0

Heat Transfer in Manufacturing and Materials Processing

2003;():1-11. doi:10.1115/HT2003-47003.

In this work, large-scale molecular dynamics simulation is conducted to explore nanoscale manufacturing with laser-assisted scanning tunneling microscope. Employing a super parallel computer, more than 100 million atoms are modeled to provide substantial details about how the localized thermal and mechanical perturbations result in surface nanostructures. It is found that thermal equilibrium cannot be established due to the small number of atoms. Extremely localized stress accumulation beneath the sample surface results in an explosion of the melted/vaporized material, leaving a nanoscale hole on the sample surface. Normal and shear stress development is observed. Stress propagation in space is strongly influenced by the anisotropic nature of the crystal. The high pressure in the melted/vaporized region pushes the melt adjacent to the solid to move, thereby forming a protrusion at the edge of the hole. More importantly, visible structural destruction is observed in the region close to the bottom of the sample. These destructions are along the direction of 45 degrees with respect to the axial direction, and are attributed to the strong tensile stress. Atomic dislocation is observed in the destructed regions.

Commentary by Dr. Valentin Fuster
2003;():13-21. doi:10.1115/HT2003-47007.

A comprehensive two-dimensional numerical model, which accounts for heat/mass transfer, solidification, and electromagnetic field, has been developed to simulate the silicon tube growth by the Edge-defined film-fed (EFG) method. A multi-block grid system has been employed to yield a high accuracy in the vicinity of die tip with relatively low CPU time, and the solution procedure is satisfied the flux conservation at the block interface. Selected results of magnetic and temperature fields have been presented for the silicon tube growth system of 30cm in diameter and 0.3mm in thickness. Two local models have also been developed to study the effect of the size of window opening and tube thickness on the maximum growth rate using the inner and outer heater temperature profiles as boundary conditions.

Commentary by Dr. Valentin Fuster
2003;():23-27. doi:10.1115/HT2003-47030.

HDP-CVD reactors are used for Shallow Trench Isolation (STI), Inter Metal Dielectric (IMD) and Inter Layer Dielectric (ILD) applications for logic and memory device fabrication. As device dimension shrinks, the trend has been to use lower pressure and higher plasma density for gap-fill with higher aspect ratio (AR). Higher AR gapfill in addition to higher throughput is achieved by running multiple wafers between a chamber clean, present a unique set of challenges for heat and mass-transfer in an HDP-CVD reactor. This paper describes some of the new state-of-the-art hardware innovations specifically developed to meet these challenges. In particular, heat transfer to plasma facing materials, fluid mechanics, and transport of sub-micron sized particles in the plasma environment of an HDP-CVD reactor are explored.

Commentary by Dr. Valentin Fuster
2003;():29-35. doi:10.1115/HT2003-47032.

In the casting of metal matrix composite, different processing parameters need to be controlled in order to promote the formation of primary alpha phase around the reinforcement. It has been shown [1–3] that when the reinforcement is allowed to be extended out of the cast mold and cooled by a heat sink, the microstructure of the composite can be improved due to faster heat extraction through the reinforcement. Thermal management of the reinforcement can eliminate a large portion of the eutectic phase during solidification, leading to an altered microstructure at the interface between the matrix and the reinforcement with the possibility of improved material properties. A companion paper [4] shows a comparison of the numerical simulation result of the casting of MMC by squeeze infiltration technique to the experimental work. The authors assumed that the solidification process started after the liquid metal has completely infiltrated the reinforcement. The simulation result gives a reasonable prediction to the experimentally measured cooling temperature profile. In this work, the effects of other processing parameters are analyzed to study the impregnation depth during squeeze infiltration. These processing parameters include the thermal conductivity of fibers, the initial (or preheat) mold temperature, the volume fraction of fibers, and the heat sink temperature. The study is based on the finite volume method for enthalpy formulated heat equation.

Commentary by Dr. Valentin Fuster
2003;():37-46. doi:10.1115/HT2003-47056.

Interfacial thermal contact resistance between the impinging flow of a molten droplet and a substrate, which is qualified by thermal contact conductance, plays an important role in the spreading and solidification of a droplet. In the present study, a simple correlation for the thermal contact conductance in the rapid contact solidification process was developed. With this correlation being directly used in numerical simulation, for the first time, a variable thermal contact resistance was taken into consideration to simulate both the dynamics and phase change responses during a molten droplet impingement. Numerical results were compared with that of the cases when thermal contact resistance was zero or a constant. The changes in spread factor with time and thermal contact conductance indicated that predictions from the computer simulation were sensitive to the values of thermal contact resistance. Experiment was conducted to demonstrate the validity of the present study. Comparison results showed that rather than using a constant average value, better agreement between the experimental and numerical results would be obtained if a variable thermal contact resistance were used in the numerical simulation.

Commentary by Dr. Valentin Fuster
2003;():47-52. doi:10.1115/HT2003-47066.

This numerical study investigates heat generation and cure during the unsaturated flow of thermosetting resins in woven, stitched or braided fiber mats during mold filling in liquid composite molding (LCM), a popular technology to manufacture polymer-matrix composites. This study is relevant to those mats, which can be characterized as a dual-scale porous medium. An iterative, control-volume approach, based on energy and cure balances in a two-layer model representing fiber tows and gaps between tows, is used for developing discretized equations for average temperatures and cures in the tows and gaps respectively. A significant difference in the temperatures and cures of the gap and tow regions is observed. The proposed model deviates significantly from the conventional single-scale model used in most LCM simulations and highlights the need to adopt a different approach in modeling cure and temperature in dual-scale fiber mats.

Commentary by Dr. Valentin Fuster
2003;():53-61. doi:10.1115/HT2003-47068.

A significant obstacle in ultrafast laser micromachining of multi-layer or heterogeneous micro-structures is the lack of an online diagnostic method to determine which material is being ablated during the material removal process. This problem arises because ultrafast lasers are generally insensitive to the material being processed. One promising technique to address this problem is the use of laser-induced breakdown spectroscopy (LIBS) by which the plasma generated during the laser-material interaction can be collected and analyzed to provide information regarding the elemental composition of the mate-rial being ablated. In this work, a real-time feedback control system for the ultrafast laser micromachining process based on the LIBS technique is built. The ultrafast LIBS signal is first characterized to prove the feasibility. Characteristics of spectral emission, temporal evolution, spatial heterogeneity of the ultrafast LIBS signal, effects from laser machining factors, etc., are discussed. Comparison methods for identifying the material emission patterns are then studied. Effective algorithms from the study are implemented into the control system software, SPECOMP, developed in the laboratory. Issues on the real-time control process are discussed. The real-time controlled machining process has then been applied to the machining of micro-structures on thermal sprayed material. Compared to the passive machining process without any such feedback control, SPECOMP system provides several important advantages including less damage to the substrate layer, shortened machining time, and more uniform feature sizes.

Commentary by Dr. Valentin Fuster
2003;():63-70. doi:10.1115/HT2003-47069.

Fabrication of structural and functional parts and components, especially at the micro and nano scales, is crucial to a wide range of applications in the electronics, communications, medical, aerospace, and military industries. This work presents an innovative conformal direct-write technique for rapid prototyping and manufacturing novel sensors. The technique combines thermal spray, which, as an additive process, produces blanket depositions of films and coatings, with ultrafast laser micromachining, a subtractive process to produce functional patterns. Several kinds of sensing components, such as microheaters and strain gauges, have been successfully fabricated in this work with thermal spray technology and a femtosecond laser, which demonstrates the feasibility and advantages of the proposed technique. The electrical and thermal property characterization of the sensors was also performed, and shows promise for sensors in micro-sensing systems. With minor modification to pattern design and processing procedures, various sensing structures and electronic components, for example, precision resistors and interdigitated capacitors, can be readily fabricated using the presented technique.

Commentary by Dr. Valentin Fuster
2003;():71-77. doi:10.1115/HT2003-47071.

A mathematical modeling of argon thermal plasma flows inside a rod type cathode (RTC) plasma torch based on electromagnetic principle is presented. Numerical investigations on the model are carried out for a better understanding of its output characteristics in terms of the plasma velocity, temperature, mass fraction and electro-magnetic field calculations. A direct comparison of the present numerical results with a reported experimental data is made and found to be in fair agreement.

Commentary by Dr. Valentin Fuster
2003;():79-86. doi:10.1115/HT2003-47076.

In this paper, we report the heat generation and temperature field in a multilayer device consisting of thin and weakly absorbing materials subject to pico-second to nano-second pulsed-laser heating. The interference effects due to the internal reflection and refraction are considered. A tracking method based on electromagnetic optics and wave optics is used to determine the two-dimensional electrical and magnetic fields. These fields are then used to calculate the heat generation and the accompanying temperature distribution. For demonstration, we apply this method to determine the temperature field in a ZnSe interference filter subject to inclined laser incidence on its side. The simulation results show strong localized heating in a narrow region along the side where laser power is incident. The localized heating produces several high power spots, aligned obliquely to the side surface. The results show that the pure absorption model is not valid for even small incident angles. Surface absorption may be a valid approximation but the heat flux distribution is not uniform.

Commentary by Dr. Valentin Fuster
2003;():87-93. doi:10.1115/HT2003-47096.

