Search NASA⌕ Search

SEARCH · Search NASA

Results for “normalizing flows”

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 577 records · Page 32
All sourcesNASA NTRSDOE OSTI
NASA NTRS2012

Error Estimates of the Ares I Computed Turbulent Ascent Longitudinal Aerodynamic Analysis

Numerical predictions of the longitudinal aerodynamic characteristics for the Ares I class of vehicles, along with the associated error estimate derived from an iterative convergence grid refinement, are presented. Computational results are based on an unstructured grid, Reynolds-averaged Navier-Stokes analysis. The validity of the approach to compute the associated error estimates, derived from a base grid to an extrapolated infinite-size grid, was first demonstrated on a sub-scaled wind tunnel model at representative ascent flow conditions for which the experimental data existed. Such analysis at the transonic flow conditions revealed a maximum deviation of about 23% between the computed longitudinal aerodynamic coefficients with the base grid and the measured data across the entire roll angles. This maximum deviation from the wind tunnel data was associated with the computed normal force coefficient at the transonic flow condition and was reduced to approximately 16% based on the infinite-size grid. However, all the computed aerodynamic coefficients with the base grid at the supersonic flow conditions showed a maximum deviation of only about 8% with that level being improved to approximately 5% for the infinite-size grid. The results and the error estimates based on the established procedure are also presented for the flight flow conditions.

Abdol-Hamid, Khaled S.↗
NASA NTRS1987

Boundary-layer and wake measurements on a swept, circulation-control wing

Wind-tunnel measurements of boundary-layer and wake velocity profiles and surface static-pressure distributions are presented for a swept, circulation-control wing. The model is an aspect-ratio-four semispan wing mounted on the tunnel side wall at a sweep angle of 45 deg. A full-span, tangential, rearward-blowing, circulation-control slot is located ahead of the trailing edge on the upper surface. Flow surveys were obtained at mid-semispan at freestream Mach numbers of 0.425 and 0.70. Boundary-layer profiles measured on the forward portions of the wing are approximately streamwise and two-dimensional. The flow in the vicinity of the jet exit and in the near wake is highly three-dimensional. The jet flow near the slot on the Coanda surface is directed normal to the slot. Near-wake surveys show large outboard flows at the center of the wake. At Mach 0.425 and a 5-deg angle of attack, a range of jet blowing rates was found for which an abrupt transition from incipient separation to attached flow occurs in the boundary layer upstream of the slot. The variation in the lower-surface separation location with blowing rate was determined from boundary-layer measurements at Mach 0.425.

Spaid, Frank W.↗
NASA NTRS1987

Boundary-layer and wake measurements on a swept, circulation-control wing

Wind-tunnel measurements of boundary-layer and wake velocity profiles and surface static pressure distributions are presented for a swept, circulation-control wing. The model is an aspect-ratio-four semispan wing mounted on the tunnel side wall at a sweep angle of 45 deg. A full-span, tangential, rearward blowing, circulation-control slot is located ahead of the trailing edge on the upper surface. Flow surveys were obtained at mid-semispan at freestream Mach numbers of 0.425 and 0.70. Boundary-layer profiles measured on the forward portions of the wing are approximately streamwise and two dimensional. The flow in the vicinity of the jet exit and in the near wake is highly three dimensional. The jet flow near the slot on the Coanda surface is directed normal to the slot. Near-wake surveys show large outboard flows at the center of the wake. At Mach 0.425 and a 5-deg angle of attack, a range of jet-blowing rates was found for which an abrupt transition from incipient separation to attached flow occurs in the boundary layer upstream of the slot. The variation in the lower-surface separation location with blowing rate was determined from boundary-layer measurements at Mach 0.425.

Spaid, Frank W.↗
NASA NTRS1990

The growth and development of a turbulent junction vortex system

The growth and development of a vortex system in an incompressible three-dimensional flow generated by a streamlined cylinder mounted with its axis normal to a flat surface was investigated using data from surface flow visualizations, surface pressure measurements, and mean flow measurements. The results from velocity measurements revealed the presence of a single large dominant vortex and a small vortex in the junction between the cylinder and the flat surface. This second vortex is tentatively identified as being connected with a corner separation occurring just downstream of the body at its thickest cross section. It is shown that, as the principal horseshoe vortex develops downstream, the strength of the vortex decreases, while its length scale increases.

