SEARCH · Engineering Papers

Results for “Iteration method”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 793 records · Page 44

All sources NASA NTRS DOE OSTI

NASA NTRS1978

Semidirect calculation of steady two- and three-dimensional flows

This paper describes a semidirect method for rapidly solving steady-state viscous flows described by the complete Navier-Stokes equations. The current results are for two-dimensional incompressible flows in general channels at arbitrary Reynolds numbers, but work in progress on compressible and three-dimensional flows is also described. The basic concept of semidirect methods is to use the recently developed fast (direct, or noniterative) linear solvers to solve linearized equations, which are then iterated to solve the nonlinearity. The method used here is an extension of the Split NOS method (Roache, 1975)

Roache, P. J.↗

NASA NTRS1991

Automated pattern eduction from turbulent flow diagnostics

The development of an automated technique for the eduction of 3-D spatial patterns in vector or scalar diagnostics was completed. The method is based on an iterative convolution between a trial pattern and the data field. It was applied to the analysis of low Reynolds number turbulent channel flow and homogeneous shear flow. The results yielded new information on the dominant flow structures in these flows, particularly with respect to the spatial relationships between various forms of organized motion. A particular application of the pattern eduction method, which is tentatively referred to as an 'adaptive wavelet transformation', is proposed with the objective of investigating the way turbulence structure changes with scale. Preliminary results using data from homogeneous turbulent shear flow simulations are presented. At the low Reynolds numbers of the simulations, there is no evidence of scale similarity. The small scales appear to be associated with the edges of the larger scale vortical structures.

Stretch, Derek D.↗

NASA NTRS1990

An alternative formulation of the global sensitivity equations

To optimize the performance of any system, the sensitivity derivatives of the system's output variables with respect to its input variables must be readily available. It is also desirable that these derivatives be inexpensive to calculate as the optimization process requires many evaluations of the output variables and their derivatives. Optimization methods that have been developed for use in automated structural design programs may not be extended for use in integrated multidisciplinary design programs until adequate means of calculating accurate sensitivity derivatives of complex, internally coupled systems have been developed. Until the development of the Global Sensitivity Equations (GSE), the only method of determining the sensitivity derivatives of coupled systems has been by using finite differences. Analytical or semi-analytical derivatives do not exist as there is no analytical solution to the coupled problem. Also, difficulties arise because the finite difference method is expensive as the system has to iterate to a converged solution for each incremental input variable. The method may not be accurate, and the choice of the input variable increment may cause the difference in the output variable to be insignificant compared to computer numerical error if the choice is too small, or the process may not predict the true value of the output variable if the increment is too large. The GSE allow the system's sensitivity derivatives to be calculated as functions of the component subsystem's (local) sensitivity derivatives. These local sensitivity derivatives are calculated from specifically decoupled subsystems, whereas the GSE account for total system coupling. Since the subsystems are decoupled, it may be possible for the local derivatives to be calculated by analytical or semi-analytical methods, which generally reduce cost and improve accuracy. Several academic problems have been solved using GSE and have demonstrated encouraging results.

James, Benjamin B.↗

NASA NTRS1977

An efficient numerical method for solving the incompressible Navier-Stokes equations

This paper describes an efficient numerical method for solving the steady incompressible Navier-Stokes equations. The method is a fully implicit method based on the generalized Galerkin method, and the resulting system of equations is solved in a sweeping mode by iterative line relaxation. Results of the present method are compared with published results for separating and reattaching flows, and parametric studies showing the effects of step size, boundary locations, and Reynolds number are presented. The present method is substantially faster than previously published methods; typical run times range from 10 sec to 1 min of 7600 CPU time; and, based on results obtained to date, it is stable at any Reynolds number.

