Search NASA⌕ Search

Engineering topics

Morel, J. E.

Publications and source records attributed to Morel, J. E..

The equilibrium-diffusion limit for radiation hydrodynamics

The equilibrium-diffusion approximation (EDA) is used to describe certain radiation-hydrodynamic (RH) environments. When this is done the RH equations reduce to a simplified set of equations. The EDA can be derived by asymptotically analyzing the full set of RH equations in the equilibrium-diffusion limit. Here, we derive the EDA this way and show that it and the associated set of simplified equations are both first-order accurate with transport corrections occurring at second order. Having established the EDA’s first-order accuracy we then analyze the grey nonequilibrium-diffusion approximation and the grey Eddington approximation and show that they both preserve this first-order accuracy. Further, these approximations preserve the EDA’s first-order accuracy when made in either the comoving-frame (CMF) or the lab-frame (LF). And while analyzing the Eddington approximation, we found that the CMF and LF radiation-source equations are equivalent when neglecting O(β 2 ) terms and compared in the LF. Of course, the radiation pressures are not equivalent. It is expected that simplified physical models and numerical discretizations of the RH equations that do not preserve this first-order accuracy will not retain the correct equilibrium-diffusion solutions. As a practical example, we show that nonequilibrium-diffusion radiative-shock solutions devolve to equilibrium-diffusion solutions when the asymptotic parameter is small.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Nonrelativistic grey S n -transport radiative-shock solutions

We present semi-analytic radiative-shock solutions in which grey Sn-transport is used to model the radiation, and we include both constant cross sections and cross sections that depend on temperature and density. These new solutions solve for a variable Eddington factor (VEF) across the shock domain, which allows for interesting physics not seen before in radiative-shock solutions. Comparisons are made with the grey nonequilibrium-diffusion radiative-shock solutions of Lowrie and Edwards [1], which assumed that the Eddington factor is constant across the shock domain. It is our experience that the local Mach number is monotonic when producing nonequilibrium-diffusion solutions, but that this monotonicity may disappear while integrating the precursor region to produce Sn-transport solutions. For temperature- and density-dependent cross sections we show evidence of a spike in the VEF in the far upstream portion of the radiative-shock precursor. We show evidence of an adaptation zone in the precursor region, adjacent to the embedded hydrodynamic shock, as conjectured by Drake [2, 3], and also confirm his expectation that the precursor temperatures adjacent to the Zel’dovich spike take values that are greater than the downstream post-shock equilibrium temperature. We also show evidence that the radiation energy density can be nonmonotonic under the Zel’dovich spike, which is indicative of anti-diffusive radiation flow as predicted by McClarren and Drake [4]. We compare the angle dependence of the radiation flow for the Sn-transport and nonequilibriumdiffusion radiation solutions, and show that there are considerable differences in the radiation flow between these models across the shock structure. Lastly, we analyze the radiation flow to understand the cause of the adaptation zone, as well as the structure of the Sn-transport radiation-intensity solutions across the shock structure.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

COMPARISONS OF THE FINITE-ELEMENT-WITH-DISCONTIGUOUS-SUPPORT METHOD TO CONTINUOUS-ENERGY MONTE CARLO FOR PIN-CELL PROBLEMS

The standard multigroup (MG) method for energy discretization of the transport equation can be sensitive to approximations in the weighting spectrum chosen for cross-section averaging. As a result, MG often inaccurately treats important phenomena such as self-shielding variations across a material. From a finite-element viewpoint, MG uses a single fixed basis function (the pre-selected spectrum) within each group, with no mechanism to adapt to local solution behavior. In this work, we introduce the Finite-Element-with-Discontiguous-Support (FEDS) method, whose only approximation with respect to energy is that the angular flux is a linear combination of unknowns multiplied by basis functions. A basis function is non-zero only in the discontiguous set of energy intervals associated with its energy element. Discontiguous energy elements are generalizations of bands and are determined by minimizing a norm of the difference between snapshot spectra and their averages over the energy elements. We begin by presenting the theory of the FEDS method. We then compare to continuous-energy Monte Carlo for one-dimensional slab and two-dimensional pin-cell problem. We find FEDS to be accurate and efficient at producing quantities of interest such as reaction rates and eigenvalues. Results show that FEDS converges at a rate that is approximately first-order in the number of energy elements and that FEDS is less sensitive to weighting spectrum than standard MG.