An experimental study is performed in this paper to verify the concept of thermal management of using a heat pipe in the drilling process. The basic idea is to insert a heat pipe at the center of the drill tool with the evaporator located close to the drill tip, and condenser located at the end of the drill. In this way, heat accumulated in the drill tip can be transported to the remote section of the drill and remove it there to the tool holder, which attaches the drill. Temperatures at the drill tip as well as tool wear can be reduced significantly. In this paper, experimental investigations on a heat pipe drill for various heat flux inputs, inclination angles and rotating speeds are presented. The effect of contact resistance and tool holder (acting as heat sink) on heat pipe performance will also be demonstrated. The results presented in this paper may be used for important design and practical implementation considerations.

Commentary by Dr. Valentin Fuster
2003;():95-102. doi:10.1115/HT2003-47145.

Thermal management using heat pipes is gaining significant attention in past decades. This is because of the fact that it can be used as an effective heat sink in very intricate and space constrained applications such as in electronics cooling or turbine blade cooling where high heat fluxes are involved. Extensive research has been done in exploring various possible applications for the use of heat pipes as well as understanding and modeling the behavior of heat pipe under those applications. One of the possible applications of heat pipe technology is in machining operations, which involves a very high heat flux being generated during the chip generation process. Present study focuses on the thermal management of using a heat pipe in a drill for a drilling process. To check the feasibility and effectiveness of the heat pipe drill, structural and thermal analyses are performed using Finite Element Analysis. Finite Element Software ANSYS was used for this purpose. It is important for any conceptual design to be made practical and hence a parametric study was carried out to determine the optimum geometry size for the heat pipe for a specific standard drill.

Commentary by Dr. Valentin Fuster
2003;():103-108. doi:10.1115/HT2003-47149.

Following solidification, an aluminum alloy microstructure is highly segregated. The microstructure consists of cored dendrites with various soluble and insoluble phases present in the dendritic regions. The solidification rate has a marked effect on the amount of coring that an alloy experiences. Understanding the effects of the solidification rate is important in explaining differences in microstructures. Subsequent heat treatments are performed to homogenize the microstructure. The microstructure evolution after each processing step is dependent upon the previous microstructures. The variation in local chemical composition may promote or hinder precipitation of new phases. A large volume fraction of coarse insoluble phases can lead to the occurrence of recrystallized grains via particle stimulated nucleation, while inhomogeneous solute distribution can lead to the precipitation of an uneven distribution of dispersoid phases. The effect of solidification rate and subsequent thermal treatments on the microstructure of an Al-4Cu alloy will be investigated and experimental and numerical results will be presented.

Topics: Alloys
Commentary by Dr. Valentin Fuster
2003;():109-116. doi:10.1115/HT2003-47154.

Melting and re-solidification of the substrate plays an important role in thermal spray coating. A good understanding of this phenomenon will help us to achieve better bonding. A numerical model is developed to investigate the solidification of the droplet, and melting and re-solidification of the substrate. The solidification interface movement is obtained by applying a rapid solidification model on the solid/melt interface. Numerical simulations have been used to study the influence of materials and temperatures of the splat and substrate on substrate melting and re-solidification. In the corresponding experiments, the molybdenum powder is sprayed onto a stainless steel, brass (70%Cu) or aluminum substrate by atmospheric plasma spraying system. The crater depth of the substrates has been measured. Experimental results show that the material properties of the splat and substrate and melting temperature of the substrate play important roles on substrate melting and maximum melting depth. A dimensionless parameter, temperature factor, has been proposed from analysis and can be used as an indicator whether a substrate melting will occur for a certain combination of the droplet and substrate, and this parameter can be correlated with the maximum melting depth of the substrate.

Commentary by Dr. Valentin Fuster
2003;():117-125. doi:10.1115/HT2003-47180.

Buoyancy plays a detrimental role in chemical vapor deposition reactors employed for thin film deposition. Buoyancy driven fluid flow causes complex flow patterns which alter the transport of the precursor gases to the substrate, and leads to nonuniform deposition patterns. Consequently, many CVD reactors operate under low pressure to mitigate these flow patterns. However, the growth rates at such pressures are relatively low. Operating a CVD reactor under vacuum conditions is also inconvenient because of the associated hardware that is required. In the present work, we have numerically explored the performance of a new type of stagnation flow CVD reactor at pressures close to atmospheric pressure. The new geometry resembles that of a pancake reactor, but the gases are supplied through a long vertical inlet. The annular wall above the substrate is maintained at a low temperature to avoid deposition on this surface. The substrate is also rotated to improve the hydrodynamic patterns and provide azimuthal symmetry. We report results of a number of high-resolution calculations in this reactor to demonstrate its merits for operation at sub-atmospheric and atmospheric pressures. It is shown that the growth rate is significantly large, in addition to a high degree of film uniformity.

Commentary by Dr. Valentin Fuster
2003;():127-133. doi:10.1115/HT2003-47199.

A model of a partially ionized, high pressure plasma in stagnation flow as it melts a nonhomogeneous solid is presented. It encompasses both the analysis of the multi-fluid plasma to ascertain its bulk temperature and the heat flux profile, as well as its interaction with a receding melt interface in and around the stagnation domain. The model examined in this study couples the plasma motion, bulk energy, electron and ion densities and temperatures, with impinging jet theory to determine the amount of heat transfer into the particular substrate material — soil. “Multi-fluid” equations are derived for an axially symmetric plasma from the Boltzmann equations for Maxwellian velocity distributions. By examining the dominant effects, the equations are scaled and the roles of the driving dimensionless parameters are established. For specified values of these parameters, various numerical methods are used to couple and solve the two distinct models. The first one, to ascertain the moving boundary phase change heat transfer characteristics, is developed by adopting a form of the enthalpy method. The second model, characterizing the plasma jet is solved via and adaptation of the commercially available code, CHEMKIN, developed by the Sandia National Laboratories. A parametric study is performed, leading to evaluation of such important torch characteristics including mass flow rate of the Argon gas, temperature of the plasma bulk, and proximity of the plasma torch to the surface, as it influences the substrate melt zone. The extremely high temperatures produced by the plasma irreversilby changes the material structure of the sample. This new structure, when cooled, forms a predominantly glassy product. Such a vitrification process has been proven to improve the construction properties of the soil and to reduce a toxic sample of the soil into a leachable solid. From the calculations of solid/liquid interfacial location, radii of the melt zones, and depths of the melt zones an overall perspective of the vitrification process is assessed.

Commentary by Dr. Valentin Fuster
2003;():135-142. doi:10.1115/HT2003-47215.

This paper reports an experimental investigation on irregular interface morphology patterns developed in thin-film unidirectional solidification of pure succinonitrile. Solidification experiments have been conducted under various temperature gradients and interface velocities. Several irregular patterns have been observed including titled dendrites, degenerate dendrites, and seaweed. It is found that as the temperature gradient increases, steady titled dendrites may evolve into degenerate dendrites and eventually to seaweed. With the same grain orientation, the irregular patterns may transform from one form to another as the growth condition changes. Observations demonstrate that normal dendrites exhibit a higher growth rate than seaweed pattern and would overgrow them. Irregular pattern may also become strongly dynamic and different patterns may evolve into each other during growth within the same experiment. These results should shed a light into the understanding of the interface morphology development during solidification.

Topics: Solidification
Commentary by Dr. Valentin Fuster
2003;():143-150. doi:10.1115/HT2003-47221.

Laser surface coating of Mo, WC and Mo-WC powders on the surface of Ti6Al4V alloys using a 2kW Nd-YAG laser was performed. The dilution effect, microstructure, microhardness and wear resistance of the fabricated MMC coating were investigated. With a constant thickness of pre-placed powder, the dilution levels of the alloyed layers were found to be increased with the incident laser power. The fabricated MMC layer was metallurgically bonded to the Ti6Al4V substrate. The microhardness of the fabricated surface layer was found to be inversely proportional to the dilution level. The EDAX and XRD spectra results show that new intermetallic compounds and alloy phases were formed in the laser fabricated layer. With increasing weight percentage content of WC particles in the Mo-WC pre-pasted powder, the microhardness and sliding wear resistance of the laser surface coating were increased by 87% and 150 times respectively as compared with the Ti6Al4V alloy.

Commentary by Dr. Valentin Fuster
2003;():151-156. doi:10.1115/HT2003-47222.

This paper presents a comprehensive experimental study on the thermal aspects in resistance welding of thermoplastic composites. A special test set-up was developed to perform the experiments. Glass fiber reinforced polyetherimide was the material used for manufacturing the welding specimens. Stainless steel mesh was used for production of heating elements. The temperature distribution was monitored using type-K thermocouples connected to a data acquisition system. The main objective of the study was investigating a possible solution for the edge effect. Temperature profiles over the weld length and over the weld width were monitored. The focus was on the transient temperature profiles at the edges of the weld. The temperature distribution through the weld thickness was also monitored. The influence of factors like insulation and power level was investigated. Finally, conclusions are drawn and options for improving the temperature distribution and modification of the models are being discussed.

Commentary by Dr. Valentin Fuster
2003;():157-160. doi:10.1115/HT2003-47241.

