On the fluid–structure interaction of a splitter plate: vibration modes and Reynolds number effects

Previous work (Eloranta et al. in Exp Fluids 39:841–855, 2005) has shown that flow separation from the trailing edge of a splitter plate in a convergent channel involves a fluid–structure interaction (FSI), which modifies the fundamental instability related to vortex shedding. Under certain conditions, the FSI induces cellular vortex shedding from the trailing edge. This paper reports detailed measurements of the plate vibration mode and studies the effect of the Reynolds number on the FSI. Experimental techniques including laser vibrometer and digital imaging are used to measure the response of the plate and particle image velocimetry is used to measure the flow field in the near wake. Combining data from these techniques, the development of the vibration frequency and mode can be addressed together with the imprint of the vibration mode in the flow. The results show that over most of the Reynolds numbers measured, the plate vibrates in a distinct mode characterized by a spanwise standing wave along the plate trailing edge. The vibration frequency and the spacing between the nodes of the standing wave depend on the Reynolds number. As the Reynolds number is increased, the frequency of the dominant vibration mode does not increase linearly. The plot of the vibration frequency as a function of the Reynolds number shows that the vibration tends to lock to a rather constant frequency over of range of Reynolds numbers. After certain Reynolds number if threshold is exceeded, the frequency jumps to a new level, which also involves a new vibration mode.     

Numerical simulation of the turbulent wake behind a normal flat plate

The three-dimensional wake flow behind a flat plate placed normal to the free stream has been investigated by means of direct numerical simulations. The Reynolds number Re based on the homogeneous inflow velocity and the uniform width d of the plate was 750. Coherent vortices were alternately shed from the sides of the plate with a frequency corresponding to a Strouhal number 0.168. The wake was distinctly turbulent downstream of the plate whereas the mean recirculation bubble extended 1.96d downstream. A steady 2D mean flow and the accompanying Reynolds stresses were obtained by averaging in time and along the span of the plate. These Reynolds-averaged statistics exhibited the same qualitative features as corresponding data from cylinder wakes.     

Silicon based flow sensors used for mean velocity and turbulence measurements

Small and directional sensitive silicon based sensors for velocity measurements have been designed and fabricated using microelectronic technology. Single-chip as well as double-chip sensors for the determination of mean velocity and turbulent stresses have been developed. To determine the performance of these silicon sensors, comparisons with conventional hot-wire sensors were done in a well-defined two-dimensional turbulent flat plate boundary layer at a constant Reynolds number of 4.2 · 10⁶. All the silicon sensors were found to have a spatial and frequency resolution that makes them suitable for turbulence measurements. In the studied flow field the measured profiles of mean velocities and Reynolds stresses of all silicon sensors show the same accuracy as corresponding hot-wire measurements. The silicon sensors are also shown to operate with good resolution even when the temperature of the heated part of the chip is reduced considerably.     

Die turbulente Grenzschicht an einer stark geheizten ebenen Platte bei Unterschallströmung

Experiments for air flowing over a flat plate heated up to 250°C with velocities of 10 to 30 m/s, which have been made at the DFVLR-AVA, are briefly reviewed and a new analysis of the data is given. The analysis is based on an analytical representation of the velocity and temperature profiles. Close to the wall, a law of the wall approximation is used, which includes the effect of density and viscosity variation. The whole velocity profile is constructed by adding Coles' law of the wake to the law of the wall. In a similar way, the temperature profile is obtained from the law of the wall and an auxiliary distribution. The integrals of momentum and heat flux for two-dimensional flow are used in conjunction with a similarity assumption, to derive a relation between rate of heat transfer from the plate and skin friction. A maximum likelihood procedure has been applied to determine skin friction and rate of heat transfer from the measured dynamic pressure profiles.—The analytical velocity and temperature profiles are found in good agreement with the experimental data, except for the stations near the leading edge of plate. The skin friction coefficients and the Stanton numbers decrease slightly in downstream direction as a consequence of growing local Reynolds number, and decrease with increasing ratio of plate to free stream temperature. The latter fact is in qualitative agreement with the behavior of turbulent boundary layers in supersonic flow. The ratio of Stanton number to half of skin friction coefficient (Reynolds analogy factor) varies with increasing local boundary layer Reynolds number from 1.23 to 1.16.     

