Reynolds‐stress modelling of M = 2.25 shock‐wave/turbulent boundary‐layer interaction

M = 2.25 shock‐wave/turbulent‐boundary‐layer interactions over a compression ramp for several angles (8, 13 and 18°) at Reynolds‐number Re=7 × 10³ were simulated with three low‐Reynolds second‐moment closures and a linear low‐Reynolds standard k–ε model. A detailed assessment of the turbulence closures by comparison with both mean‐flow and turbulent experimental quantities is presented. The Reynolds‐stress model which is wall‐topology free and which uses an optimized redistribution closure, is in good agreement with experimental data both for wall‐pressure and mean‐velocity profiles. Detailed analysis of three components of the Reynolds‐stress tensor (comparison with measurements and transport‐equation budgets) provides a critical evaluation of full Reynolds‐stress models for the separated supersonic compression ramp. The discrepancy observed in the shock‐wave foot region, between computations and measurements for the Reynolds‐stresses profiles, could be explained by considering the experimental shock‐wave oscillation and directions for future modelling work are indicated. Copyright © 2007 John Wiley & Sons, Ltd.     

Effect of Partial Closure of the Channel on the Pressure Pulse in a Shock Wave Emerging from a Duct

The temporal-spatial development of the flow pattern behind a shock wave emerging from open-ended and half-closed ducts at M₀=1.15 to 3.0 Mach number is investigated both experimentally and numerically and the action of the diffracted wave on a barrier placed at different distances from the duct exit is studied. Flow toeplerograms are obtained and the pressure on the barrier is measured. The distinctive features of the interaction between the shock wave and a barrier mounted perpendicular to the duct axis are established. It is found that partial closing of the duct results in a decrease in the barrier pressure in the case of a strong shock (M₀ > 2.2) and leads to a pressure decrease when a weak shock is diffracted (M₀=1.1 to 1.7). A dependence characterizing the dynamic action of the shock wave on the barrier and specifying the threshold value of a combination of the shock Mach number and the distance from the barrier determining whether the pressure pulse on the barrier increases or decreases, is obtained.     

Dual-element directional shock wave attenuators

This study explores the interaction between shock waves and dual-element porous plates used to ameliorate the hazardous effects of these waves. Tests were performed in a shock tube to determine the effects that a pair of porous plates with directional resistance properties had on the initial peak pressure and impulse amelioration experienced by an end wall. Mild steel test specimens, ranging in porosity values from 6.6% to 41.1%, were mounted two at a time at different spacings in the test section. Each plate had directional properties, i.e. resistance to flow was different for flow coming from either side. Four plates were used, and 48 plate configurations were tested. Side wall and end wall pressure measurements and schlieren photographs were taken of the interactions. Tests were run at Mach numbers of 1.23, 1.35 and 1.42. The separation distances between the plate specimens were varied between 30 mm and 60 mm; however, the distance between the downstream plate and the end wall was kept constant at 140 mm for all tests.Both the initial peak pressure and impulse amelioration values were found to be dependant on the plate combination porosity. As the porosity of the combination increased, the amelioration values decreased. Complementary plate combinations produced differing results as different wave interactions occur when plate positions were interchanged. The porosity of the combined plates was found to have an overriding influence on the end wall peak pressure and impulse amelioration values when compared to the effect that plate arrangement (i.e. geometrical influences) had. Impulse amelioration values were found to increase as the separation distance between plates were increased. The amplitude of the end wall pressure trace was found to increase as the incident Mach number was increased. The significant attenuation of the incident shock wave obtained during this study is attributed to the system of multiple reflected and transmitted waves that are produced by the presence of the plate specimens in series. This increases the frequency of shock wave and barrier interactions, when compared to just using a single barrier.     

Interaction theory of hypersonic laminar near-wake flow behind an adiabatic circular cylinder

The separation and shock wave formation on the aft-body of a hypersonic adiabatic circular cylinder were studied numerically using the open source software OpenFOAM. The simulations of laminar flow were performed over a range of Reynolds numbers (\(8\times 10^3 < Re < 8\times 10^4\)) at a free-stream Mach number of 5.9. Off-body viscous forces were isolated by controlling the wall boundary condition. It was observed that the off-body viscous forces play a dominant role compared to the boundary layer in displacement of the interaction onset in response to a change in Reynolds number. A modified free-interaction equation and correlation parameter has been presented which accounts for wall curvature effects on the interaction. The free-interaction equation was manipulated to isolate the contribution of the viscous–inviscid interaction to the overall pressure rise and shock formation. Using these equations coupled with high-quality simulation data, the underlying mechanisms resulting in Reynolds number dependence of the lip-shock formation were investigated. A constant value for the interaction parameter representing the part of the pressure rise due to viscous–inviscid interaction has been observed at separation over a wide range of Reynolds numbers. The effect of curvature has been shown to be the primary contributor to the Reynolds number dependence of the free-interaction mechanism at separation. The observations in this work have been discussed here to create a thorough analysis of the Reynolds number-dependent nature of the lip-shock.     