This experimental study focused mainly on the solidification of a binary mixture of ammonium chloride and water (NH4 Cl-H2 O) in a differentially heated cavity. One vertical wall is cooled at temperature TC , and the opposite vertical wall is kept at constant temperature TH = +20°C. The effect on the solidification process of the initial concentration of ammonium chloride and cooling conditions is examined. Particle image velocimetry (PIV) is used for the visualization of the dynamic field during the solidification process. The temperature distribution at discrete locations in the solution and on the vertical cooling wall was monitored using thermocouples. The convection flow patterns, the ice thickness, and the temperature distribution were obtained for various initial concentrations of ammonium chloride ranging from 0wt% to 20wt% (sub-eutectic and near-eutectic growth). The convection patterns obtained for different initial concentrations showed significant differences. The results showed that the process of solidification is slower with an increase in the initial concentration levels of the binary solution. The ice growth rate was almost double at the bottom of the cavity.

Commentary by Dr. Valentin Fuster
2003;():161-171. doi:10.1115/HT2003-47261.

A study of the outside vapor deposition (OVD) process for the manufacture of fiber optic sleeve tubes is presented. High purity silicon dioxide (SiO2 ) is deposited on the outside of a rotating substrate via flame hydrolysis of silicon tetrachloride (SiCl4 ). Three double-flame burners hydrolyze the precursor forming streams of nominal 50 nm particulate, which are driven to the substrate surface via thermophoresis. The partial-premix burners utilize two concentric combustion chambers to provide fine control of the hydrolyzation process and heat flux to the preform surface. The bulk average deposition rate and efficiency to create a full-size sample are 4.93 gm/min/burner and 28%, respectively. The peak surface temperatures hover around 980 deg C at the bare quartz substrate surface, but then rapidly increase to 1200 deg C as the first four layers of SiO2 are deposited. These peak surface temperatures then monotonically decrease as the circumference and surface area of the porous preform increase. Similarly, the SiO2 layer density is 0.96 gm/cm3 at the substrate surface, but then decreases to 0.28 gm/cm3 as the porous preform grows to a diameter of 174 mm.

Commentary by Dr. Valentin Fuster
2003;():173-182. doi:10.1115/HT2003-47289.

Melting of a subcooled powder bed with the finite thickness that contains a mixture of two metal powders with significantly different melting points is investigated analytically. Shrinkage induced by melting is taken into account in the physical model. The temperature distributions in the liquid and solid phases were obtained using an exact solution and an integral approximate solution, respectively. The effects of porosity, Stefan number, and subcooling on the surface temperature and solid-liquid interface are also investigated. The present work built solid foundation to investigate the complex three-dimensional selective laser sintering (SLS) process.

Commentary by Dr. Valentin Fuster
2003;():183-188. doi:10.1115/HT2003-47295.

Directed Metal/Material Deposition (DMD) process is one of additive manufacturing processes based on laser cladding process. A full understanding of laser cladding process is a must to make the DMD process consistent and robust. A two dimensional mathematical model of laser cladding was developed to understand the influence of fluid flow to the mixing, dilution, and deposition dimension, incorporating melting, solidification, and evaporation phenomena. The fluid flow in the melt pool driven by thermal capillary convection and energy balance at liquid-vapor and solid-liquid interface was investigated and the impact of the droplets on the melt pool shape and ripple was studied. Dynamic motion, development of melt pool and the formation of cladding layer were simulated.

Commentary by Dr. Valentin Fuster
2003;():189-198. doi:10.1115/HT2003-47298.

A full 3-D transient model is developed for the ablation phenomena and thermal stress evolution during laser cutting and/or drilling of ceramic plates. The computational methodology is based on the Galerkin finite element method along with the use of a fixed grid algorithm to treat the thermal ablation resulting from an applied laser source. The present model is able to model any complex ablation operations involving discontinuity in geometries, as encountered in laser cutting and laser drilling operations. This is an advantage over the front tracking method by which the ablation moving interface is precisely tracked in time and which is useful for simple geometries. The laser ablation model is coupled with a thermal stress model to predict the evolution of thermal stresses, which arise due to a rapid change in thermal gradient near the laser beams. Model predictions compare well with the available data in literature for a simple configuration. Results obtained from model for both dual pulsed laser cutting and single laser drilling are discussed.

Commentary by Dr. Valentin Fuster
2003;():199-209. doi:10.1115/HT2003-47308.

This paper presents a numerical procedure for achieving desired features of a melt undergoing solidification by applying an external magnetic field whose intensity and spatial distribution are obtained by the use of a hybrid optimization algorithm. The intensities of the magnets along the boundaries of the container are described as B-splines. The inverse problem is then formulated as to find the magnetic boundary conditions (the coefficients of the B-splines) in such a way that the gradients of temperature along the gravity direction are minimized. For this task, a hybrid optimization code was used that incorporates several of the most popular optimization modules; the Davidon-Fletcher-Powell (DFP) gradient method, a genetic algorithm (GA), the Nelder-Mead (NM) simplex method, quasi-Newton algorithm of Pshenichny-Danilin (LM), differential evolution (DE), and sequential quadratic programming (SQP). Transient Navier-Stokes and Maxwell equations were discretized using finite volume method in a generalized curvilinear non-orthogonal coordinate system. For the phase change problems, an enthalpy formulation was used. The code was validated against analytical and numerical benchmark results with very good agreements in both cases.

Commentary by Dr. Valentin Fuster
2003;():211-217. doi:10.1115/HT2003-47340.

Existing models for the solute redistribution during solidification have been reviewed. The typical models are applied for the numerical simulation of heat and mass transfer with phase change for the new inverse casting technology and the traditional CSP technology. The results show that the effect of micro mass transfer models on the perfection of continuous solidification processes for Fe-C alloy cannot be omitted for coupled heat and mass transfer phenomenon.

Commentary by Dr. Valentin Fuster
2003;():219-226. doi:10.1115/HT2003-47371.

Hydrothermal solution growth is an important technique to grow high quality piezoelectric single crystals. In industry hydrothermal crystal growth, an autoclave is divided into two chambers by a baffle located in the middle height. Industrial practice found that better quality crystals could be grown under certain baffle hole openings or using a multi-hole baffle. This paper presents a numerical study of the effects of the baffle opening, as well as the arrangement of holes on the baffle, on the fluid flow and temperature fields in an industry-size autoclave. A wide range of baffle hole openings from 2% to 25%, together with five hole-arrangements, is investigated. Computational results indicate that changing the baffle hole opening and number of holes on the baffle are effective ways to control the temperature uniformity in the upper growing chamber. With a single-hole baffle, a smaller hole-opening leads to a weaker flow field and more uniform temperature in the growing chamber. With the same opening area, a multi-hole baffle will perform better than a single-hole baffle. The number of holes in a multi-hole baffle shows a strong effect on thermal condition in the upper chamber with 8-hole baffles working better than both baffles with 4 and 16 holes. The hole-arrangement, however, does have significant effect on thermal condition in the growing chamber.

Commentary by Dr. Valentin Fuster
2003;():227-228. doi:10.1115/HT2003-47382.

A high voltage cable consists of a conductor (typically made of either aluminium or copper) which is covered by several layers of other materials with different properties and functions (Thue 1999).

Commentary by Dr. Valentin Fuster
2003;():229-237. doi:10.1115/HT2003-47403.

This article presents the phenomena of melt flow, heat transfer, and solidification in Czochralski (CZ) melt growth processes of optical crystals, with emphasis on the effect of internal radiative heat transfer on the temperature distributions in oxide melt and crystal, melt convection, and melt-crystal interface shape. An integrated numerical model has been developed for simulating the physical phenomena in generic CZ furnaces, which includes the models for electromagnetic induction in crucible, surface exchange radiation in furnace, internal radiation in semi-transparent oxide melt and crystal, Marangoni convection in the melt, and solidification. Each developed model compares well with available analytical solutions. Numerical simulations were carried out for the prediction of fluid flow and heat transfer in furnaces. The simulation results show that the variation in optical properties of melt and crystal strongly impact their temperature distributions. It also affects the melt flow profile and intensity. The interface shape becomes more deeply convex toward the melt, as the optical thickness of the melt increases. However, the optical thickness of the crystal exhibits a minor impact on the interface shape. The results also show that the natural convection is dominated in the melt and the Marangoni flow enforces the natural convection.

Topics: Crystals
Commentary by Dr. Valentin Fuster
2003;():239-249. doi:10.1115/HT2003-47424.

A rigorous electromagnetic model is developed to predict the radiative properties of patterned silicon wafers. For nonplanar structures with characteristic length close to the wavelength of incident radiation, Maxwell’s equations must be used to describe the associated radiative interaction and they are solved by the finite difference time-domain (FDTD) method. In the die area, only one period of the structure is modeled due to its periodicity in geometry. To truncate a computational domain, both the Mur condition and perfectly matched layer (PML) technique are available to absorb outgoing waves. With the steady state time-harmonic electromagnetic field known, the Poynting vector is used to calculate the radiative properties. Due to its importance, the reflection error is checked at first for two absorbing boundary conditions. As expected, the PML technique yields much lower errors than the Mur condition and it is thus used in this study. To validate the present model, radiative interactions with a planar structure and a nonplanar structure are investigated, and predicted reflectivities are found to match available other solutions very well. To demonstrate the importance of the present study, a patterned wafer consisting of periphery and die area is also investigated. While the thin film theory is accurate for the wafer periphery, the rigorously electromagnetic model described in this study is found to be necessary to accurately predict the radiative properties in the die area.