Two-dimensional unsteady laminar flow of a power law fluid across a square cylinder

The two-dimensional and unsteady free stream flow of power law fluids past a long square cylinder has been investigated numerically in the range of conditions 60≤Re≤160 and 0.5≤n≤2.0. Over this range of Reynolds numbers, the flow is periodic in time. A semi-explicit finite volume method has been used on a non-uniform collocated grid arrangement to solve the governing equations. The global quantities such as drag coefficients, Strouhal number and the detailed kinematic variables like stream function, vorticity and so on, have been obtained for the above range of conditions. While, over this range of Reynolds number, the flow is known to be periodic in time for Newtonian fluids, a pseudo-periodic flow regime displaying more than one dominant frequency in the lift is observed for shear-thinning fluids. This seems to occur at Reynolds numbers of 120 and 140 for n=0.5 and 0.6, respectively. Broadly speaking, the smaller the value of the power law index, lower is the Reynolds number of the onset of the pseudo-periodic regime. This work is concerned only with the fully periodic regime and, therefore, the range of Reynolds numbers studied varies with the value of the power law index. Not withstanding this aspect, in particular here, the effects of Reynolds number and of the power law index have been elucidated in the unsteady laminar flow regime. The leading edge separation in shear-thinning fluids produces an increase in drag values with the increasing Reynolds number, while shear-thickening fluid behaviour delays this separation and shows the lowering of the drag coefficient with the Reynolds number. Also, the preliminary results suggest the transition from the steady to unsteady flow conditions to occur at lower Reynolds numbers in shear-thinning fluids than that in Newtonian fluids.     

Vortex shedding and three-dimensional behaviour of flow past a cylinder confined in a channel

A numerical investigation of the flow past a circular cylinder centred in a two-dimensional channel of varying width is presented. For low Reynolds numbers, the flow is steady. For higher Reynolds numbers, vortices begin to shed periodically from the cylinder. In general, the Strouhal frequency of the shedding vortices increases with blockage ratio. In addition, a two-dimensional instability of the periodic vortex shedding is found, both empirically and by means of a Floquet stability analysis. The instability leads to a beating behaviour in the lift and drag coefficients of the cylinder, which occurs at a Reynolds number higher than the critical Reynolds number for the three-dimensional mode A-type instability, but lower than a Reynolds number for any mode B-type instability.     

Non-Isotropic Length Scales During the Compression Stroke of a Motored Piston Engine

For the case of axial compression the two-point velocity correlation equations of axisymmetric homogeneous turbulence are derived. Appropriate integrations then lead to equations for the components of the Reynolds stress tensor as well as to those for the two independent integral length-scales characterizing axisymmetric homogeneous turbulence. These equations contain a certain number of empirical constants. Values for these constants are taken from the literature, or were adjusted from the present data.The resulting model is validated using data from a motored piston engine. The flow field, which has negligible swirl and tumble, has been measured using particle image velocimetry (PIV). Since turbulence is axisymmetric and homogeneous in the counter region, two-dimensional PIV provides the time history of the axial and radial length-scales. The experimental data are compared with the mathematical model.     

Characteristic Regimes of Transitional Separation Bubbles in Unsteady Flow

This experimental investigation deals with transition phenomena of a separated boundary layer under unsteady inlet flow conditions. The main purpose of this investigation is to understand the influence of the rotor-stator interaction in turbomachinery on the subsequent, highly loaded boundary layer. The research project is divided into two phases. In the first phase, which has been completed recently, only the variation of mean velocity caused by upstream blades was simulated in the experiments while the free-stream turbulence intensity was retained at a constant low level. The experiments are carried out in an Eifel-type wind tunnel to investigate the laminar separated boundary layer of a flat plate under oscillating inlet conditions. The adverse pressure gradient, similar to that of turbomachines, is generated by the contoured upper wall. The unsteadiness is produced by a rotating flap located downstream of the test section. The reduced frequency, the amplitude and the mean Reynolds number are varied to simulate the conditions prevailing in turbomachines. In addition to the Kelvin–Helmholtz instability of the separated shear layer, a lower frequency instability was observed. This is frequently referred to as `free shear layer flapping' and results in two distinctly different ways of re-attachment, depending primarily on the Reynolds number. For low momentum thickness Reynolds numbers at the separation point, large-scale vortices locked to the frequency of the unsteady main flow are identified. They originate nearly at the top of the separation bubble and are ejected downstream. A fully turbulent boundary layer develops after these vortices mix out. For higher Reynolds numbers, transition is completed within a short length of the free shear layer and there-attachment region. The characteristic momentum thickness Reynolds number separating these two regimes in unsteady flow is about 125. The Strouhal number (reduced frequency) does not appear to have any significant effect. Based on the experimental results, this behaviour is discussed in some detail. This revised version was published online in July 2006 with corrections to the Cover Date.     

