The Effect of Geometrical Features on Air-Side Heat Transfer and Friction Characteristics, and Optimization of a Large-Size Plain Fin-and-Tube Heat Exchanger by 3-D Numerical Simulation

The effect of geometrical features on the air-side heat transfer and friction characteristics of an industrial plain fin-and-tube heat exchanger is investigated by 3-D numerical modeling and simulations. The heat exchanger has been designed and employed as an intercooler in a gas power plant and is a large-size compact heat exchanger. Most of the available design correlations developed so far for plain fin–and–tube heat exchangers have been prepared for small-size exchangers and none of them fits completely to the current heat exchanger regarding the geometrical limitations of correlations. It is shown that neglecting these limitations and applying improper correlations may generate considerable amount of error in the design of such a large-size heat exchanger. The geometry required for numerical modeling is produced by Gambit® software and the boundary conditions are defined regarding the real operating conditions. Then, three-dimensional simulations based on the SIMPLE algorithm in laminar flow regime are performed by FLUENT™ code. The effect of fin pitch, tube pitch, and tube diameter on the thermo-hydraulic behavior of the heat exchanger is studied. Some variations in the design of the heat exchanger are suggested for optimization purposes. It is finally concluded that the current numerical model is a powerful tool to design and optimize of large-size plain fin-and-tube heat exchangers with acceptable accuracy.

Axial Heat Conduction in Parallel Flow Microchannel Heat Exchangers

This paper deals with the effect of axial heat conduction on the hot and cold fluid effectiveness of a balanced parallel flow microchannel heat exchanger. The ends of wall separating the fluids are subjected to Dirichlet boundary condition. This leads to heat transfer between the microscale heat exchanger and its surroundings and thereby leading to axial heat conduction through the wall separating the fluids. Three one dimensional energy equations were formulated, one for each of the fluids and one for the wall. These equations were solved using finite difference method. The effectiveness of the fluids depends on the NTU, axial heat conduction parameter, and the temperature of the ends of the wall separating the fluids. With decrease in temperature of the end wall at the inlet section of the fluids, while keeping the temperature of the other end wall constant, the effectiveness of the hot and cold fluid increased and decreased, respectively. When the temperature at the ends of the wall separating the heat exchanger is average of the inlet temperature of the fluids then there is no axial heat conduction through the heat exchanger. The effectiveness of a counter flow microchannel heat exchanger is better than that of a parallel flow microchannel heat exchanger subjected to similar operating conditions, i.e. axial heat conduction parameter and end wall temperatures.

Particle Dispersion and Deposition in a Curved Duct

Particle transport in ducts of square cross-section with constant streamwise curvature is studied using numerical simulations. The flow is laminar, with Reynolds numbers of Reτ = 40 and 67, based on the friction velocity and duct width. The corresponding Dean numbers for these cases are 82.45 and 184.5, respectively, where De = Rea/R, a is the duct width and R is the radius of curvature. A Lagrangian particle tracking method is used to account for the particle trajectories, with the particle volume fraction assumed to be low such that inter-particle collisions and two-way coupling effects are negligible. Four particle sizes are studied, τp+ = 0.01, 0.05, 0.1, and 1. Particle dispersion patterns are shown for each Dean number, and the steady-state particle locations are found to be reflective of the Dean vortex structure. Particle deposition on the walls is shown to be dependent upon both the Dean number and particle response time, with the four-cell Dean vortex pattern able to prevent particle deposition along the center of the outer wall.