Cross sections↗

Neutron transport in Eulerian coordinates with bulk material motion

A consistent, numerically stable algorithm for the solution of the neutron transport equation in the presence of a moving material background is presented for one-dimensional spherical geometry. Manufactured solutions are used to demonstrate the correctness and stability of our numerical algorithm. The importance of including moving material corrections is shown for the r-process in proto-neutron stars. (authors)

97 MATHEMATICS AND COMPUTING↗

A new least-squares transport equation compatible with voids

We define a new least-squares transport equation that is applicable in voids, can be solved using source iteration with diffusion-synthetic acceleration, and requires only the solution of an independent set of second-order self-adjoint equations for each direction during each source iteration. We derive the equation, discretize it using the S{sub n} method in conjunction with a linear-continuous finite-element method in space, and computationally demonstrate various of its properties. (authors)

97 MATHEMATICS AND COMPUTING↗

Exponentially-convergent Monte Carlo for the 1-D transport equation

We define a new exponentially-convergent Monte Carlo method for solving the one-speed 1-D slab-geometry transport equation. This method is based upon the use of a linear discontinuous finite-element trial space in space and direction to represent the transport solution. A space-direction h-adaptive algorithm is employed to restore exponential convergence after stagnation occurs due to inadequate trial-space resolution. This methods uses jumps in the solution at cell interfaces as an error indicator. Computational results are presented demonstrating the efficacy of the new approach. (authors)

97 MATHEMATICS AND COMPUTING↗

A new indirect measure of diffusion model error

We define a new indirect measure of the diffusion model error called the diffusion model error source. When this model error source is added to the diffusion equation, the transport solution for the angular-integrated intensity is obtained. This source represents a means by which a transport code can be used to generate information relating to the adequacy of diffusion theory for any given problem without actually solving the diffusion equation. The generation of this source does not relate in any way to acceleration of the iterative convergence of transport solutions. Perhaps the most well-known indirect measure of the diffusion model error is the variable-Eddington tensor. This tensor provides a great deal of information about the angular dependence of the angular intensity solution, but it is not always simple to interpret. In contrast, our diffusion model error source is a scalar that is conceptually easy to understand. In addition to defining the diffusion model error source analytically, we show how to generate this source numerically relative to the S{sub n} radiative transfer equations with linear-discontinuous spatial discretization. This numerical source is computationally tested and shown to reproduce the Sn solution for a Marshak-wave problem. (authors)

97 MATHEMATICS AND COMPUTING↗

Second-order discretization in space and time for radiation hydrodynamics

We present a method for solving the equations of radiation hydrodynamics that is second-order accurate in space and time. This method combines the MUSCL-Hancock method for solving the Euler equations with the TR/BDF2 scheme in time for solving the equations of radiative transfer. We use an LDFEM to discretize the radiative transfer equations in space, which, though uncommon for radiation diffusion calculations, is a standard for radiation transport applications. We address the challenges inherent to using different spatial discretizations for the hydrodynamics and radiation and demonstrate how these may be overcome. We define our method for a 1-D model of compressible fluid dynamics coupled with grey radiation diffusion. Using the method of manufactured solutions, we show that the method is second-order accurate in space and time for both the equilibrium diffusion and streaming limit. (authors)

97 MATHEMATICS AND COMPUTING↗

Characterization of high order spatial discretizations and lumping techniques for discontinuous finite element SN transport

