Search NASA⌕ Search

SEARCH · Search NASA

Results for “backward”

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 271 records · Page 15
All sourcesNASA NTRSDOE OSTI
NASA NTRS1992

Planning Assembly Of Large Truss Structures In Outer Space

Report dicusses developmental algorithm used in systematic planning of sequences of operations in which large truss structures assembled in outer space. Assembly sequence represented by directed graph called "assembly graph", in which each arc represents joining of two parts or subassemblies. Algorithm generates assembly graph, working backward from state of complete assembly to initial state, in which all parts disassembled. Working backward more efficient than working forward because it avoids intermediate dead ends.

De Mello, Luiz S. Homem↗
NASA NTRS1992

Asymptotic integration algorithms for first-order ODEs with application to viscoplasticity

When constructing an algorithm for the numerical integration of a differential equation, one must first convert the known ordinary differential equation (ODE), which is defined at a point, into an ordinary difference equation (O(delta)E), which is defined over an interval. Asymptotic, generalized, midpoint, and trapezoidal, O(delta)E algorithms are derived for a nonlinear first order ODE written in the form of a linear ODE. The asymptotic forward (typically underdamped) and backward (typically overdamped) integrators bound these midpoint and trapezoidal integrators, which tend to cancel out unwanted numerical damping by averaging, in some sense, the forward and backward integrations. Viscoplasticity presents itself as a system of nonlinear, coupled first-ordered ODE's that are mathematically stiff, and therefore, difficult to numerically integrate. They are an excellent application for the asymptotic integrators. Considering a general viscoplastic structure, it is demonstrated that one can either integrate the viscoplastic stresses or their associated eigenstrains.

Freed, Alan D.↗
NASA NTRS1992

A fast, uncoupled, compressible, two-dimensional, unsteady boundary layer algorithm with separation for engine inlets

A finite difference boundary layer algorithm was developed to model viscous effects when an inviscid core flow solution is given. This algorithm solved each boundary layer equation separately, then iterated to find a solution. Solving the boundary layer equations sequentially was 2.4 to 4.0 times faster than solving the boundary layer equations simultaneously. This algorithm used a modified Baldwin-Lomax turbulence model, a weighted average of forward and backward differencing of the pressure gradient, and a backward sweep of the pressure. With these modifications, the boundary layer algorithm was able to model flows with and without separation. The number of grid points used in the boundary layer algorithm affected the stability of the algorithm as well as the accuracy of the predictions of friction coefficients and momentum thicknesses. Results of this boundary layer algorithm compared well with experimental observations of friction coefficients and momentum thicknesses. In addition, when used interactively with an inviscid flow algorithm, this boundary layer algorithm corrected for viscous effects to give a good match with experimental observations for pressures in a supersonic inlet.

Roach, Robert L.↗
NASA NTRS1992

Digital control algorithms for microgravity isolation systems

New digital control algorithms were developed to achieve the desired acceleration transmissibility function. The attractive electromagnets have been taken as actuators. The relative displacement and the acceleration of the mass were used as feedback signals. Two approaches were developed to find that controller transfer function in Z-domain, which yields the desired transmissibility at each frequency. In the first approach, the controller transfer function is obtained by assuming that the desired transmissibility is known in Z-domain. Since the desired transmissibility H sub d(S) = 1/(tauS+1)(exp 2) is given in S-domain, the first task is to obtain the desired transmissibility in Z-domain. There are three methods to perform this task: bilinear transformation, and backward and forward rectangular rules. The bilinear transformation and backward rectangular rule lead to improper controller transfer functions, which are physically not realizable. The forward rectangular rule does lead to a physically realizable controller. However, this controller is found to be marginally stable because of a pole at Z=1. In order to eliminate this pole, a hybrid control structure is proposed. Here the control input is composed of two parts: analog and digital. The analog input simply represents the velocity (or the integral of acceleration) feedback; and the digital controller which uses only relative displacement signal, is then obtained to achieve the desired closed-loop transfer function. The stability analysis indicates that the controller transfer function is stable for typical values of sampling period. In the second approach, the aforementioned hybrid control structure is again used. First, an analog controller transfer function corresponding to relative displacement feedback is obtained to achieve the transmissibility as 1/(tauS+1)(exp 2). Then the transfer function for the digital control input is obtained by discretizing this analog controller transfer function via bilinear transformation. The stability of the resulting Z-domain closed loop system is analyzed. Also, the frequency response of the Z-domain closed-loop transfer function is determined to evaluate the performance of the control system.

