Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Conditional Normalizing Flow”

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 307 records · Page 17
All sourcesNASA NTRSDOE OSTI
NASA NTRS1976

Supersonic Flow of Chemically Reacting Gas-Particle Mixtures. Volume 2: RAMP - A Computer Code for Analysis of Chemically Reacting Gas-Particle Flows

A computer program written in conjunction with the numerical solution of the flow of chemically reacting gas-particle mixtures was documented. The solution to the set of governing equations was obtained by utilizing the method of characteristics. The equations cast in characteristic form were shown to be formally the same for ideal, frozen, chemical equilibrium and chemical non-equilibrium reacting gas mixtures. The characteristic directions for the gas-particle system are found to be the conventional gas Mach lines, the gas streamlines and the particle streamlines. The basic mesh construction for the flow solution is along streamlines and normals to the streamlines for axisymmetric or two-dimensional flow. The analysis gives detailed information of the supersonic flow and provides for a continuous solution of the nozzle and exhaust plume flow fields. Boundary conditions for the flow solution are either the nozzle wall or the exhaust plume boundary.

Penny, M. M.↗
NASA NTRS1976

Improved computational treatment of transonic flow about swept wings

Relaxation solutions to classical three-dimensional small-disturbance (CSD) theory for transonic flow about lifting swept wings are reported. For such wings, the CSD theory was found to be a poor approximation to the full potential equation in regions of the flow field that are essentially two-dimensional in a plane normal to the sweep direction. The effect of this deficiency on the capture of embedded shock waves in terms of (1) the conditions under which shock waves can exist and (2) the relations they must satisfy when they do exist is emphasized. A modified small-disturbance (MSD) equation, derived by retaining two previously neglected terms, was proposed and shown to be a consistent approximation to the full potential equation over a wider range of sweep angles. The effect of these extra terms is demonstrated by comparing CSD, MSD, and experimental wing surface pressures.

Ballhaus, W. F.↗
NASA NTRS1993

On the assumption of vanishing temperature fluctuations at the wall for heat transfer modeling

Boundary conditions for fluctuating wall temperature are required for near-wall heat transfer modeling. However, their correct specifications for arbitrary thermal boundary conditions are not clear. The conventional approach is to assume zero fluctuating wall temperature or zero gradient for the temperature variance at the wall. These are idealized specifications and the latter condition could lead to an ill posed problem for fully-developed pipe and channel flows. In this paper, the validity and extent of the zero fluctuating wall temperature condition for heat transfer calculations is examined. The approach taken is to assume a Taylor expansion in the wall normal coordinate for the fluctuating temperature that is general enough to account for both zero and non-zero value at the wall. Turbulent conductivity is calculated from the temperature variance and its dissipation rate. Heat transfer calculations assuming both zero and non-zero fluctuating wall temperature reveal that the zero fluctuating wall temperature assumption is in general valid. The effects of non-zero fluctuating wall temperature are limited only to a very small region near the wall.