Commentary by Dr. Valentin Fuster
2003;():251-258. doi:10.1115/HT2003-47425.

In liquid composite molding technologies such as Resin Transfer Molding (RTM), a thermoset resin is injected into a mold cavity with a pre-placed preform made of fiber mats to create a cured part. In order to improve the physics of resin flow in dual-scale (woven, stitched or braided) fiber mats, the authors carried out many transient 1-D mold-filling experiments to investigate the onset of unsaturated flow through the inlet-pressure history. Their study revealed that the measured pressure history, which droops downwards for dual-scale fiber mats, is at a variance with the linear pressure profile predicted by state-of-the-art Liquid Composite Molding (LCM) mold-filling simulation physics. It was also observed that the drooping of the inlet pressure increases with an increase in the compression of fiber mats. In this paper, the correlation between a previously proposed dimensionless number pore volume ratio and the droop in the inlet pressure history has been sought. Studying the micrographs of composite samples, pore volume ratio is measured for various fiber mat compression. It is observed that the droop in the inlet pressure profiles increase with an increase in the pore volume ratio. This is the first attempt to quantitatively validate the previous theories on the unsaturated flow.

Commentary by Dr. Valentin Fuster
2003;():259-268. doi:10.1115/HT2003-47445.

In this study, the force required to draw a polymer preform into optical fiber is predicted and measured, along with the resultant free surface shape of the polymer, as it is heated in an enclosed cylindrical furnace. The applied drawing force affects the degree of chain alignment within the polymer. Chain alignment causes orientational birefringence, an unwanted property that attenuates any propagating optical signal. The draw force is a function of the highly temperature dependent polymer viscosity. Therefore accurate prediction of the drawing force requires a detailed investigation of the heat transfer within the furnace. In this investigation, the full axi-symmetric conjugate problem (including both natural convection and thermal radiation) was solved using the commercial finite element package FIDAP. In addition, the location of the polymer/air interface was solved for as part of the problem and was not prescribed beforehand. Results show that thermal radiation accounts for approximately 70% of the total heating experienced by the deforming polymer, but only 15% of the cooling. The draw force is very sensitive to both the furnace wall temperature and to the feed rate of the polymer. Numerical results compared well with the experimentally measured draw tension and neck-down profiles for several preform diameters, draw speeds, and furnace temperatures. The predicted draw forces were typically within 20% of the experimentally measured values.

Commentary by Dr. Valentin Fuster
2003;():269-274. doi:10.1115/HT2003-47446.

The feasibility of using pyrolytic Laser-Induced Chemical Vapor Deposition (LCVD) to deposit carbon coatings on moving fused quartz rods have been investigated in this study. This LCVD system uses a CO2 laser to locally heat substrates in open air to create a hot spot. Pyrolysis of hydrocarbon species occurs and subsequently deposits a layer of carbon film onto the substrate surface. The results of this study indicate that the deposition rate of carbon film increases exponentially within the range of laser power, while an increase in traverse velocity of the substrate will also increase the deposition rate until a maximum deposition rate is reached, and further increases in the traverse velocity will decrease the deposition rate. We suspect that this optimal deposition rate is caused by substrate motion, which affects the substrate surface temperature, and consequently the effective surface area available for film deposition.

Commentary by Dr. Valentin Fuster
2003;():275-281. doi:10.1115/HT2003-47453.

Impinging jet is widely used in both traditional industrial and new high-tech fields. High efficiency heat transfer in impinging jet cooling makes it an important method for heat transfer enhancement, in particular in cooling of electronic devices with high heat density. This paper presents an experimental study of heat transfer by an impinging circular water jet. A Constantan foil with the size of 5 mm × 5 mm was used to simulate a microelectronic chip with heat generated by passing an electrical current through the foil. A high heat flux over 106 W/m2 was achieved. The surface temperature was measured by a thermocouple glued onto the back surface of the foil. Both a free surface jet and a submerged jet were investigated. Effect of the nozzle-to-surface spacing as well as the jet speed at the exit of the nozzle on cooling was examined. By positioning the jet away from the center of the heating foil surface, the radial variation of the heat transfer coefficients over the foil was also investigated. Quantitative heat transfer data have been obtained and analyzed.

Topics: Heat transfer , Water
Commentary by Dr. Valentin Fuster
2003;():283-288. doi:10.1115/HT2003-47470.

An investigation was conducted to understand the contact line movement and associated contact angle phenomena. Contact line was supposed to move on a thin precursor film caused by molecular interaction between solid and liquid and asperity of solid surface. It is expected that contact line has a velocity and is subject to viscous stress on the film or geometrically on the solid surface. With the introduction of a characteristic parameter, λ′, the movement of contact line and contact angle phenomena were very well described in both physics and mathematics. The viscous shearing stress exerted by liquid on solid surface was derived, and the behavior of dynamic contact angle was recognized on rough solid surfaces. The analyses indicate that characteristic parameter, λ′, is dependent upon solid wall intrinsic property and mechanical performance, not liquid property. The comparison of theoretical predictions with available experimental data in open literature showed a quite good agreement with each other.

Commentary by Dr. Valentin Fuster
2003;():289-294. doi:10.1115/HT2003-47477.

This paper presents an experimental study of copper dissolution in molten tin and tin-silver (Sn-Ag) solders and the formation and presence of the Cu-Sn intermetallic compound at solder/copper interfaces. During the experiments, copper (99.9% pure) samples, coated with a RMA flux, were dipped vertically in a molten solder for different time periods ranging from 5 seconds to 10 minutes. The molten solder was maintained at temperatures of 232°C, 250°C and 300°C for pure tin and 221°C, 250°C, and 300°C for Sn-3.5%Ag respectively. The samples were then cut, cleaned and cold mounted in epoxy at ambient temperature. Mechanical grinding, finish polishing, etching, and optical metallographic procedures were utilized for examining the microstructures of the polished and etched samples. The average thickness of the intermetallic compound and the amount of copper dissolved was determined. Experimental results indicate the temperature of molten solder to control the rate of dissolution of copper and the formation and presence of intermetallic compounds at the interfaces. At a given temperature of the solder temperature, the rate of dissolution of copper in the solder revealed a rising trend with an increase in dwell time of copper in the solder. For short contact time periods, the dissolution rate is low and the thickness of the intermetallic compound is small. With an increase in dwell time, the dissolution rate of copper rapidly increases and eventually reaches a plateau. Initiation of dissolution of copper causes a layer of the Sn-Cu intermetallic compound to form around the copper substrate. This in turn prevents direct contact of the copper substrate with the molten solder. The rate of formation of the layer of intermetallic compound reveals a similar trend. Based on experimental results, the kinetic parameters involved in governing the growth of the intermetallic were determined for the two solders. The parameters can be used to estimate the kinetics of copper dissolution and intermetallic compound formation during soldering.

Commentary by Dr. Valentin Fuster
2003;():295-302. doi:10.1115/HT2003-47497.

Hydrothermal growth is the most common technique to grow piezoelectric single crystals such as quartz. Due to a high-temperature and high-pressure growth condition, hydrothermal autoclaves are designed to operate as a closed system. During operation, the only control mechanism that crystal growers have is adjusting the power input of the heaters, based on the temperature readings obtained by the thermocouples along the centerline inside the autoclaves. The power adjusting process, however, is purely experience dependent, and, normally, uniform heating conditions from electric heaters are employed along the autoclave wall. This study develops an inverse algorithm, with which the required heat flux distributions from the heaters can be obtained for a desired growth environment inside an autoclave. The algorithm involves solving three sub-models step by step. The first step is to solve a two-dimensional axisymmetric model of solution in the autoclave to obtain the temperature and heat flux on the solution/wall interface. Using these temperature and heat flux conditions as thermal boundary conditions, the second step solves an inverse heat conduction problem in the metal wall. The solution provides the heat flux and temperature on the outer surface of the metal wall. The final step is to solve a heat conduction problem in the insulation layer to obtain the heat flux on the inner surface of the insulation layer. The heat flux distributions for heaters are then determined by the heat flux on the outer surface of the metal wall and heat flux on the inner surface of the insulation layer. The paper describes the details of each model. As an example, the method is used to find the required heat flux distributions of heaters for the growth environment predicted by a 2-D isothermal wall model. The result is then used to develop a two-patch heater for industry autoclaves.

Topics: Heating
Commentary by Dr. Valentin Fuster
2003;():303-312. doi:10.1115/HT2003-47502.

A comprehensive axisymmetric model of the coupled thermal-electrical-mechanical analysis predicting weld nugget development and residual stresses for the resistance spot welding process of Al-alloys is developed. The model estimates the heat generation at the faying surface, the workpiece-electrode interface, and the Joule heating of the workpiece and electrode. The phase change due to melting in the weld pool is considered. The contact area and its pressure distribution at both the faying surface and the electrode-workpiece interface are determined from a coupled thermal-mechanical model using a finite element method. The knowledge of the interface pressure provides accurate prediction of the interfacial heat generation. For the numerical model, temperature dependent thermal, electrical and mechanical properties are used. The proposed model can successfidly calculate the nugget diameter and thickness, and predict the residual stresses and the elastic-plastic deformation history. The calculated nugget shape and the deformation of sheets based on the model are compared with the experimental data. The computed residual stresses approach the distribution of experimental measurement of the residual stress.

