A numerical investigation of a single highly confined bubble moving through a millimeter-scale channel in the absence of phase change is presented. The simulation includes thermal boundary conditions designed to match those of completed experiments. The channel is horizontal with a uniform-heat-generation upper wall and an adiabatic lower boundary condition. The use of a Lagrangian framework allows for the simulation of a channel of arbitrary length using a limited computational domain. The liquid phase is a low-Reynolds-number unsteady laminar flow, and the phase interactions are modeled using the Volume-of-Fluid method with full geometric reconstruction of the liquid/gas interface. Results are presented for two bubble sizes, two liquid flow rates, and two Prandtl numbers. The paper focuses on heat transfer in the rearward wake of the bubble. Nusselt numbers for the higher Prandtl number case are shown to follow a power law relationship with distance behind the bubble. Important dynamical structures include a pair of vortical structures at the rear of the bubble associated with cold fluid being brought near the wall and fluid jets oriented in the transverse direction to either side of the bubble.

In the wake of utilization of novel materials in various thermal applications open cell metal foams have received attention due to their inherent properties such as large surface area to volume ratio and higher thermal conductivity. Additionally, complex tetradecahedron structure promotes mixing and makes them a good candidate for heat transfer applications. In this paper, a relative comparison has been made between the thermal-hydraulic performance of aluminum and copper metal foam heat exchangers with the same geometry under dry and wet operating conditions. Heat exchanger consisting of round tube with annular layer of metal foam have been considered. Experiments have been conducted using a closed-loop wind tunnel to measure the heat transfer performance and pressure drop. The impact of base metal (aluminum and copper) on the heat transfer rate has been evaluated at varying air flow rates and upstream relative humidity. It has been found that due to similar geometry (flow depth, face area, pore size) both aluminum and copper foam samples have comparable pressure drop under dry conditions. However, the pressure gradient was noticeably different for two samples under wet operating conditions. An obvious difference in heat transfer rate for aluminum and copper metal foam heat exchangers was observed under both dry and wet operating conditions. The findings have been explained in terms of the impact of the thermal conductivity of base metal and condensate retention.

The effective use of coolant in maintaining gas turbine components below failure limits is becoming increasingly necessary in view of increasing combustor exit temperatures. Incorporating coolant impingement in a pin-finned channel provides an avenue for increased heat transfer. Two types of cooling configuration are modeled and compared: case 1: with the jet impinging in a cut out region behind a pin-fin (thus sheltering the jet from crossflow effects), and case 2: with the jet impinging further downstream along the next row of pin-fin and downstream of the wake region of the upstream pin-fin. In the current work, a Large Eddy Simulation (LES) study is conducted for a Reynolds number of 7500. All other geometrical and flow parameters are kept similar in both the configurations to enable a direct comparison. Since the stagnation point Nusselt number is not affected by the cross-flow in the sheltered configuration, there is enhanced cooling as compared to the configuration in which the jet impinges further downstream and is not protected from the crossflow. In this paper, the flow field, the jet-crossflow interactions and the heat transfer behavior are discussed for both the configurations.

Sliding bubbles are known to augment heat transfer rates on the surface on which they slide. The pre-cursor problem — the bubble approaching an inclined superheated wall provides the initial flow and thermal field for the sliding bubble problem. An FC-87 vapor bubble rising in a thermally stratified flow field is simulated along with the bubble wall interaction effects. The simulation is conducted on a dynamic octree grid for improved accuracy and efficiency. The evolution of the bubble shape and the wake behind the rising bubble is captured in a three-dimensional model, which takes into account bubble growth due to superheat at the liquid-vapor interface and the effect of interface heat flux on the local saturation temperature. After the first bubble-wall interaction, a microlayer tens of microns thick forms between the bubble and the wall; a thermal wake develops behind the bubble as it begins to slide against the wall. The predicted shapes, Re and Weber numbers and microlayer thicknesses show excellent agreement in comparison to experimental data from other researchers. Evolution of the flow and temperature fields were examined with the aid of contours of vapor volume fraction and iso-lines of mixture temperature superimposed on three-dimensional shapes of the bubble. Overall bubble dynamics and microlayer dynamics, including microlayer thickness and microlayer heat flux, are presented as functions of time. Using the wall, microlayer and wake heat transfer rates, an enhancement of the total wall heat flux was found to be on the order of 6 times the background heat flux. This work describes the bubble evolution through the first rebounding in detail, but the dynamic octree adaption algorithm lends itself to study of the long-term dynamics well into the sliding regime. The technique can also be used to investigate other multiphase flow phenomena — especially bubble coalescence and breakup.