Shedding patterns of the near-wake vortices behind a circular cylinder

The unsteady incompressible Navier-Stokes equations have been accurately solved for the laminar flow past a circular cylinder in the Reynolds number range 50–200. A direct elliptic solver called the SEVP is used to rapidly advance the streamfunction in time, facilitating the overall convergence to the fully periodic or quasi-steady state. A new integral-series method is developed for the far-field streamfunction condition on a finite two-dimensional computational domain. The use of fourth-order Hermitian relations for the convection terms in the conservation-form vorticity transport equation has also contributed to the good comparison of the present results with the earlier experimental data. The vortex-shedding patterns visualized by the experimentalist are numerically reproduced here in the given Reynolds number range. Discussions that may be helpful in interpreting the behaviour of the shedding frequency are presented in the main text.     

Hydrodynamic and heat transfer characteristics of laminar flow past a parabolic cylinder with constant heat flux

 Steady, two-dimensional, symmetric, laminar and incompressible flow past parabolic bodies in a uniform stream with constant heat flux is investigated numerically. The full Navier–Stokes and energy equations in parabolic coordinates with stream function, vorticity and temperature as dependent variables were solved. These equations were solved using a second order accurate finite difference scheme on a non-uniform grid. The leading edge region was part of the solution domain. Wide range of Reynolds number (based on the nose radius of curvature) was covered for different values of Prandtl number. The flow past a semi-infinite flat plate was obtained when Reynolds number is set equal to zero. Results are presented for pressure and temperature distributions. Also local and average skin friction and Nusselt number distributions are presented. The effect of both Reynolds number and Prandtl number on the local and average Nusselt number is also presented. Received on 5 July 2000     

Experimental study of the response of a flexible plate to a harmonic forcing in a flow

Most aquatic animals propel themselves by flapping flexible appendages. To gain insight into the effect of flexibility on the swimming performance, we have studied experimentally an idealized system. It consists of a flexible plate whose leading edge is forced into a harmonic heave motion, and which is immersed in a uniform flow. As the forcing frequency is gradually increased, resonance peaks are evidenced on the plate's response. In addition to the forcing frequency, the Reynolds number, the plate rigidity and the forcing amplitude have also been varied. In the range of parameters studied, the main effect on the resonance is due to the forcing amplitude, which reveals that non-linearities are essential in this problem.     

Flow induced oscillation of two rigid rectangular plates in a side-by-side configuration

The present work is an experimental study of two oscillating rigid plates placed in side-by-side configuration, hinged at their leading edges, subjected to low subsonic flow. This problem is investigated using smoke-wire flow visualization, hot-wire anemometry, and time resolved particle image velocimetry. It is found that beyond a critical Reynolds number, the plates set into oscillatory motion. This critical Reynolds number depends on the gap between the plates. It is also seen that this value of Reynolds number, at lower values of gap to thickness ratio (<7) is significantly higher than that of the single plate configuration value. The frequency and amplitude of the oscillating plates at various gaps and Reynolds numbers have been studied and compared with the characteristics of an oscillating single plate. It is also found that depending on the gap and acceleration of the free-stream, there exist two modes of oscillation - (i) in-phase and (ii) out-of-phase. For gap to thickness ratio less than 10, only in-phase oscillations take place for all values of free-stream velocity considered in the present work, whereas, when this ratio is greater than 10, the mode of oscillation depends on the initial conditions up to a certain free-stream velocity, beyond which the plates switch to in-phase mode. Smoke wire flow visualization technique along with time resolved particle image velocimetry reveal that the vorticity distributions around the plates are responsible for the initiation of the two modes of oscillations.     

