[en] Many efforts have been made to reveal the jet-noise generation mechanisms for more than half a century. Although jet-noise phenomena of some specific types have been well understood, the mechanism of the most fundamental one, i.e. mixing noise, has not been revealed, or at least none of the claims or hypotheses have been widely accepted. To overcome this hurdle, recent acoustic- and flow-diagnostic techniques have been applied to near- and far-field measurements, and the relation between large-scale flow structures and far-field sound as well as properties of jet-noise sources have been investigated in many studies. In this paper, these diagnostic studies are reviewed, particularly focusing on the multi-point measurements including phased-array techniques, and the studies on subsonically convecting round jets, i.e. jets whose phase velocity is less than the speed of sound, are summarized. (invited review)

[en] Pile fabrics are known to be effective at reducing aerodynamic noise. However, it is not clear which kind of pile fabric is most effective at reducing noise and what the noise-reducing mechanism is. In this paper, simple experiments using a cylinder with and without pile fabrics clarify that the flow resistance of pile fabrics is an important parameter in reducing aerodynamic noise. Light resistance is preferable and the optimum filling rate of pile fabrics is around 1%. The noise-reducing mechanism of pile fabrics is discussed and clarified by comparing measured flow properties with computational fluid dynamics (CFD) results, which simulate pile fabrics by the porous media model with light flow resistance. (invited paper)

[en] Aerodynamic sound derived from bluff bodies can be considerably reduced by flow control. In this paper, the authors propose a new method in which porous material covers a body surface as one of the flow control methods. From wind tunnel tests on flows around a bare cylinder and a cylinder with porous material, it has been clarified that the application of porous materials is effective in reducing aerodynamic sound. Correlation between aerodynamic sound and aerodynamic force fluctuation, and a surface pressure distribution of cylinders are measured to investigate a mechanism of aerodynamic sound reduction. As a result, the correlation between aerodynamic sound and aerodynamic force fluctuation exists in the flow around the bare cylinder and disappears in the flow around the cylinder with porous material. Moreover, the aerodynamic force fluctuation of the cylinder with porous material is less than that of the bare cylinder. The surface pressure distribution of the cylinder with porous material is quite different from that of the bare cylinder. These facts indicate that aerodynamic sound is reduced by suppressing the motion of vortices because aerodynamic sound is induced by the unstable motion of vortices. In addition, an instantaneous flow field in the wake of the cylinder is measured by application of the PIV technique. Vortices that are shed alternately from the bare cylinder disappear by application of porous material, and the region of zero velocity spreads widely behind the cylinder with porous material. Shear layers between the stationary region and the uniform flow become thin and stable. These results suggest that porous material mainly affects the flow field adjacent to bluff bodies and reduces aerodynamic sound by depriving momentum of the wake and suppressing the unsteady motion of vortices. (invited paper)

[en] A method for simulating the hole-tone feedback cycle (Rayleigh's bird-call), based on an axisymmetric discrete vortex method, is described. Evaluation of the sound generated by self-sustained flow oscillations is based on the Powell-Howe theory of vortex sound combined with a thin-plate boundary element theory, to account for scattering from the end plate. A model for acoustic feedback is developed. Several numerical examples are presented and discussed. These examples give an understanding of the relative strengths of contributions from monopole, dipole and quadrupole source terms. It is found that the contribution from monopole sources is largest, followed by the dipole contribution. The quadrupole contribution is quite small in comparison. The effect of acoustic feedback is carefully studied. The acoustic feedback velocity components are too small to alter the fundamental hole-tone dynamics. Still, their effect can be seen on the sound pressure frequency spectra. It is found that they do not alter the peak at the fundamental hole-tone frequency f₀, but reinforce the peaks at higher harmonics 2f₀, 3f₀, ... , and certain combination frequency peaks. (invited paper)

[en] An investigation on flow-induced response characteristics of two identical circular cylinders in staggered arrangement is conducted in the fluid mechanics laboratory of Kitami Institute of Technology, Japan. Each cylinder was two-dimensional, spring mounted, and allowed to vibrate independently in the cross-flow direction. Measurements were conducted at stagger angle α = 5 deg., 10 deg., 15 deg., 25 deg., 45 deg. and 60 deg., L/D ranging from 0.1 to 3.2, with ΔL/D = 0.1, where L is the gap width between the cylinders, and D is the diameter of a cylinder. At each position (α, L/D) of the cylinders, dependence of vibration-amplitude-to-diameter ratio a/D on reduced velocity U_r (= U_∞/(f_nD)) is examined, where U_∞ is the free-stream velocity and f_n is the natural frequency of the cylinder. There are seven cylinder-response patterns, depending on whether vortex-excited and/or galloping vibrations of the cylinders are generated, in the range α = 5⁰-60⁰, L/D = 0.1-3.2 and U_r = 1.5-26. Pattern I corresponds to no generation of vortex-excited or galloping vibration of either cylinder. In Pattern II the upstream cylinder does not experience vortex-excited or galloping vibration, but the downstream one experiences a galloping vibration. Pattern III involves both vortex-excited and galloping vibrations of the downstream cylinder and only galloping vibration of the upstream cylinder. Pattern IV is associated with generation of vortex-excited vibration of both cylinders at the same U_r range. Pattern V refers to the case where the downstream cylinder experiences vortex-excited vibration, but the upstream cylinder does not. Pattern VI is characterized by vortex-excited vibration of the downstream cylinder in two regimes of U_r, whereas that of the upstream cylinder occurs in one regime only. In Pattern VII the upstream and downstream cylinders experience vortex-excited vibration at two different U_r regimes, respectively. The L/D regime of the vibration patterns generated at each α is identified.

