SEARCH · Search NASA
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.
Buxbaum has reported on three algorithms for computing least squares estimates that are based on fixed amounts of data. In this correspondence, the filter is arranged as a point-deleting Kalman filter concatenated with the standard point-inclusion Kalman filter. The resulting algorithm is couched in a square root framework for greater numerical stability, and special attention is given to computer implementation.
A class of linear multistep methods is proposed for the solution of the equations of motion of certain dynamical systems encountered in celestial mechanics and astrodynamics. These methods are distinguished from the classical predictor-corrector methods in that they permit 'back-corrections' of the solution to be made. As the integration advances in time, the numerical solution is corrected or improved at certain points in the past. The enhanced numerical stability of these methods allows the meaningful application of high-order algorithms. Consequently, step sizes larger than those attainable with the classical methods may be adopted, and greater overall efficiency may be realized. These methods are applied to the problem of determining the orbit of an artificial satellite, and the results are compared with those obtained using classical methods.
A detailed description is given of how the decoupling approximation known as the doubly asymptotic approximation is implemented with NASTRAN to solve shock problems for submerged structures. The general approach involves locating the nonsymmetric terms (which couple structural and fluid variables) on the right hand side of the equations. This approach results in coefficient matrices of acceptable bandwidth but degrades numerical stability, requiring a smaller time step size than would otherwise be used. It is also shown how the structure's added (virtual) mass matrix, is calculated with NASTRAN.
The design and usage of two main subprograms using direct methods to solve large linear complex systems, of the form Ax = b, whose coeffficient matrices are too large to be stored in core are described. The first main subprogram is for systems whose coefficient matrices are of a particular sparse structure, namely, the matrix A can be written in the form B + D, where B is a block-banded system, and D has only a few columns of nonzeros. Key elements of the algorithms used in the subprograms include: the data structure, the strategy for preserving numerical stability, the adaptability of the algorithms for dense systems as well as for block-profile systems.
An implicit, shock-capturing finite-difference code which is used to calculate two-dimensional inlet flow fields in a supersonic free stream is explained. The Euler equations are subjected to general nonorthogonal transformation and a body-fitted coordinate system is employed. The mathematical formulation of the problem is given along with the numerical algorithm. Initial and boundary conditions, numerical stability, program limitations, and accuracy is discussed. An overall program logic as well as instructions for program use and operation are also furnished.
Modifications to explicit finite difference schemes for solving the shallow water equations for meteorological applications by increasing the time step for the fast gravity waves are analyzed. Terms associated with the gravity waves in the shallow water equations are treated on a coarser grid than those associated with the slow Rossby waves, which contain much more of the available energy and must be treated with higher accuracy, enabling a several-fold increase in time step without degrading the accuracy of the solution. The method is presented in Cartesian and spherical coordinates for a rotating earth, using generalized leapfrog, frozen coefficient, and Fourier filtering finite difference schemes. Computational results verify the numerical stability of the approach.
A technique of deriving Hovenier's exit function equation (1978) for multiple scattering, starting with a set of invariant imbedding equations, is presented. The feasibility of the exit function equation as a means to obtain reflection and transmission functions is investigated for isotropic, Rayleigh, and Henyey-Greenstein phase functions with emphasis on the numerical stability, accuracy and timing. It is possible to compute the reflection function rather accurately with an efficiency comparable to that of the standard doubling technique for the same phase functions with moderate anisotropy. The resulting transmission is slightly less accurate than the reflection function, but it may be acceptable for practical purposes.
An additional approach is being taken to address the problem of how to decide where to truncate spherical harmonic representations of geomagnetic data. The idea is to simply evaluate some aspect of interest at successively higher levels of truncation and then to chose that value of truncation level (N) which gives the most reasonable appearance of convergence. The preliminary determination of the onset of relative numerical stability at a truncation level of N=7 in two independent studies is thought to be a potentially significant result of practical value for field modeling. It requires further substantiation and explanation before dissemination.
Inlet flow fields for airbreathing missiles are calculated by the adaptation of a two dimensional computational method developed for the flow around airfoils. A supersonic free stream is assumed to allow the forebody calculation to be uncoupled from the inlet calculation. The inlet calculation employs an implicit, time marching finite difference procedure to solve the thin layer Navier-Stokes equations formulated in body fitted coordinates. The mathematical formulation of the problem and the solution algorithm are given. Numerical stability and accuracy as well as the initial and boundary conditions used are discussed. Instructions for program use and operation along with the overall program logic are also given.
An implicit finite difference scheme for an efficient computation of unsteady potential flow about airfoils is presented. The formulation uses density and velocity potential as dependent variables, and is cast in conservation form to assure the theoretically correct determination of shockwave location and speed. To enable boundary conditions to be imposed directly on the airfoil surface, a time varying sheared rectilinear coordinate transformation is employed. Calculated time history solutions on a pulsating airfoil are compared with the results of another unsteady transonic code. It is concluded that the method has excellent numerical stability and gives accurate solutions with sharply resolved shocks.