We examine several possible methods of mass matrix lumping for discontinuous finite element discrete ordinates transport using a Lagrange interpolatory polynomial trial space. Though positive outflow angular flux is guaranteed with traditional mass matrix lumping in a purely absorbing 1-D slab cell for the linear discontinuous approximation, we show that when used with higher degree interpolatory polynomial trial spaces, traditional lumping does yield strictly positive outflows and does not increase in accuracy with an increase in trial space polynomial degree. As an alternative, we examine methods which are 'self-lumping'. Self-lumping methods yield diagonal mass matrices by using numerical quadrature restricted to the Lagrange interpolatory points. Using equally-spaced interpolatory points, self-lumping is achieved through the use of closed Newton-Cotes formulas, resulting in strictly positive outflows in pure absorbers for odd power polynomials in 1-D slab geometry. By changing interpolatory points from the traditional equally-spaced points to the quadrature points of the Gauss-Legendre or Lobatto-Gauss-Legendre quadratures, it is possible to generate solution representations with a diagonal mass matrix and a strictly positive outflow for any degree polynomial solution representation in a pure absorber medium in 1-D slab geometry. Further, there is no inherent limit to local truncation error order of accuracy when using interpolatory points that correspond to the quadrature points of high order accuracy numerical quadrature schemes. (authors)

97 MATHEMATICS AND COMPUTING↗

S{sub 2}SA preconditioning for the S{sub n} equations with strictly non negative spatial discretization

Preconditioners based upon sweeps and diffusion-synthetic acceleration have been constructed and applied to the zeroth and first spatial moments of the 1-D S{sub n} transport equation using a strictly non negative nonlinear spatial closure. Linear and nonlinear preconditioners have been analyzed. The effectiveness of various combinations of these preconditioners are compared. In one dimension, nonlinear sweep preconditioning is shown to be superior to linear sweep preconditioning, and DSA preconditioning using nonlinear sweeps in conjunction with a linear diffusion equation is found to be essentially equivalent to nonlinear sweeps in conjunction with a nonlinear diffusion equation. The ability to use a linear diffusion equation has important implications for preconditioning the S{sub n} equations with a strictly non negative spatial discretization in multiple dimensions. (authors)

97 MATHEMATICS AND COMPUTING↗

Application of nonlinear Krylov acceleration to radiative transfer problems

The iterative solution technique used for radiative transfer is normally nested, with outer thermal iterations and inner transport iterations. We implement a nonlinear Krylov acceleration (NKA) method in the PDT code for radiative transfer problems that breaks nesting, resulting in more thermal iterations but significantly fewer total inner transport iterations. Using the metric of total inner transport iterations, we investigate a crooked-pipe-like problem and a pseudo-shock-tube problem. Using only sweep preconditioning, we compare NKA against a typical inner / outer method employing GMRES / Newton and find NKA to be comparable or superior. Finally, we demonstrate the efficacy of applying diffusion-based preconditioning to grey problems in conjunction with NKA. (authors)

97 MATHEMATICS AND COMPUTING↗

Hybrid S{sub n}-Diffusion and S{sub n}-P{sub 3} transport calculations

We investigate coupling and preconditioning techniques for 1-D hybrid neutronics calculations. Each problem is represented by two spatial regions with S{sub n} used in one region and either diffusion or P{sub 3} used in the other. For each of these two cases we define one coupling scheme and two preconditioned systems. These systems are solved with both fixed-point iteration and the GMRES Krylov method. The solution techniques are compared in terms of iteration count and CPU time per iteration. Preconditioning with a global diffusion operator is found to be very effective for the most difficult problems. (authors)

97 MATHEMATICS AND COMPUTING↗

Comparison of two Galerkin quadrature methods

We compare two methods for generating Galerkin quadrature for problems with highly forward-peaked scattering. In Method 1, the standard Sn method is used to generate the moment-to-discrete matrix and the discrete-to-moment is generated by inverting the moment-to-discrete matrix. In Method 2, which we introduce here, the standard Sn method is used to generate the discrete-to-moment matrix and the moment-to-discrete matrix is generated by inverting the discrete-to-moment matrix. Method 1 has the advantage that it preserves both N eigenvalues and N eigenvectors (in a pointwise sense) of the scattering operator with an N-point quadrature. Method 2 has the advantage that it generates consistent angular moment equations from the corresponding S{sub N} equations while preserving N eigenvalues of the scattering operator with an N-point quadrature. Our computational results indicate that these two methods are quite comparable for the test problem considered. (authors)

97 MATHEMATICS AND COMPUTING↗