SEARCH · Engineering Papers
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.
The present paper will compare temperature field measurements from selected cases in these investigations with distributions calculated with an empirical model based on assumed vertical profile similarity and superposition and with a 3-D elliptic code using a standard K-E turbulence model. The results will show the capability (or lack thereof) of the models to predict the effects of the principle flow and geometric variables.
Due to complexity of the program which used a hydrodynamic, axially and radially deformable disk-ring model and the resulting long computing time only the output gap was investigated. Results from independent studies were used to initiate the starting conditions for the electrons and the RF voltage using our program. Although this method of computation is less exact than processing the entire klystron interaction in 3-Dimensions it is shown that, for a confined flow focused throughout the penultimate cavity, radial velocities remain very small and the beam is highly laminar. It is concluded that possible errors resulting from treating only the output cavity in 3-D would remain small.
An interactive computer code, written for a microcomputer, is presented which displays 2-D and 3-D oblique plots of the temperature distribution downstream of jets mixing with a confined crossflow, for either single-side or opposed jet injection. Temperature profiles calculated with this routine are presented to show the effects of flow and geometric variables on the mixing. Examples are shown to illustrate the different perspectives on the mixing available by exercising various view options. In addition, the program is used to calculate profiles for opposed rows of jets with their centerlines in-line, by assuming that the confining effect of an opposite wall is equivalent to that of a plane of symmetry between opposed jets.
A procedure for using an efficient axisymmetric code to generate downstream pressure input for more costly Euler codes is discussed. Two and three dimensional inviscid solutions for the flow within a transonic axial compressor rotor at design speed are compared to laser anemometer measurements at maximum flow and near stall operating points. Computational details of the 2-D axisymmetric stream function solution and the 3-D full Euler solution are described. Relative Mach number contours, shock location, and shock strength as measured and as predicted by the 3-D code are compared. Downstream of the rotor the inviscid computations agree with each other but predict higher pressure ratios than those measured. Euler codes require a downstream pressure as input. Since that pressure controls the computed mass flow and shock system, it must be consistent with an inviscid solution.
This paper compares temperature field measurements from selected experiments on a single row, and opposed rows, of jets injected into a ducted crossflow with profiles calculated using an empirical model based on assumed vertical profile similarity and superposition, and distributions calculated with a 3-D elliptic code using a standard K-E turbulence model. The empirical model predictions are very good within the range of the generating experiments, and the numerical model resultings, although exhibiting too little mixing, correctly describe the effects of the principal flow and geometric variables.
This paper compares temperature field measurements from selected experiments on a single row, and opposed rows, of jets injected into a ducted crossflow with profiles calculated using an empirical model based on assumed vertical profile similarity and superposition, and distributions calculated with a 3-D elliptic code using a standard K-E turbulence model. The empirical model predictions are very good within the range of the generating experiments, and the numerical model results, although exhibiting too little mixing, correctly describe the effects of the principal flow and geometric variables.
A model is presented in which externaly driven reconnection is simulated by solving the MHD equations in an initially plane current sheet. Magnetic reconnection in the earth's magnetotail is widely believed to be the direct cause of magnetospheric substorms. Both 2-D and 3-D versions of the model have been developed. It is postulated that connection in the tail is triggered by a local compression of the plasma sheet which results from an invasion of the solar wind into the magnetotail. Thus, the simulation is started by introducing flow from the lobes normal to the plasma sheet. When resistivity is generated in a local region of the neutral sheet, reconnection develops and magnetic energy is converted into plasma bulk flow. Although the driven reconnection model is highly simplified, it can aid in understanding many features of substorms in the tail; in particular, results show that rapid flows both earthward and tailward of the neutral line and the nightside substorm current system are natural consequences of driven magnetic reconnection.
A hybrid numerical algorithm, developed to solve the full three-dimensional Navier-Stokes equations, is applied to the computation of the flowfield in a simulated three-dimensional high speed aircraft inlet at a Mach number of 2.5 and Reynolds number of 1.4 x 10 to the 7th based on inlet length. The numerical algorithm incorporates a coordinate transformation in order to handle general flow geometries, and utilizes the algebraic turbulent eddy viscosity model of Baldwin and Lomax. The hybrid algorithm has been vectorized on the CDC CYBER 203 computer using the SL/1 vector programming language developed at NASA Langley. The computed results are compared with experimental measurements of the ramp and cowl static pressures, and boundary layer pitot profiles. The results are also compared with a previous two-dimensional Navier-Stokes computation of the same configuration. The agreement with the experimental data is generally good; however, additional improvements in turbulence modeling are needed.
This paper is concerned with the propagation of laser light through a slab of a randomly varying medium. A theoretical analysis is presented which relates the spectrum of the recorded-intensity field some distance downstream of the medium to the spectrum of the index-of-refraction field. For a homogeneous and isotropic random field, the 3-D spectrum of the medium is obtained from the 2-D spectrum of the photograph by dividing each component of the spectrum by the frequency raised to the fourth power. Free-space propagation outside the random medium is accounted for by a scaling factor. Experimental results are presented which support the theoretical analysis. The nonintrusive diagnostic technique presented here is applicable to photographs which contain partially developed caustic networks.
An approach to solution adaptive grid generation for use with finite difference techniques, previously demonstrated on model problems in one space dimension, has been extended to multidimensional problems. The method is based on the popular elliptic steady grid generators, but is 'dynamically' adaptive in the sense that a grid is maintained at all times satisfying the steady grid law driven by a solution-dependent source term. Testing has been carried out on Burgers' equation in one and two space dimensions. Results appear encouraging both for inviscid wave propagation cases and viscous boundary layer cases, suggesting that application to practical flow problems is now possible. In the course of the work, obstacles relating to grid correction, smoothing of the solution, and elliptic equation solvers have been largely overcome. Concern remains, however, about grid skewness, boundary layer resolution and the need for implicit integration methods. Also, the method in 3-D is expected to be very demanding of computer resources.