Murphy, J. D.↗

NASA NTRS2003

Fan Beam Emission Tomography for Estimating Scalar Properties in Laminar Flames

A new method of estimating temperatures and gas species concentrations (CO2 and H2O) in a laminar flame is reported. The path-integrated, spectral radiation intensities emitted from a laminar flame at multiple wavelengths and view angles are calculated using a narrow band radiation model. Synthetic data, in the form of radial profiles of temperature and gas concentrations, are used in these calculations. The calculations mimic measurements that would theoretically be obtained using a mid-infrared spectrometer with a scanner. The path integrated spectral radiation intensities are deconvoluted using a maximum likelihood estimation method in conjunction with an iterative scheme. The deconvolution algorithm accounts for the self-absorption of radiation by the intervening gases, and provides the local temperature and gas species concentrations. The deconvoluted temperatures and gas concentrations are compared with the synthetic data used for calculating the spectral radiation intensities. The deconvoluted temperatures and gas species concentrations are within 0.5 % of the synthetic data. The deconvolution algorithm is expected to provide combustion researchers with an easy method of obtaining the radial profiles of major gas species concentrations and temperatures in laminar flames non-intrusively using a mid-infrared spectrometer with a scanner.

Lim, Jongmook↗

NASA NTRS1991

Engineering calculations of three-dimensional inviscid hypersonic flow fields

An approximate solution technique has been developed for three-dimensional, inviscid, hypersonic flows. The method uses Maslen's explicit pressure equation and the assumption of approximate stream surfaces in the shock layer. This approximation represents a simplification of Maslen's asymmetric method. The solution procedure involves iteratively changing the shock shape in the subsonic-transonic region until the correct body shape is obtained. Beyond this region, the shock surface is determined by using a marching procedure. Results are presented herein for a paraboloid and elliptic cone at angle of attack. Calculated surface pressure distributions, shock shapes, and property profiles are compared with experimental data and finite-difference solutions of the Euler equations. Comparisons of the results of the present method with experimental data and detailed predictions are very good. Since the present method provides a very rapid computational procedure, it can be used for parametric or preliminary design applications. One useful application would be to incorporate a heating procedure for aerothermal studies.

Riley, Christopher J.↗

NASA NTRS1932

Calculation of potential flow past airship bodies in yaw

An outline of Von Karman's method of computing the potential flow of airships in yaw by means of partially constant dipolar superposition on the axis of the body is followed by several considerations for beginning and end of the superposition. Then this method is improved by postulating a continuous, in part linearly variable dipolar superposition on the axis. The second main part of the report brings the calculation of the potential flow by means of sources and sinks, arranged on the surface of the airship body. The integral equation which must satisfy this surface superposition is posed, and the core reduced to functions developed from whole elliptical normal integrals. The functions are shown diagrammatically. The integration is resolvable by iteration. The consequence of the method is good. The formulas for computing the velocity on the surface and of the potential for any point conclude the report.

Lotz, I↗

NASA NTRS1993

Sensitivity analysis of discrete systems

A selection is presented of methods used in the calculation of sensitivity derivatives of behavior and response of discretely modeled structure. Attention is given to methods for the selection of finite-difference step sizes for structural constraint derivatives, and techniques for finite difference derivatives in iteratively-solved problems. Also noted are methods for sensitivity of linear static response, and methods for second derivatives of structural response, derivatives of stresses, and nonlinear response; an emerging area of interest concerns the derivatives of transient response.

Adelman, Howard M.↗

NASA NTRS1996

Conversion of Osculating Orbital Elements to Mean Orbital Elements

Orbit determination and ephemeris generation or prediction over relatively long elapsed times can be accomplished with mean elements. The most simple and efficient method for orbit determination, which is also known as epoch point conversion, performs the conversion of osculating elements to mean elements by iterative procedures. Previous epoch point conversion methods are restricted to shorter elapsed times with linear convergence. The new method presented in this paper calculates an analytic initial guess of the unknown mean elements from a first order theory of secular perturbations and computes a transition matrix with accurate numerical partials. It thereby eliminates the problem of an inaccurate initial guess and an identity transition matrix employed by previous methods. With a good initial guess of the unknown mean elements and an accurate transition matrix, converging osculating elements to mean elements can be accomplished over long elapsed times with quadratic convergence.