This study presents a transient three-dimensional numerical study on fluid flow and heat transfer of flat-tube array using large eddy simulation (LES) covering both laminar and turbulent flow regimes. The simulations were performed in a rectangular region containing only one tube with periodic conditions specified on all boundaries. A staggered flat-plate array was first studied, and an existing solution was used for validation purpose. The numerical models were then applied to an in-line array composed of flat tubes with an aspect ratio of 0.25 and fixed tube spacings. By varying the in-flow velocity, the tube array was studied over a wide range of Reynolds number (600–12000). Temperature, velocity, and turbulent kinetic energy distributions as well as the interactions between them are presented and analyzed. Furthermore, the local heat transfer rate was analyzed along the various parts of the tube (leading edge, flat-top and wake or trailing-edge regions). Heat transfer correlation for each region of the tube and the entire tube array is proposed.

This research focuses on the computational fluid dynamics simulation of near wind turbine wake. Three dimensional wind turbine model is built based on S809 airfoil data [1]. Three different turbulence models are used and compared. They are Realizable k-ε model, SST k-ω model, and Large Eddy Simulation (LES) model. The simulation results from different turbulence models are compared with the NREL Phase VI experiment data. Different boundary conditions, including neutral and unstable conditions, were adopted in the simulation to analyze their influence on wake flow. Updraft and downdraft are considered in this part. Overall numerical results match well with the experiment data. The discussion also compares wind turbine wake under different atmospheric boundary conditions.

The present study aimed to investigate the effect of an unsteady wake on the heat transfer for the endwall surface of a linear turbine blade cascade. A naphthalene sublimation method was implemented to obtain the detailed heat/mass transfer distributions on the endwall surface. Tests were conducted on a five-passage linear cascade in a low-speed wind tunnel. The effects of unsteady wakes were simulated in the facility by a wake generator consisting of circular rods that were traversed across the inlet flow. The test conditions were fixed at a Reynolds number of 70,000 based on the inlet velocity and chord length. The flow coefficients were varied from 1.3 to 4.2. and range of Strouhal number was 0.1 to 0.3. The results showed that the heat transfer distributions differed between steady and unsteady test cases. The overall heat transfer for the unsteady test cases was higher, and the heat transfer was enhanced with increasing the Strouhal number due to the resulting thin boundary layer and high turbulence intensity. Therefore, a cooling system for the endwall of a rotor should focus on decreasing the high temperatures of the endwall surface induced by the unsteady wakes.

Film cooling experiments were run at the High-Speed Cascade Wind-Tunnel of the University of the Federal Armed Forces Munich. The investigations were carried out on a linear cascade of aerodynamically highly loaded turbine blades. The main targets of the tests were to assess the film cooling effectiveness and the heat transfer in zones with main flow separation. The cascade was designed to have a large zone with flow separation on the pressure side starting at the leading edge and reaching up to approximately half of the axial chord. Film cooling is provided on the pressure side at the front part of the blade in order to reduce flow separation and to provide effective film cooling inside the separation bubble and at the reattachment zone. The studies comprise the measurement of adiabatic film cooling effectiveness and heat transfer coefficient for the different cascades under a set of different Mach and Reynolds numbers at engine relevant levels. Periodic wakes, generated by bars moving upstream of the cascade, simulate the rotor stator interaction. The results show that film cooling is detrimental for cooling the surface inside the flow separation zone, whereas after reattachment improved overall film cooling effectiveness is obtained with periodic unsteady inflow compared to the cases with homogeneous inflow.