双颗粒沉降的动力学分析

采用格子Boltzmann-虚拟区域方法对雷诺数范围为50≤Re≤200的双颗粒自由沉降进行了直接数值模拟。首先,研究发现双颗粒最终沉降的位置分别在通道的1/4和3/4位置附近,且与颗粒的初始间距、雷诺数以及通道宽度无关。其次,重点研究了颗粒沉降过程中所受的侧向力(与沉降方向垂直),首次揭示了侧向力的振动频率与雷诺数呈二次关系,且单颗粒的结果始终小于双颗粒的结果;研究还发现侧向力的振动频率与通道宽度近似呈幂律函数关系,且幂律指数与雷诺数有关,雷诺数越大,幂律指数的绝对值越小。最后还研究了雷诺数及通道宽度对侧向力振动振幅的影响。     

Numerical solution of viscous flows using integral equation methods

A formulation of the boundary element method for the solution of non-zero Reynolds number incompressible flows in which the non-linear terms are lumped together to form a forcing function is presented. Solutions can be obtained at low to moderate Reynolds numbers. The method was tested using the flow of a fluid in a two-dimensional converging channel (Hamel flow) for which an exact solution is available. An axisymmetric formulation is demonstrated by examining the drag experienced by a sphere held stationary in uniform flow. Performance of the method was satisfactory. New results for an axisymmetric free jet at zero Reynolds number obtained using the boundary element method are also included. The method is ideal for this type of free-surface problem.     

Micro-Fourier rheometer: Inertial effects

The inertial effects in a random squeezing rheometer are examined, both theoretically and experimentally. The rheometer is based on small amplitude random squeezing between two parallel plates, where the upper plate is driven by a random displacement with a broad band spectrum. A fast Fourier transform is used to deliver the complex modulus (or viscosity) of the fluid in a single brief test, over more than two decades of frequency. The inertia of the fluid is shown to produce an error factor, which is also a function of the frequency. The correction factor can be well approximated by a first-order correction in the Reynolds number, for a very large range of Reynolds number, making the inertial correction a very simple procedure for light fluids.     

Transverse galloping of circular cylinders fitted with solid and slotted splitter plates

The galloping response of a circular cylinder fitted with three different splitter plates and free to oscillate transverse to a free stream has been investigated considering variations in plate length and plate porosity. Models were mounted in a low mass and damping elastic system and experiments have been carried out in a recirculating water channel in the Reynolds number range of 1500 to 16 000. Solid splitter plates of 0.5 and 1.0 diameter in length are shown to produce severe galloping responses, reaching displacements of 1.8 diameters in amplitude at a reduced velocity of around 8. Fitting a slotted plate with a porosity ratio of 30% also caused considerable vibration, but with a reduced rate of increase with flow speed. All results are compared with the typical vortex-induced vibration response of a plain cylinder. Force decomposition in relation to the body velocity and acceleration indicates that a galloping mechanism is responsible for extracting energy from the flow and driving the oscillations. Visualisation of the flow field around the devices performed with PIV reveal that the reattachment of the free shear layers on the tip of the plates is the hydrodynamic mechanism driving the excitation.     

Taylor instability of a fluid film on a body

The Taylor instability develops in a parallel flows when the body force acts in the direction from the heavier fluid toward the lighter [1]. It has been suggested that an increase in flow vorticity may have a stabilizing influence on the Taylor instability [2]. In studying the hydrodynamic stability of a viscous film on a body in a flow of a low-viscosity fluid [3], the author noted some stabilization of the Taylor instability with increase in Reynolds number, and suggested that cases of complete stabilization of the flow with respect to two-dimensional disturbances are possible with some increase in Reynolds number. In the present investigation, calculations revealed cases in which with increase in Reynolds number the Taylor instability goes over into a Helmholtz instability, which increases with increase in Reynolds number, and also cases in which the Taylor instability completely disappears at some value of the Reynolds number before a Helmholtz instability has developed, i.e., cases of complete stabilization of the flow with respect to two-dimensional disturbances as a result of an increase in Reynolds number.     