Der, Gim J.↗

NASA NTRS1989

Applications of a transonic wing design method

A method for designing wings and airfoils at transonic speeds using a predictor/corrector approach was developed. The procedure iterates between an aerodynamic code, which predicts the flow about a given geometry, and the design module, which compares the calculated and target pressure distributions and modifies the geometry using an algorithm that relates differences in pressure to a change in surface curvature. The modular nature of the design method makes it relatively simple to couple it to any analysis method. The iterative approach allows the design process and aerodynamic analysis to converge in parallel, significantly reducing the time required to reach a final design. Viscous and static aeroelastic effects can also be accounted for during the design or as a post-design correction. Results from several pilot design codes indicated that the method accurately reproduced pressure distributions as well as the coordinates of a given airfoil or wing by modifying an initial contour. The codes were applied to supercritical as well as conventional airfoils, forward- and aft-swept transport wings, and moderate-to-highly swept fighter wings. The design method was found to be robust and efficient, even for cases having fairly strong shocks.

Campbell, Richard L.↗

NASA NTRS1974

Adjustment technique without explicit formation of normal equations /conjugate gradient method/

For a simultaneous adjustment of a large geodetic triangulation system, a semiiterative technique is modified and used successfully. In this semiiterative technique, known as the conjugate gradient (CG) method, original observation equations are used, and thus the explicit formation of normal equations is avoided, 'huge' computer storage space being saved in the case of triangulation systems. This method is suitable even for very poorly conditioned systems where solution is obtained only after more iterations. A detailed study of the CG method for its application to large geodetic triangulation systems was done that also considered constraint equations with observation equations. It was programmed and tested on systems as small as two unknowns and three equations up to those as large as 804 unknowns and 1397 equations. When real data (573 unknowns, 965 equations) from a 1858-km-long triangulation system were used, a solution vector accurate to four decimal places was obtained in 2.96 min after 1171 iterations (i.e., 2.0 times the number of unknowns).

Saxena, N. K.↗

NASA NTRS1990

Iterative computation of the optimal H(infinity) norm by using two-Riccati-equation method

The two-Riccati-equation method solution to a standard H(infinity) control problem can be used to characterize all possible stabilizing optimal or suboptimal H(infinity) controllers if the optimal or suboptimal H(infinity) norm is available in the literature. An iterative algorithm for computing the optimal H(infinity) norm is proposed. The algorithm employs fixed-point, double secant and bisection to guarantee a super linear convergence.

Chang, B. C.↗

NASA NTRS1975

A study of the effects of vertical resolution and measurement errors on an iteratively inverted temperature profile

A direct inversion method for inverting the temperature profile from satellite-measured radiation is discussed. The nth power of the weighting function in the integral radiative-transfer equation is used as the weight in the averaging process. The vertical resolution of the inverted temperature profile and the response of the inverted temperature profile to the measurement errors are examined in terms of n. It is found that for smaller values of n, the vertical resolution and the effect of measurement errors are reduced. When n = 0, both the vertical resolution and error effect are minimum. The temperature profile is adjusted by a constant; any structure different from the initial shape cannot be resolved. This is equivalent to the case where the entire atmosphere is treated as one layer with a fixed shape of temperature profile. When n approaches infinity, both the vertical resolution and error effect are maximum. This is equivalent to the case where the entire atmosphere is divided into m (the number of spectral channels) layers. Within each layer, the temperatures are adjusted by a constant, and any structure different from the initial shape cannot be resolved. Also, the shape of the final solution is closer to the initial profile if the value of n is smaller.

Chow, M.-D.↗

NASA NTRS1979

A dual method for maximum entropy restoration

A simple iterative dual algorithm for maximum entropy image restoration is presented. The dual algorithm involves fewer parameters than conventional minimization in the image space. Minicomputer test results for Fourier synthesis with inadequate phantom data are given.