Sinha, Alok↗
NASA NTRS1991

Sufficient conditions for a local minimum of the Bolza problem with multiple terminal point constraints

Sufficient conditions for a weak relative minimum are derived for a form of the Bolza problem of variational calculus. The derivation ties together first-order and second-order conditions in a unified consistent manner, addresses controllability considerations in some detail, and clarifies some small inconsistencies in earlier work. The resulting second-order conditions of optimality involve the integration of fewer backward-sweep matrix elements than the standard conditions in the literature for problems with an unspecified final time. As a result, the backward-sweep matrices have the same general structure and dimensionality whether the final time is specified or free, with the terminal values of two of three sweep matrices being more complicated in the latter case.

Wood, Lincoln J.↗
NASA NTRS1992

A fast, uncoupled, compressible, two-dimensional, unsteady boundary layer algorithm with separation for engine inlets

A finite difference boundary layer algorithm was developed to model viscous effects when an inviscid core flow solution is given. This algorithm solved each boundary layer equation separately, then iterated to find a solution. Solving the boundary layer equations sequentially was 2.4 to 4.0 times faster than solving the boundary layer equations simultaneously. This algorithm used a modified Baldwin-Lomax turbulence model, a weighted average of forward and backward differencing of the pressure gradient, and a backward sweep of the pressure. With these modifications, the boundary layer algorithm was able to model flows with and without separation. The number of grid points used in the boundary layer algorithm affected the stability of the algorithm affected the stability of the algorithm as well as the accuracy of the predictions of friction coefficients and momentum thicknesses. Results of this boundary layer algorithm compared well with experimental observations of friction coefficients and momentum thicknesses. In addition, when used interactively with an inviscid flow algorithm, this boundary layer algorithm corrected for viscous effects to give a good match with experimental observations for pressures in a supersonic inlet.

Roach, Robert L.↗
NASA NTRS1992

Diabatic dynamic initialization

A generalized dynamical adjustment procedure was applied to a diabatic model to produce balanced initial conditions. Namely, backward adiabatic model integration is followed by forward diabatic model integration, with a high frequency (low pass) filter in the form of the Euler-backward time differencing scheme being applied throughout the whole integration. As a result of the application of such a diabatic dynamic initialization procedure within the Goddard Laboratory for Atmospheres (GLA) 4-D data assimilation system, the following properties of forecasts from initialized fields are achieved right from the beginning of the usual forecast integration: (1) the forecast tendencies (and fields) are free of any noise due to imbalance in initial conditions; (2) the shocks related to an initial imbalance between model physics and dynamics, and especially the substantial initial imbalance of precipitation and evaporation fields, or the initial spin-up effect, are practically removed. Diabatic dynamic initialization has been compared with implicit nonlinear normal mode initialization, and found to be superior in removing the initial spin-up effect, and also in improving the tropical structure. The diabatic dynamic initialization procedure was successfully tested for the GLA system with the use of all conventional data and the GLA satellite data retrievals. It allows a smooth data insertion without any shocks or imbalances, which is highly desirable for efficient functioning of 4-D data assimilation systems. The developed initialization procedure is computationally efficient and in principle easily applicable to different forecast models.

Fox-Rabinovitz, Michael S.↗
NASA NTRS1993

Diabatic dynamic initialization

A version of the dynamical adjustment procedure is generalized in order to make the procedure applicable to a diabatic model. In this new initialization procedure, backward adiabatic model integration is followed by forward diabatic model integration, with a high-frequency (low-pass) time filter in the form of the Euler backward time-differencing scheme applied throughout the integration. The new procedure was applied to the Goddard Laboratory for Atmospheres 4D data-assimilation system. It was found that, right from the beginning of the forecast integration, the forecast tendencies (and fields) were free of any noise due to imbalance in initial conditions, and the shocks related to an initial imbalance between model physics and dynamics, or the initial spinup effect, were practically removed.