The necessity of performing an unsteady simulation for the purpose of predicting the heat transfer on the endwall surfaces of a turbine passage is addressed. This is measured by the difference between the two solutions obtained from a steady simulation and time average of an unsteady simulation. The heat transfer coefficient (Nusselt number) based on the adiabatic wall temperature is used as the basis of the comparison. As there is no film cooling in the proposed case, a heat transfer coefficient so computed should be a better measure of such difference than say, wall heat flux. Results show that the effect of unsteadiness due to wake passage on the pressures and recovery temperatures on both the hub and casing is negligible. Heat transfer on the endwalls, however, is affected by the unsteady wake; the time-average results yield higher heat transfer; in some regions up to 15% higher. The results for the endwall heat transfer were compared to results in the open literature and are found to be comparable.

The aim of this study is to analyze the phenomena of heat transfer enhancement between two periodic sinusoidal walls for a single gas flow. The experimental set-up is characterized by a few geometrical parameters: amplitude of wall waviness (A); channel height (H); fin wavy-channel width (ω); fin period (l) and total wavy length (L). Combination of these ones is reduced to: the wall aspect ratio γ, the cross-section aspect ratio α and the channel spacing ratio ε. The Reynolds number defined on the hydraulic diameter and the bulk velocity is greater than 4000. A constant heat flux is maintained on the second lateral wall. For Re = 5700, we observe an entrance region from the first to the fourth period; beyond, the velocity profiles are autosimilar. A shear layer is generated just downstream of the crest and develops in its wake up to the concavity area. Thermal experimental approach is performed by local measurements of convective heat transfer coefficient along the walls, within the viscous sublayer. The heat transfer profile presents an increasing from the crest of 15%, and the maximum is located at the first quarter of the period, close to the separated point. Beyond, the value of heat transfer decreases of 50% and the minimum is located close to the reattachment point. Then the heat transfer increases up to the next crest. The same phenomenon is observed in the next periods of the channel. To explain theses results, we calculate the turbulence terms obtained from the classical equations of fluid mechanics. The turbulence production (P) presents a maximum in the core of the shear layer, where the Reynolds constraints and the heat transfer are maxima. A good correlation is obtained between turbulence production and heat transfer. The flow pattern (mean, fluctuating and turbulence terms) are performed with PIV technique in order to analyze the vortices that develop in the shear layer, based on 1000 pairs of images.

The film cooling effectiveness on the surface of a high pressure turbine blade is measured using the Pressure Sensitive Paint (PSP). Four rows of fan-shaped, laid-back compound angled cooling holes are distributed on the pressure side while two such rows are provided on the suction side of the blade. The coolant is only injected to either the pressure side or suction side of the blade at five average blowing ratios from 0.4 to 1.5. Presence of wake due to upstream vanes is simulated by placing a periodic set of rods upstream of the test blade. The wake rods can be clocked by changing their stationary positions to simulate a progressing wake. Effect of wake is recorded at four phase locations with equal intervals along the pitch-wise direction. The free stream Reynolds number, based on the axial chord length and the exit velocity, is 750,000 and the inlet and the exit Mach numbers are 0.27 and 0.44, respectively, resulting in a blade pressure ratio of 1.14. Results reveal that the tip leakage vortices and endwall vortices sweep the coolant film on the suction side to the midspan region. The fan-shaped, laid-back compound angled holes produce good coolant film coverage on the suction side except for those regions affected by the secondary vortices. Due to the concave surface, the coolant trace is short and effectiveness level is low on the pressure surface. However, the pressure side acquires relatively uniform film coverage with the design of multiple rows of cooling holes. The presence of stationary upstream wake results in lower film cooling effectiveness on the blade surface. Variation of blowing ratio from 0.4 to 1.5 shows steady increase in effectiveness on the pressure side or the suction side for a given wake rod phases locations. The compound angle shaped holes outperform the compound angle cylindrical holes by the elevated film cooling effectiveness particularly at higher blowing ratios.