Pierce, F. J.↗
NASA NTRS1946

Investigation of Flow in a Centrifugal Pump

The investigation of the flow in a centrifugal pump indicated that the flow patterns in frictional fluid are fundamentally different from those in frictionless fluid. In particular, the dead air space adhering to the section side undoubtedly causes a reduction of the theoretically possible delivery head. The velocity distribution over a parallel circle is also subjected to a noticeable change as a result of the incomplete filling of the passages. The relative velocity on the pressure side of the vane, which for passages completely filled with active flow would differ little from zero even at comparatively lower than normal delivery volume, is increased, so that no rapid reverse flow occurs on the pressure side of the vane even for smaller delivery volume. It was established, further, that the flow ceases to be stationary for very small quantities of water. The inflow to the impeller can be regarded as radial for the operating range an question. The velocity triangles at the exit are subjected to a significant alteration in shape ae a result of the increased peripheral velocity, which may be of particular importance in the determination of the guide vane entrance angle.

Fischer, Karl↗
NASA NTRS2007

Physics of the Dayside Magnetosphere: New Results From a Hybrid Kinetic Code

We use a global hybrid code kinetic model to demonstrate how kinetic processes at the bow shock and within the foreshock can dramatically modify the solar wind just before its interaction with the magnetosphere. During periods of steady radial interplanetary magnetic field (IMF) orientation, the foreshock fills with a diffuse population of suprathermal ions. The ions generate cavities marked by enhanced temperatures, depressed densities, and diminished magnetic field strengths that convect antisunward into the bow shock with the solar wind flow. Tangential discontinuities marked by inward-pointing electric fields and normals transverse to the Sun-Earth line generate hot flow anomalies marked by hot tenuous plasmas bounded by outward propagating shocks. When the motional electric field in the magnetosheath points inward towards the Earth, a solitary bow shock appears. For typical IMF orientations, the solitary shocks should appear at poorly sampled high latitudes, but for strongly northward or southward IMF orientations the solitary shocks should appear on the flanks of the magnetosphere. Although quasi-perpendicular, solitary shocks should be marked by turbulent magnetosheath flows, often directed towards the Sun-Earth line, and abrupt spike-like enhancements in the density and magnetic field strength at the shock. Finally,we show how flux transfer events generated between parallel subsolar reconnection lines are destroyed upon encountering the magnetopause at latitudes above the cusp.

Siebeck, D. G.↗
NASA NTRS1999

The value of assessing pulmonary venous flow velocity for predicting severity of mitral regurgitation: A quantitative assessment integrating left ventricular function

Although alteration in pulmonary venous flow has been reported to relate to mitral regurgitant severity, it is also known to vary with left ventricular (LV) systolic and diastolic dysfunction. There are few data relating pulmonary venous flow to quantitative indexes of mitral regurgitation (MR). The object of this study was to assess quantitatively the accuracy of pulmonary venous flow for predicting MR severity by using transesophageal echocardiographic measurement in patients with variable LV dysfunction. This study consisted of 73 patients undergoing heart surgery with mild to severe MR. Regurgitant orifice area (ROA), regurgitant stroke volume (RSV), and regurgitant fraction (RF) were obtained by quantitative transesophageal echocardiography and proximal isovelocity surface area. Both left and right upper pulmonary venous flow velocities were recorded and their patterns classified by the ratio of systolic to diastolic velocity: normal (>/=1), blunted (<1), and systolic reversal (<0). Twenty-three percent of patients had discordant patterns between the left and right veins. When the most abnormal patterns either in the left or right vein were used for analysis, the ratio of peak systolic to diastolic flow velocity was negatively correlated with ROA (r = -0.74, P <.001), RSV (r = -0.70, P <.001), and RF (r = -0.66, P <.001) calculated by the Doppler thermodilution method; values were r = -0.70, r = -0.67, and r = -0.57, respectively (all P <.001), for indexes calculated by the proximal isovelocity surface area method. The sensitivity, specificity, and predictive values of the reversed pulmonary venous flow pattern for detecting a large ROA (>0.3 cm(2)) were 69%, 98%, and 97%, respectively. The sensitivity, specificity, and predictive values of the normal pulmonary venous flow pattern for detecting a small ROA (<0.3 cm(2)) were 60%, 96%, and 94%, respectively. However, the blunted pattern had low sensitivity (22%), specificity (61%), and predictive values (30%) for detecting ROA of greater than 0.3 cm(2) with significant overlap with the reversed and normal patterns. Among patients with the blunted pattern, the correlation between the systolic to diastolic velocity ratio was worse in those with LV dysfunction (ejection fraction <50%, r = 0.23, P >.05) than in those with normal LV function (r = -0.57, P <.05). Stepwise linear regression analysis showed that the peak systolic to diastolic velocity ratio was independently correlated with RF (P <.001) and effective stroke volume (P <.01), with a multiple correlation coefficient of 0.71 (P <.001). In conclusion, reversed pulmonary venous flow in systole is a highly specific and reliable marker of moderately severe or severe MR with an ROA greater than 0.3 cm(2), whereas the normal pattern accurately predicts mild to moderate MR. Blunted pulmonary venous flow can be seen in all grades of MR with low predictive value for severity of MR, especially in the presence of LV dysfunction. The blunted pulmonary venous flow pattern must therefore be interpreted cautiously in clinical practice as a marker for severity of MR.