Fox-Rabinovitz, Michael S.↗
NASA NTRS1993

Large eddy simulation of a boundary layer with concave streamwise curvature

One of the most exciting recent developments in the field of large eddy simulation (LES) is the dynamic subgrid-scale model. The dynamic model concept is a general procedure for evaluating model constants by sampling a band of the smallest scales actually resolved in the simulation. To date, the procedure has been used primarily in conjunction with the Smagorinsky model. The dynamic procedure has the advantage that the value of the model constant need not be specified a priori, but rather is calculated as a function of space and time as the simulation progresses. This feature makes the dynamic model especially attractive for flows in complex geometries where it is difficult or impossible to calibrate model constants. The dynamic model was highly successful in benchmark tests involving homogeneous and channel flows. Having demonstrated the potential of the dynamic model in these simple flows, the overall direction of the LES effort at CTR shifted toward an evaluation of the model in more complex situations. The current test cases are basic engineering-type flows for which Reynolds averaged approaches were unable to model the turbulence to within engineering accuracy. Flows currently under investigation include a backward-facing step, wake behind a circular cylinder, airfoil at high angles of attack, separated flow in a diffuser, and boundary layer over a concave surface. Preliminary results from the backward-facing step and cylinder wake simulations are encouraging. Progress on the LES of a boundary layer on a concave surface is discussed. Although the geometry of a concave wall is not very complex, the boundary layer that develops on its surface is difficult to model due to the presence of streamwise Taylor-Gortler vortices. These vortices arise as a result of a centrifugal instability associated with the convex curvature.

Lund, Thomas S.↗
NASA NTRS1993

New techniques for reversible compression of multispectral data

While spatial correlations are adequately exploited by standard lossless image compression techniques, little success has been attained in exploiting spectral correlations when dealing with multispectral image data. In this paper, we present some new lossless image compression techniques that capture spectral correlations as well as spatial correlation in a simple and elegant manner. The schemes are based on the notion of a prediction tree, which defines a non-causal prediction model for an image. We present a backward adaptive technique and a forward adaptive technique. We then give a computationally efficient way of approximating the backward adaptive technique. The approximation gives good results and is extremely easy to compute. Simulation results show that for high spectral resolution images, significant savings can be made by using spectral correlations in addition to spatial correlations. Furthermore, the increase in complexity incurred in order to make these gains is minimal.

Memon, N. D.↗
NASA NTRS1995

Simulation study of the role of ion kinetics in low-frequency wave train evolution

The evolution of uniform, parallel propagating, low-frequency (less than or approx. ion cyclotron) wave trains is followed with a one-dimensional hybrid numerical code with fluid electrons and particle ions. We show that moderate amplitude (delta B/B less than 1/2) wave trains give instabilities and saturated states which differ completely from pure fluid evolution. This is most clearly seen when beta greater than 1 and instability exists for wavenumbers both below and above the wavenumber of an inital, left-handed wave train or pump wave. For corresponding parameters a fluid theory gives only a narrow range of instability above the pump wavenumber where decay and beat instabilities can occur. in simulations wave energy inverse cascades to smaller wavenumbers and into a greater number of forward than backward going waves. In fluids energy by decay goes mostly to backward ones of smaller wavenumber, and energy by beat goes mostly to forward ones of larger wavenumber. Neither fluid instability explains simulation results. The instability is saturated by thermalizing ions and sometimes exciting small wavenumber electrostatic or acoustic modes. In contrast, saturation in fluids first occurs by generating the harmonics of the growing linear modes. Harmonic generation is mostly absent in simulations. Simulations are carried out to long times and mostly reach a limit beyond which no further significant evolution can occur. Application to Alfvenic fluctuations in the solar wind is discussed.

Vasquez, Bernard J.↗
NASA NTRS1994

Beat, modulational, and decay instabilities of a circularly polarized Alfven wave

