Myung Kyoon Chung

A KAPPA-EPSILON-GAMMA EQUATION TURBULENCE MODEL

By considering the entrainment effect on the intermittency in the free boundary of shear layers, a set of turbulence model equations for the turbulent kinetic energy k , the dissipation rate e, and the intermittency factor γ is proposed. This enables us to incorporate explicitly the intermittency effect in the conventional K –e turbulence model equations. The eddy viscosity ν t is estimated by a function of K , e and γ. In contrast to the closure schemes of previous intermittency modelling which employ conditional zone averaged moments, the present model equations are based on the conventional Reynolds averaged moments. This method is more economical in the sense that it halves the number of partial differential equations to be solved. The proposed K –e–γ model has been applied to compute a plane jet, a round jet, a plane far wake and a plane mixing layer. The computational results of the model show considerable improvement over previous models for all these shear flows. In particular, the spreading rate, the centreline mean velocity and the profiles of Reynolds stresses and turbulent kinetic energy are calculated with significantly improved accuracy.

Approximation of turbulent conditional averages by stochastic estimation

Conditional averages of turbulent flow quantities can be approximated in terms of unconditional correlation data by means of stochastic estimation. The validity and accuracy of this procedure are investigated by comparing stochastic estimates to conditional averages measured in four turbulent flows: grid turbulence, the axisymmetric shear layer of a round jet, a plane shear layer, and pipe flow. Comparisons are made for quantities that are separated from the conditional data in time or space, and for turbulent pressures, as well as turbulent velocities. In each case, the linear estimate accurately represents large scale structure. Nonlinear quadratic estimation shows little improvement over linear estimation, because the second‐order terms are small for probable values of the turbulent fluctuations.

Convective mass transfer from a square cylinder and its base plate

Abstract The mass transfer from a square cylinder and from the plate on which the cylinder is mounted vertically, is investigated with the naphthalene sublimation technique. The general pattern of local mass transfer is somewhat different from that with a circular cylinder. A comparison with heat transfer measurements on a square cylinder in the two-dimensional flow region, using the heat/mass transfer analogy, shows good agreement in average transfer rates, but slight differences in local values. A dramatic change of mass transfer rates is found both on the cylinder and the base plate around the plate-cylinder junction region due to a horseshoe vortex system. Multiple vortices, which include the primary horseshoe vortex, the corner vortex and two pairs of counterrotating vortices, affect the mass transfer process. Variation of either Reynolds number or initial boundary layer thickness does not significantly change the location of the peaks created by the horseshoe vortex system, but only affects the magnitude of the local mass transfer rate. Visualization of the surface flow on the base plate is included to supplement the mass transfer measurements.

Analysis of heat transfer in a pipe carrying two-phase gas-particle suspension

Abstract A ‘two-fluid model’ using the thermal eddy diffusivity concept and Lumleys drag reduction theory, is proposed to analyse heat transfer of the turbulent dilute gas-particle flow in a vertical pipe with constant wall heat flux. The thermal eddy diffusivity model is derived to be a function of the ratio of the heat capacity-density products \ grC p of the gaseous phase and the particulate phase and also of the ratio of the thermal relaxation time scale to that of turbulence. Lumleys theory is applied to find the variation of the viscous sublayer thickness depending on the particle loading ratio Z and the relative particle size dp/D. At low loading ratio, the size of the viscous sublayer thickness is important for suspension heat transfer, while at higher loading, the effect of the ratio \ gr P C p P / \ gr f C pf . is dominant. The major cause of decrease in the suspension Nusselt number at low loading ratio is found to be due to the increase of the viscous sublayer thickness caused by the suppression of turbulence near the wall by the presence of solid particles. Predicted Nusselt numbers using the present model are in satisfactory agreement with available experimental data both in the pipe entrance and the fully developed regions.