Non-NASA Center↗
NASA NTRS2004

Validation of an Impedance Education Method in Flow

This paper reports results of a research effort to validate a method for educing the normal incidence impedance of a locally reacting liner, located in a grazing incidence, nonprogressive acoustic wave environment with flow. The results presented in this paper test the ability of the method to reproduce the measured normal incidence impedance of a solid steel plate and two soft test liners in a uniform flow. The test liners are known to be locally react- ing and exhibit no measurable amplitude-dependent impedance nonlinearities or flow effects. Baseline impedance spectra for these liners were therefore established from measurements in a conventional normal incidence impedance tube. A key feature of the method is the expansion of the unknown impedance function as a piecewise continuous polynomial with undetermined coefficients. Stewart's adaptation of the Davidon-Fletcher-Powell optimization algorithm is used to educe the normal incidence impedance at each Mach number by optimizing an objective function. The method is shown to reproduce the measured normal incidence impedance spectrum for each of the test liners, thus validating its usefulness for determining the normal incidence impedance of test liners for a broad range of source frequencies and flow Mach numbers. Nomenclature

Watson, Willie R.↗
NASA NTRS1982

A basic code for the prediction of transient three-dimensional turbulent flowfields

A primitive pressure-velocity finite difference code has been developed to predict transient three-dimensional turbulent flow. The code is a simplified yet effective prediction procedure for use by persons with little experience in computational fluid dynamics, and into which user-oriented complexities can easily be added. The method is based on the transient two-dimensional Los Alamos SOLA prediction technique for laminar flows. Turbulence is simulated by means of a two-equation turbulence model; species diffusion and buoyancy are also included. Two applications of the code are presented to local destratification near the release structure of a reservoir and to the deflection of a jet entering normally into a uniform cross-flow. Predicted results exhibit good agreement with experimental data, showing that a useful characterization of fully three-dimensional flows is now available.

Busnaina, A. A.↗
NASA NTRS1990

An LDA investigation of the normal shock wave boundary layer interaction

Nonintrusive measurements have been made of two normal shock wave-boundary layer interactions. Two-dimensional measurements were made throughout the interaction region while three-dimensional measurements were made in the vicinity of the shock wave. The measurements were made in the corner of the test section of a continuous flow supersonic wind tunnel in which a normal shock wave had been stabilized. LDA, surface pressure measurement and flow visualization techniques were employed for two freestream Mach number test cases: 1.6 and 1.3. The former contained separated flow regions and a system of shock waves. The latter was found to be far less complicated. The reported results define the flowfield structure in detail for each case.