Numerical simulation of turbulent free surface flow with two‐equation k–ε eddy‐viscosity models

This paper presents a finite difference technique for solving incompressible turbulent free surface fluid flow problems. The closure of the time‐averaged Navier–Stokes equations is achieved by using the two‐equation eddy‐viscosity model: the high‐Reynolds k–ε (standard) model, with a time scale proposed by Durbin; and a low‐Reynolds number form of the standard k–ε model, similar to that proposed by Yang and Shih. In order to achieve an accurate discretization of the non‐linear terms, a second/third‐order upwinding technique is adopted. The computational method is validated by applying it to the flat plate boundary layer problem and to impinging jet flows. The method is then applied to a turbulent planar jet flow beneath and parallel to a free surface. Computations show that the high‐Reynolds k–ε model yields favourable predictions both of the zero‐pressure‐gradient turbulent boundary layer on a flat plate and jet impingement flows. However, the results using the low‐Reynolds number form of the k–ε model are somewhat unsatisfactory. Copyright © 2004 John Wiley & Sons, Ltd.     

Experimental study on oscillating grid turbulence and free surface fluctuation

This paper analyses the interaction between the turbulence and free surface. The phenomenon takes place in many natural flows and industrial processes. In the present experiments, turbulence is generated by a vertically oscillating grid moving beneath the free surface. Fluid velocity has been measured through a hot-film anemometer, and the free surface elevation has been measured by an ultrasonic sensor. Integral length scales and several turbulence estimators have been computed. In order to detect the generation of turbulence near the free surface, the correlation between free surface elevation and the underneath flow velocity has been studied, as well as the time lag between turbulence and free surface. The free surface dynamics has been characterized by a velocity scale and a length scale. The kinetic energy associated with the free surface fluctuations increases with the Reynolds number at a rate depending on the frequency of the grid movement. For Reynolds number larger than ≈1000, however, the relationships collapse to a single curve characterized by a lower rate. The present experiments do not achieve the inertial sub-range in the vertical velocity fluctuations, and the estimated spectrum decays with an exponent smaller than ?3, which is the typical value for the two-dimensional turbulence in the inertial sub-range. The macro length scale, estimated by using the Taylor’s frozen turbulence hypothesis, experiences a decay away from the grid, which follows reasonably well the profile of Thompson and Turner (J Fluid Mechanics 67: 349–368, 1975). The micro length scale reduces immediately beneath the free surface, which can be interpreted by the increase of dissipation rate in the subsurface layer. The classification diagram by Brocchini and Peregrine (J Fluid Mech 449: 225–254, 2001) indicates that most tests fall in the weak turbulence domain, but some tests fall in the wavy domain. The vertical velocity fluctuations and the free surface level show a significant correlation with a negative phase lag, that is, the free surface fluctuations are ahead of the vertical velocity fluctuations.     

Experimental investigation on a turbulence generation system with high-blockage plates

An experimental study was conducted to develop and characterize systematically a new turbulence generator system to yield large turbulent Reynolds numbers in a compact configuration. The effect of the geometric parameters of two families of high-blockage plates on the resulting turbulent flow field was systematically studied: one series of plates was characterized by the number and distribution of circular openings; a second series had non-circular opening(s) with different shapes, distribution and position of the opening(s). The plates were placed upstream of a contoured contraction and the near field at the centerline of the resulting turbulent free jet was characterized by hot-wire anemometry in terms of mean axial velocity, turbulence intensity, turbulence length scales and corresponding Reynolds numbers. The plate with a central, non-circular opening produced the best compromise of highest turbulence levels along with excellent uniformity in average velocity and turbulence intensity, as evidenced by scan in the transverse direction. It appears to be the most promising one. By comparison with more traditional approaches to turbulence generation, we increased the turbulent Reynolds numbers based on the integral length scale to values on the order of 1000, which was one of the design objectives. Other plate geometries also yielded intense turbulence, but, in some cases, exhibited spurious frequency peaks in their power spectrum. The turbulent generation approach is to be adapted to combustion studies to reproduce conditions typical of practical system in relatively small experimental set-ups that are well-suited for bench-top experiments.