Sommer, T. P.↗
NASA NTRS1995

A test of a vortex method for the computation of flap side edge noise

Upon approach to landing, a major source location of airframe noise occurs at the side edges of the part span, trailing edge flaps. In the vicinity of these flaps, a complex arrangement of spanwise flow with primary and secondary tip vortices may form. Each of these vortices is observed to become fully three-dimensional. In the present study, a numerical model is developed to investigate the noise radiated from the side edge of a flap. The inherent three-dimensionality of this flow forces us to carefully consider a numerical scheme which will be both accurate in its prediction of the flow acoustics and also computationally efficient. Vortex methods have offered a fast and efficient means of simulating many two and three-dimensional, vortex dominated flows. In vortex methods, the time development of the flow is tracked by following exclusively the vorticity containing regions. Through the Biot-Savart law, knowledge of the vorticity field enables one to obtain flow quantities at any desired location during the flow evolution. In the present study, a numerical procedure has been developed which incorporates the Lagrangian approach of vortex methods into a calculation for the noise radiated by a flow-surface interaction. In particular, the noise generated by a vortex in the presence of a flat half plane is considered. This problem serves as a basic model of flap edge flow. It also permits the direct comparison between our computed results and previous acoustic analyses performed for this problem. In our numerical simulations, the mean flow is represented by the complex potential W(z) = Aiz(exp l/2), which is obtained through conformal mapping techniques. The magnitude of the mean flow is controlled by the parameter A. This mean flow has been used in the acoustic analysis by Hardin and is considered a reasonable model of the flow field in the vicinity of the edge and away from the leading and trailing edges of the flap. To represent the primary vortex which occurs near the flap, a point vortex is introduced just below the flat half plane. Using a technique from panel methods, boundary conditions on the flap surface are satisfied by the introduction of a row of stationary point vortices along the extent of the flap. At each time step in the calculation, the strength of these vortices is chosen to eliminate the normal velocity at intermediary collocation points. The time development of the overall flow field is then tracked using standard techniques from vortex methods. Vortex trajectories obtained through this computation are in good agreement with those predicted by the analytical solution given by Hardin, thus verifying the viability of this procedure for more complex flow arrangements. For the flow acoustics, the Ffowcs Williams-Hawkings equation is numerically integrated. This equation supplies the far field acoustic pressure based upon pressures occurring along the flap surface. With our vortex method solution, surface pressures may be obtained with exceptional resolution. The Ffowcs Williams-Hawkings equation is integrated using a spatially fourth order accurate Simpson's rule. Rational function interpolation is used to obtain the surface pressures at the appropriate retarded times. Comparisons between our numerical results for the acoustic pressure and those predicted by the Hardin analysis have been made. Preliminary results indicate the need for an improved integration technique. In the future, the numerical procedure developed in this study will be applied to the case of a rectangular flap of finite thickness and ultimately modified for application to the fully three-dimensional problem.

Martin, James E.↗
NASA NTRS1992

Preliminary characterization of a water vaporizer for resistojet applications

A series of tests was conducted to explore the characteristics of a water vaporizer intended for application to resistojet propulsion systems. The objectives of these tests were to (1) observe the effect of orientation with respect to gravity on vaporizer stability, (2) characterize vaporizer efficiency and outlet conditions over a range of flow rates, and (3) measure the thrust performance of a vaporizer/resistojet thruster assembly. A laboratory model of a forced-flow, once-through water vaporizer employing a porous heat exchange medium was built and characterized over a range of flow rates and power levels of interest for application to water resistojets. In a test during which the vaporizer was rotated about a horizontal axis normal to its own axis, the outlet temperature and mass flow rate through the vaporizer remained steady. Throttlability to 30 percent of the maximum flow rate tested was demonstrated. The measured thermal efficiency of the vaporizer was near 0.9 for all tests. The water vaporizer was integrated with an engineering model multipropellant resistojet. Performance of the vaporizer/thruster assembly was measured over a narrow range of operating conditions. The maximum specific impulse measured was 234 s at a mass flow rate and specific power level (vaporizer and thruster combined) of 154 x 10(exp-6)kg/s and 6.8 MJ/kg, respectively.

Morren, W. Earl↗
NASA NTRSDate not supplied

Wall Modeled Large Eddy Simulations for NASA’s Jet Noise Consensus Database of Single-Flow, Round, Convergent Jets

A campaign of wall-modeled large-eddy simulations (WMLES) using structured curvilinear overlapping grids has been performed with the Launch Ascent and Vehicle Aerodynamics(LAVA) computational fluid dynamics (CFD) software to predict jet noise for single-stream axisymmetric round jets. The simulations address the new Prediction Uncertainty Reduction(PUR) technical challenge within the context of NASA’s Commercial Supersonic Technology(CST) project. The goal of PUR is to quantify and reduce uncertainties from scale-resolving simulations to assess noise characteristics of next generation quiet supersonic commercial jets during takeoff and landing conditions where the noise from the exhaust jet dominates. The focus of this effort is to generate a simulation database for single-stream axisymmetric round nozzles at several conditions both for static (no ambient co-flow), which is the focus of this article, and in-flight (flight stream co-flow) conditions, which are beyond the current scope. Nine different flow conditions ranging in jet exit Mach number from 0.38 to 1.0 with nozzle temperature ratios (NTR) from 0.84 to 2.7 have been conducted. Details of the structured overset grids, numerical discretization and wall-model are provided. Near-field comparisons to PIV data show great agreement for both velocity and normal stresses, however a systematic TKE overshoot at the nozzle exit is seen in the lip line shear-layer. A permeable Ffowcs Williams Hawkings (FWH) surface, enclosing the jet, is used to predict far-field noise from the simulated flow-field. Comparison of CFD predictions to microphone array measurements demonstrate excellent agreement within the resolved frequency range. A systematic under-prediction of far-aft observer angles larger than 150 degrees has been observed across all simulations. We achieved a cost reduction of an order of magnitude for these WMLES compared to an earlier study of this configuration due to algorithmic and software improvements. The accuracy of the results and short turnaround time demonstrate that WMLES within the LAVA framework is a cost-effective approach for jet noise predictions that could soon be incorporated into the design cycle of jet noise reduction technologies.