Film cooling is extensively used to provide protection against the severe thermal environment in gas turbine engines. Most of the computational studies on film cooling flow have been done using steady Reynolds-Averaged Navier-Stokes (RANS) calculation procedures. However, the turbulent stress field is highly anisotropic in the wake region of the coolant jet, and the inherent unsteadiness of the coolant jet-crossflow interactions may have important implications in the cooling performance. In this paper, a computational investigation about the unsteady behavior of jet-in-crossflow applications is performed using DES. Detailed computation of a single row of 35 degree round holes on a flat plate has been obtained for a blowing ratio of 1.0 and a density ratio of 2.0. Firstly, time step size, grid resolution tests have been conducted. Comparison of the time-averaged DES prediction with the measured film cooling effectiveness shows that DES prediction is reasonable. From present simulations, the typical coherent vortical structures of the jet-in-crossflows can be seen. The unsteady physics of jet-in-crossflow interactions and a jet liftoff in film cooling flows have been explored.

The computational study investigates different pin fin arrangements at low Reynolds numbers, which would typically be prevalent in mini-micro-channels used in enhancing heat as well as mass transfer. The effect of pin density, span-wise pitch, and stream-wise pitch is investigated on friction and heat transfer over a range 5<Re D <400. High density pins with small span-wise pitches were found to provide the highest augmentation in heat transfer capacity (conductance), whereas low density pins with or without a large stream-wise pitch were found to provide the least heat transfer benefits in the low Reynolds number range studied. Friction factor decreases considerably as the pin density decreases. The effect of decreasing span-wise pitch increases the friction factor in the low Reynolds number regime (Re D <200) but decreases it beyond Re D = 200 by delaying wake instabilities and the associated increase in form drag. Increasing the stream-wise pitch decreases the friction factor at low Re D <200, but increases it at Re D >200 due to the formation of larger recirculating wakes. Overall it is concluded that a high density arrangement with a small span-wise pitch provides the best thermal performance.

The time-dependent temperature distribution on an inclined, thin-foil uniform-heat-generation heater was used to infer the surface heat transfer enhancement caused by the passage of an FC-87 bubble sliding beneath the lower surface of the heater. A two-camera system was used: one camera recorded color images of a liquid crystal layer applied to the upper (dry) side of the heater while a second camera simultaneously recorded the position, size, and shape of the bubble from below. The temperature response of the heater could then be correlated directly to the bubble characteristics at any given time during its passage. Data along the line bisecting the bubble wake from 9 bubbles comprising 54 bubble images were analyzed. Heat transfer in the wake behind sliding cap-shaped bubbles is very effective compared to the natural convection that occurs before the passage of the bubble. Maximum values of heat transfer coefficient in the range of 2500 W/m 2 K were produced in very sharply peaked curves. The point of maximum cooling measured as a fraction of the local driving temperature difference before the bubble passage was identified and correlated with some success to the streamwise length of the bubble. The location of the maximum heat transfer coefficient was reasonably correlated to bubble width. The level of the maximum heat transfer coefficient when cast as a Nusselt number based on bubble width grew to a saturation value as the bubble moved across the plate. A constant value of Nusselt number requires that the heat transfer coefficient falls as the bubble grows past some critical bubble size. This behavior was observed for the larger bubbles.

The use of an interrupted plate fin with surface roughness in the form of split-dimples is investigated. Time-dependent high-fidelity simulations are conducted for laminar, early turbulent, and fully turbulent flows, Re H = 360, 800, and 2000. Detailed analysis of the domain’s flow structure, turbulent statistics, and heat transfer distribution is presented. Regions of high heat transfer occur at the fin and protrusion leading edges, at flow impingement on the protrusion faces, and flow acceleration region between protrusions. Flow separation and large wakes induced by the large protruding surfaces of the split-dimples, increase friction losses and reduce heat transfer from the fin. The split-dimple fin has a heat conductance 60–175% higher than that of the plate fin, but at 4–8 times the pressure drop.