Commentary by Dr. Valentin Fuster
2003;():313-316. doi:10.1115/HT2003-47503.

The ability to predict and control microstructure in laser deposited materials requires an understanding of the thermal conditions at the onset of solidification. To this end, the focus of this work is the development of thermal process maps relating solidification cooling rate and thermal gradient (the key parameters controlling microstructure) to laser deposition process variables (e.g., laser power and velocity). Results presented herein are based on the Rosenthal solution for a moving point heat source traversing an infinite substrate. Ongoing work includes the effect of a distributed laser power through superposition of the Rosenthal solution, while the effects of finite geometry, temperature dependent properties and latent heat of transformation are included through thermal finite element modeling of the laser deposition process.

Topics: Lasers
Commentary by Dr. Valentin Fuster
2003;():317-323. doi:10.1115/HT2003-47504.

As an innovative technique, the lost foam casting (LFC) process has drawn great attention from both academia and industry in recent years. The key feature of LFC process is that a desired shape pattern made of expandable polystyrene (EPS) foam is buried in unbonded sand and replaced by advancing molten metal. The heat and mass transfer between the molten metal front and the EPS foam pattern plays an important role in the soundness of the product in the LFC process. The present study focuses on determining the characterization of heat and mass transfer during the EPS pattern degradation process. A unique experimental system using a cylindrical quartz window and heated steel block simulating the hot molten metal front has been constructed to make measurements and visualize the process. The foam pattern is 88 mm in diameter and 254 mm long. It is coated twice with DCH Ashland refractory material and the average coating thickness is 1.2 mm. The heat flux and pressure between the moving steel block and the EPS pattern are measured. The process variables studied during this experiment include foam density and steel block speed. It was found that unlike the fluidity of the molten metal which is highly dependent on the density of the foam patterns, foam density has marginal effect on the heat flux from the steel block to the foam pattern. The heat flux increases about 37% during a one-minute process under steel block velocity of 4.4 mm/s using different EPS foam density of 24 kg/m3 and 27 kg/m3 . Flow visualization shows a gaseous gap formed between the steel block and the foam pattern. The phase change and degradation of EPS foam pattern and the heat and mass transfer in the gap are crucial to characterize the mold filling process which decides the quality of casting products. The maximum pressures measured in the gap using steel block velocity of 4.4 mm/s are 1.1 kPa and 1.4 kPa for EPS foam density of 24 kg/m3 and 27 kg/m3 , respectively. Under a slower steel block velocity of 3.6 mm/s the gap peak pressure using 24 kg/m3 density EPS foam pattern is 0.43 kPa. It is concluded that higher foam density and faster steel block speed give rise to larger gas pressure between the steel block and foam pattern. The measured pressure values confirm data reported in literature.

Commentary by Dr. Valentin Fuster
2003;():325-334. doi:10.1115/HT2003-47511.

Atmospheric pressure dc plasma microdischarges were generated between a thin cylindrical electrode and a flat surface. The glow discharges were characterized both experimentally and numerically. As the applied voltage is increased, the plasma potential in the plasma microdischarge remains constant while the plasma current increases proportionally. For the types of electrode materials and the range of the electrode diameters considered, the current-voltage characteristics of the plasma microdischarge are almost independent of the electrode material and the diameter of the electrodes. Predictions from a numerical model of the argon microdischarge compare favorably with the experimental measurements and the visualization studies. The atmospheric pressure plasma microdischarge can be used for deposition of thin films and for microfabrication.

Commentary by Dr. Valentin Fuster
2003;():335-341. doi:10.1115/HT2003-47519.

A significant amount of aluminum is processed by melting secondary aluminum that contains small amounts of magnesium. A major drawback of aluminum production in secondary melt furnaces is the formation of dross or aluminum oxide by the oxidation of the molten metal. Since aluminum scrap forms a major source of the metal in secondary aluminum processing, the presence of alloying elements plays a key role in the oxidation process. Here, we consider the early stage during the oxidation of Al-Mg alloys during which the primary oxidation is that of magnesium to magnesium oxide occurs. We have simulated the processes in an aluminum melting furnace and considered the metal oxidation to be limited by one–dimensional diffusion. Our results predict the temporal variation of the oxygen distribution and the rate of metal evaporation and formation of the metal oxide. The effects of melt composition, gas temperature and oxygen concentration in the gas are discussed.

Commentary by Dr. Valentin Fuster
2003;():343-351. doi:10.1115/HT2003-47534.

In order to control the impurity distribution and remove defects in a crystal grown in Czochralski growth for high quality crystals of silicon, it is necessary to study and control the melt-crystal interface shape, which plays an important role in control of the crystal quality. The melt-crystal interface interacts with and is determined by the convective thermal flow of the melt in the crucible. Application of magnetic field in the Czochralski system is an effective tool to control the convective thermal flow in the crucible. Therefore, the shape of the melt-crystal interface can be modified accordingly. Numerical study is performed in this paper to understand the effect of magnetic field on the interface deflection in Czochralski system. Comparisons have been carried out by computations for four arrangements of the magnetic field: without magnetic field, a vertical magnetic field and two types of cusp-shaped magnetic field. The velocity, pressure, thermal and electromagnetic fields are solved with adaptation of the mesh to the iteratively modified interface shape. The multi-block technique is applied to discretize the melt field in the crucible and the solid field of silicon crystal. The unknown shape of the melt-crystal interface is achieved by an iterative procedure. The computation results show that the magnetic fields have obvious effects on both the pattern and strength of the convective flow and the interface shape. Applying magnetic field in the Czochralski system, therefore, is an effective tool to control the quality of bulk crystal in Czochralski growth process.

Commentary by Dr. Valentin Fuster
2003;():353-354. doi:10.1115/HT2003-47564.

In SiC vapor growth, micropipes and dislocations that originate at the seed/boule interface can continuously propagate into the newly grown crystal. They adversely affect the quality of the crystals. The defect density can be reduced by the method of growing a large diameter crystal from a small seed through lateral growth under controlled thermal environment. In this paper, SiC growth processes with varying thermal conditions have been simulated. The effects of operational parameters such as axial and radial temperature gradients, and the presence of polycrystal are also investigated. The current finding can also help in the design of AlN/GaN growth system.

Commentary by Dr. Valentin Fuster
2003;():355-359. doi:10.1115/HT2003-47585.

An experimental study was prepared to find the relationship between Leidenfrost temperature and droplet size and velocity of impinging jets. The study is done for the case of steel surface cooling with two-phase nozzles. The sprayed surface moves under the spray at a velocity of 1 m/min. Cooling experiments were done for initial temperature of 1250°C. Thermal experiments are transient: internal temperature is measured and surface temperature and heat transfer coefficient distribution is computed by the inverse task. Droplet size and velocity of the impinging jet was modified by setting water and air pressures at the input of the nozzle. Spray parameters for each pressure combination was measured using a laser-doppler anemometer. The paper shows the results of the thermal and fluid flow experiment and the correlation between Leidenfrost temperature and flow parameters.The application of obtained results is expected for high temperature cooling especially in continuous casting.

Commentary by Dr. Valentin Fuster
2003;():361-365. doi:10.1115/HT2003-47586.

This is a novel application of Computational Fluid Dynamics (CFD), in the vacuum De-zincing process. The complete modeling process would involve the solution of the following equations: a) Navier-Stokes Equations; b) The Energy Equation; c) The Solution of the Species Concentration. The aim of this research as a novel approach in vacuum Dezincing process has been to gain an understanding in terms of the actual complicated physics involved de-zincing process such, as phase change and solidification. The results in this paper have contributed to a better understanding of the vacuum De-zincing process, hence identifying parameters, which would aid the efficient recovery of the zinc from the molten metal bath.

Commentary by Dr. Valentin Fuster
2003;():367-372. doi:10.1115/HT2003-47596.

Molecular Dynamics (MD) and Finite Difference (FD) methods are applied to investigate femtosecond laser ablation of copper. Laser induced heat transfer, melting, vaporization, and material ablation are studied. Phase change relevant parameters, such as the velocity of solid-liquid and liquid-vapor interfaces are reported. It is shown that results of MD and FD are in good agreement before strong material removal occurs. However, only MD is capable of capturing the ablation phenomena accurately.

Commentary by Dr. Valentin Fuster

Heat Transfer in Electronic Equipment

2003;():373-378. doi:10.1115/HT2003-47014.

A network method for quickly calculating the temperature distributions in an LSI chip with silicon-on-insulator (SOI) transistors and multi-layered lines has been developed. Its calculation time is less than 1/1000 of that of the finite element method, and its error is within 15%. The developed fast calculation method can be used in the case of more than 300 heating devices and more than 1000 lines in an LSI chip. It is thus a practical tool for designing the optimum layout of devices to prevent local temperature increases in an LSI chip.

Commentary by Dr. Valentin Fuster
2003;():379-387. doi:10.1115/HT2003-47015.

A study of thermal management of a harsh environment power electronics system is presented. The thermal environments were found to be between 65 °C and 90 °C that is considerably higher than many traditional electronics applications. A modular, low cost, and passive air-cooling system was desired. An analytical model was developed to obtain the heat transfer characteristics. Further performance verification of the thermal management solution was completed using a commercially available CFD tool. A small footprint area for thermal design of the power electronics connected with an electrically isolating low-conductivity material to the heat sink increased the challenge. A further thermal performance enhancement was achieved with the addition of a heat spreader between power electronics and the heat sink, and optimization of the heat spreader was achieved by utilizing FEM technique.