A circulary polarized low-frequency electomagnetic pump wave propoagating along an ambient magnetic field is known to be unstable to the growth of several parallel-propagating parametric instabilties. If ion-cyclotron effects are retained in a two-fluid description, the dispersion relation is a sixth-order polynomial. We present a series of new analytical approximations to this dispersion relation. We emphasize new results for the beat instability that occurs as an interaction of the forward prpagating upper sideband with the backward propagating lower sideband. The nature of the beat instabitlity depends on beta = (v(sub sound)/v(sub A)(exp 2) and on the sense of polarization of the pump wave. The beat and decay instabilities can occur together if the pump is left-handed (i.e., ion resonant) and if beta is less than or approximately 1, but they cannot occur together if the pump is right-handed. For a left-handed pump the beat mode is the only instability if beta is greater than or approximately 1. If the pump is right-handed and beta is greater than or approximately 1, then the beat instability exists only when the pump amplitude exceeds a threshold value, and the beat will be the only instability if the pump amplitude is large enough to stabilize the modulational instability. If the pump is left-handed and beta is less than or approximately 1, then the beat mode is stabilized when the pump amplitude becomes sufficiently large. The beat instability primarily produces a forward propagating transverse wave in the upper sideband. Thus if beta is greater than or approximately 1, the instabilities considered here do not produce the backward propagating waves which are thought to affect turbulence and the evolution of cross helicity in the solar wind. New analytical results are presented also for the decay and modulational instabilites when beta is approximately equal to 1.

Hollweg, Joseph V.↗
NASA NTRS1993

Parametric instabilities of parallel-propagating Alfven waves: Some analytical results

We consider the stability of a circularly polarized Alfven wave (the pump wave) which propagates parallel to the ambient magnetic field. Only parallel-propagating perturbations are considered, and we ignore dispersive effects due to the ion cyclotron frequency. The dissipationless MHD equations are used throughout; thus possibibly important effects arising from Landau and transit time damping are omitted. We derive a series of analytical approximations to the dispersion relation using A = (Delta B/B(sub O))(exp 2) as a small expansion parameter; Delta B is the pump amplitude, and B(sub O) is the ambient magnetic field strength. We find that the plasma beta (the square of the ratio of the sound speed to the Alfven speed) plays a crucial role in determining the behavior of the parametric instabilities of the pump. If 0 less than beta less than 1 we find the familiar result that the pump decays into a forward propagating sound wave and a backward propagating Alfven wave with maximum growth rate gamma(sub max) varies A(sup 1/2), but beta cannot be too close to 0 or to 1. If beta approx. 1, we find gamma(sub max) varies A(sup 3/4), if beta greater than 1, we find gamma(sub max) varies A(sup 3/2), while if beta approx. 0, we obtain gamma(sub max) varies A(sup 1/3); moreover, if beta approx. 0 there is a nearly purely growing instability. In constrast to the familiar decay instability, for which the backward propagating Alfven wave has lower frequency and wavenumber than the pump, we find that if beta greater than or approx. equal to 1 the instability is really a beat instability which is dominated by a transverse wave which is forward propagating and has frequency and wavenumber which are nearly twice the pump values. Only the decay instability for 0 less than beta less than 1 can be regarded as producing two recognizable normal modes, namely, a sound wave and an Alfven wave. We discuss how the different characteristics of the instabilities may affect the evolution of Alfven waves in the solar wind. However, for a solar wind in which beta approx. 1 the growth times of the instabilities are probably too long for these instabilities to have an appreciable effect inside 1 AU.

Jayanti, V.↗
NASA NTRS1995

Inertial acceleration as a measure of linear vection: An alternative to magnitude estimation

The present study focused on the development of a procedure to assess perceived self-motion induced by visual surround motion - vection. Using an apparatus that permitted independent control of visual and inertial stimuli, prone observers were translated along their head x-axis (fore/aft). The observers' task was to report the direction of self-motion during passive forward and backward translations of their bodies coupled with exposure to various visual surround conditions. The proportion of 'forward' responses was used to calculate each observer's point of subjective equality (PSE) for each surround condition. The results showed that the moving visual stimulus produced a significant shift in the PSE when data from the moving surround condition were compared with the stationary surround and no-vision condition. Further, the results indicated that vection increased monotonically with surround velocities between 4 and 40/s. It was concluded that linear vection can be measured in terms of changes in the amplitude of whole-body inertial acceleration required to elicit equivalent numbers of 'forward' and 'backward' self-motion reports.

