[en] Highlights: • Measuring everywhere does not improve control when limited to one actuator plane. • Actuating everywhere does not improve control when limited to one sensor plane. • The distribution of optimal forcing strongly depends on the control setup. • We employ the eddy-viscosity-enhanced linearized Navier-Stokes equations. • We look at the largest structures in an incompressible turbulent channel flow. We consider linear feedback flow control of the largest scales in an incompressible turbulent channel flow at a friction Reynolds number of $R{e}_{\tau }=2000$. A linear model is formed by linearizing the Navier–Stokes equations about the turbulent mean and augmenting it with an eddy viscosity. Velocity perturbations are then generated by stochastically forcing the linear operator. The objective is to reduce the kinetic energy of these velocity perturbations at the largest scales using body forces. It is shown that a control set-up with a well-placed array of sensors and actuators performs comparably to either measuring the flow everywhere (while limiting actuators to a single wall height) or actuating the flow everywhere (while limiting sensors to a single wall height). In this way, we gain insight (at low computational cost) into how the very large scales of turbulence are most effectively estimated and controlled.

[en] The basis of many turbulence models in computational fluid dynamics is the linear Boussinesq hypothesis that assumes an alignment between the mean strain rate tensor and the Reynolds stress tensor. The validity of this main assumption is analyzed for the test case of an infinite tube bundle with periodic boundary conditions at ${\mathit{Re}}_b\approx 34,800$. This work focuses on the application of five methods based on the Reynolds-averaged Navier–Stokes equations and two scale resolving methods and their ability to accurately reproduce the mean velocity field and the Reynolds stresses. In addition, their capability to predict the anisotropic behavior of the turbulent flow is analyzed. The results indicate that only scale resolving methods are able to predict the turbulent flow field properly for the case under investigation. The results of the large eddy simulation are used to further analyze the distribution of the validity parameter ${\rho }_{\mathit{RS}}$ and the anisotropic turbulence field. Only small areas between the tubes are identified in which alignment occurs and the linear Boussinesq hypothesis is therefore fully valid.

[en] Highlights: • The evolution of a wall-attached plume in a confined box. • The entrainment coefficient of wall plume was reduced to half of that in the free plume. • Plume spreading across the top wall to form a buoyant fluid layer and eventually moving downwards. The evolution of a wall-attached plume in a confined box is studied here with the aid of three dimensional direct numerical simulations (DNS). The plume originates from a local line heat source of length, L, placed at the bottom left corner of the box. The Reynolds number of the wall plume, based on box height and buoyant velocity scale, is ${\mathit{Re}}_H=14530$ and boxes of two different aspect ratios (ratio of box width to height) for a particular value of L are simulated. We observe that the plume develops along the vertical sidewall while remaining attached to it before spreading across the top wall to form a buoyant fluid layer and eventually moving downwards and filling the whole box. The original filling box model of Baines and Turner (1969) is modified to incorporate the wall shear stress, and the results from the DNS are compared against the new model. In modelling plumes, we find that the entrainment coefficient ($\alpha$) for wall-attached plumes is reduced to approximately half of that in the free plume, and the main reason is a diminished contribution of turbulence production to $\alpha$ resulting from a restricted ability of the large-scale eddies to transport momentum. Also, unlike the free plume where away from the source inertial forces balances buoyancy forces, here in our simulations of wall-attached plumes this balance is marginally off, likely due to wall friction. A reasonable agreement is observed between our model and DNS data for the volume and momentum fluxes in the quiescent uniform environment and also for the time-dependent buoyancy profile calculated far away from the plume.

[en] Highlights: • Behavior of round fountains in a cylindrical container is studied experimentally. • Transition of flow behavior and secondary flows increase with Froude number. • Secondary flows change from buoyancy-inertial regime to buoyancy-viscous regime. • Stratification rate increases with Froude number, but decreases with Reynolds number. In this paper, high-speed cameras and flow visualization techniques are used to investigate the behavior of the ‘fountain filling box’ flow resulted from releasing a round fountain in a homogeneous quiescent fluid in a cylindrical container over the ranges of $1.0⩽\mathit{Fr}⩽20.0,102⩽\mathit{Re}⩽1502$, and $27.9⩽\lambda ⩽48.75$, where $\mathit{Fr},\mathit{Re}$ and $\lambda$ are the Froude number, the Reynolds number, and the dimensionless radius of the container, respectively, with $\lambda$ non-dimensionalized by the fountain source radius. The results show the transition of the flow behavior of the fountain and its secondary flows (i.e., the intrusion, reversed flow and stratification) from laminar to turbulent with increasing Fr, and turbulence of the flow strengthened with increasing Re. For intermediate (e.g., $\mathit{Fr}=3.0$) and forced turbulent fountains (e.g., $\mathit{Fr}=5.0$, 8.0, 15.0) with a specific $\lambda$, the non-dimensionalized time-scale for the intrusion front to impinge upon the sidewall, ${\tau }_w$, is nearly constant for $\mathit{Re}\gtrsim 500$. This is because the secondary intrusion flow is dominated by the wall-jet and buoyancy-inertial regimes where the non-dimensionalized intrusion front velocity, $v_i$, is only time-dependent ($v_i~{\tau }^{-1/2}$) or time-dependent but also under the influence of Fr ($v_i~{\mathit{Fr}}^{-1/2}{\tau }^{-1/4}$). However, ${\tau }_w$ for the fountains of $\mathit{Re}\lesssim 204$ is significantly different, which may result from the change in the dominant regime for the intrusion or the interaction between the upflow, downflow of the fountain and the ambient fluids. Furthermore, it is found that the non-dimensionalized quasi-steady development rate of the stratification, $v_s$, increases with Fr, but decreases with Re, since the diffusion effect is suppressed with decreasing Fr or increasing Re.