Commentary by Dr. Valentin Fuster
2003;():389-395. doi:10.1115/HT2003-47041.

This paper summarizes the practical use of CFD (Computational Fluid Dynamics) using a commercially available package, FLOTHERM [1], in a tight and highly competitive marketplace to produce a functional pre-production piece of telecom gear with no prototyping for thermal issues. The paper highlights the direct production, noprototype, analytical thermal performance verification of a small CMTS (Cable Modem Termination System) used in telecom applications.

Commentary by Dr. Valentin Fuster
2003;():397-410. doi:10.1115/HT2003-47050.

Knowledge of flow pattern and flow pattern transitions is essential to the development of reliable predictive tools for pressure drop and heat transfer in two-phase micro-channel heat sinks. In the present study, experiments were conducted with adiabatic nitrogen-water two-phase flow in a rectangular micro-channel having a 0.406 × 2.032 mm cross-section. Superficial velocities of nitrogen and water ranged from 0.08 to 81.92 m/s and 0.04 to 10.24 m/s, respectively. Flow patterns were first identified using high-speed video imaging, and still photos were then taken for representative patterns. Results reveal that the dominant flow patterns are slug and annular, with bubbly flow occurring only occasionally; stratified and churn flow were never observed. A flow pattern map was constructed and compared with previous maps and predictions of flow pattern transition models. Annual flow is identified as the dominant flow pattern for conditions relevant to two-phase micro-channel heat sinks, and forms the basis for development of a theoretical model for both pressure drop and heat transfer in micro-channels. Features unique to two-phase micro-channel flow, such as laminar liquid and gas flows, smooth liquid-gas interface, and strong entrainment and deposition effects are incorporated into the model. The model shows good agreement with experimental data for water-cooled heat sinks.

Commentary by Dr. Valentin Fuster
2003;():411-431. doi:10.1115/HT2003-47051.

The thermal contact resistance (TCR) problem is categorized into three different problems: geometrical, mechanical, and thermal. Each problem includes a macro and micro scale sub-problem; existing theories and models for each part are reviewed. Empirical correlations for microhardness, and the equivalent (sum) rough surface approximation are discussed. Suggested correlations for estimating the mean absolute surface slope are summarized and compared with experimental data. The classical conforming rough contact models, i.e elastic and plastic, as well as elastoplastic models are reviewed. A set of scale (dimensionless) relationships are derived for the contact parameters, i.e. the mean microcontact size, number of micro-contacts, density of microcontacts, and the external load as functions of dimensionless separation, for the above models. These scale relationships are plotted; it is graphically shown that the behavior of these models, in terms of the contact parameters, are similar. The most common assumptions of existing thermal analysis are summarized. As basic elements of thermal analysis, spreading resistance of a circular heat source on a half-space and flux tube are reviewed, also existing flux tube correlations are compared. More than 400 TCR data points collected by different re-searchers during last forty years are grouped into two limiting cases: conforming rough, and elasto-constriction. Existing TCR models are reviewed and compared with the experimental data at these two limits. It is shown that the existing theoretical models do not cover both of the above-mentioned limiting cases.

Commentary by Dr. Valentin Fuster
2003;():433-443. doi:10.1115/HT2003-47078.

The present study investigates temperature distribution and heat transfer characteristics of synchronous generators to attain more efficient design and better cooling performance by using thermal network analysis and CFD. Under all circumstances, the temperature of rotor and stator windings must be kept below the maximum permissible temperature of insulations to maintain reliability and prolong durability of a machine. In this paper, the temperature rise of rotor winding, stator winding, and associated end-winding was calculated for some generators in a range of machine sizes. The temperature rise of the coolant at discrete points in its flow path was also predicted. Finally, the calculation results were compared with real test results. From the comparison, it was found that there was a good agreement between the analytical calculations and the experimental results.

Commentary by Dr. Valentin Fuster
2003;():445-452. doi:10.1115/HT2003-47124.

An experimental study of a pin fin heat sink was carried out in support of the development of heat sink optimization methods requiring more detailed measurements be made. Measurements of heat flux and temperature are used to separately determine heat transfer coefficients for the pins and the base region between the pins. Three pitch to diameter ratios (distance from pin center to pin center measured diagonally) were studied: P/d = 3/1, 9/4, 3/2. Heat generation was accomplished using cartridge heaters inserted into a copper block. The high thermal conductivity of the copper ensured that the surface beneath the heat sink would be at a constant temperature. The cooling fluid was air and the experiments were conducted with a Reynolds numbers based on a porous media type hydraulic diameter ranging from 500 to 25000. The channel had a shroud that touches the fin tips, eliminating any flow bypass. The pin surface heat transfer coefficients match the values reported by Kays and London and by Zukauskas. The base region heat transfer coefficients were, surprisngly, larger than the pin values.

Commentary by Dr. Valentin Fuster
2003;():453-459. doi:10.1115/HT2003-47129.

Numerical prediction and experimental verification of the temperature rise for a single-phase and a three-phase gas-insulated bus bar with current flow are investigated. Various heat generation rates possibly produced in the gas-insulated bus bar are calculated. To estimate the power loss caused by eddy current, the magnetic field analysis is carried out. The heat balance calculation solving the differential form of an energy balance equation with empirical relations is conducted by using the 5th order Runge-Kutta method. The various cases representing different geometries and current values are investigated by conducting the heat balance calculation. Three-dimensional numerical flow field analysis using finite volume method is performed for the different type of the bus bars. From the flow field analysis based on laminar natural convection, the temperature gradient in the current flowing direction caused by contact heat source is found for both single-phase and three-phase bus bars. In the experiments, temperature rises in each of conductor, contact part, and external tank are measured for a full-scale gas-insulated bus bar. The comparisons of the predicted values of the heat balance calculation and the numerical analyses to results of the experiments are made. From the comparisons, it is concluded that the temperature rise of a bus bar can be predicted quite well by performing laminar natural convection flow analyses.

Commentary by Dr. Valentin Fuster
2003;():461-468. doi:10.1115/HT2003-47162.

Impinging jets for cooling of electronic equipment have been used by many researchers. Only few studies using arrays composed of a small number of jets are available in the literature. When very small jet diameters are used, the jet Reynolds number becomes quite small and no data are available for Reynolds number values below 500. In this work attention has been focused on circular arrays of free surface micro jets. Experiments were conducted by employing three jet pitches, 1, 2 and 3 mm and four jet diameters 50, 100, 150 and 250 μm and two different fluids, DI water and FC 40. The jet Reynolds number range was varied between 90 and 2000 while the Prandtl number varied from 6 to 84. Heat fluxes as high as 250 W/cm2 could be removed when water was utilized. Experimental data have been correlated within ±20%.

Commentary by Dr. Valentin Fuster
2003;():469-476. doi:10.1115/HT2003-47245.

Spot-cooling of discrete electronic packages mounted on a printed wiring board (PWB) is achieved by the impingement of an axisymmetric synthetic jet when the jet actuator is attached to one board and cools target integrated circuit (IC) on the opposite board. Present work demonstrates that even when the jet outflow is confined between two closely-spaced boards, the jet entrains ample volume flow rate of cooler ambient air, induces effective cooling by strong mixing near the heated surface, and ultimately dispenses the heated air to ambient. The cooling performance of the jet module is investigated experimentally in a scaled up model that enables high-resolution thermal and flow measurements. The test setup comprises of two circular parallel plates (D = 158.8mm) where one plate contains an integrated jet actuator and the opposite plate includes a target heater (dh = 86mm). The spacing between the plates is variable between D/12 and D/3. The flow within the gap is mapped using particle image velocimetry (PIV). It is found that confined jet impingement induces a countercurrent radial flow within the gap that includes a layer of cool ambient air entrained along the actuator plane and a layer of heated air that is dispensed along the target surface. Particle pathlines demonstrate significant mixing between the countercurrent streams and strong entrainment into the vortex rings that synthesize the jet. Heat transfer coefficient attains a local maximum away from the stagnation point that can be attributed to strong vorticity diffusion into the thermal boundary layer and enhanced mixing that accompanies the vortex ring impingement. Although the jet Reynolds number does not exceed 3300, the stagnation heat transfer coefficient is about 90 W/m2 K for H/D = 0.32.

Topics: Cooling
Commentary by Dr. Valentin Fuster
2003;():477-487. doi:10.1115/HT2003-47282.

The flow modeling approaches employed in Computational Fluid Dynamics (CFD) codes dedicated to the thermal analysis of electronic equipment are generally not specific for the analysis of forced airflows over populated Printed Circuit Boards. This limitation has been previously highlighted [1], with component junction temperature prediction errors of up to 35% reported. This study evaluates the predictive capability of candidate turbulence models more suited to the analysis of electronic component heat transfer. Significant improvements in component junction temperature prediction accuracy are obtained, relative to a standard high-Reynolds number k-e model, which are attributed to better prediction of both board leading edge heat transfer and component thermal interaction. Such improvements would enable parametric analysis of product thermal performance to be undertaken with greater confidence in the thermal design process, and the generation of more accurate temperature boundary conditions for use in Physics-of-Failure based reliability prediction methods. The case is made for vendors of CFD codes dedicated to the thermal analysis of electronics to consider the adoption of eddy viscosity turbulence models more suited to board-level analysis.

