Global stability of separated flows: subsonic flow past corners

The triple-deck equations for the steady subsonic flow past a convex corner are solved numerically using a novel technique based on Chebychev collocation in the direction normal to the body combined with finite differences in the direction along the flow. The resulting set of nonlinear algebraic equations are solved with Newton linearization and using the GMRES method for the solution of the linear system of equations. The stability of the computed steady flows is then examined using global stability analysis. It is found that for small corner angles, the Tollmien?CSchlichting modes are globally unstable and these persist to larger corner angles. Multiple steady state solutions also exist beyond a critical corner angle but these are globally unstable because of the presence of the Tollmien?CSchlichting modes.     

A numerical investigation of the influence of the aspect ratio in three‐dimensional separated flows

The influence of aspect ratio in three‐dimensional, numerical experiments of separated flows is studied in the case of the backward‐facing step at Reynolds numbers 600, 800, and 950. The computational domain is designed as an actual laboratory experiment. The governing equations are the steady state, isothermal, and incompressible Navier–Stokes equations. The expansion ratio of the computational domain is 1:2. The aspect ratio varies from 1:10 to 1:40. The results of the computations focus on the spanwise variations of the length and the strength of the two eddies along the lower and upper wall. It is concluded that both numerical and laboratory experiments should be designed with an aspect ratio of at least 1:20, if only the accuracy of the position of the detachment and the re‐attachment points matters. If the accuracy of the shear‐stress distributions is also taken into account, then an aspect ratio of at least 1:30 should be chosen. Finally, if the magnitudes of the vortex centers are also considered, then only the aspect ratio of 1:40 qualifies for a realization of two‐dimensional flow conditions in the plane of symmetry. This is contrary to the common practice in the field, at least from the side of laboratory experiments, where an aspect ratio of 1:10 is still considered adequate for this purpose. Copyright © 2011 John Wiley & Sons, Ltd.     

An experimental investigation of gas-particle flows through diffusers in the freeboard region of fluidized beds

The phenomenon of particulate loss (elutriation) from fluidized beds is important in many industrial processes. Results reported in Kale & Eaton (1984a) showed that very-wide-angle diffusers located in the freeboard above a fluidized bed substantially reduce elutriation—a result that was contrary to intuition. The present experiment was designed to explain these results. The same fluidized bed apparatus (Kale & Eaton 1984a) was used—150 mm square in cross section with a variable-angle diffuser in the freeboard region. Glass beads (nominally 50–100 μm in diameter) were fluidized by air at atmospheric pressure in the bubbling regime. Gas-phase velocity measurements were made using a single-component laser-Doppler anemometer. Four diffuser configurations (0, 20, 40 and 60° full opening angle) were studied. One set of measurements was made with the bed in place and a second set with the bed material removed. The flow structure was drastically altered by the presence of the fluidized bed below the diffuser. The single-phase flow was separated in the diffuser for the 20, 40 and 60° cases. However, the flow did not separate in the presence of the bed, and the peak fluid velocities were lower than those in the separated flow. This behavior is responsible for the decrease in the elutriation rate with increasing diffuser opening angle. A simple analysis suggests that suspended particles in the diffuser flow are responsible for the change in the flow structure. Momentum loss from the gas to the suspended particles reduces the pressure gradient, thereby eliminating the tendency to separate.     

Stability of subsonic gasdynamic flows

A system of differential equations describing small perturbations of the steady flow of a non-viscous ideal gas in a channel of variable cross section is analyzed in this paper. The equations of nonsteady flow and the boundary conditions are linearized, and the solution of the linearized equations is sought in the form v(x)expt t, where v(x) is an eigenfunction while is the natural frequency for the boundary problem being studied. With such an approach the problem is reduced to finding the solutions to ordinary differential equations with variable coefficients which depend on the parameter . Analytical solutions of this system are obtained for small values of and for values of ¦¦1. The results can be used to calculate the growth of high-frequency and low-frequency perturbations imposed on subsonic, supersonic, and mixed (i.e., with transitions through the velocity of sound) gasdynamic flows, to analyze the stability of subsonic sections, and to verify and supplement various numerical methods for calculating unsteady flows and numerical methods for studying stability in gasdynamics. The application of the solutions found for small and large is demonstrated on a study of flow stability behind a shock wave (a direct compression shock in the present formulation). Analytical expressions are obtained for the determination of from which it follows that the flow stability behind a shock essentially depends on the shape of the channel at the place where the shock is located in the steady flow, which was noted earlier in [1], and on the conditions of the reflection of small perturbations in the exit cross section of the channel, which was first pointed out in [2].Translated from Izvestiya Akademii Nauk SSSR, Mekhanika Zhidkosti i Gaza, No. 1, pp. 90–97, January–February, 1978.In conclusion, the author thanks A. G. Kulikovskii and A. N. Sekundov for helpful discussions of the work.     