Numerical Investigation of a Liquid Droplet Impinging on a Heated Surface

Previous studies, most of them experimental, reveal that the cooling effectiveness of a water drop impinging on a heated surface depends on the wall temperature, droplet shape and velocity. All previous studies focus on the behavior of a droplet falling in a quiescent environment, such as still air. Evidence in the literature also shows that gas assisted droplet sprays, in which a gas phase propels the droplets, are more efficient in heat removal than sprays consisting of droplets alone. It is conjectured that this is due to an increase in the maximum droplet spreading diameter upon impact, a thinner film, and consequently an increase in the overall heat transfer coefficient. Recent experiments in the author’s group [1, 2] show that the carrier gas jet strongly influences droplet spreading dynamics by imposing normal and shear forces on the liquid surface. The heat transfer is greatly augmented in the process, compared to a free falling droplet. To date, there has been no fundamental investigation of the physics of gas assisted spray cooling. To begin to understand the complicated process, this paper reports on a fundamental problem of a single liquid droplet that impinges on a heated surface. This paper contributes a numerical investigation of the problem using the volume of fluid (VOF) technique to capture droplet spreading dynamics and heat transfer in a single drop event. The fluid mechanics is investigated and compared to the experimental data. The greatest uncertainty in the simulation is in the specification of the contact angle of the advancing or receding liquid front, and in capturing the onset of the three-dimensional fingering phenomena.

Simulation of Multi-Mode Film Boiling Using Level Set Method

2D transient multi-mode film boiling simulation of water near critical pressure (p = 0.99pc = 21.9 MPa) on a heated horizontal surface is carried out using an in-house Level Set (LS) method based semi-explicit finite volume method code. The influence of initial vapor film thickness (yo) on the dominant instability mode is evaluated by carrying out simulations on domain having width greater than most dangerous Taylor wavelength i.e. LX = 4λd with y0 = 0.0425λd and 0.125λd at low wall superheat (ΔT = 2K). For lower initial film thickness, the viscous force dominated Rayleigh-Taylor instability is captured and the average bubble spacing is found close to the prediction made using lubrication theory i.e. λP = 2λc = 0.816λd. However, for higher initial film thickness, the inertia force dominated Taylor-Helmholtz mode of instability is found with the average bubble spacing close to λd. Simulations are carried out to check the existence of Rayleigh-Taylor instability on various domain width LX = 2λd, 3λd, 4λd and 6λd at yo = 0.0425λd and ΔT = 2K. The average bubble spacing for all domain widths is found to be less than 2λc indicating that the Rayleigh-Taylor instability is dominant.

Design Optimization of an L-Shaped Extrusion Die

It is an important problem in the polymer extrusion of complex profiles to balance the flow at the die exit. In this paper, we employ simulated annealing-kriging meta-algorithm to optimize the geometric parameters of a die channel to obtain a uniform exit velocity distribution. Design variables for our optimization problem involve the suitable geometric parameters for the die design, which are the thickness of the large channel and the length of the narrow channel. Die balance is based on the deviation of the velocity with respect to the average velocity at the die exit. So the cost function for the optimization problem involves the minimization of this deviation. For the design of numerical experiments, we use Latin Hypercube Sampling (LHS) to construct the kriging model. Then, based on the LHS points, the numerical solutions are performed using Polyflow, a commercial software based on the finite element method and is specifically designed to simulate the flow and heat transfer of non-newtonian, viscoelastic fluids. In our simulations, a HDPE (high density polyethylene) is used as extrusion material. Having obtained numerical simulations for N = 60 LHS points in two-dimensional parameter space (t and L), the optimization of these parameters is carried out by Simulated Annealing (SA) method in conjunction with kriging model. We show that kriging model employed in SA algorithm can be used to optimize the die geometry.

Theoretical Optimization of Cascading Magnets’ Structure for Oxygen Enrichment by Gradient Magnetic Fields

Oxygen molecules are paramagnetic while nitrogen molecules are diamagnetic. In the same gradient magnetic field, the magnetizing forces on oxygen molecules are stronger than those on nitrogen molecules, which in opposite directions. The intercepting effect on oxygen molecules by gradient magnetic field can be used for oxygen enrichment from air. The structure, which is called multi-channel cascading magnets array frame in the paper, are optimized by additional yokes. By comparison of distributions of magnetic field in multi-channel array without yokes and that with yokes, the additional yokes can eliminate the differences among different magnetic spaces in multi-channel cascading magnets’ arrays and enhances the magnetic flux densities in spaces. Joining magnets together in the length direction can make the air stay longer in the ‘magnetic sieve’ and raise the oxygen concentration of air flowing out from the optimized multi-channel cascading magnets’ arrays. The inside additional yoke can used to avoid the gradient magnetic field at the joints of the magnets and get near uniform magnetic field along length direction. The optimized multi-channel cascading magnets’ array frames can effectively promote the development of oxygen enrichment from air by “magnetic sieve”.

