown only one time before it becomes operational (Artemis 1). The Orion spacecraft Crew Module (CM) will have been tested twice, once on a Delta IV launch vehicle (Exploration Flight Test 1) and once as a fully integrated system with the SLS launch vehicle (Artemis 1). The Orion Launch abort system will have been tested twice, once in a pad abort scenario (Pad Abort 1) and once in an inflight abort scenario (Ascent Abort 2) on a modified Peacekeeper booster. Both of these latter tests involve only a boiler plate CM, not a functional Orion spacecraft. Thus, unlike aircraft, there is very little opportunity for engineers to assess and evaluate their preflight predictions. Instead, space vehicle designers rely on Monte Carlo flight simulations with detailed dispersions of predicted nominal flight behavior to determine how robust their design is to errors and uncertainties in the flight conditions their vehicle may encounter. These Monte Carlo analyses entail thousands of trajectory simulations to demonstrate that the vehicle can meet design requirements at a specified level of reliability. From an aerodynamics and aerothermodynamics perspective, these trajectory simulations are fueled by an extensive aerodynamic database that covers the complete range of expected flight conditions, vehicle configurations, and flight attitudes expected in a given mission. Today, these databases amount to a table of engineering parameters that can be quickly interrogated by the trajectory simulator. The aerodynamic and aerothermodynamic databases are assembled via a series of ground tests, empirical and analytical analysis, physics-based computational analysis, applicable past flight performance data, and in some cases, engineering judgment. These databases generally take years to assemble for a new space vehicle system and in the case of SLS/Orion, over a decade of test and analysis have been expended to develop the extensive databases required to cover the myriad of configurations and potential flight conditions required for the system. Recently, it has been proposed that Computational Fluid Dynamic (CFD) and computing capability may be reaching a point where it is foreseeable that CFD could be integrated directly into the production trajectory simulation tools used to design NASA’s space vehicles. To demonstrate this, NASA has embarked on two demonstrations of this type of capability, one where six degree of freedom flight trajectory simulation equations are embedded in an existing CFD solver and another where a production CFD solver is loosely coupled with a production trajectory simulation tool. These efforts represent an initial demonstration of a future approach to flight trajectory simulation, but they are a far cry from the capability required to perform a full-up CFD-in-the-loop Monte Carlo trajectory simulation. Therefore, this represents a viable grand challenge for computational methods addressing space vehicle design and development. The final paper/presentation will discuss the many hurdles, beyond simply raw computational power, to realizing this grand challenge and how they map directly to the CFD Vision 2030 ojectives. Among these are the wide range of flight conditions, including accelerating/decelerating flight, encountered by a space vehicle during launch and/or entry. The vehicle can also encounter numerous configuration changes, some of which can be quite drastic, during the course of its flight, so robust, automated geometry modeling, grid generation, and adaptation will play a huge role in reaching this goal. Multiply this by 1000’s of trajectory simulations occurring simultaneously in a given Monte Carlo analysis, and the problem readily scales to absorb virtually any size of supercomputer envisioned today. The concept of CFD-in-the-loop Monte Carlo trajectory simulation poses a formidable challenge for emerging and future computing systems, and it has the potential to shave years off the development cycle for aerodynamic and aerothermodynamic performance predictions as compared to today’s space vehicle design approach.

CFD 2030↗

On the Development of an Efficient Parallel Hybrid Solver with Application to Acoustically Treated Aero-Engine Nacelles

A finite element solution to the convected Helmholtz equation in a nonuniform flow is used to model the noise field within 3-D acoustically treated aero-engine nacelles. Options to select linear or cubic Hermite polynomial basis functions and isoparametric elements are included. However, the key feature of the method is a domain decomposition procedure that is based upon the inter-mixing of an iterative and a direct solve strategy for solving the discrete finite element equations. This procedure is optimized to take full advantage of sparsity and exploit the increased memory and parallel processing capability of modern computer architectures. Example computations are presented for the Langley Flow Impedance Test facility and a rectangular mapping of a full scale, generic aero-engine nacelle. The accuracy and parallel performance of this new solver are tested on both model problems using a supercomputer that contains hundreds of central processing units. Results show that the method gives extremely accurate attenuation predictions, achieves super-linear speedup over hundreds of CPUs, and solves upward of 25 million complex equations in a quarter of an hour.

