Search NASA⌕ Search

SEARCH · Search NASA

Results for “splitting methods”

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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
NASA NTRS1987

Grid adaption for hypersonic flow

The methods of grid adaption are reviewed and a method is developed with the capability of adaption to several flow variables. This method is based on a variational approach and is an algebraic method which does not require the solution of partial differential equations. Also the method has been formulated in such a way that there is no need for any matrix inversion. The method is used in conjunction with the calculation of hypersonic flow over a blunt nose body. The equations of motion are the compressible Navier-Stokes equations where all viscous terms are retained. They are solved by the MacCormack time-splitting method. A movie has been produced which shows simultaneously the transient behavior of the solution and the grid adaption.

Abolhassani, Jamshid S.↗
NASA NTRS1993

A strategy for the efficient simulation of viscous compressible flows using a multi-domain pseudospectral method

The viscous compressible flow in a quasi-1D nozzle is presently simulated by means of a novel multidomain pseudospectral method in which interface-point flow variables are advanced in time on the basis of the second-order time-marching scheme used for the interior points. A spectrally accurate modification of Van Leer's (1982) flux-vector flux-splitting method is used to evaluate the spatial derivatives of the inviscid flux vector; these derivatives are alternately evaluated from neighboring subdomains, and the scheme is found to be spectrally accurate and to resolve shock waves without oscillations.

Hanley, Patrick↗
NASA NTRS1992

A comparison of artificial compressibility and fractional step methods for incompressible flow computations

We have applied and compared the efficiency and accuracy of two commonly used numerical methods for the solution of Navier-Stokes equations. The artificial compressibility method augments the continuity equation with a transient pressure term and allows one to solve the modified equations as a coupled system. Due to its implicit nature, one can have the luxury of taking a large temporal integration step at the expense of higher memory requirement and larger operation counts per step. Meanwhile, the fractional step method splits the Navier-Stokes equations into a sequence of differential operators and integrates them in multiple steps. The memory requirement and operation count per time step are low, however, the restriction on the size of time marching step is more severe. To explore the strengths and weaknesses of these two methods, we used them for the computation of a two-dimensional driven cavity flow with Reynolds number of 100 and 1000, respectively. Three grid sizes, 41 x 41, 81 x 81, and 161 x 161 were used. The computations were considered after the L2-norm of the change of the dependent variables in two consecutive time steps has fallen below 10(exp -5).

Chan, Daniel C.↗
NASA NTRS1989

On the application of subcell resolution to conservation laws with stiff source terms

LeVeque and Yee recently investigated a one-dimensional scalar conservation law with stiff source terms modeling the reacting flow problems and discovered that for the very stiff case most of the current finite difference methods developed for non-reacting flows would produce wrong solutions when there is a propagating discontinuity. A numerical scheme, essentially nonoscillatory/subcell resolution - characteristic direction (ENO/SRCD), is proposed for solving conservation laws with stiff source terms. This scheme is a modification of Harten's ENO scheme with subcell resolution, ENO/SR. The locations of the discontinuities and the characteristic directions are essential in the design. Strang's time-splitting method is used and time evolutions are done by advancing along the characteristics. Numerical experiment using this scheme shows excellent results on the model problem of LeVeque and Yee. Comparisons of the results of ENO, ENO/SR, and ENO/SRCD are also presented.

Chang, Shih-Hung↗
NASA NTRS2017

Aerosol Data Assimilation at GMAO

This presentation presents an overview of the aerosol data assimilation work performed at GMAO. The GMAO Forward Processing system and the biomass burning emissions from QFED are first presented. Then, the current assimilation of Aerosol Optical Depth (AOD), performed by means of the analysis splitting method is briefly described, followed by some results on the quality control of observations using a Neural Network trained using AERONET AOD. Some applications are shown such as the Mount Pinatubo eruption in 1991 using the MERRA-2 aerosol dataset. Finally preliminary results on the EnKF implementation for aerosol assimilation are presented.

Splitting↗
NASA NTRS2019

Assimilation of Aerosol Observations in the NASA GEOS Model

In the GEOS near real-time system, as well as in MERRA-2 which is the latest reanalysis produced at NASA's Global Modeling Assimilation Office (GMAO), the assimilation of aerosol observations is performed by means of a so-called analysis splitting method. The prognostic model is based on the GEOS model radiatively coupled to GOCART aerosol module and includes assimilation of bias-corrected Aerosol Optical Depth (AOD) at 550 nm from various space-based remote sensing platforms. In line with the transition of the GEOS meteorological data assimilation system to a hybrid Ensemble-Variational formulation, we are updating the aerosol component of our assimilation system to a variational ensemble type of scheme. In this talk we will examine the impact of replacing the current analysis splitting scheme with this new approach. Starting with the assimilation of satellite based single-channel retrievals; we will discuss the impact of this aerosol data assimilation technique on the 3D aerosol distributions by means of innovation statistics and verification against independent datasets such as the Aerosol Robotic Network (AERONET) and surface PM2.5. We will also present preliminary results related to the introduction of new aerosol data types in GEOS, including multi-spectral AOD retrievals.

