Engineering topics
Publications and source records attributed to Samimy, M..
Broad Flow Control Issues: a) Understanding flow physics. b) Specific control objective(s). c) Actuation. d) Sensors. e) Integrated active flow control system. f) Development of design tools (CFD, reduced order models, controller design, understanding and utilizing instabilities and other mechanisms, e.g., streamwise vorticity).
The idea of manipulating flow to change its characteristics is over a century old. Manipulating instabilities of a jet to increase its mixing and to reduce its radiated noise started in the 1970s. While the effort has been successful in low-speed and low Reynolds number jets, available actuators capabilities in terms of their amplitude, bandwidth, and phasing have fallen short in control of high-speed and high Reynolds number jets of practical interest. Localized arc filament plasma actuators have recently been developed and extensively used at Gas Dynamics and Turbulence Laboratory (GDTL) for control of highspeed and high Reynolds number jets. While the technique has been quite successful and is very promising, all the work up to this point had been carried out using small high subsonic and low supersonic jets from a 2.54 cm diameter nozzle exit with a Reynolds number of about a million. The preliminary work reported in this paper is a first attempt to evaluate the scalability of the technique. The power supply/plasma generator was designed and built in-house at GDTL to operate 8 actuators simultaneously over a large frequency range (0 to 200 kHz) with independent control over phase and duty cycle of each actuator. This allowed forcing the small jet at GDTL with azimuthal modes m = 0, 1, 2, 3, plus or minus 1, plus or minus 2, and plus or minus 4 over a large range of frequencies. This power supply was taken to and used, with minor modifications, at the NASA Nozzle Acoustic Test Rig (NATR). At NATR, 32 actuators were distributed around the 7.5 in. nozzle (a linear increase with nozzle exit diameter would require 60 actuators). With this arrangement only 8 actuators could operate simultaneously, thus limiting the forcing of the jet at NATR to only three azimuthal modes m = plus or minus 1, 4, and 8. Very preliminary results at NATR indicate that the trends observed in the larger NASA facility in terms of the effects of actuation frequency and azimuthal modes are similar in both small GDTL and larger NASA jets. However, the actuation authority seems to fall short in the larger jet at higher Mach numbers, resulting in decreased amplitude response compared to the small jet, which is attributed at this point to the lack of sufficient number of actuators. The preliminary results seem also to suggest that amplitude of actuation tones is similar in both the small and larger jets.
The results of an experimental investigation on the effect of vortex generators, in the form of small tabs at the nozzle exit on the evolution of a jet, are reported in this paper. Primarily tabs of triangular shape are considered, and the effect is studied up to a jet Mach number of 1.8. Each tab is found to produce a dominant pair of counter-rotating streamwise vortices having a sense of rotation opposite to that expected from the wrapping of the boundary layer. This results in an inward indentation of the mixing layer into the core of the jet. A triangular-shaped tab with its apex leaning downstream, referred to as a delta tab, is found to be the most effective in producing such vortices, with a consequential large influence on the overall jet evolution. Two delta tabs, spaced 180 deg apart, completely bifurcate the jet. Four delta tabs stretch the mixing layer into four 'fingers,' resulting in a significant increase in the jet mixing downstream. For six delta tabs the mixing layer distortion settles back to a three finger configuration through an interaction of the streamwise vortices. The tabs are found to be equally effective in jets with turbulent or laminar initial boundary layers. Two sources of streamwise vorticity are postulated for the flow under consideration. One is the upstream 'pressure hill,' generated by the tab, which constitutes the main contributor of vorticity to the dominant pair. Another is due to vortex filaments shed from the sides of the tab and reoriented downstream by the mean shear of the mixing layer. Depending on the orientation of the tab, the latter source can produce a vortex pair having a sense of rotation opposite to that of the dominant pair. In the case of the delta tab, vorticity from the two sources add, explaining the strong effect in that configuration.
The effect of vortex generators, in the form of small tabs projecting normally into the flow at the nozzle exit, on the characteristics of an axisymmetric jet is investigated experimentally over the jet Mach number range of 0.3-1.81. The tabs eliminate screech noise from supersonic jets and alter the shock structure drastically. They distort the jet cross section and increase the jet spread rate significantly. The distortion produced is essentially the same at subsonic and underexpanded supersonic conditions. Thus, the underlying mechanism must be independent of compressibility effects. A tab with a height as small as 2 percent of the jet diameter, but larger than the efflux boundary-layer thickness, is found to produce a significant effect. Flow visualization reveals that each tab introduces an 'indentation' into the high speed side of the shear layer via the action of streamwise vortices. These vortices are inferred to be of the 'trailing vortex' type rather than of the 'necklace vortex' type. It is apparent that a substantial pressure differential must exist between the upstream and the downstream sides of the tab to effectively produce these trailing vortices. This explains why the tabs are ineffective in the overexpanded flow, as in that case an adverse pressure gradient exists near the nozzle exit which reduces the pressure differential produced by the tab.