Experimental analysis of unsteady separated flows in a supersonic planar nozzle

The unsteady aspects of shock-induced-separation patterns have been investigated inside a Mach 2 planar nozzle. The mean location of the shock can vary by changing, relatively to the nozzle throat, the height of the second throat which is positioned downstream of the square test section. This study focuses on the wall pressure fluctuations spectra and the unsteady behaviour of the shock. Symmetric shock configurations appear both for the largest openings of the second throat, and for the smallest openings. For an intermediate opening the shock system exhibits asymmetrical configurations. A coating with roughnesses sticked on the throat part of the nozzle in order to modify the state of the incoming boundary layers (from smooth to rought turbulent statement) is a driver for the asymmetry. The fluctuating displacements of the shock patterns were analysed by using an ultra fast shadowgraph visualization technique. A spectral analysis of the unsteady wall pressure measurements has revealed low frequency phenomena governed by large structure dynamics in the separated flows. Communicated by K. Takayama PACS 02.60.Cb; 05.10.Ln; 47.11.+j; 47.15.Cb; 47.40.Nm     

Wake development in turbulent subsonic axisymmetric flows

The development of the wake velocity and turbulence profiles behind a cylindrical blunt based body aligned with a subsonic uniform stream was experimentally investigated as a function of the momentum thickness of the approaching boundary layer and the transfer of mass into the recirculating region. Measurements were made just outside of the recirculating region at distances of 1.5, 2 and 3 diameters downstream of the cylinder. Results indicate that, even at these short distances from the cylinder base, the velocity profiles are similar. They also show that the width of the wake increases with the thickness of the boundary layer while the velocity at the centerline decreases. Near wake mass transfer was found to alter centerline velocities while the width of the wake was not significantly altered. Wake centerline velocity development as a function of boundary layer thickness is presented for distances up to three diameters from the base. This work was supported in part by the ‘Xunta de Galicia’ under Project No. XUGA20611B93.     

Local subsonic zones in supersonic jet flows

The results of the experimental investigation of supersonic turbulent jets with local subsonic zones of forward and reverse flow exhausting into the ambient atmosphere or an outer stream, either parallel or transverse to the jet, are presented. Some gasdynamic features of the flows containing these zones, which have not been adequately addressed in the literature, are analyzed. Thus, supersonic flows with back pressure, e.g., highly overexpanded and underexpanded jet flows, and those upstream and downstream of a jet on the leeward side of a cone in a supersonic gas stream, are studied. Moscow. Translated from Izvestiya Rossiiskoi Akademii Nauk, Mekhanika Zhidkosti i Gaza, No. 1, pp. 143–150, January–February, 1998.     

On the use of adaptive hierarchical meshes for numerical simulation of separated flows

This paper describes the use of adaptive hierarchical grids to predict incompressible separated flow at low Reynolds number. The grids consist of a quadtree system of hierarchical Cartesian meshes which are generated by recursive subdivision about seeding points. The governing equations are discretized in collocated primitive variable form using finite volumes and solved using a pressure correction scheme. The mesh is locally adapted at each time step, with panel division or removal dependent on the vorticity magnitude. The resulting grids have fine local resolution and are economical in array size. Results are presented for unidirectional, impulsively started flow past a circular and a square cylinder at various Reynolds numbers up to 5000 and 250 respectively. It is clear that hierarchical meshes may offer gains in efficiency when applied to complex flow domains or strongly sheared flows. However, as expected, the stepped approximation to curved boundaries resulting from the Cartesian quadtree representation adversely affects the accuracy of the results for flow past a circular cylinder. © 1998 John Wiley & Sons, Ltd.     

Experimental investigation of return currents in electrogasdynamic flows

An investigation is made of return electric currents in electrogasdynamic flows for laboratory sources of unipolar charged particles. These currents play an important role in the process of airplane electrification as a result of the work of jet engines. Models have been built, making it possible to study the behavior of return currents outside and inside an axisymmetric electrogasdynamic flow, in the absence (single-contour source) and the presence (double contour source) of an external annular neutral jet. It is shown that a rise in the return current J^– outside an electrogasdynamic jet is accompanied by a decrease in the take-off current J °. A decrease in the relative distance L from the source to an external grounded surface and an increase in the ratiov of the velocity of the external neutral jet to the velocity of the electrogasdynamic flow lowers J^– in both grounded and insulated models; in the latter case, where J ° J°0, there is an appreciable return current outside the jet. With an increase in the potential of the source from =0 to the floating potential, the current J^– rises, attaining a maximum, and then decreases. This effect is observed also when J^–=0 in both grounded and insulated models. For the case L–1,v=1, the theoretical and experimental dependences of J^– on the potential of the source , retarding the charged particles of the flow under transitional conditions, are in satisfactory agreement.Translated from Izvestiya Akademii Nauk SSSR, Mekhanika Zhidkosti i Gaza, No. 1, pp. 127–134, January–February, 1978.In conclusion, the authors thank A. B. Vatazhin for his interest in the work, and A. P. Strekal and V. F. Kudryashov for their participation in the experiments.     