CST↗
NASA NTRS1975

Steady subsonic flow around finite-thickness wings

The general method for analyzing steady subsonic potential aerodynamic flow around a lifting body having arbitrary shape is presented. By using the Green function method, an integral representation for the potential is obtained. Under small perturbation assumption, the potential at any point, P, in the field depends only upon the values of the potential and its normal derivative on the surface of the body. Hence if the point P approaches the surface of the body, the representation reduces to an integral equation relating the potential and its normal derivative (which is known from the boundary conditions) on the surface. The question of uniqueness is examined and it is shown that, for thin wings, the operator becomes singular as the thickness approaches zero. This fact may yield numerical problems for very thin wings. However, numerical results obtained for a rectangular wing in subsonic flow show that these problems do not appear even for thickness ratio tau = .001. Comparison with existing results shows that the proposed method is at least as fast and accurate as the lifting surface theories.

Kuo, C. C.↗
NASA NTRS2017

Navier-Stokes Simulation of UH-60A Rotor/Wake Interaction Using Adaptive Mesh Refinement

Time-dependent Navier-Stokes simulations have been carried out for a flexible UH-60A rotor in forward flight, where the rotor wake interacts with the rotor blades. These flow conditions involved blade vortex interaction and dynamic stall, two common conditions that occur as modern helicopter designs strive to achieve greater flight speeds and payload capacity. These numerical simulations utilized high-order spatial accuracy and delayed detached eddy simulation. Emphasis was placed on understanding how improved rotor wake resolution affects the prediction of the normal force, pitching moment, and chord force of the rotor. Adaptive mesh refinement was used to highly resolve the turbulent rotor wake in a computationally efficient manner. Moreover, blade vortex interaction was found to trigger dynamic stall. Time-dependent flow visualization was utilized to provide an improved understanding of the numerical and physical mechanisms involved with three-dimensional dynamic stall.

UH-60A RotorWake↗
NASA NTRS2014

DNS of Laminar-Turbulent Transition in Swept-Wing Boundary Layers

Direct numerical simulation (DNS) is performed to examine laminar to turbulent transition due to high-frequency secondary instability of stationary crossflow vortices in a subsonic swept-wing boundary layer for a realistic natural-laminar-flow airfoil configuration. The secondary instability is introduced via inflow forcing and the mode selected for forcing corresponds to the most amplified secondary instability mode that, in this case, derives a majority of its growth from energy production mechanisms associated with the wall-normal shear of the stationary basic state. An inlet boundary condition is carefully designed to allow for accurate injection of instability wave modes and minimize acoustic reflections at numerical boundaries. Nonlinear parabolized stability equation (PSE) predictions compare well with the DNS in terms of modal amplitudes and modal shape during the strongly nonlinear phase of the secondary instability mode. During the transition process, the skin friction coefficient rises rather rapidly and the wall-shear distribution shows a sawtooth pattern that is analogous to the previously documented surface flow visualizations of transition due to stationary crossflow instability. Fully turbulent features are observed in the downstream region of the flow.

Duan, L.↗
NASA NTRS1977

Diffraction of a shock wave by a compression corner. II - Single Mach reflection

The two-dimensional time-dependent Euler equations which govern the flow field resulting from the interaction of a planar shock with a compression corner are solved for initial conditions which result in single Mach reflection of the incident planar shock. The Euler equations are first transformed to include the self-similarity of the flow field. A second transformation is employed to normalize the distances between the ramp and the reflected shock and between the wall and the Mach stem. The resulting equations in strong conservation-law form are solved using a second-order discontinuity-fitting finite-difference approach. The results are compared with experimental interferograms and existing first-order shock-capturing numerical solutions.