An existing three-dimensional compressible integral boundary-layer method was modified to account for mean dilatation effects, to model transition properly, and to provide better numerical stability near computational boundaries. Results of this method were compared against those from a three-dimensional finite-difference boundary-layer method on a difficult test case. An interaction procedure was developed to couple this integral method with a number of wing-alone and wing-body transonic potential codes to account for viscous effects. A strip wake model, including thickness and curvature effects, was developed and incorporated into this interaction procedure. Results from this procedure were compared against experimental data and results from previous procedures, on test cases where viscous effects were large.
A sample heat transfer analysis is demonstrated which includes the heat of fusion. The method can be used to analyze a system with nonconstant specific heat. The enthalpy is introduced as an independent degree of freedom at each node. The user input consists of a curve of temperature as a function of enthalpy, which may include a constant temperature phase change. The basic NASTRAN heat transfer capability is used to model the effects of latent heat with existing direct matrix output and nonlinear load data cards. Although some user care is required, the numerical stability of the integration is quite good when the given recommendations are followed. The theoretical equations used and the NASTRAN techniques are shown.
Turbulence modeling methods for the compressible Navier-Stokes equations, including several zero- and two-equation eddy-viscosity models, are described and applied. Advantages and disadvantages of the models are discussed with respect to mathematical simplicity, conformity with physical theory, and numerical compatibility with methods. A new two-equation model is introduced which shows advantages over other two-equation models with regard to numerical compatibility and the ability to predict low-Reynolds-number transitional phenomena. Calculations of various transonic airfoil flows are compared with experimental results. A new implicit upwind-differencing method is used which enhances numerical stability and accuracy, and leads to rapidly convergent steady-state solutions.
A new, directionally dependent, hybrid numerical algorithm for solving the unsteady, three-dimensional Navier-Stokes equations has been developed and used to compute the viscous supersonic flow over complex configurations, which may generate local regions of embedded subsonic or streamwise separated flows or both. The new hybrid implicit-explicit algorithm is derived from the more general implicit Beam-Warming algorithm and is particularly suitable for viscous computations in which the grid spacing in the direction outward from the body is considerably smaller than the spacing in the other two directions. Numerical results obtained from both the hybrid and implicit schemes are presented and compared on the basis of numerical stability, convergence history, and computer and core memory requirements.
Some of the methods of obtaining approximate numerical solutions to boundary value problems that arise in elastohydrodynamic lubrication are reviewed. The highlights of four general approaches (direct, inverse, quasi-inverse, and Newton-Raphson) are sketched. Advantages and disadvantages of these approaches are presented along with a flow chart showing some of the details of each. The basic question of numerical stability of the elastohydrodynamic lubrication solutions, especially in the pressure spike region, is considered. Computers used to solve this important class of lubrication problems are briefly described, with emphasis on supercomputers.
A finite element computer program is developed for evaluating the transient behavior of surface temperature in high performance spur gears. The time dimension is implemented using two and three point finite difference schemes. The different schemes are provided for the purpose of numerical stability and convergence studies. A detailed explanation of the gear cooling process leading to the establishment of a modified Blok model is also included. Other conventional models for approximating the heat transfer coefficients are available for comparison. Preliminary results are given showing snap shots of gear temperature contours at the initial stages of tooth engagement.
A numerical procedure is described for predicting the viscoelastic response of general laminates. A nonlinear compliance model is used to predict the creep response of the individual laminae. A biaxial delayed failure model predicts ply failure. The numerical procedure, based on lamination theory, increases by increments through time to predict creep compliance and delayed failures in laminates. Numerical stability problems and experimental verification are discussed. Although the program has been quite successful in predicting creep of general laminates, the assumptions associated with lamination theory have resulted in erroneous bounds on the predicted material response. Delayed failure predictions have been conservative. Several improvements are suggested to increase the accuracy of the procedure.
A nonuniform transonic airfoil code is developed for applications in analysis, inverse design and direct optimization involving an airfoil immersed in propfan slipstream. Problems concerning the numerical stability, convergence, divergence and solution oscillations are discussed. The code is validated by comparing with some known results in incompressible flow. A parametric investigation indicates that the airfoil lift-drag ratio can be increased by decreasing the thickness ratio. A better performance can be achieved if the airfoil is located below the slipstream center. Airfoil characteristics designed by the inverse method and a direct optimization are compared. The airfoil designed with the method of direct optimization exhibits better characteristics and achieves a gain of 22 percent in lift-drag ratio with a reduction of 4 percent in thickness.