Commentary by Dr. Valentin Fuster
2003;():489-494. doi:10.1115/HT2003-47287.

Recent developments in natural graphite composites by GrafTech International, Ltd. have led to novel highly conductive and lightweight materials with thermally non-isotropic structures [1]. The in-plane and out-of-plane thermal conductivities for such materials differ by a factor of up to 50. The non-isotropy in thermal conductivities is found to influence heat dissipation in different ways. This paper investigates the influence of thermal conductivity on the thermal performance of graphite heat sinks. In addition to addressing technology challenges encountered in design and optimization of heat sinks, this paper also discusses the efficient ways to utilize graphite composite materials, which can be expected to play an important role in future electronics cooling applications.

Commentary by Dr. Valentin Fuster
2003;():495-500. doi:10.1115/HT2003-47297.

Experiments of dielectric liquid spray mist cooling of a row of 9, 38 mm square heaters located along one side of a 400 mm channel formed by two adjacent parallel boards were carried out. The objective was to determine the heat removal capability of the approach as a function of heater downstream distance from a single atomizer for channel height (board pitch) and orientation effects. Fundamental characteristics of the spray field and heat transfer were sought and no attempt was made to optimize the system for high heat transfer rate. The fluid used was FC-72 which as delivered into the gap between the boards by a full-cone TG0.3 atomizer oriented parallel to the row of heaters. The results indicate that heat transfer in the upstream region is dominated by droplet impingement which depends on the board pitch. The heat removal at the leading heater is up to 4 W/cm2 at 60°C uniform heater temperature. The heat transfer rate in the downstream region is determined by the flow pattern of the liquid film on the boards, which in turn depends on droplet behavior between the boards. For the horizontal orientation, gravitational forces cause the droplets to fall onto the heated board and improve heat transfer in the middle region along the heaters. This effect was enhanced and heat transfer coefficient was improved with the increase of liquid temperature which is consistent with the phenomena observed in the experiments. For the vertical orientation, gravitational effects do not appear to have an effect on heat transfer. Flow pattern of liquid film in the downstream region is heavily affected by the liquid distribution in the upstream region in the vertical orientation. In this case, the liquid film pattern was preserved and shifted downstream when the power to the upstream heaters was turned off. The heat removal at the “leading” heater at a specific uniform heater temperature is shown to depend on the subcooling of liquid for both orientations.

Topics: Cooling , Lumber , Sprays
Commentary by Dr. Valentin Fuster
2003;():501-505. doi:10.1115/HT2003-47332.

The heat generated by microprocessors is steadily increasing as die size shrinks. The trend towards increasing microprocessor heat flux is forcing thermal engineers to consider alternative system cooling technologies. Greater thermal control is also an issue during the engineering test of microprocessors and package characterization. The thermal challenges posed during test and debug are significant as power levels are higher than system power and temperature control requirements are very tight. Thermo-electric cooling technology (TEC) integrated with liquid cooling has evolved significantly in the past few years as a thermal management technique for engineering test and debug. In this study, an experimental approach was taken to characterize a TEC-based thermal control unit (TCU). The TCU thermal resistance was calculated as a function of the water temperature and device heat outputs. Several tests evaluated the suitability of using the TCU to control a device at low and elevated package case temperatures. Test results indicated that the TEC-Liquid system can operate in a cold or hot mode and that the thermal management capability of the system is a strong function of the liquid bulk temperature on the hot side of the TEC surface. The water bulk temperature at which the TCU fails to maintain the required device temperature is also reported in this paper.

Commentary by Dr. Valentin Fuster
2003;():507-517. doi:10.1115/HT2003-47349.

A stable numerical procedure is developed to analyze the transient performance of flat heat pipes for large input heat fluxes and high wick conductivity. Computation of flow and heat transfer in a heat pipe is complicated by the strong coupling among the velocity, pressure and temperature fields with phase change at the interface between the vapor and wick. A structured collocated finite volume scheme is used in conjunction with the SIMPLE algorithm to solve the continuity, energy and momentum equations. In addition, system pressurization is computed using overall mass balance. The stability of the standard sequential procedure is improved by accounting for the coupling between the evaporator/condenser mass flow rate and the interface temperature and pressure as well as the system pressure. The improved numerical scheme is applied to a flat two-dimensional heat pipe and shown to perform well. Parametric studies are performed by varying the vapor core thickness of the heat pipe and the heat input at the evaporator. The model predictions are validated by comparing the heat pipe wall temperatures against experimental values.

Commentary by Dr. Valentin Fuster
2003;():519-527. doi:10.1115/HT2003-47358.

This paper examines the performance of a silicon nozzle on the droplet impingement cooling of a simulated electronic chip with a dielectric coolant. A silicon nozzle plate with 49 orifices on a 1 inch square area is bonded with swirl chips. A compact test beds with embedded cartridge heaters and thermocouples at specified locations, is built and integrated with a condenser to test the cooling performance at a given heat flux. Through the system testing, the importance of vapor/excess-liquid exits, orientation due to gravity, and the temperature control of recirculating liquid are revealed. The use of Swiss-roll micronozzles have provided up to 45 W/cm2 uniform heat removal at 23 cc/min cm2 liquid flux.

Commentary by Dr. Valentin Fuster
2003;():529-533. doi:10.1115/HT2003-47374.

Experiments have been carried out to investigate the convective heat transfer characteristics from triangular folded fin heat sinks in a suction-type fan duct. The dimension of the triangular folded fin heat sinks is 62 mm in height with a 12 mm thick base plate, 292 mm in width, and 447 mm in length. The inlet flow velocity is varied in the range of 0.6–5.3 m/s. Thermal performance of triangular folded fin heat sinks is evaluated in terms of thermal resistance of heat sinks according to flow velocity and fan power. The results obtained show that the present triangular folded-fin heat sink shows a higher thermal performance compared to a conventional extruded plate-fin heat sink. Especially, a perforated slit folded-fin heat sink displays a lower thermal resistance. As the number of slit fabricated on the perforated folded fins increases, thermal performance is more pronounced.

Topics: Ducts , Heat sinks
Commentary by Dr. Valentin Fuster
2003;():535-539. doi:10.1115/HT2003-47376.

The thermal interaction of an electrical and an optical component located on the same vertical circuit board is studied experimentally. The effects of component proximity and convective flow rate on overall power dissipation from each component are analyzed. The components are represented by isothermal heat sources mounted to a standard 1.59mm (0.0625 in) thick FR4 circuit board. In natural convection situations, when the spacing between components is great enough that the component thermal footprints do not interfere, the power dissipation reaches a maximum “plateau” value that is independent of spacing. If the components are located close enough together that their thermal footprints interfere then the total power dissipation is highly dependent on component spacing (relative location of the electrical source and the geometric positioning of both sources). In forced convection, the total power dissipated increases with both Reynolds number and component spacing. As in natural convection, the relative location of the electrical sources and the positioning of the sources are found to have a strong influence on power dissipation.

Commentary by Dr. Valentin Fuster
2003;():541-550. doi:10.1115/HT2003-47419.

Natural convection has important implications in many applications like cooling of electronic equipment due to its low cost and easy maintenance. In the present study, two-dimensional natural convection heat transfer to air from multiple identical protruding heat sources, which simulate electronic components, located in a horizontal channel has been studied numerically. The fluid flow and temperature profiles, above the heating elements placed between an adiabatic lower plate and an isothermal upper plate, are obtained using numerical simulation. The effects of source temperatures, channel dimensions, openings, boundary conditions, and source locations on the heat transfer from and flow above the protruding sources are investigated. Different configurations of channel dimensions and separation distances of heat sources are considered and their effects on natural convection heat transfer characteristics are studied. The results show that the channel dimensions have a significant effect on fluid flow. However, their effects on heat transfer are found to be small. The separation distance is found to be an important parameter affecting the heat transfer rate. The numerical results of temperature profiles are compared with the experimental measurements performed using Filtered Rayleigh Scattering (FRS) technique in an earlier study, indicating good agreement. It is observed that adiabatic upper plate assumption leads to better temperature predictions than isothermal plate assumption.

Commentary by Dr. Valentin Fuster
2003;():551-556. doi:10.1115/HT2003-47481.

Heat sink is commonly found in electronic systems. For its optimization, numerical computation is introduced. However, narrow gaps between the fins of heat sink have been a troubling factor. That increases the number of grid excessively, and results in increased computation time. The quality of grid can be poor and that halt the accuracy of computed numerical solution. To avoid these problems, many simplification methods are proposed by simplifying complex heat sink. The most popular example is regarding the array of fins as flow resistance from hydraulic point of view and working fluid with different thermal conductivity for thermal equivalence [1]. Its thermal conductivity can be determined according to well-known relationship between Nu, Re, and Pr (see [2, 3]). This simplification presents many advantages but it is not applicable to natural convection. In this paper, a modified model is suggested to extend the simplification to natural convection, which is still popularly applied to electro cooling systems. With the results of [4], thermal conductivity of flow resistance region is iteratively. The modified model is verified by computing flow and thermal fields of PDP. Applying this model to fanless PDP, the number of total grid is reduced by 38.5% percents and corresponding computation time was saved while the accuracy of computed solution is kept undamaged.