Attenuation of shock waves propagating in polyurethane foams

Shock wave attenuation in polyurethane foams is investigated experimentally and numerically. This study is a part of research project regarding shock propagation in polyurethane foams with high-porosities = 0.951 ~ 0.977 and low densities of ρc = 27.6 ~55.8 kg/m³. Sixty Millimeter long cylindrical foams with various cell numbers and foam insertion condition were installed in a horizontal shock tube of 50 mm i.d. and 5.4 mm in length. Results of pressure measurements in air/foam combination are compared with CFD simulation solving the one-dimensional Euler equations. In the case of a foam B fixed on shock tube wall, pressures at the shock tube end wall increases relatively slowly comparing to non-fixed foam, free to move and a foam A fixed on shock tube wall. This implies that elastic inertia hardly contributes to pressure build up. Pressures behind a foam C fixed on shock tube wall decrease indicating that shock wave is degenerated into compression wave. Dimensionless impulse and attenuation factor decrease as the initial cell number increases. The momentum loss varies depending on cell structure and cell number.     

Shock wave attenuation by grids and orifice plates

The interaction of weak shock waves with porous barriers of different geometries and porosities is examined. Installing a barrier inside the shock tube test section will cause the development of the following wave pattern upon a head-on collision between the incident shock wave and the barrier: a reflected shock from the barrier and a transmitted shock propagating towards the shock tube end wall. Once the transmitted shock wave reaches the end wall it is reflected back towards the barrier. This is the beginning of multiple reflections between the barrier and the end wall. This full cycle of shock reflections/interactions resulting from the incident shock wave collision with the barrier can be studied in a single shock tube test. A one-dimensional (1D), inviscid flow model was proposed for simulating the flow resulting from the initial collision of the incident shock wave with the barrier. Fairly good agreement is found between experimental findings and simulations based on a 1D flow model. Based on obtained numerical and experimental findings an optimal design procedure for shock wave attenuator is suggested. The suggested attenuator may ensure the safety of the shelter’s ventilation systems.     

One-dimensional model for microscale shock tube flow

A one-dimensional model for the numerical simulation of transport effects in small-scale, i.e., low Reynolds number, shock tubes is presented. The conservation equations have been integrated in the lateral directions and three-dimensional effects have been introduced as carefully controlled sources of mass, momentum and energy, into the axial conservation equations. The unsteady flow of gas behind the shock wave is reduced to a quasi-steady flow by choosing a coordinate system attached to the shock. The boundary layer problem is thereby reduced to a laminar solution, similar to the Blasius solution, with the exception that the wall velocity can be nonzero. The resulting one-dimensional equations are then solved numerically using a two-step Lax-Wendroff/ MacCormack scheme with flux correction transport. For validation purposes, comparisons are performed against previously published shock structure and low Reynolds number shock tube experiments; good agreement is observed. The model has been used to predict the performance of a 10μm shock tube and the result of this simulation shows the possibility of shock wave disappearance at lower pressure ratios for a micro-scale shock tube.      

Hydrodynamic and thermal fields analysis in gas-solid two-phase flow

The present work aims to investigate numerically the flowfield and heat transfer process in gas-solid suspension in a vertical pneumatic conveying pipe. The Eulerian-Lagrangian model is used to simulate the flow of the two-phases. The gas phase is simulated based on Reynolds Average Navier-Stokes equations (RANS) with low Reynolds number k-ε model, while particle tracking procedure is used for the solid phase. An anisotropic model is used to calculate the Reynolds stresses and the turbulent Prandtl number is calculated as a function of the turbulent viscosity. The model takes into account the lift and drag forces and the effect of particle rotation as well as the particles dispersion by turbulence effect. The effects of inter-particles collisions and turbulence modulation by the solid particles, i.e. four-way coupling, are also included in the model. Comparisons between different models for turbulence modulation with experimental data are carried out to select the best model. The model is validated against published experimental data for velocities of the two phases, turbulence intensity, solids concentration, pressure drop, heat transfer rates and Nusselt number distribution. The comparisons indicate that the present model is able to predict the complex interaction between the two phases in non-isothermal gas-solid flow in the tested range. The results indicate that the particle-particle collision, turbulence dispersion and lift force play a key role in the concentration distribution. In addition, the heat transfer rate increases as the mass loading ratio increases and Nusselt number increases as the pipe diameter increases.     