Experimental investigation of helical structures in swirling flows

In this investigation the flow in a generic swirl tube with a tangential double-inlet swirl generator and variable exit orifices was experimentally investigated. Using magnetic resonance velocimetry (MRV) three-dimensional, three-component velocity fields were measured for two different Reynolds numbers: 10,000 and 15,000, and for three different exit orifices. The swirl generator had a fixed geometry producing an initial swirl number of 1.6 for all cases. One major observation is the occurrence of a three-layered flow structure. An annular main flow was surrounded by a recirculation zone, as reported in previous literature. However, this recirculation zone – also of an annular shape – exhibited a third layer inside: a thin, high speed jet in the center of the tube with the same flow direction as the main flow. Therefore, the conventional classification of swirling flows into ring and recirculation zone, has to be extended by a core zone. This three-layered flow structure develops independently of the exit configuration. Helical structures were observed in the near-wall region for all cases investigated. Applying an eccentric exit orifice results in the occurrence of strong stationary helical structures not only in the near-wall region but also in the center of the tube. The results, deviating significantly from previous results in the literature, underline the need for more detailed research on the topic of cyclone type flows.     

An investigation of LDV velocity bias correction techniques for high-speed separated flows

An experimental study of the effects of velocity bias in single realization laser Doppler velocimetry measurements in a high-speed, separated flow environment is reported. The objective of the study is to determine a post-facto correction method which reduces velocity bias after individual realization data have been obtained. Data are presented for five velocity bias correction schemes: inverse velocity magnitude weighting, interarrival time weighting, sample and hold weighting, residence time weighting, and the velocity-data rate correlation method. These data were compared to a reference measurement (saturable detector sampling scheme); the results show that the interarrival time weighting method compares favorably with the reference measurement under the present conditions.     

Numerical investigation of unsteady subsonic viscous gas flows in a plane channel with a sudden expansion

The nonlinear processes of development of instability in an unsteady subsonic viscous gas flow in a plane channel with a sudden expansion are investigated numerically with allowance for acoustico-vortex interaction over a broad interval of the characteristic parameters. Effects associated with the acoustic self-excitation of the jet flowing into the wider part of the channel are determined. Approximate relations are obtained for the resonance conditions of self-excitation. The effect of the inlet mean-velocity profiles on the evolution of the flow is estimated. The processes of formation and subsequent interaction of the coherent structures are analyzed.Translated from Izvestiya Akademii Nauk SSSR, Mekhanika Zhidkosti i Gaza, No. 4, pp. 32–41, July–August, 1988.     

Numerical solution of subsonic and transonic cascade flows

In this work a study of the application of the finite element method to transonic flows in axial turbomachines is undertaken. Solution techniques capable of accurately predicting flows from the incompressible regime up to the establishment of shocks in the transonic regime are presented. In the subsonic and shockless transonic regimes a local linearization method capable of very rapid convergence is used. In the full transonic regime the artificial compressibility method is employed to exclude downstream influences in the supersonic regions. The two approaches can be combined in a unified package and appropriate switches introduced to select the relevant method in any flow regime.     

Three-dimensional reduced Navier–Stokes solutions for subsonic separated and non-separated flows,using a global pressure relaxation procedure

The reduced Navier–Stokes and thin layer approximations to the Navier–Stokes equations are used to obtain solutions for viscous subsonic three-dimensional flows. A spatial marching method is combined with a direct sparse matrix solver to obtain successive solutions in a global relaxation process. Results have been obtained for flow fields with and without regions of flow reversal.     

Wave disturbances in transonic separated flows

The results of an investigation of the nature and role of wave disturbances in transonic separated flows are presented. The effect of these disturbances on flow formation and stability, as well as on the characteristics of the pressure fluctuations and the time-average flow parameters, is considered.     

Shock structure in separated nozzle flows

In the case of high overexpansion, the exhaust jet of the supersonic nozzle of rocket engines separates from nozzle wall because of the large adverse pressure gradient. Correspondingly, to match the pressure of the separated flow region, an oblique shock is generated which evolves through the supersonic jet starting approximately at the separation point. This shock reflects on the nozzle axis with a Mach reflection. Thus, a peculiar Mach reflection takes place whose features depend on the upstream flow conditions, which are usually not uniform. The expected features of Mach reflection may become much difficult to predict, depending on the nozzle shape and the position of the separation point along the divergent section of the nozzle.      

A numerical method for the calculation of incompressible,steady, separated flows around aerofoils

The present work deals with the numerical calculation of the incompressible turbulent flow around aerofoils. An orthogonal curvilinear grid of ‘C’ type is used for the solution of the time averaged equations and Reynolds stresses are modelled according to the κ-ε turbulence model. PISO and SIMPLE algorithms are used to solve the strongly coupled system of the derived finite volume equations and convergence is improved by applying the method of variable local underrelaxation factors. Comparisons between the calculated and measured pressure distributions are presented for NACA 0012 and NACA 4412 wing sections. The formation of separation bubbles according to calculations is also shown.