[en] A third-order asymptotic analysis is conducted to study the three-dimensional resonant interaction between a monochromatic progressive surface wave and two oblique interfacial waves in an open, lightly viscous, two-layer fluid of intermediate depth. By solving the evolution equations of the waves, the short- and long-term behaviors of the interfacial waves are studied. The analysis provides a correction to the second-order theory. The results indicate that the third-order analysis predicts a much lower limit on the growth of the interfacial waves than the second-order theory. Furthermore, in the long term, viscous effects cause the interfacial wave amplitudes to approach a constant value. The effects of viscosity, surface wave frequency, surface wave amplitude, density difference of the layers and relative thickness of the two layers on the dynamics of the waves are examined. The theory is in qualitative agreement with laboratory observations. (paper)

[en] Two-layer systems heated from above with buoyancy as the driving force may show a rather paradoxical instability mechanism called anticonvection. This transition from a conductive to a convective state is determined by the interface as well as bulk properties (buoyancy forces) of the two fluids. In this paper, we derive the equations for perturbated fields from the basic hydrodynamic equations. An analytical expression for the control parameter at threshold is presented for a vertically infinitely extended system. Further on, we perform a linear stability analysis of a vertically bounded mercury–water system and compare these results with the vertically infinitely extended system. Finally, we show some three-dimensional simulations of the fully nonlinear equations and the resulting patterns for an anticonvective mercury–water system. (paper)

[en] This paper deals with the numerical study of bifurcations in the two-dimensional (2D) lid-driven cavity (LDC). Two specific geometries are considered. The first geometry is the two-sided non-facing (2SNF) cavity: the velocity is imposed on the upper and the left side of the cavity. The second geometry is the four-sided (4S) cavity where all the sides have a prescribed motion. For the first time, the linear stability analysis is performed by coupling two specific algorithms. The first one is dedicated to the computation of the stationary bifurcations and the bifurcated branches. Then, a second algorithm is dedicated to the computation of Hopf bifurcations. In this study, for both problems, it is shown that the flow becomes asymmetric via a stationary bifurcation. The critical Reynolds numbers are close to 1070 and 130, respectively, for the 2SNF and the 4S cavity. Following the stationary bifurcated branches, supplementary results concerning the stability are found. Firstly, for both examples, a second stationary bifurcation appears on the unstable solution, for a Reynolds number equal to 1890 and 360, respectively, for the 2NSF and the 4S cavity. Secondly, a second stationary bifurcation is found on the stable solutions of the 4S LDC for a critical Reynolds number close to 860. Nevertheless, no Hopf bifurcation has been found on this stable bifurcated branch for Reynolds numbers between 130 and 1000. Concerning the 2SNF LDC, Hopf bifurcation points have been determined on these stable bifurcated solutions. The first bifurcation occurs for a Reynolds number close to 3000 and a Strouhal number equal to 0.47. (paper)

[en] This paper presents the results of a detailed experimental and theoretical investigation on the viscous propagation of non-Newtonian gravity currents. Laboratory gravity currents are generated in a horizontal rectangular tank by releasing a constant flux of high-concentration fluid mud suspensions that exhibit profound non-Newtonian (shear thinning) behavior. Experimental observations on the propagation of fluid mud gravity currents revealed that viscous propagation of these currents was typically preceded by two phases as expected: an initial momentum-driven horizontal buoyant wall jet and a buoyancy-driven inertial phase. The experimental transition times, t**, and positions, x**, at which fluid mud gravity currents transition into viscous propagation phase were determined. The experimental data that correspond to the viscous propagation of fluid mud gravity currents (i.e. experimental time, t ⩾ t**, and front position, x_N ⩾ x**) were used to evaluate the predictive capabilities of two well-known mathematical modeling approaches: the lubrication theory approximation and the box-model approaches. Regarding the lubrication theory approximation, a recently developed self-similarity solution for viscous propagation of power-law gravity currents that has not been experimentally evaluated was used. Regarding the box-model approach, a viscous box-model solution for two-dimensional (2D) non-Newtonian gravity currents was developed. The evaluation of these models using experimental data revealed that both models were in good agreement with the experimental observations, despite several simplifying assumptions embedded in each. Given its more advanced mathematical development, the lubrication theory approximation model provides a more complete description of a gravity current (i.e. shape and velocity variation along the gravity current) than the box model at the expense of a relatively simple computational effort. (paper)

[en] The boundary region equations (BREs) are applied for the simulation of the nonlinear evolution of a spanwise periodic array of streaks in a flat plate boundary layer. The well-known BRE formulation is obtained from the complete Navier–Stokes equations in the high Reynolds number limit, and provides the correct asymptotic description of three-dimensional boundary layer streaks. In this paper, a fast and robust streamwise marching scheme is introduced to perform their numerical integration. Typical streak computations present in the literature correspond to linear streaks or to small-amplitude nonlinear streaks computed using direct numerical simulation (DNS) or the nonlinear parabolized stability equations (PSEs). We use the BREs to numerically compute high-amplitude streaks, a method which requires much lower computational effort than DNS and does not have the consistency and convergence problems of the PSE. It is found that the flow configuration changes substantially as the amplitude of the streaks grows and the nonlinear effects come into play. The transversal motion (in the wall normal-streamwise plane) becomes more important and strongly distorts the streamwise velocity profiles, which end up being quite different from those of the linear case. We analyze in detail the resulting flow patterns for the nonlinearly saturated streaks and compare them with available experimental results. (paper)