Effects of inlet guide-vane number on flowfields in a side-dump combustor

The flowfields of a side-dump combustor with various number of side-inlet guide-vane are measured using laser-Doppler velocimetry (LDV). The Reynolds number based on the bulk mean velocity and combustor diameter was 2.6×10⁴. Quantities such as mean velocity, turbulence intensity, turbulent kinetic energy, vorticity, friction factor, and wall static pressure oscillation are used to characterize the fluid flow. In the dome region of the inlet-jet plane, there is one pair of counter-rotating vortices for the no-vane, one-vane, and two-vane cases and two pairs of counter-rotating vortices for the three-vane case, respectively. This trend is reversed in the impinging plane. The combustor flowfield downstream of the X_c^*=2.5 station is found to be insensitive to the variation of inlet guide-vane number. In addition, the guide-vane number which provides the least pressure loss and the lowest pressure oscillation is identified for the first time. Based on the presented data, a better guide-vane number for practical reference is suggested.     

A Semi-empirical Correlation for Pressure Drop in Packed Beds of Spherical Particles

The effect of a confining wall on the pressure drop of fluid flow through packed beds of spherical particles with small bed-to-particle diameter ratios was investigated to develop an improved pressure drop correlation. The dependency of pressure loss on both wall friction and increased porosity near the wall was accounted for by using a theoretical approach. A semi-empirical model was created based upon the capillary-orifice model, which included a wall correction factor for the inertial pressure loss. In this model, packed beds were treated as a bundle of capillary tubes whose orifice diameter in the core region was different from that of the wall region. Using this model, a new pressure drop correlation was obtained, based on the Ergun equation and applicable for a wide range of Reynolds numbers (10⁻²–10³). The proposed correlation was compared with previous correlations, as well as with experimental data. This correlation showed close agreement with the experimental data for both low- and high-Reynolds number regimes and for a wide range of bed-to-particle diameter ratios. The ratio of the pressure drop in finite packing to that in homogeneous packing was then calculated. This ratio clearly shows how the wall effect depends on the Reynolds number and the bed-to-particle diameter ratio.     

Investigation of the flowfield immediately upstream of a hot film probe

The laminar flowfield in a rectangular channel immediately upstream of a hot film gradient probe with two parallel films was investigated in the range of Reynolds number Re _pr= 6 to 95, with the Reynolds number based on the probe diameter and the local flow velocity. For this study a photochromic dye flow visualization technique was used. The results show that the smaller the Reynolds number Re _prthe larger the influence of the probe is upon the flowfield. No distinct influence of the probe location relative to the channel walls on the flow deceleration process immediately upstream of the probe was observed.List of symbols a distance between the hot films - d _h hydraulic diameter - d _pr diameter of the probe body - Reynolds number based on hydraulic diameter and mean flow velocity - Reynolds number based on probe diameter and the undisturbed flow velocity at the centerline of probe - u flow velocity in x-direction - u ₀ undisturbed velocity in the center of the channel - undisturbed mean flow velocity - u(x,y) velocity at position (x,y) - averaged velocity gradient - x coordinate in main flow direction - y coordinate normal to the larger wall of the rectangular channel - z coordinate normal to x and y - v kinematic viscosity     

Heat transfer of gas-particle flow in a supersonic convergent-divergent nozzle

Summary Heat flux, wall heat transfer coefficients, and wall pressures are determined for high velocity flow of gas-solid mixtures in a converging-diverging nozzle. Flow separation accompanied with oblique shock formation occurs in the diverging section of the nozzle. The shock strength is reduced upon the addition of solid particles. The wall pressure in the convergent section of the nozzle appears unaffected by the presence of solid particles. In the divergent section, however, the wall pressure is slightly lowered. At the maximum ratio of solid to air flow used in the experiments (3.7) increases in the heat transfer rate of up to 20 and 50 percent are obtained in the convergent and separated (divergent) regions of the nozzle, respectively. Slightly larger increases in the wall heat transfer coefficients are also obtained. It is concluded that the wall heat flux and heat transfer coefficients are influenced strongly by the presence of disturbances upstream of the nozzle inlet.Nomenclature W _a air flow rate - W _s solids flow rate - x axial distance from nozzle entrance - L axial length of nozzle - specific heat ratio of fluid - A _e exit cross section of flow - A _* throat cross section of flow - P ₀ inlet pressure - P _s wall separation pressure - P _a ambient exhaust pressure - shock wave angle - shock wave deflection angle - M ₁ Mach number upstream of shock wave - Mach number normal to shock wave - q heat flux - k _f thermal conductivity of fluid - T _wi inside wall temperature - T _wo outside wall temperature - T _ad adiabatic wall temperature - h wall heat transfer coefficient - C nozzle constant - A local cross section of flow - c _p specific heat of fluid - Pr Prandtl number - viscosity of fluid - r _c throat radius of curvature - factor accounting for variation of and Units absolute temperature °R(ankine) °F+459.7 - conductivity 1 BTU (hr ft °F)^–1 4.137×10^–3 cal (s cm °C)^–1 - specific heat 1 BTU (1b °F)^–1 1 cal (g °C)^–1 - absolute pressure 1 psia 0.0680 atm Supported in part by aid provided by the UCLA Space Science Center (Grant NsG 236-62 Libby).Listed for readers not familiar with the units adopted in this paper (editor).     