Streamwise vortices previously observed in underexpanded jets were studied using pitot pressure measurements and flow visualizations. A simple model was developed that gives reasonable agreement with the pressure measurements. A convergent nozzle and a convergent-divergent nozzle of design Mach number 1.5 were employed to generate jet flows of equivalent Mach numbers up to 2.5. By operating the nozzles fully expanded, overexpanded, and underexpanded, insight into both the occurrence and the cause of formation of the vortices was gained.
The results of a continuing investigation of the effect of vortex generators, in the form of small tabs at the nozzle exit, on the evolution of a jet are reported. Primarily, tabs of triangular shape are considered, and the effect is studied up to an equivalent jet Mach number of 1.8. By changing the orientation of the tab with respect to the nozzle exit plane, streamwise vortex pairs of opposite sign were generated. This resulted in either an outward selection of jet core fluid into the ambient or an inward indentation of the mixing layer into the core of the jet. A triangular shaped tab with its apex leaning downstream, referred to as a delta tab, was found to be the most effective in influencing the jet evolution. Two delta tabs, spaced 180 degrees apart, completely bifurcated the jet. Four delta tabs increased jet mixing substantially, more than by various other methods tried previously; the mass flux at fourteen jet diameters downstream from the nozzle increased by about 50 percent over that for the no tab case. The tabs were found to be effective in jets with laminar or turbulent boundary layers as well as in jets with low or high core turbulence intensities.
Results are presented of experiments in which a Rayleigh-scattering-based technique was used to make quantitative planar measurements in the compressible free shear layers. To this end, the absorption characteristics of the iodine molecular filter were investigated, and initial planar density measurements in M(c) = 0.51 shear layer were performed. The preliminary results presented here demonstrate the validity of the measurement technique.
Space-time correlations were used to study compressibility effects on large structures in mixing layers. Two high-Reynolds number mixing layers with M(c) = 0.51 (case 1) and 0.86 (case 2) were studied. The results indicate that the structures in case 1 are similar to those in the incompressible case, but less organized. The structures in case 2 are highly three-dimensional, with a good spatial but a poor temporal orgnization. The streamwise correlations showed a decay rate four to five times greater for case 2 relative to case 1. While the spanwise correlations for case 1 showed trends similar to incompressible mixing layers, the behavior of case 2 was very different. The pressure fluctuations in the fully developed region of case 2 displayed significant rms variation in the spanwise direction with a well-defined pattern. Based on these measurements, the structures in case 2 seem to be of a horseshoe type, transversely spanning the mixing layer with the head in the low-speed side and the legs inclined in both the x-y and the x-z planes.
The results of a continuing investigation of the effect of vortex generators, in the form of small tabs at the nozzle exit, on the evolution of a jet are reported. Primarily, tabs of triangular shape are considered, and the effect is studied up to an equivalent jet Mach number of 1.8. By changing the orientation of the tab with respect to the nozzle exit plane, streamwise vortex pairs of opposite sign were generated. This resulted in either an outward election of jet core fluid into the ambient or an inward indentation of the mixing layer into the core of the jet. A triangular shaped tab with its apex leaning downstream, referred to as a delta tab, was found to be the most effective in influencing the jet evolution. Two delta tabs, spaced 180 degrees apart, completely bifurcated the jet. Four delta tabs increased jet mixing substantially, more than by various other methods tried previously; the mass flux at fourteen jet diameters downstream from the nozzle increased by about 50 percent over that for the no tab case. The tabs were found to be effective in jets with laminar or turbulent boundary layers as well as in jets with low or high core turbulence intensities.
The effects of the inertia of a particle on its flow-tracking accuracy and particle dispersion are studied using direct numerical simulations of 2D compressible free shear layers in convective Mach number (Mc) range of 0.2 to 0.6. The results show that particle response is well characterized by tau, the ratio of particle response time to the flow time scales (Stokes' number). The slip between particle and fluid imposes a fundamental limit on the accuracy of optical measurements such as LDV and PIV. The error is found to grow like tau up to tau = 1 and taper off at higher tau. For tau = 0.2 the error is about 2 percent. In the flow visualizations based on Mie scattering, particles with tau more than 0.05 are found to grossly misrepresent the flow features. These errors are quantified by calculating the dispersion of particles relative to the fluid. Overall, the effect of compressibility does not seem to be significant on the motion of particles in the range of Mc considered here.
