Large eddy simulation of flow around a NACA-0012 airfoil at a Reynolds number of 50,000 has been used to study the behavior of a laminar separation bubble near stall. The effects of the subgrid-scale model and explicit filtering were studied for a test case in which direct numerical simulation results were available. It was found that a method incorporating a mixed-time-scale model in addition to explicit filtering gave improved results compared with a method with filtering alone. Statistical results as well as snapshots of the flow below stall exhibit good agreement with the direct numerical simulations. For a configuration near stall, the effect of the spanwise domain width was investigated by increasing the spanwise length from 20 to 50% chord. Two-point velocity correlations showed a significant improvement for the wider computational domain, in which the simulation was able to capture a lowfrequency flow oscillation, in which intermittent bursting of the bubble was observed. The bubble bursting observed here is more irregular than in experiments at higher Reynolds number. The amplitude and frequency are compared with experimental results and with an unsteady viscous-inviscid interaction method which is shown to be capable of capturing unsteady behavior during stall.inviscid fluxes vector f = frequency of the flow oscillation G = filtered viscous fluxes vector H = shape factor H = L z = domain length in spanwise direction M = reference Mach number Pr = Prandtl number p = filtered pressure Q = conservative flow variables vector or second invariant of the velocity gradient tensor Q = filtered conservative flow variables vector q = flow variable q = filtered flow variablẽ q = Favre filtered flow variable q SGS = velocity scale for unresolved flow q 0 = small scale of the flow variable Re = Reynolds number based on the freestream conditions and chord length St = Strouhal number St fc sin =U S ij = strain rate tensor T= filtered temperature U = freestream velocity u = friction velocity t = eddy viscosity = angle of attack incidence = ratio of specific heats = filter size ij = Kronecker delta = displacement thickness = momentum thickness = dynamic viscosity = fluid kinematic viscosity = filtered fluid density ij = subgrid scale stress tensor kk = isotropic part of the subgrid scale Reynolds stress tensor w = shear wall stress h i = averaging over z

No abstract

This paper reports the effect of setting density on flow uniformity, pressure drop, pumping power, and convective heat transfer coefficients (CHTCs). High-density setting (HDS) comprises 768 bricks, and low-density setting (LDS) contains 512 bricks are tested for different inlet air velocities using both local and average approaches. The investigation is carried out using a 3D-computational fluid dynamics (CFD) model with k–ω turbulence model. Both settings are validated against experimental data reported in the literature with errors less than 1.9% for pressure drop and −1.0% for brick surface temperature. The reported results indicated that the LDS has distinct benefits over the HDS as it enhances the flow uniformity, particularly in the stack channels. Also, LDS attains lower pressure drop, pumping power, and firing time than HDS by 45.93%, 50%, and 35%, respectively. In addition, LDS produces larger CHTCs, rates of heat transfer for individual bricks, and the ratio of heat transfer to pumping power than HDS by 24.53%, 35%, and 34%, respectively. Moreover, LDS produces more homogenous heating of the setting bricks than HDS as the maximum difference of CHTCs between bricks is about 4.39% for LDS and 19.62% for HDS. The best performance of the firing process is accomplished at low inlet air velocity (3 m/s), whereas the highest productivity is achieved at high inlet air velocity (9 m/s).

This paper presents the effect of brick roughness and surface roughness of tunnel walls, ceiling, and floor on the fluid flow, pressure drop, and convection and radiation heat transfer in tunnel kilns. Surface roughness values of 0-4mm are investigated for bricks and tunnel boundary. Moreover, another wall roughness of 10mm is considered to explore the effect of major defects in the tunnel boundary. The study is conducted using a three-dimensional CFD-model based on the finite volume method with k-ω turbulence model. The convection heat transfer coefficients enhance by 45 and 97%, and the pressure drop increase by 25.1 and 80.4% as the brick roughness is increased from 0-1mm and 0-4mm, respectively. The ratio of the rate of heat transfer to pumping power reaches its maximum at brick roughness of 2mm. These results provide important knowledge about the acceptable range of brick roughness for manufacturers. The limited increase in heat transfer rates (1.34-3.88%) and pressure drops (7.5-18.2%) are experienced for the tested tunnel boundary roughness (1-10mm). These results are supportive of scheduling the maintenance of tunnel kilns’ interior structure. Moreover, the enhancement of the radiation heat transfer depends on the brick emissivity and the area ratio of rough to smooth surfaces.

No abstract