Shankar, V.↗
NASA NTRS1983

Heat-transfer distributions on biconics at incidence in hypersonic-hypervelocity He, N2, air, and CO2 flows

Laminar heat transfer rates were measured on spherically blunted, 13 deg/7 deg on axis and bent biconics (fore cone bent 7 deg upward relative to aft cone) at hypersonic hypervelocity flow conditions in the Langley Expansion Tube. Freestream velocities from 4.5 to 6.9 km/sec and Mach numbers from 6 to 9 were generated using helium, nitrogen, air, and carbon dioxide test gases, resulting in normal shock density ratios from 4 to 19. Angle of attack, referenced to the axis of the aft cone, was varied from 0 to 20 deg in 4 deg increments. The effect of nose bend, angle of attack, and real gas phenomena on heating distributions are presented along with comparisons of measurement to prediction from a code which solves the three dimensional parabolized Navier-Stokes equations.

Miller, C. G.↗
NASA NTRS1983

Heat-transfer distributions on biconics at incidence in hypersonic-hypervelocity He, N2, air, and CO2 flows

Laminar heat-transfer rates were measured on spherically blunted, 13 degrees/F degrees on-axis and bent biconics (fore cone bent 7 degrees upward relative to aft cone) at hypersonic-hypervelocity flow conditions in the Langley Expansion Tube. Freestream velocities from 4.5 to 6.9 km/sec and Mach numbers from 6 to 9 were generated using helium, nitrogen, air, and carbon dioxide test gases, resulting in normal shock density ratios from 4 to 19. Angle of attack, referenced to the axis of the aft cone, was varied from zero to 20 degrees in 4 degree increments. The effect of nose bend, angle of attack, and real-gas phenomena on heating distributions are presented along with comparisons of measurement to prediction from a code which solves the three-dimensional 'parabolized Navier-Stokes' equations.

Miller, C. G.↗
NASA NTRS1991

AIAA/IKI Microgravity Science Symposium, 1st, Moscow, USSR, May 13-17, 1991, Proceedings

The state of the art of microgravity science is examined in reviews and reports. Consideration is given to fluid physics; modeling and numerical analysis; crystal growth of electronic materials, proteins, metals, alloys, glasses, and ceramics; biotechnology, combustion science, microgravity hardware, and fundamental phenomena; and education. Particular attention is given to containerless science and technology, a low-gravity two-phase flow with heat transfer, numerical investigation on Benard instability in a finite liquid layer, synthesis and crystallization of refractory compounds from solutions in metallic melts under microgravitation conditions, the long duration protein crystallization experiment aboard the Mir Space Station, directional solidification in immiscible systems, formation of periodic layer structure in oxide melts under different gravity values, opposed flow flame spread in normal, enhanced, and reduced gravity, automatic equipment for semiconductor production in space, and consort and joust sounding rocket missions.

Source record↗
NASA NTRS1993

Flame spread across liquid pools

For flame spread over liquid fuel pools, the existing literature suggests three gravitational influences: (1) liquid phase buoyant convection, delaying ignition and assisting flame spread; (2) hydrostatic pressure variation, due to variation in the liquid pool height caused by thermocapillary-induced convection; and (3) gas-phase buoyant convection in the opposite direction to the liquid phase motion. No current model accounts for all three influences. In fact, prior to this work, there was no ability to determine whether ignition delay times and flame spread rates would be greater or lesser in low gravity. Flame spread over liquid fuel pools is most commonly characterized by the relationship of the initial pool temperature to the fuel's idealized flash point temperature, with four or five separate characteristic regimes having been identified. In the uniform spread regime, control has been attributed to: (1) gas-phase conduction and radiation; (2) gas-phase conduction only; (3) gas-phase convection and liquid conduction, and most recently (4) liquid convection ahead of the flame. Suggestions were made that the liquid convection was owed to both vuoyancy and thermocapillarity. Of special interest to this work is the determination of whether, and under what conditions, pulsating spread can and will occur in microgravity in the absence of buoyant flows in both phases. The approach we have taken to resolving the importance of buoyancy for these flames is: (1) normal gravity experiments and advanced diagnostics; (2) microgravity experiments; and (3) numerical modelling at arbitrary gravitational level.