The hypersonic, laminar flow around the Space Shuttle Orbiter has been computed for both an ideal gas (gamma = 1.2) and equilibrium air using a real-gas, parabolized Navier-Stokes code. This code employs a generalized coordinate transformation; hence, it places no restrictions on the orientation of the solution surfaces. The initial solution in the nose region was computed using a 3-D, real-gas, time-dependent Navier-Stokes code. The thermodynamic and transport properties of equilibrium air were obtained from either approximate curve fits or a table look-up procedure. Numerical results are presented for flight conditions corresponding to the STS-3 trajectory. The computed surface pressures and convective heating rates are compared with data from the STS-3 flight.
During entry of the Space Shuttle Orbiter, the convective heating within inter-tile gaps of the Thermal Protection System (TPS) material produces elevated tile sidewall temperatures in regions of high surface pressure gradient. Arc-jet tests have been conducted recently to obtain a measure of the gap heating down the TPS tile sidewalls at test conditions representative of Orbiter flight environments. The object of this paper is to present the gap heating correlations that were developed from a thermal analysis for 3-D curved and flat TPS tile segments. Predictions of gap sidewall temperature were obtained within 30 F of test data on both Wing Glove and Double Wedge models. Derived heating ratios were obtained for a range of test conditions (pressure, pressure gradient, enthalpy, boundary layer thickness, gap width, surface temperature, etc.). The results of the study, which showed that heating ratios varied with the pressure gradient times the square root of the surface pressure, are being used to provide an assessment of gap filler requirements on Orbiter forward fuselage/chine and wing glove regions.
The PLANS system is described which is a finite element program for nonlinear analysis. The system represents a collection of special purpose computer programs each associated with a distinct physical problem class. Modules of PLANS specifically referenced and described in detail include: (1) REVBY, for the plastic analysis of bodies of revolution; (2) OUT-OF-PLANE, for the plastic analysis of 3-D built-up structures where membrane effects are predominant; (3) BEND, for the plastic analysis of built-up structures where bending and membrane effects are significant; (4) HEX, for the 3-D elastic-plastic analysis of general solids; and (5) OUT-OF-PLANE-MG, for material and geometrically nonlinear analysis of built-up structures. The SATELLITE program for data debugging and plotting of input geometries is also described. The theoretical foundations upon which the analysis is based are presented. Discussed are the form of the governing equations, the methods of solution, plasticity theories available, a general system description and flow of the programs, and the elements available for use.
A three-dimensional (3-D) chimera grid-embedding technique is described. The technique simplifies the construction of computational grids about complex geometries. The method subdivides the physical domain into regions which can accommodate easily generated grids. Communication among the grids is accomplished by interpolation of the dependent variables at grid boundaries. The procedures for constructing the composite mesh and the associated data structures are described. The method is demonstrated by solution of the Euler equations for the transonic flow about a wing/body, wing/body/tail, and a configuration of three ellipsoidal bodies.
The problem of the potential incompressible flow about a helicopter rotor blade is solved using an unsteady vortex-panel method where the mutual interaction between the blade and the distorting free wake is taken into account. The present method alleviates the need to rely upon measured-wake geometries or p5escribed-wake models in order to calculate the airloads. A computer program has been developed which is capable of predicting the geometry of the time-dependent three-dimensional (3-D) wake and the instantaneous loadings for a single blade in hover, climb, and forward flight. The solution is obtained by using a time-accurate step-by-step procedure. The complex-wake geometry at any time is presented graphically with a computer graphics system. Calculated results are compared with published data for a rotor blade in both hover and forward flight. The code has also been applied to the study of the effect of changing blade tip geometry.
Rarefied flow fields and heating rates are calculated by a Monte Carlo direct simulation technique for several shuttle orbiter configurations at angles of attack between 40 and 60 deg. Continuum inviscid flow fields are determined by 2-D unsteady and 3-D steady finite-difference/artificial viscosity techniques used in combination with a 2-D shock layer analysis method. Continuum viscous flow fields and heating rates are computed by a boundary-layer integral matrix technique and an aerodynamic heating method.
A numerical minimization scheme is used in conjunction with two-dimensional and three-dimensional inviscid transonic flow analysis codes to provide procedures for wing leading edge aerodynamic design. The procedures are demonstrated in the design of a new leading edge to improve C-141 cruise performance. For the high aspect ratio moderately swept C-141 wing, the 2-D procedure is shown to yield results which are in close agreement with those obtained using the 3-D technique. Although the 2-D approach uses much less computation time than the 3-D technique, the latter requires fewer manhours than the former. Comparisons of predicted and wind tunnel measured performance improvements are presented which verify the design procedures.
This paper discusses a number of algorithms for solving the transonic full-potential equation in conservative form on a vector computer, such as the CDC STAR-100 or the CRAY-1. Recent research with the 'artificial density' method for transonics has led to development of some new iteration schemes which take advantage of vector-computer architecture without suffering significant loss of convergence rate. Several of these more promising schemes are described and 2-D and 3-D results are shown comparing the computational rates on the STAR and CRAY vector computers, and the CYBER-175 serial computer. Schemes included are: (1) Checkerboard SOR, (2) Checkerboard Leapfrog, (3) odd-even vertical line SOR, and (4) odd-even horizontal line SOR.