The paper describes two- and three-dimensional computer simulations which are used to study fundamental flow and thermal phenomena in multilouvered fins used for air-side heat transfer enhancement in compact heat exchangers. Results pertaining to flow transition, thermal wake interference, and fintube junction effects are presented. It is shown that a Reynolds number based on flow path rather than louver pitch is more appropriate in defining the onset of transition, and characteristic frequencies in the louver bank scale better with a global length scale such as fin pitch than with louver pitch or thickness. With the aid of computer experiments, the effect of thermal wakes is quantified on the heat capacity of the fin as well as the heat transfer coefficient, and it is established that experiments which neglect accounting for thermal wakes can introduce large errors in the measurement of heat transfer coefficients. Further, it is shown that the geometry of the louver in the vicinity of the tube surface has a large effect on tube heat transfer and can have a substantial impact on the overall heat capacity.

The effect of buoyancy on the wake instability of a heated circular cylinder in a contra flow arrangement is investigated experimentally in the present study. A novel optical diagnostic technique, Molecular Tagging Velocimetry and Thermometry (MTV&T), was used for qualitative flow visualization and quantitative simultaneous measurements of velocity and temperature fields in the wake of the heated cylinder. The experiment was conducted in a water channel with the temperature and Reynolds number of the approaching flow being held constant at T ∞ = 24° C and Re = ρ ∞ U ∞ D / μ ∞ = 130. The temperature of the heated cylinder varied between 24° C (unheated cylinder) and 85° C , corresponding to a Richardson number (Ri) varying between zero (unheated) and unity. The heat transfer process around the heated cylinder changes from the forced convection regime to the mixed convection regime. With increasing Richardson number, significant modifications of the wake instability were revealed from both qualitative flow visualization images and quantitative simultaneous velocity and temperature fields. The effect of buoyancy on the wake instability of the heated cylinder was discussed in the terms of vortex shedding pattern, vortex shedding frequency, turbulent heat flux distribution, the wake closure length and averaged Nusselt number of the heated cylinder.

Experiments were conducted to determine the relationship between wake structure and heat transfer for an oscillating circular cylinder in cross-flow. An internally heated cylinder was suspended in a water tunnel and oscillated transverse to the freestream. The cylinder’s heat transfer coefficient was measured over a wide range of oscillation amplitudes and frequencies. By comparing these results to the known wake mode regions in the amplitude-frequency plane, relationships between wake mode and heat transfer were identified. Representative cases were investigated further by using digital particle image thermometry/velocimetry (DPIT/V) to simultaneously measure the temperature and velocity fields in the near-wake. This revealed more detail about the mechanisms of heat transfer enhancement. The dynamics of the vortex formation process, including the trajectories of the vortices during roll-up, are the primary cause of the heat transfer enhancement.

The development of a buoyant plane wake has been investigated experimentally. A water channel has been used as a statistically steady experiment to investigate the plane wakes. Parallel streams of hot and cold water are initially separated by a splitter plate. The streams are oriented such that the cold fluid is above the hot fluid, resulting in an unstable stratification. At the end of the splitter plate, the two streams are allowed to mix and a buoyancy driven mixing layer develops. Downstream of the splitter plate, growth of the turbulent buoyancy-driven mix is disrupted by a cylinder. The cylinder is located at the centerline of the mixing layer and associated wake. As a result the dynamic flows of the plane wake and buoyancy driven mixing layer interact. Particle image velocimetry (PIV), and a high-resolution thermocouple system are used to measure the response of the plane wake to buoyancy driven turbulence. Velocity and density measurements are used as a basis from which we describe the transition, and return to equilibrium, of the buoyancy driven mixing layer. We found for wakes where buoyancy is driving the motion, a remarkably fast recovery of a Rayleigh-Taylor mix in the wake region.

The sedimentation of a sphere through a fluid is a standard testbed problem in non-Newtonian fluid mechanics. The experimentally determined velocity fields may be compared with numerical predictions obtained using finite element simulations. In this work, the influence of an adjacent wall, in addition to elastic and shear thinning effects, on the velocity field in the fluid surrounding the sphere is studied. Three different test fluids are employed: a Newtonian reference fluid, a constant shear viscosity (elastic) Boger fluid, and a shear thinning elastic fluid. All three fluids have similar zero shear viscosities. For all experiments, the terminal velocity is achieved before measurements begin. Significant differences in both the location and magnitude of the recirculation zones are observed for the different test fluids. In addition, the shape of the wake is qualitatively different for the various fluids.