Ross, Howard↗
NASA NTRS1994

Matching Inviscid/Boundary Layer Flowfields

New boundary-layer equations are developed and the solutions match all the boundary-layer properties, except for the normal velocity, exactly with the corresponding inviscid properties. The numerical procedure solves tridiagonal matrices at each marching station. As part of the solution an inviscid transpiration velocity at the surface is calculated from the boundary-layer solution. This transpiration velocity could be used as a boundary condition to calculate a new inviscid solution. If the inviscid/boundary layer solutions are iterated, then the normal velocity from the boundary layer solution will match the inviscid values exactly. Solutions are calculated for shear flows over a flat plate. Results from the present method compared well with Navier-Stokes solutions for incompressible constant shear flow and sinusoidal shear flow. Iterations of the inviscid/boundary layer solutions were not necessary. Compressible flow of an exponential shear at Mach 4 was calculated for cold, adiabatic, and hot walls. It was found that inviscid shear flows have a significant effect on skin friction and heating rates. The present method should yield accurate boundary-layer solutions at Reynolds numbers lower than the traditional boundary-layer equations.

DeJarnette, Fred R.↗
NASA NTRS2019

Influence of Source Propagation Direction and Shear Flow Profile in Impedance Eduction of Acoustic Liners

The acoustic impedance of liners is a key parameter for their design, and depends on the flow conditions, i.e., the sound pressure level and the presence of a grazing flow. The surface impedance of a locally reacting liner is defined as a local intrinsic property relating the acoustic pressure to the normal acoustic particle velocity at the liner surface. Impedance eduction techniques are now widely used to retrieve the impedance of liners in aeroacoustic facilities in the presence of a shear grazing flow. While surface impedance is intrinsic by definition, the educed impedance has recently been shown to depend on the direction of the incident waves relative to the mean flow. Different studies have investigated this issue by considering different acoustic propagation models used in the education process in the hope of matching the educed values. The purpose of the present work is to continue the previous investigations by evaluating the influence of the shear flow profile on the educed impedance, while considering a Bayesian inference process in order to evaluate the uncertainty on the educed values. The identified uncertainties were not able to totally account for the observed discrepancies between educed impedances.

Roncen, R.↗
NASA NTRS1992

BLSTA: A boundary layer code for stability analysis

A computer program is developed to solve the compressible, laminar boundary-layer equations for two-dimensional flow, axisymmetric flow, and quasi-three-dimensional flows including the flow along the plane of symmetry, flow along the leading-edge attachment line, and swept-wing flows with a conical flow approximation. The finite-difference numerical procedure used to solve the governing equations is second-order accurate. The flow over a wide range of speed, from subsonic to hypersonic speed with perfect gas assumption, can be calculated. Various wall boundary conditions, such as wall suction or blowing and hot or cold walls, can be applied. The results indicate that this boundary-layer code gives velocity and temperature profiles which are accurate, smooth, and continuous through the first and second normal derivatives. The code presented herein can be coupled with a stability analysis code and used to predict the onset of the boundary-layer transition which enables the assessment of the laminar flow control techniques. A user's manual is also included.

Wie, Yong-Sun↗
NASA NTRS1976

Diffraction of a shock wave by a compression corner; regular and single Mach reflection

The two dimensional, time dependent Euler equations which govern the flow field resulting from the injection of a planar shock with a compression corner are solved with initial conditions that result in either regular reflection or single Mach reflection of the incident planar shock. The Euler equations which are hyperbolic are transformed to include the self similarity of the problem. A normalization procedure is employed to align the reflected shock and the Mach stem as computational boundaries to implement the shock fitting procedure. A special floating fitting scheme is developed in conjunction with the method of characteristics to fit the slip surface. The reflected shock, the Mach stem, and the slip surface are all treated as harp discontinuities, thus, resulting in a more accurate description of the inviscid flow field. The resulting numerical solutions are compared with available experimental data and existing first-order, shock-capturing numerical solutions.

Vijayashankar, V. S.↗