AERONET↗
NASA NTRS2015

Determination of Parachute Joint Factors using Seam and Joint Testing

This paper details the methodology for determining the joint factor for all parachute components. This method has been successfully implemented on the Capsule Parachute Assembly System (CPAS) for the NASA Orion crew module for use in determining the margin of safety for each component under peak loads. Also discussed are concepts behind the joint factor and what drives the loss of material strength at joints. The joint factor is defined as a "loss in joint strength...relative to the basic material strength" that occurs when "textiles are connected to each other or to metals." During the CPAS engineering development phase, a conservative joint factor of 0.80 was assumed for each parachute component. In order to refine this factor and eliminate excess conservatism, a seam and joint testing program was implemented as part of the structural validation. This method split each of the parachute structural joints into discrete tensile tests designed to duplicate the loading of each joint. Breaking strength data collected from destructive pull testing was then used to calculate the joint factor in the form of an efficiency. Joint efficiency is the percentage of the base material strength that remains after degradation due to sewing or interaction with other components; it is used interchangeably with joint factor in this paper. Parachute materials vary in type-mainly cord, tape, webbing, and cloth -which require different test fixtures and joint sample construction methods. This paper defines guidelines for designing and testing samples based on materials and test goals. Using the test methodology and analysis approach detailed in this paper, the minimum joint factor for each parachute component can be formulated. The joint factors can then be used to calculate the design factor and margin of safety for that component, a critical part of the design verification process.

Mollmann, Catherine↗
NASA NTRS1994

Accretion disk coronae in high-luminosity systems

We present the results of self-consistent models of Compton-heated accretion disk coronae. The models are calculated using a new method for computing monochromatic radiative transfer n two dimensions. The method splits the radiation into direct and scattered components. The direct radiation is computed by calculating the optical depth along rays, while transfer of the scattered radiation is approximated by flux-limited diffusion. The resulting code agrees with more accurate treatments to within 50%, and is highly efficient, making it practical for use in large hydrodynamic simulations. The coronal models are used to confirm the results of earlier work, and to extend it to higher luminosities. In contrast to earlier work, which found the outer disks to be shadowed by the inner corona at high luminosities, we find our results to form an almost continuous extension of the models at lower luminosities. This is due to the presence of multiply scattered radiation, which acts to partially offset the loss of direct radiation from the central source. Although the analytic methods derived at lower luminosities cannot be used to derive the coronal structure for L/L(sub Edd) approx. greater than 0.1, the results of the models are amenable to semiempirical fits. We also discuss possible observational consequences of the results for coronal veiling and line fluorescence from the disk.

Murray, Stephen D.↗
NASA NTRS1982

A zonal approach to solution of the Euler equations

A technique for the solution of the one- and two-dimensional Euler equations in a partitioned flow field is presented. The field is divided into distinct 'zones', each of which is computed separately. An implicit boundary procedure based on the characteristic propagaion of information and flux splitting methods is applied at the zonal interfaces. Numerical results are presented for both quasi-one-dimensional nozzle flows with shock waves and the unsteady shock-tube problem. These calculations demonstrate the capability of shock propagation through arbitrarily located zonal boundaries in a stable, conservative, and accurate manner. Two-dimensional results include the zonal computation of flows over blunt bodies and airfoils.

Hessenius, K. A.↗
NASA NTRS1984

Evaluation of a velocity-split solution of the Navier-Stokes equations for fighter forebodies

A velocity-splitting method of solving the three-dimensional Navier-Stokes equations, originally designed for afterbody flows, is examined for its applicability for predicting fighter forebody flows. Results from the AFTEND Code are compared with wind tunnel data for two fighter configurations at a Mach number of 0.9 and angles-of-attack from 0 deg to 20 deg. Results compare well with data, and in areas where data do not exist, the viscous AFTEND results show realistic effects of viscosity compared to inviscid predictions. The inviscid results themselves are in general superior to results obtained from a small-disturbance transonic potential code.

Yaros, S. F.↗
NASA NTRS1987

Research in computational fluid dynamics

The numerical integration of quasi-one-dimensional unsteady flow problems which involve finite rate chemistry are discussed, and are expressed in terms of conservative form Euler and species conservation equations. Hypersonic viscous calculations for delta wing geometries is also examined. The conical Navier-Stokes equations model was selected in order to investigate the effects of viscous-inviscid interations. The more complete three-dimensional model is beyond the available computing resources. The flux vector splitting method with van Leer's MUSCL differencing is being used. Preliminary results were computed for several conditions.