Quadratic programming algorithm for wall slip and free boundary pressure condition

Wall slip is often observed in a highly sheared fluid film in a solid gap. This makes a difficulty in mathematical analysis for the hydrodynamic effect because fluid velocity at the liquid–solid interfaces is not known a priori. If the gap has a convergent–divergent wedge, a free boundary pressure condition, i.e. Reynolds pressure boundary condition, is usually used in the outlet zone in numerical solution. This paper, based on finite element method and parametric quadratic programming technique, gives a numerical solution technique for a coupled boundary non‐linearity of wall slip and free boundary pressure condition. It is found that the numerical error decreases with the number of elements in a negative power law having an index larger than 2. Our method does not need an iterative process and can simultaneously gives rise to fluid film pressure distribution, wall slip velocity and surface shear stress. Wall slip always decreases the hydrodynamic pressure. Large wall slip even causes a null hydrodynamic pressure in a pure sliding solid gap. Copyright © 2005 John Wiley & Sons, Ltd.     

Experimental study of planar shock wave interactions with dense packed sand wall

A dense packed sand wall is impacted by a planar shock wave in a horizontal shock tube to study the shock-sand wall interaction. The incident shock Mach number ranges from 2.18 to 2.38. A novel device for actively rupturing diaphragm is designed for the driver section of the shock tube. An apparatus for loading particles is machined by the electrical discharge cutting technique to create a dense packed particle wall. High-speed schlieren imaging system and synchronized pressure measurement system are used together to capture the wave structures and particle cloud velocity. The dynamic evolution model from dense packed particles to dense gas–solid cloud at the initial driving stage is established. The blockage and permeation effects of the sand wall work together and influence each other. The high pressure gas behind the incident shock wave blocked by the sand wall pushes the upstream front of the wall forward like a piston. Meanwhile, the high speed gas permeating through the sand wall drags the sands of the most downstream layer forward. The incident shock strength, initial sand wall thickness and particle diameter are varied respectively to investigate the shock attenuation and the wall acceleration. Increasing the sands diameter or mixing in small diameter sands can significantly attenuate the incident shock. The smaller particles or the particles in thinner wall can be dispersed into a larger range in the process of transform from dense packed particles to dense gas–solid cloud. Moreover, the stronger incident shock can disperse the particles into a larger region.     

Shock train generated turbulence inside a nozzle with a small opening angle

The flow inside an over expanding rectangular nozzle with a small opening angle of 1.6° is investigated by means of high-speed schlieren and shadowgraph photography, pressure probes and hot-wire anemometry on the nozzle centre line in order to measure the turbulent fluctuations generated by the occurring shock wave/boundary layer interaction. Additionally, an optical shock capturing tool is deployed to measure the amplitude and frequency of the shock train oscillation. Varying the back pressure, the pre-shock Mach number is changed between Ma ₁ = 1.1 and 2.1. Two different modes of turbulence generation and distribution are detected. For a single normal shock and a normal shock train, the normal Reynolds stress [`(u^￠2 )] \overline{{u^{\prime 2} }} on the channel axis is only slightly increased compared to the free stream value, whereas for the cases of a lambda foot shock train and an x-type shock train, a strong intensification by forming a turbulent mixing zone can be observed.     