Experiments were conducted to explore the effects of vortex generators, in the form of tabs projecting normally into the jet, on the mixing and the far-field noise characteristics of a jet. A converging-diverging nozzle with a design Mach number of 1.36 was used in the experiments. The flow regimes from subsonic to highly underexpanded supersonic conditions were studied. One, two, and four tabs were used and some of the findings of previous investigators were examined and confirmed. The tabs eliminated screech noise from moderately overexpanded cases to highly underexpanded cases. Detailed flow visualizations and measurements showed that two tabs bifurcated the jet at all Mach numbers. While the effect of two tabs was persistent and the jet remained bifurcated, the distortions produced by one and four tabs disappeared by a streamwise distance of approximately 16 jet diameters. Two and four tabs significantly increased the entrainment of ambient air into the jet.
The effect of vortex generators, in the form of small tabs at the nozzle exit, on the evolution of an axisymmetric jet was investigated experimentally over a jet Mach number range of 0.34 to 1.81. The effects of one, two, and four tabs were studied in comparison with the corresponding case without a tab. Each tab introduced an indentation in the shear layer, apparently through the action of streamwise vortices which appeared to be of the trailing vortex type originating from the tips of the tab rather that of the necklace vortex type originating from the base of the tab. The resultant effect of two tabs, placed at diametrically opposite locations, was to essentially bifurcate the jet. The influence of the tabs was essentially the same at subsonic and supersonic conditions indicating that compressibility has little to do with the effect.
The effect of vortex generators, in the form of small tabs at the nozzle exit, on the evolution of an axisymmetric jet was investigated experimentally over a jet Mach number range of 0.34 to 1.81. The effects of one, two, and four tabs were studied in comparison with the corresponding case without a tab. Each tab introduced an indentation in the shear layer, apparently through the action of streamwise vortices which appeared to be of the trailing vortex type originating from the tips of the tab rather than of the necklace vortex type originating from the base of the tab. The resultant effect of two tabs, placed at diametrically opposite locations, was to essentially bifurcate the jet. The influence of the tabs was essentially the same at subsonic and supersonic conditions indicating that compressibility has little to do with the effect.
A preliminary experimental study has been conducted to gain insight into the temporal and spatial contents of largescale structures in a convective Mach number = 0.51 high Reynolds number planar two-dimensional compressible free shear layer. Power spectra, coherence, and space-time correlations were obtained using single- and two-point pressure measurements. Both developing and fully developed regions of the flow were investigated. The passage frequency of structures were found to be 10-15 kHz in the developing region and 4-8 kHz in the fully developed region. The convective velocity obtained from the space-time correlation in the middle of the shear layer was close to the theoretical value but varied toward the edge of the shear layer. The structures were shown to be surprisingly three-dimensional even in this low compressibility level shear layer. The structures size and spacing were also determined and discussed.
Explore the source record for details and available documents.
A high Reynolds number two-dimensional constant pressure compressible shear layer was formed at the trailing edge of an 0.5 mm-thick splitter plate. Convective Mach numbers of 0.51 and 0.64 were investigated using a two-component coincident LDV for the measurements. For the lower convective Mach number case, the nondimensionalized shear-layer and vorticity thickness growth rates were over 20 percent higher and the momentum thickness growth rate was over 30 percent higher than those of the higher convective Mach number case. The results seen to indicate that both small scale and large scale mixing are reduced with increasing convective Mach number.
High-Reynolds-number compressible free shear layers were studied experimentally to explore the effects of compressibility on the turbulence field. A reduction in both the level and the lateral extent of turbulence fluctuations with increasing convective Mach number (Mc) (reported earlier for Mc of 0.51 and 0.64) is much higher at Mc of 0.86. The higher-order moments of turbulence fluctuations such as skewness and flatness show that the intermittency due to the excursion of large-scale structures into the free streams at the edge of shear layers was significantly reduced (both in the level and the extent) due to increased Mc.
Two compressible free shear layers with convective Mach numbers of .51 and .86 were studied as baseline configurations to investigate the effects of compressibility on the turbulence characteristics. These shear layers were then disturbed by the placement of an obstruction in the shear layer in an attempt to enhance the shear layer growth rate. These models produced a curved shock in the supersonic side of the shear layer. The results indicate a significant reduction in turbulence levels with increased compressibility. However, there are not any significant changes due to the bow shock interaction with the shear layer.