A modified theory for the flow mechanism in a regenerative flow pump

Abstract The regenerative flow pump (RFP) and regenerative flow compressor (RFC) are turbomachines capable of developing high pressure ratios in a single stage. They are also known by other names, such as peripheral, side channel, turbine, traction and vortex compressor/pump. Even though the efficiency of RFP/RFC is usually less than 50 per cent based on past design experience, they have found wide applications in automotive and aerospace fuel pumping, booster systems, water supply, agricultural industries, shipping and mining, chemical and food stuffs industries, and regulation of lubrication and filtering. RFCs have been proposed for use in hydrogen gas pipelines and as helium compressors for cryogenic applications in space vehicles. RFTs are used as accessory drives on aircraft and missiles. With the aim of improving the performance and efficiency of an RFP, this paper proposes an improved and modified theoretical model that can explain the change in the circulatory velocity caused by variation in channel area. All previous works concentrated on the fully developed flow region in the RFP and this work expands consideration to the developing region. Furthermore, in order to make the above-suggested model a closed problem, several loss models were assumed and the results of predictions were compared with experimental and CFD data.

Study on Hydrodynamic Torque of a Butterfly Valve

Since knowledge on hydrodynamic torque of a butterfly valve is very important for butterfly valve design, its hydrodynamic torque is investigated theoretically. For this, a recently developed two-dimensional butterfly valve model is solved through the free-streamline theory with a newly devised iterative scheme and the resulting two-and three-dimensional torque coefficients are compared with previous theoretical results based on the conventional butterfly valve model and experiments. Comparison shows that the improvement due to the new butterfly valve model is marginal. That is, the three-dimensional torque coefficient is well represented by the new model. Otherwise, the two-dimensional torque coefficient is well predicted by the conventional model. In spite this fact, the present results can be used in further researches on butterfly valves because the improved butterfly valve model is mathematically correct and reflects physical reality more correctly than the conventional valve model.

A nonlinear return‐to‐isotropy model with Reynolds number and anisotropy dependency

A new computational model for the return to isotropy is presented. In order to reproduce the significant role of the third invariant IIIb(=bijbjkbki) of the Reynolds stress anisotropy bij[=uiuj/(2k)−(1/3)δij] in the return‐to‐isotropy process, a nonlinear return‐to‐isotropy model is formulated by a Taylor series expansion up to fifth power of bij. Then the strong realizability condition for non‐negativity of the component energies is utilized to reduce the number of model constants produced. Correction for the low Reynolds number effect is then included by investigating an energy‐weighted average time scale of eddies over the three‐dimensional energy spectrum. Superiority of the proposed model performance is exemplified by a number of test computations of homogeneous relaxing turbulence in a wide range of turbulence Reynolds number and IIIb.

Effects of angle of attack on mass transfer from a square cylinder and its base plate

Abstract A naphthalene sublimation technique is employed to investigate the convective mass transfer process from a square cylinder and its base plate in a flow of air. Distributions of local mass transfer coefficients on each face of the cylinder change dramatically with the angle of attack. The average Sherwood number has a minimum value at α = 12–13°, and a maximum value at α=20–25° where a is the angle of attack. A comparison of the present mass transfer measurement with earlier heat transfer measurements, using the heat/mass transfer analogy, shows good agreement at average transfer rates, but notable differences in local values, especially near the reattachment point. A remarkable enhancement of mass transfer due to the horseshoe vortex system is observed on the square cylinder and on the base plate near the plate-cylinder junction. As the angle of attack increases, the mass transfer peak created by counterrotating vortices diminishes. On the other hand, the influence of the corner vortex remains strong regardless of the angle of attack. Surface flow on the base plate is visualized using the oil-lampblack technique. Streaks of the corner vortex and the counterrotating vortices in the visualization picture compare well with the trails of the peak of mass transfer created by these vortices.

Curvature-dependent two-equation model for prediction of turbulent recirculating flows

Amelioration de la precision des previsions en incorporant un modele dependant de la courbure dans le modele k-e

ANALYSIS OF TURBULENT GAS-SOLID SUSPENSION FLOW IN A PIPE.

The mixing length theory is extended to close the relevant momentum equations for two-phase turbulent flow at a first-order closure level. It is assumed that the mass fraction of the particles is on the order of unity, that the particle size is so small that the particles are fully suspended in the primary fluid, and that the relaxation time scale of the particles is sufficiently small compared with the time scale of the energy containing eddies so that the suspended particles are fully responsive to the fluctuating turbulent field. Bulk motion of the particles is treated as a secondary fluid flow with its own virtual viscosity. The proposed closure is applied to a fully developed gas-solid pipe flow in which the particles are assumed to be uniformly distributed across the pipe section. Predicted velocity profiles and the friction factors are in good agreement with available experimental data.