Watson, Willie R.↗

A three-dimensional laser ray-tracing methodology for radiation-hydrodynamics simulations

We report on a methodology for performing laser ray-tracing in three spatial dimensions for radiation-hydrodynamics simulation codes. Our method, which is an extension of that developed in Haines et al., Comput. Fluids 201, 104478 (2020), utilizes an automatically generated separate mesh for the laser ray-tracing from the radiation-hydrodynamics mesh. This enables the laser mesh to be tailored to minimize ray noise with significantly fewer rays than would be required when the ray-tracing is performed on the radiation-hydrodynamics mesh, primarily by allowing the use of high-aspect-ratio cells that are not suitable for hydrodynamics solvers. For a planar target, we show that our method provides a ≈ 100× reduction in computational expense to achieve a fixed level of ray noise relative to ray-tracing directly on the radiation-hydrodynamics mesh. The relatively low ray requirement also enables efficient computation of cross-beam energy transfer. Each cell in the logically cubic laser mesh is a non-convex dodecahedron with triangular sides, and numerical integration of the ray trajectories and inverse bremsstrahlung is performed by mapping each cell to the unit cube. We will describe our methodology in detail as well as its implementation in the xRAGE radiation-hydrodynamics code, discuss performance, and present the results from applying the methodology to test problems with analytic solutions for laser ray-tracing through a quadratic density gradient with an analytic solution as well as for a laser-driven heat front. In 3D radiation-hydrodynamics simulations of laser-driven experiments performed on the National Ignition Facility, laser ray-tracing with our methodology uses less than 1% of total computational time while introducing acceptably low levels of ray noise.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Co-design for Particle Applications at Exascale

Co-design across the Exascale Computing Project (ECP) has been critical for both enabling science applications and bringing disparate communities together. Developing and porting applications to the various high-performance computing (HPC) architectures on pre-exascale and exascale computers has been quite challenging due to the diversity of hardware features and software stacks. The Co-design Center for Particle Applications (CoPA) has developed and enhanced the Cabana and PROGRESS/BML libraries to facilitate the creation of new particle applications, make existing particle applications exascale capable, and allow teams to explore new capabilities. Particle methods from atomistic, mesoscale, continuum, through cosmological scales have been built with Cabana, along with new possibilities for application coupling. Similarly, the PROGRESS/BML library has enabled quantum particle applications with linear algebra solvers to use advanced hardware. Across these CoPA-developed libraries, the co-design abstraction layer combines performance portability with math library support to facilitate separation of concerns and directly support science runs.

97 MATHEMATICS AND COMPUTING↗

CAMEO: A Co-design Architecture for Multi-objective Energy System Optimization (Project Report)

CAMEO (Codesign Architecture for Multi-objective Energy System Optimization) is a modular workflow management framework that abstracts co-design problems as Directed Acyclic Graphs (DAG). The framework employs JSON-based workflow specifications that enable systematic decomposition of complex optimization problems into reusable, interchangeable components including data loaders, scenario generators, optimization solvers, and result summarizers.

97 MATHEMATICS AND COMPUTING↗

3-D CFD in a day - The laser digitizer project

The computation of airflow over complex configurations requires a complete description of the geometry. This can be obtained from CAD data, from blueprints, or from actual models. In any case, the time required is currently estimated at 4 to 6 months. It is proposed to shorten this time by a factor of 10 to 100 through the use of automated software, a fast, highly parallel computer and a three-dimensional laser digitizer. This device can provide (x,y,z) coordinates of surface points at rates exceeding 14,500/sec. Thus, it is possible to digitize an entire model in a few minutes. The accuracy of measurement on a flat white surface is better than 0.005 inches. Higher accuracy is available at higher cost. This work discusses the challenges which remain to be addressed. In particular, the surface point data need to be converted into a surface description, the surface description needs to be made into a surface grid, and the surface grid used to make a volume grid for the flow solver. Algorithms are kept in place or in mind for all of these problems. Integration of the more mature flow solution and visualization algorithms then allows generation of solution graphics directly from a wind tunnel model.

Merriam, Marshal↗

CFD analysis on control of secondary losses in STME LOX turbines with endwall fences

The rotor blade in the newly designed LOX turbine for the future Space Transportation Main Engine (STME) has a severe flow turning angle, nearly 160 degrees. The estimated secondary loss in the rotor alone accounts for nearly 50 percent of the total loss over the entire stage. To reduce such a loss, one of the potential methods is to use fences attached on the turbine endwall (hub). As a prelude to examining the effects of endwall fence with actual STME turbine configuration, the present study focuses on similar issues with a different, but more generic, geometry - a rectangular duct with a 160-degree bend. The duct cross-section has a 2-to-1 aspect ratio and the radii of curvature for the inner and outer wall are 0.25 and 1.25 times the duct width, respectively. The present emphasis lies in examining the effects of various fence-length extending along the streamwise direction. The flowfield is numerically simulated using the FDNS code developed earlier by Wang and Chen. The FDNS code is a pressure based, finite-difference, Navier-Stokes equations solver.

Chyu, Mingking K.↗