Murman, Earll M.↗
NASA NTRS1987

Numerical solutions of Navier-Stokes equations for a Butler wing

The flow field is simulated on the surface of a Butler wing in a uniform stream. Results are presented for the Mach number 3.5 and Reynolds number of 2,000,000. The simulation is done by integrating the viscous Navier-Stokes equations. These equations govern the unsteady, viscous, compressible and heat conducting flow of an ideal gas. The equations are written in curvilinear coordinates so that the wing surface is represented accurately. O-type and H-type grids have been used for this study, and results are compared. The governing equations are solved by the McCormack time-split method, and the results are compared with other theoretical and experimental results. The codes are written in FORTRAN, vectorized and currently run on the CDC Vector Processing System (VPS-32) computer.

Abolhassani, Jamshid S.↗
NASA NTRS1987

A computational analysis of flow separation over five different airfoil geometries at high angles-of-attack

The flux splitting method of Van Leer (1982) and an implicit, upwind-biased, finite-volume scheme of Rumsey (1985) for solving the two-dimensional Reynolds-averaged Navier-Stokes equations were applied to analyze five airfoil geometries (the NACA 0012, 12-percent Joukowski, NACA 4412, NACA 65-015, and LS1-0417 sections), at high angles of attack. The effect of grid density and grid extent on the Strouhal number of the unsteady flow was computed, as well as the effect of variations in Mach number, Reynolds number, and angle of attack. It was found, that past an angle of attack of 20 deg, Strouhal numbers were independent of angle of attack (at a constant value of 0.115), agreeing well with the experimental findings of Tyler (1931).

Rumsey, Christopher L.↗
NASA NTRS1987

Numerical solutions of Navier-Stokes equations for a Butler wing

The flow field is simulated on the surface of a Butler wing in a uniform stream. Results are presented for Mach number 3.5 and Reynolds number of 2,000,000. The simulation is done by integrating the viscous Navier-Stokes equations. These equations govern the unsteady, viscous, compressible and heat conducting flow of an ideal gas. The equations are written in curvilinear coordinates so that the wing surface is represented accurately. The O-type and H-type grids have been used for this study, and results are compared. The governing equations are solved by the MacCormack time-split method, and the results are compared with other theoretical and experimental results. The codes are written in FORTRAN, vectorized and currently run on the CDC Vector Processing System (VPS-32) computer.

Abolhassani, Jamshid S.↗
NASA NTRS1988

A Study of Flow Separation in Transonic Flow Using Inviscid and Viscous Computational Fluid Dynamics (CFD) Schemes

A comparison of flow separation in transonic flows is made using various computational schemes which solve the Euler and the Navier-Stokes equations of fluid mechanics. The flows examined are computed using several simple two-dimensional configurations including a backward facing step and a bump in a channel. Comparison of the results obtained using shock fitting and flux vector splitting methods are presented and the results obtained using the Euler codes are compared to results on the same configurations using a code which solves the Navier-Stokes equations.

Rhodes, J. A.↗
NASA NTRS1987

Numerical simulation of homogeneous, isotropic, compressible turbulence

A new numerical algorithm is developed and applied to the direct simulation of compressible, homogeneous turbulent flows at low Mach numbers. A split method in time first solves a subset of the equations explicitly, followed by an implicit treatment of the pressure terms. The definition of an average sound speed over the entire field allows the implicit equations to be solved analytically, while relaxing the severe time limit imposed by the large sound speeds. Results from direct simulations on 96(3) grids provide a data base against which a new subgrid-scale model for compressible homogeneous turbulence is tested. This model reduces to the linear combination model in the absence of compressibility.

Erlebacher, G.↗
NASA NTRS1988

A numerical study of ENO and TVD schemes for shock capturing

The numerical performance of a second-order upwind-based total variation diminishing (TVD) scheme and that of a uniform second-order essentially non-oscillatory (ENO) scheme for shock capturing are compared. The TVD scheme used is a modified version of Liou, using the flux-difference splitting (FDS) of Roe and his superbee function as the limiter. The construction of the basic ENO scheme is based on Harten, Engquist, Osher, and Chakravarthy, and the 2-D extensions are obtained by using a Strang-type of fractional-step time-splitting method. Numerical results presented include both steady and unsteady, 1-D and 2-D calculations. All the chosen test problems have exact solutions so that numerical performance can be measured by comparing the computer results to them. For 1-D calculations, the standard shock-tube problems of Sod and Lax are chosen. A very strong shock-tube problem, with the initial density ratio of 400 to 1 and pressure ratio of 500 to 1, is also used to study the behavior of the two schemes. For 2-D calculations, the shock wave reflection problems are adopted for testing. The cases presented in this report include flows with Mach numbers of 2.9, 5.0, and 10.0.

Chang, Shih-Hung↗
NASA NTRS1989

Flux-split algorithms for the multi-dimensional Euler equations with real gases

Upwind algorithms are developed for the numerical solution of the multidimensional Euler equations for real gases. Flux-splitting methods are derived which account for a general equation of state. Approximations to the state equation based on physical arguments result in simplified algorithms which may be implemented into existing perfect-gas codes. Applications of the method to several high-Mach-number high-temperature flows are presented for two and three space dimensions.

Grossman, B.↗