[en] Highlights: • An improved hybrid immersed boundary (IB)/wall-model LES approach is proposed. • SGS viscosity at the border of IB is modified to match the wall shear stress. • High-Reynolds number turbulent flows are simulated to validate the proposed method. We propose an improved hybrid immersed boundary (IB)/wall-model approach for high-Reynolds-number large-eddy simulations (LES). A preliminary test shows that the existing model that modifies sub-grid viscosity based on the mixing-length model leads to deviation of the wall shear stress when implemented at high Reynolds numbers. To correct the deviation of the wall shear stress at high Reynolds numbers, the sub-grid viscosity at the border of the IB force is reconstructed based on the near-wall balance of shear stress. A correct value for wall shear stress can thus feed back to the border of the IB force and LES solution. The proposed method is tested on several numerical examples, namely a turbulent pipe-flow with friction Reynolds numbers up to $1.0\times {10}^5$, turbulent flow over a circular cylinder with Reynolds numbers up to $1.4\times {10}^5$, turbulent flow over a sphere with Reynolds numbers up to $1.0\times {10}^4$, and turbulent flow over a pitching airfoil with a Reynolds number of $4.8\times {10}^4$. The results provide reasonable predictions compared with previous studies.

[en] Highlights: • The heat transfer of a sweeping jet impinging on concave and convex surfaces is investigated. • Planar particle image velocimetry is conducted to find key flow parameters. • Heat transfer rate is dependent on the magnitude of surface curvature and has a peak in the moderate surface curvature. • Heat transfer performance is correlated with the phase-averaged velocity profile of the wall jet after impingement and the turbulence kinetic energy for both concave and convex surfaces. -- Abstract: The application of an impinging sweeping jet, which oscillates periodically with a large angle, to convective heat transfer has received attention owing to its capability to provide a more spatially uniform and enhanced heat removal rate when compared to a steady jet. Herein, we study how the surface curvature affects the heat transfer performance of a sweeping jet and couple it with the representative flow characteristics. Heat transfer measurement and quantitative flow visualization are conducted experimentally for concave and convex surfaces as well as a flat surface. Whereas concave surfaces have a better heat transfer rate than a flat surface, the enhancement of the heat transfer is relatively small for a convex surface. For both concave and convex surfaces, the Nusselt number does not increase monotonically with the curvature magnitude but has a peak for a moderate curvature. The variation in heat transfer performance with the surface curvature is correlated with the phase-averaged velocity profile of the wall jet deflected after an impingement and the turbulence kinetic energy inside the jet. For both concave and convex surfaces, the wall jet becomes thinner than a flat surface in general, which contributes to improved heat transfer. However, whereas the turbulence kinetic energy is significantly larger for a concave surface of a moderate curvature than that of a flat surface, the turbulence kinetic energy for a convex surface is reduced from that of a flat surface, resulting in degradation of the heat transfer performance.

[en] We studied numerically the heat transfer in flow over a rotationally oscillating cylinder at a subcritical Reynolds number ($Re=1.4\times {10}^5$) that is an order of magnitude higher than previously reported in the literature. This paper is a follow-up of the earlier study of hydrodynamics and drag force in a range of forcing frequencies and amplitudes (Palkin et al., 2018). This time we focus on heat transfer and its correlation with the observed flow field and vortical patterns. Four forcing frequencies $f=f_e/f_0=0,1,2.5,4$ for two forcing amplitudes $\Omega ={\Omega }_{e}D/2U_{\infty }=1$ and 2 are considered, where f₀ is the natural vortex-shedding frequency, U_∞ the free-stream velocity and D the cylinder diameter. The parametric study was performed by solving three-dimensional unsteady Reynolds-averaged Navier–Stokes (URANS) equations closed by a wall-integrated second-moment (Re-stress) model, verified earlier by Large-eddy simulations and experiments in several reference cases including flows over a stagnant, as well as rotary oscillating cylinders at the same Re number. The thermal field, treated as a passive scalar, was obtained from the simultaneous solution of the energy equation, closed by the standard (GGDH) anisotropic eddy-diffusivity model. The computations showed that for the unforced cylinder heat transfer is characterized by very high local rates due to a strong thinning of the thermal boundary layer as a result of the impact and interactions of large coherent structures with the wall. The overall average Nusselt number does not change much for the forced cylinder but its time-averaged, phase-averaged and instantaneous circumferential profiles show some profound differences compared to the stationary cylinder. The distribution of Nu on the back surface becomes more uniform with less frequent occurrence of high values, especially for the higher frequencies $f=2.5$ and $f=4$. This is attributed to diminishing of the mean-recirculation zone as well as to the overall suppression of turbulent fluctuations. The rotary oscillation of the cylinder appears potentially efficient in achieving a more uniform circumferential distribution of Nu and avoiding local overheats and hot spots.