Carpenter-Smith, Theodore R.↗
NASA NTRS1996

Assessment of the Subgrid-Scale Models at Low and High Reynolds Numbers

Accurate SGS models must be capable of correctly representing the energy transfer between GS and SGS. Recent direct assessment of the energy transfer carried out using direct numerical simulation (DNS) data for wall-bounded flows revealed that the energy exchange is not unidirectional. Although GS kinetic energy is transferred to the SGS (forward scatter (F-scatter) on average, SGS energy is also transferred to the GS. The latter energy exchange (backward scatter (B-scatter) is very significant, i.e., the local energy exchange can be backward nearly as often as forward and the local rate of B-scatter is considerably higher than the net rate of energy dissipation.

Horiuti, K.↗
NASA NTRS1997

A Near-Wall Reynolds-Stress Closure without Wall Normals

With the aid of near-wall asymptotic analysis and results of direct numerical simulation, a new near-wall Reynolds stress model (NNWRS) is formulated based on the SSG high-Reynolds-stress model with wall-independent near-wall corrections. Only one damping function is used for flows with a wide range of Reynolds numbers to ensure that the near-wall modifications diminish away from the walls. The model is able to reproduce complicated flow phenomena induced by complex geometry, such as flow recirculation, reattachment and boundary-layer redevelopment in backward-facing step flow and secondary flow in three-dimensional square duct flow. In simple flows, including fully developed channel/pipe flow, Couette flow and boundary-layer flow, the wall effects are dominant, and the NNWRS model predicts less degree of turbulent anisotropy in the near-wall region compared with a wall-dependent near-wall Reynolds Stress model (NWRS) developed by So and colleagues. The comparison of the predictions given by the two models rectifies the misconception that the overshooting of skin friction coefficient in backward-facing step flow prevalent in those near-wall, models with wall normal is caused by he use of wall normal.

Yuan, S. P.↗
NASA NTRS1999

Atmospheric Chemical Transport Based on High Resolution Model- Derived Winds: A Case Study

Flight 10 of NASA's Subsonic Assessment (SASS) Ozone and Nitrogen Oxide Experiment (SONEX) extended southwest of Lajes, Azores. A variety of chemical signatures were encountered. These signatures are examined in detail, relating them to meteorological data from a high resolution numerical model having horizontal grid spacing of 30 and 90 km and 26 vertical levels. The meteorological output at hourly intervals is used to create backward trajectories from the locations of the chemical signatures. Four major categories of chemical signatures are discussed-stratospheric, lightning, continental pollution, and a transition layer. The strong stratospheric signal is encountered just south of the Azores in a region of depressed tropopause height. Three chemical signatures at different altitudes in the upper troposphere are attributed to lightning. Backward trajectories arriving at locations of these signatures are related to locations of cloud-to-ground lightning. Results show that the trajectories pass through regions of lightning 1-2 days earlier over the eastern Gulf of Mexico and off the southeast coast of the United States. The lowest leg of the flight exhibits a chemical signature consistent with continental pollution. Trajectories arriving at this signature are found to pass over the highly populated Northeast Corridor of the United States. Surface based pollution apparently is lofted to the altitudes of the trajectories by convective clouds along the East Coast that did not contain lightning. Finally, a chemical transition layer is described. Its chemical signature is intermediate to those of lightning and continental pollution. Trajectories arriving in this layer pass between the trajectories of the lightning and pollution signatures. Thus, they probably are impacted by both sources.

Hannan, John R.↗
NASA NTRS1999

The Yearly Variation in Fall-Winter Arctic Winter Vortex Descent

Using the change in HALOE methane profiles from early September to late March, we have estimated the minimum amount of diabatic descent within the polar which takes place during Arctic winter. The year to year variations are a result in the year to year variations in stratospheric wave activity which (1) modify the temperature of the vortex and thus the cooling rate; (2) reduce the apparent descent by mixing high amounts of methane into the vortex. The peak descent amounts from HALOE methane vary from l0km -14km near the arrival altitude of 25 km. Using a diabatic trajectory calculation, we compare forward and backward trajectories over the course of the winter using UKMO assimilated stratospheric data. The forward calculation agrees fairly well with the observed descent. The backward calculation appears to be unable to produce the observed amount of descent, but this is only an apparent effect due to the density decrease in parcels with altitude. Finally we show the results for unmixed descent experiments - where the parcels are fixed in latitude and longitude and allowed to descend based on the local cooling rate. Unmixed descent is found to always exceed mixed descent, because when normal parcel motion is included, the path average cooling is always less than the cooling at a fixed polar point.

Schoeberl, Mark R.↗