Smith, C. B.↗

NASA NTRS1988

Predictive wall adjustment strategy for two-dimensional flexible walled adaptive wind tunnel: A detailed description of the first one-step method

Following the realization that a simple iterative strategy for bringing the flexible walls of two-dimensional test sections to streamline contours was too slow for practical use, Judd proposed, developed, and placed into service what was the first Predictive Strategy. The Predictive Strategy reduced by 75 percent or more the number of iterations of wall shapes, and therefore the tunnel run-time overhead attributable to the streamlining process, required to reach satisfactory streamlines. The procedures of the Strategy are embodied in the FORTRAN subroutine WAS (standing for Wall Adjustment Strategy) which is written in general form. The essentials of the test section hardware, followed by the underlying aerodynamic theory which forms the basis of the Strategy, are briefly described. The subroutine is then presented as the Appendix, broken down into segments with descriptions of the numerical operations underway in each, with definitions of variables.

Wolf, Stephen W. D.↗

NASA NTRS1971

An analysis and comparison of several trajectory optimization methods

The sensitivities of the convergence characteristics of the methods to initially assumed parameters and trial solution, convergence times, computer logic, and storage requirements are discussed. Numerical comparison of the convergence characteristics is made by considering a minimum time, low thrust, Earth-Mars transfer trajectory. A modified quasi-linearization method reduces convergence time by approximately 70% when compared with the generalized Newton-Raphson method and allows the terminal boundary to be specified by a general function of the problem variables. A uniquely specified and easily determined, time dependent weighting matrix for the gradient techniques accelerates the shaping of the optimal control program and improves the convergence characteristics during the terminal iterations. Convergence envelopes, indicating how sensitive the convergence characteristics are to initially assumed parameters, are plotted for the perturbation and quasi-linearization methods. Several iteration schemes are proposed which increase the size of the convergence envelopes and decrease the sensitivity of the method to initially assumed parameters.

Lewallen, J. M.↗

NASA NTRS1990

Attitude determination and parameter estimation using vector observations - Application

This paper presents tests of a new method for the simultaneous estimation of spacecraft attitude and sensor biases, based on a quaternion estimation algorithm minimizing Wahba's loss function. The new method is compared with a conventional batch least-squares differential correction algorithm. The estimates are based on data from strapdown gyros and star trackers, simulated with varying levels of Gaussian noise for both inertially-fixed and earth-pointing attitudes. Both algorithms solve for the spacecraft attitude and the gyro drift rate biases. In the majority of tests performed, the two methods converge to the same estimates in the same number of iterations, but the new algorithm requires about 60 percent more computational effort. Some cases were found in which the new method converges in fewer iterations than the differential correction, and some for which the differential correction requires fewer iterations.

Markley, F. Landis↗

NASA NTRS1988

Numerical simulation of axisymmetric turbulent flow in combustors and diffusors

A method for predicting turbulent flow in combustors and diffusers is developed. The Navier-Stokes equations, incorporating a turbulence kappa-epsilon model equation, were solved in a nonorthogonal curvilinear coordinate system. The solution applied the finite volume method to discretize the differential equations and utilized the SIMPLE algorithm iteratively to solve the differenced equations. A zonal grid method, wherein the flow field was divided into several subsections, was developed. This approach permitted different computational schemes to be used in the various zones. In addition, grid generation was made a more simple task. However, treatment of the zonal boundaries required special handling. Boundary overlap and interpolating techniques were used and an adjustment of the flow variables was required to assure conservation of mass, momentum and energy fluxes. The numerical accuracy was assessed using different finite differencing methods, i.e., hybrid, quadratic upwind and skew upwind, to represent the convection terms. Flows in different geometries of combustors and diffusers were simulated and results compared with experimental data and good agreement was obtained.

Yung, Chain Nan↗