Chriss, R. M.↗
NASA NTRS1983

An experimental investigation of multiple shock wave/turbulent boundary layer interactions in a circular duct

Detailed pitot, static and wall pressure measurements have been obtained for multiple shock wave/turbulent boundary layer interactions in a circular duct at a free-stream Mach number of 1.49 and at a unit Reynolds number of 4.90 x 10 to the 6th per meter. The details of the flow field show the formation of a series of normal shock waves with successively decreasing strength and with decreasing distance between the successive shock waves. The overall pressure recovery is much lower than the single normal shock pressure recovery at the same free-stream Mach number. A one-dimensional flow model based on the boundary layer displacement buildup is postulated to explain the formation of a series of normal shock waves.

Om, D.↗
NASA NTRS1993

Experimental investigations of the time and flow-direction responses of shear-stress-sensitive liquid crystal coatings

Time and flow-direction responses of shearstress-sensitive liquid crystal coatings were explored experimentally. For the time-response experiments, coatings were exposed to transient, compressible flows created during the startup and off-design operation of an injector-driven supersonic wind tunnel. Flow transients were visualized with a focusing Schlieren system and recorded with a 1000 frame/sec color video camera. Liquid crystal responses to these changing-shear environments were then recorded with the same video system, documenting color-play response times equal to, or faster than, the time interval between sequential frames (i.e., 1 millisecond). For the flow-direction experiments, a planar test surface was exposed to equal-magnitude and known-direction surface shear stresses generated by both normal and tangential subsonic jet-impingement flows. Under shear, the sense of the angular displacement of the liquid crystal dispersed (reflected) spectrum was found to be a function of the instantaneous direction of the applied shear. This technique thus renders dynamic flow reversals or flow divergences visible over entire test surfaces at image recording rates up to 1 KHz. Extensions of the technique to visualize relatively small changes in surface shear stress direction appear feasible.

Reda, Daniel C.↗
NASA NTRS1993

Time dependent turbulence modeling and analytical theories of turbulence

By simplifying the direct interaction approximation (DIA) for turbulent shear flow, time dependent formulas are derived for the Reynolds stresses which can be included in two equation models. The Green's function is treated phenomenologically, however, following Smith and Yakhot, we insist on the short and long time limits required by DIA. For small strain rates, perturbative evaluation of the correlation function yields a time dependent theory which includes normal stress effects in simple shear flows. From this standpoint, the phenomenological Launder-Reece-Rodi model is obtained by replacing the Green's function by its long time limit. Eddy damping corrections to short time behavior initiate too quickly in this model; in contrast, the present theory exhibits strong suppression of eddy damping at short times. A time dependent theory for large strain rates is proposed in which large scales are governed by rapid distortion theory while small scales are governed by Kolmogorov inertial range dynamics. At short times and large strain rates, the theory closely matches rapid distortion theory, but at long times it relaxes to an eddy damping model.

Rubinstein, R.↗
NASA NTRS2019

Application of Swept-Sine Excitation for Acoustic Impedance Education

The NASA Langley Normal Incidence Tube (NIT) and Grazing Flow Impedance Tube (GFIT) are regularly employed to characterize the frequency response of acoustic liners through the eduction of their specific acoustic impedance. Both test rigs typically use an acoustic source that produces sine wave signals at discrete frequencies (Stepped-Sine) to educe the impedance. The current work details a novel approach using frequency-swept sine waveforms normalized to a constant sound pressure level for excitation. Determination of the sound pressure level and phase from microphone measurements acquired using swept-sine excitation is performed using a modified Vold-Kalman order tracking filter. Four acoustic liners are evaluated in the NIT and GFIT with both stepped-sine and swept-sine sources. Using these two methods, the educed impedance spectra are shown to compare favorably. However, the new (Swept-Sine) approach provides much greater frequency resolution in less time, allowing the acoustic liner properties to be studied in much greater detail.

Howerton, Brian M.↗
NASA NTRS1995

Three Dimensional Solution of Pneumatic Active Control of Forebody Vortex Asymmetry

Pneumatic active control of asymmetric vortical flows around a slender pointed forebody is investigated using the three dimensional solution for the compressible thin-layer Navier-Stokes equation. The computational applications cover the normal and tangential injection control of asymmetric flows around a 5 degree semi-apex angle cone at a 40 degree angle of attack, 1.4 freestream Mach number and 6 x 10(exp 6) freestream Reynolds number (based on the cone length). The effective tangential angle range of 67.5 approaches minus 67.5 degrees is used for both normal and tangential ports of injection. The effective axial length of injection is varied from 0.03 to 0.05. The computational solver uses the implicit, upwind, flux difference splitting finite volume scheme, and the grid consists of 161 x 55 x 65 points in the wrap around, normal and axial directions, respectively. The results show that tangential injection is more effective than normal injection.