[en] Highlights: • The first study that documents the effects of targeted wall shapes on recovery of turbulent pipe flow. • Insights into how the response and recovery differs based on particular Fourier modes of the wall shape. • One can control the turbulent pipe flow response between first and second order through the implementation of different wall shapes. • Targeted wall shapes induce flow manipulation on boundary layer structures (LSM and VLSM), the recovery from which are observed here on the core flow. The three-dimensional turbulent pipe flow response to targeted wall conditions was numerically examined at a high Reynolds number of $1.58\times 10^5$. The perturbed wall was designed based on azimuthal Fourier modes of $m=3$ (Case I) and $m=15$ (Case II), as well as their superposition at $m\in \left\{3+15\right\}$ (Case III). The mean flow and turbulent fields were investigated and compared, which depicted long-lasting changes in flow properties. A rapid decay of turbulence levels immediately past the perturbation was observed for all cases. However, Case I showed a faster overall recovery rate compared to the higher Fourier Mode cases. While the flow recovered by $19D$ in Case I, the higher Fourier Mode wall shapes (Cases II and III) showed a longer recovery length at $45D$. The flow response also differed between these wall shapes. While Case I depicted a monotonic response, the other two wall shapes exhibited a non-monotonic second-order response characteristic along with a delayed recovery trend. The second-order response of turbulence was also evident in the Reynolds stress variations for cases with a higher Fourier mode wall shapes. The overshoot of Reynolds stresses, accompanied by their rapid decay rate downstream of the modified wall shape, was observed to follow known second-order turbulent response in pipe flow. Dominant flow structures were further identified in the downstream wake of each insert which provided a qualitative description of potential mechanisms behind the observed recovery trends.

[en] Highlights: • Along the centreline, the particle axial velocities decrease as their size decreases. • All particle sizes have a velocity larger than the carrier single-phase. • Particles in polydisperse case decay faster than monodisperse case. • All particles show preferential accumulation along the centreline except the smallest particles. This study presents experimental measurements and computational modelling of particle-laden jets for a wide range of Stokes numbers to analyse the effect of polydispersity on particle volume fraction and velocity. The polydisperse two-phase turbulent jet issues from a pipe into a low velocity coflow, resulting in inlet Stokes numbers, ${\mathit{St}}_{\mathit{in}}$, ranging from 0.003 to 25. The particle velocity and volume fraction are experimentally measured using digital particle image velocimetry and planar nephelometry simultaneously. The simulations are conducted using a newly developed model based on the probability density function of the population balance equation (PDF-PBE) in a large eddy simulation framework. A stochastic Monte Carlo method is adopted to solve the PDF-PBE on an ensemble of notional Lagrangian particles, while the method of Stokes binning is employed to explicitly treat inertial effects in a computationally efficient way. There is a satisfactory agreement between the measurements and simulations. A series of monodisperse simulations were also conducted to compare with the polydisperse flows in a two-way coupled flow. The results confirm that in the monodisperse jet, small particles (with ${\mathit{St}}_{\mathit{in}}\le 0.3$) closely follow the carrier phase velocity whereas in the polydisperse jet they are affected by the presence of larger particles to reduce their axial velocity decay rate. The opposite trend is observed for large particles (${\mathit{St}}_{\mathit{in}}\ge 5.5$), confirming that in the polydisperse jet, the presence of small particles increases the large particles’ radial dispersion that lowers their volume fraction along the centreline compared to the monodisperse jet.

[en] Highlights: • In this work tests are performed for subsonic and choked flow with different NPRs. • Using pitot pressures and schlieren, the effect of crosswire on mixing is studied. • In the uncontrolled case, elliptic jets have better mixing due to asymmetry. • The % reduction of core length for elliptical jet is highest in subsonic jets. • The cross wires improves the mixing characteristics and modifies the shock-cells. The effectiveness of cross wire in controlling the mixing characteristics of a circular and an equivalent elliptic jet is investigated experimentally. While circular jets are conventional, elliptic jets have gained attention due to their better mixing characteristics and faster decay. To further explore and augment the capabilities of elliptic jets for practical utility, it is investigated whether using an elliptic jet with cross wire control gives additional benefit in terms of mixing enhancement over an axisymmetric jet. Experiments are performed for subsonic and choked flow conditions with nozzle pressure ratios ranging from $1.2$ to $7.0$. Time-averaged pitot pressures and schlieren visualization is used for diagnosis. The jet bifurcation can be seen in controlled elliptical jets at all nozzle pressure ratios (NPRs). Core length is reduced to as much as $70\%$ in the elliptical jet and $84\%$ in the case of the circular jet. The core length values estimated from the present data are compared with the previous investigations.