The Effect of Catalyst Coated Diffusion Media on PEM Fuel Cell Performance

Two fuel cell architectures, differing only by the surfaces onto which the electrodes were applied, have been analyzed to determine the root causes of dissimilarities in performance. The basic proton exchange membrane fuel cell (PEMFC) is comprised of the proton transporting membrane, platinum-containing anode and cathode electrodes, porous carbon fiber gas diffusion media (GDM), and flow fields which deliver the reactant hydrogen and air flows. As no optimal cell design currently exists, there is a degree of latitude regarding component assembly and structure. Catalyst coated diffusion media (CCDM) refers to a cell architecture option where the electrode layers are coated on the GDM layers and then hot-pressed to the membrane. Catalyst coated membrane (CCM) refers to an architecture where the electrodes are transferred directly onto the membrane. A cell with CCDM architecture has tightly bonded interfaces throughout the assembly which can result in lower thermal and electrical contact resistances. Considering the fuel cell as a 1-D thermal system, the through-plane thermal resistance was observed to decrease by 5–10% when comparing CCDM to CCM architectures. This suggests the thermal contact resistance at the electrode interfaces was significantly reduced in the hot-press process. In addition, the electrical contact resistances between the electrode and GDM were observed to be significantly reduced with a CCDM architecture. This study shows that these effects, which have a potential to increase performance, can be attributed to the hot-press lamination process and use of CCDM architecture.

Comparison of Pressure Drop and Endwall Heat Transfer Measurements to Flow Visualization Testing of Solid and Porous Pin Fin Arrays

This paper examines the local and averaged endwall heat transfer effects of a staggered array of porous aluminum pin fins with a channel blockage ratio (blocked channel area divided by open channel area) of 50%. Two sets of pins were used with pore densities of 0 (solid) and 10 pores per inch (PPI). The pressure drop through the channel was also determined for several flow rates using each set of pins. Local heat transfer coefficients on the endwall were measured using Thermochromatic Liquid Crystal (TLC) sheets recorded with a charge-coupled device (CCD) camera. Static and total pressure measurements were taken at the entrance and exit of the test section to determine the overall pressure drop through the channel and explain the heat transfer trends through the channel. The heat transfer and pressure data was then compared to flow visualization tests that were run using a fog generator. Results are presented for the two sets of pins with Reynolds numbers between 25000 and 130000. Local HTC (heat transfer coefficient) profiles as well as spanwise and streamwise averaged HTC plots are displayed for both pin arrays. The thermal performance was calculated for each pin set and Reynolds number. All experiments were carried out in a channel with an X/D of 1.72, a Y/D of 2.0, and a Z/D of 1.72.

Dynamics of Evaporation and Cooling of a Water Droplet During the Early Stage of Depressurization

This paper reports an experimental and numerical study of evaporation and cooling of a water droplet during the early stage of depressurization in a test vessel. During the experiment, a distilled water droplet was suspended on a thermocouple, which was also used to measure the droplet center temperature, and the droplet surface temperature was captured by an infrared thermograph. Experimental data indicated a large temperature difference within the droplet during the early stage of depressurization. A thermodynamic analysis of the experimental data found that the pressure reduction was not fast enough to induce liquid superheating and thus equilibrium evaporation was expected. A mathematical model was then constructed to simulate the droplet evaporation process. The model solves one-dimensional heat conduction equation for the temperature distribution inside the water droplet, with the convective heat transfer inside the droplet simplified through an effective conductivity factor. A simplified treatment was introduced to quantify the convective evaporation due to air movement and droplet swing induced by sudden opening of the electro-magnetic valve and the following air exiting. The model-predictions agree well with the measured temperature data, demonstrating the soundness of the present model.