Kandil, Osama A.↗
NASA NTRS1980

Strouhal number influence on flight effects on jet noise radiated from convecting quadrupoles

The paper reports a complementary extension of previous work to include the high-frequency features reflected in the discussion of the higher Strouhal number influence on flight effects. It is found that, in addition to the usual features of flight effects on noise from ordinary flows, the high Strouhal number flows exhibit some more interesting features which are uniquely characteristic to them. The additional features are as follows: (1) Flight effects are more favorable to hot jets than to cold jets; (2) the higher the Strouhal number of the jet flow, the lesser the forward arc amplification due to flight; (3) as the Strouhal number increases, the peak amplification angle in the forward quadrant and the peak suppression angle in the aft quadrant move toward 90 deg and get closer, thus reducing the amplification exposure to a constricted angular region; (4) the silence zone is disturbed and displaced from its normal position parallel to the jet flow to give rise to multiple crossings of flight curves with the static line; and (5) the occurrence of multiple crossings is a strange phenomenon solely characteristic of high Strouhal number with high subcritical jet flows in flight.

Dash, R.↗
NASA NTRS1994

Observation of anomalous slow-mode shock and reconnection layer in the dayside magnetospause

Plasma and magnetic field data from the International Sun-Earth Explorer ISEE 2 spacecraft recorded during an outbound crossing of the dayside, northern hemisphere magnetopause in October 29, 1979, provide evidence for a slow shock (SS) in the observed reconnection layer. This layer is found to be bounded on the magnetosheath side by the SS; near the magnetospheric side of the layer, a second current sheet is found that may have been rotational discontinuity (RD). The direction of the accelerated plasma flow, the earthward sense of the normal magnetic filed across the SS and RD, and the relative orientation of the SS and the RD all indicate that the reconection site was located south of the spacecraft. Quantitative tests show that, allowing for experiemental uncertanties, data taken upstream and downstream of the SS are consistent with coplanarity and other Rankine-Hugoniot (RH) conditions. Examiniations of the flow parameters indicates two anomalous properties of the SS: the upstream flow, viewed in the deHoffmann-Teller frame, is superalfvenic and the downstream plasma is firehose unstable. In comparison to the long-wavelength slow-mode phase speed, however, the flow in the upstream region is super slow, while in the downstream region it is subslow, as requeired for a slow-mode shock. Further properties of the shock include a large decrease in total enthalpy across it, indicating the escape of a sizable heat flux from the shock structure, and the occurrence of a polarization reversal of the tangential magnetic field within the shock layer, a feature that is predicted by linear double-polytropic Hall-MHD and results from a large increase in pressure anistropy from the upstream region, where p(sub parallel) approximately equal to p(sub perp), to the downstream region, where p(sub parrallel) greater than p(sub perp). Quantitative tests of the RD-like discontinuity show that it satisfies the necessary RH conditions within experimental uncertainties and indicate that the flow across it is earthward, that is, it is the same as for the SS. Thus the RD is propagating behind the SS in the direction away from the Earth. The reversal in the order of the discontinuities is attributed to the reversal of the slow and intermediate phase speeds produced by the pressure anistropy in the region between the RD and the SS. The linear dispersion relation and polarization properties of dispersive MHD waves in an anisotropic plasma are examined in the appendix.

Walthour, D. W↗
NASA NTRS1980

Drag reduction effects in turbulent boundary layers over wavy walls

Two dimensional incompressible flow over wavy surfaces are analyzed numerically by spectral methods. Algorithms for periodic flows (Fourier modes in the periodic flow direction and Chebycheff modes in the normal direction), and inflow-outflow boundary conditions (Chebycheff modes used in both directions) are described. Results obtained using both codes are reported for laminar flows. Comparisons with known theoretical and experimental results are made.

Balasubramanian, R.↗