Penetration of Cylindrical Projectiles into Saturated Sandy Media

The time and depth of vertical one-dimensional projectile penetration into sandy media in the near shore region are derived. A precise definition for the physical properties and for the behavior of the sandy medium following the projectile impact are evaluated. Three separate time intervals following projectile impact are identified. During the first 3 ms of penetration, the deviatoric friction stress is shown to be negligible and the integrated Mie–Grüneisen equation of state (or, equivalently, the Hugoniot-adiabat) may be applied to compute the normal penetration resistance force from the sand pressure. In order to compute sand pressure as a function of the sand density D by the integrated Mie–Grüneisen equation of state, the Mie–Grüneisen dimensionless constants γ₀ and s and the dimensional speed of sound C ₀ in the sandy medium are required. In order to illustrate the one-dimensional shock wave propagation in both wet and dry sands, Hugoniot data for wet and dry silica sands are evaluated by a three degrees of freedom algorithm to compute these required constants. The numerical results demonstrate that the amplitude of the shock wave pressure in the wet silica sand (41% porosity) is approximately one-third of the shock wave pressure amplitudes in the dry silica sands (22% and 41% porosity). In addition, the shock wave pressure dampens quicker in the wet sand than in the dry sands.     

高超声速稀薄流中横向喷流干扰特性实验研究

喷流干扰是高超声速飞行高精度控制的一种有效手段,研究者们以往大部分都主要集中于连续流条件下喷流干扰效应的机理研究,并给出了喷流干扰流场的典型结构,而稀薄流条件下喷流干扰特性的实验数据还十分匮乏.本文利用JFX爆轰激波风洞产生高超声速稀薄自由流,基于平板模型开展不同喷流压力和自由来流参数对横向喷流干扰特性影响的实验研究,采用高速纹影成像及图像处理技术,获得稀薄流条件下喷流干扰流场演化过程及流场结构的变化规律.相比于无喷流条件形成的流场,横向喷流与稀薄自由流相互作用形成的流场结构更为复杂,喷流压力由于受到稀薄来流的扰动,斜激波会短暂穿透喷流干扰流场并延伸至楔形体上部.喷流干扰流场内桶状激波的影响范围随着喷流压力的升高而逐渐变宽,位于三波点上游的斜激波空间位置不会随喷流压力的变化而改变,而位于三波点下游的弓形激波则向上游移动,当喷流压力过低时,桶状激波不会与其他两种激波交汇形成三波点.高超声速稀薄来流压力的降低同样会使桶状激波的影响范围变宽,弓形激波同样也会向上游移动,但基本不会对斜激波空间位置产生任何影响.     

Some Reynolds number effects on two- and three-dimensional turbulent boundary layers

Experimental data for a two-dimensional (2-D) turbulent boundary layer (TBL) flow and a three-dimensional (3-D) pressure-driven TBL flow outside of a wing/body junction were obtained for an approach Reynolds number based on momentum thickness of Re _θ =23,200. The wing shape had a 3:2 elliptical nose, NACA 0020 profiled tail, and was mounted on a flat wall. Some Reynolds number effects are examined using fine spatial resolution (Δy ⁺=1.8) three-velocity-component laser-Doppler velocimeter measurements of mean velocities and Reynolds stresses at nine stations for Re _θ =23,200 and previously reported data for a much thinner boundary layer at Re _θ =5,940 for the same wing shape. In the 3-D boundary layers, while the stress profiles vary considerably along the flow due to deceleration, acceleration, and skewing, profiles of the parameter correlate well and over available Reynolds numbers. The measured static pressure variations on the flat wall are similar for the two Reynolds numbers, so the vorticity flux and the measured mean velocities scaled on wall variables agree closely near the wall. The stresses vary similarly for both cases, but with higher values in the outer region of the higher Re _θ case. The outer layer turbulence in the thicker high Reynolds number case behaves similarly to a rapid distortion of the flow, since stream-wise vortical effects from the wall have not diffused completely through the boundary layer at all measurement stations. Received: 9 June 2000/Accepted: 26 January 2001     

Experimental study of the shock–bubble interaction with reshock

The interaction of a planar shock wave with a spherical density inhomogeneity is studied experimentally under reshock conditions. Reshock occurs when the incident shock wave, which has already accelerated the spherical bubble, reflects off the tube end wall and reaccelerates the inhomogeneity for a second time. These experiments are performed at the Wisconsin Shock Tube Laboratory, in a 9m-long vertical shock tube with a large square cross section (25.4×25.4 cm²). The bubble is prepared on a pneumatically retracted injector and released into a state of free fall. Planar diagnostic methods are used to study the bubble morphology after reshock. Data are presented for experiments involving two Atwood numbers (A = 0.17 and 0.68) and three Mach numbers (1.35 < M < 2.33). For the low Atwood number case, a secondary vortex ring appears immediately after reshock which is not observed for the larger Atwood number. The post-reshock vortex velocity is shown to be proportional to the incident Mach number, M, the initial Atwood number, A, and the incident shock wave speed, W _i.