Commentary by Dr. Valentin Fuster
2003;():557-564. doi:10.1115/HT2003-47524.

In this manuscript a systematic multidisciplinary electronic packaging design and optimization methodology based on the artificial neural networks technique is presented. This method is applied to a Ball Grid Array (BGA) package design as an example. Multidisciplinary criteria including thermal, structural (thermal strain), electromagnetic leakage, and cost are optimized simultaneously. A simplified routability criterion is also considered as a constraint. The artificial neural networks technique is used for thermal and structural performance predictions. Large calculation time reduction is achieved using the artificial neural networks, which also provide enough information to specify the individual weights for each design discipline within the objective function used for optimization. This methodology is able to provide the designers a clear view of the design trade-offs, which are represented in the objective function using various design parameters. This methodology can be applied to any electronic product design at any packaging level.

Commentary by Dr. Valentin Fuster
2003;():565-572. doi:10.1115/HT2003-47525.

The present study investigates experimentally and numerically the natural convection of air in square enclosures with a localized heat source from below and symmetrical cooling from the sides. The heat source was centered on the bottom wall and the study analysed the effect of the variation in the heat source length on the natural convection inside the square cavity; the length of the heat source investigated are 1/5 and 2/5 of the wall The cooling was achieved by the two vertical walls and all the other zones were adiabatic; the symmetrical cooling from the sides is expected to be an efficient cooling option while the partial heating at the lower surface simulates the electronic components such as a chip. The experimental data are obtained by measuring the temperature distribution in the air layer by the real-time and double-exposure holographic interferometry and the numerical investigation was conducted using the commercial finite volumes code Fluent 6.0. Convection was studied for Rayleigh number from 103 to 106 . Different convection forms were obtained depending on Ra and on the heat source length. The Nusselt number was evaluated on the heat source surface and it showed a symmetrical form raising near the heat source borders. Graphs with relations between average Nu, Ra and the heat source length are finally presented.

Commentary by Dr. Valentin Fuster

Computational Heat Transfer

2003;():573-579. doi:10.1115/HT2003-47006.

In the sublimation crystal growth process, source materials sublime at a higher temperature, and the resulting vapor transports to the seed crystal at a lower temperature where surface reaction and crystallization takes place. We have developed a two-dimensional numerical model to simulate the thermal and fluid fields in an AlN bulk growth system. A growth model that incorporates thermodynamic analyses, Stefan flow or diffusion transport, and growth kinetics has been developed to predict the growth rate of AlN bulk crystal. In addition, a thermomechanical stress model has also been developed to predict the thermal stress evolution in an AlN crystal.

Topics: Crystals , Stress
Commentary by Dr. Valentin Fuster
2003;():581-588. doi:10.1115/HT2003-47018.

A wide variety of inverse heat conduction problems have been studied in the last two decades for the estimation of boundary or initial conditions, thermophysical properties, geometrical parameters, or heat source intensities. In most transient heat conduction problems, the mathematical models were cast in dimensionless forms, by using a diffusion time scale. As the thermal diffusivities are usually small, the physical time scales turn to be rather long. In this way, most works show that the inverse analysis yields satisfactory results, without addressing the implications of the physical time scale. The physical time scale, in fact, influences significantly the quality of the inverse solution. We present here a unified treatment for one-dimensional, linear inverse heat conduction problems using the conjugate gradient method with an adjoint equation, and also show that there are physical limitations by the time scale on the inverse solutions.

Topics: Heat conduction
Commentary by Dr. Valentin Fuster
2003;():589-595. doi:10.1115/HT2003-47021.

Heat transfer considerations are important in almost all areas of technology. However, heat transfer analysis can be very difficult for complicated systems such as very large-scale integrated (VLSI) electric circuits and systems, making the simulation an attractive technique for studying heat transfer of those systems. This paper presents methods of heat transfer simulation using PSpice. First, typical heat transfer modes are discussed and heat transfer equations are presented. Then, equivalent electrical models are developed, and PSpice representations of those models are investigated. Finite-difference RC network models are developed and used for the simulation of complicated heat transfer problems using PSpice. Two typical heat transfer examples are studied. Simulations are performed to investigate and study the heat transfer and energy flow of the two examples using PSpice.

Commentary by Dr. Valentin Fuster
2003;():597-606. doi:10.1115/HT2003-47028.

The systematic design and precise control of the microfluidic dispenser (crossing microchannels etched into a plastic or glass plate) is key to the performance of many lab-on-a-chip devices. The fundamental understanding of the complicated electrokinetic phenomena in microfluidic dispensers therefore is necessary. In the literature, a few theoretical models studying the transport phenomena in similar crossing microchannels didn’t consider the spatial gradients of conductivity due to its complexity. A new theoretical model was developed in this paper to simulate the transport phenomena in a microfluidic dispenser with the consideration of a large range of spatial gradient of electric conductivity. This developed model was used to simulate the potential field, flow field, and concentration field of the injection processes where the conductivity of the sample-carrying buffer differs significantly from that of the driving buffer. The transport phenomena are found to be very sensitive to the conductivity difference between the sample-carrying buffer and the driving buffer. The developed model can be employed to find the optimal voltages for controlling the dispensed sample size and to provide guidance for designing such a microfluidic dispenser in lab-on-a-chip devices.

Commentary by Dr. Valentin Fuster
2003;():607-616. doi:10.1115/HT2003-47042.

The interactions of cells with dynamic blood flow as they adhere firmly to a micro-channel wall are investigated. The cell is modeled either as a rigid solid or an elastic solid protrusion on a micro-channel wall. Blood is simulated using a validated non-Newtonian blood viscosity model (Walburn and Schneck model). We determine the effects of cell deformability on the force generated on the cell as the deformation progresses. The amplifications of mechanical stress and force on an adherent cell due to blood flow in a micro-channel are predicted. This two-dimensional model is solved by the finite volume package (CFDRC, CFD Research Corporation). This study shows that the pressure drop and drag force on the adherent cell are sensitive to the cell’s morphology, especially for the large ratio of cell diameter to channel height (d/D>0.5), the stresses and forces of deformable cells can be much smaller than that predicted based on the rigid cell model. These calculations may be used to predict blood flow interactions with cells in a micro-vessel. The modeling approach is useful in understanding the behaviors in cell adhesion and rolling.

Commentary by Dr. Valentin Fuster
2003;():617-623. doi:10.1115/HT2003-47054.

Soil remediation using Heated Soil Vapor Extraction System has gained a significant attention in recent years. The process, developed by Advanced Remedial Technology**, comprises of a heat well (heat source) and an extraction well (sink). These wells are pipes, which are implanted in the soil. Heating is accomplished by circulating hot oil through the heat exchange units in heat well. The extraction well has a blower, which sucks the air, and other volatile gases that are evaporated due to heating. An analysis aimed at improving the predictability of the process using numerical tools has been carried out. The key parameters in the process can be identified as the distance between the wells, the temperature that has to be maintained in the heat well and the time required vaporizing the gases and taking them off the soil. These parameters are strongly dependent on the properties of the soil and properties of the chemical pollutants present in the soil. An attempt has been made to model the real process of heating the soil and vaporizing of chemicals in the soil. Such comprehensive analysis will be very much helpful in predicting the different parameters as discussed above and result in increase in effectiveness and efficiency of the process.

Topics: Vapors , Simulation , Soil
Commentary by Dr. Valentin Fuster
2003;():625-631. doi:10.1115/HT2003-47063.

The investigation reported in this paper includes the variation of transient and local heat transfer coefficient and heat flux in the combustion chamber of a spark ignition (SI) engine. Heat transfer characteristics are obtained from the Kiva-3v CFD (Computational Fluid Dynamics) code. Instantaneous results including the variations of mean heat transfer coefficient on the piston surface, combustion chamber, and wall of the cylinder are presented. Moreover, variations of the local heat transfer coefficient and heat flux along a centerline on the piston as well as a few locations on the combustion chamber surface are shown. It is illustrated that maximum heat transfer coefficient on the piston and combustion chamber surfaces varies with location and also it is observed that the initial high rate of increase of heat flux at any position is related to the instant of flame arrival at that position. In this work, the major focus is on the determination of the locations where heat flux and heat transfer are maximum.

Commentary by Dr. Valentin Fuster
2003;():633-640. doi:10.1115/HT2003-47065.

The study of natural convection inside a two-dimensional cavity with one and three undulations on the right vertical wall has been carried out where the top wall is heated by a spatially varying temperature and other three walls are cold walls. Non-orthogonal body-fitted coordinate system and SIMPLE algorithm with higher-order upwinding scheme are used. The streamline pattern shows two cells are formed separated at the middle vertical plane for both the configurations. The center of the cells is lifted upwards with increase in Ra. The heat rejection from the fluid to the right wall increases for the uppermost undulation whereas there is not much improvement of heat transfer for the other two in the three undulation case. For this particular configuration, the heat rejection increases with increase of Ra for the uppermost undulation and decreases with increase of Ra for the other two undulations. The overall improvement of heat rejection has been observed for three undulation case compared with one undulation case.

Commentary by Dr. Valentin Fuster
2003;():641-648. doi:10.1115/HT2003-47090.

A three-dimensional computational model is developed to analyze fluid flow in a semi-porous channel. In order to understand the developing fluid flow and heat transfer process inside the semi-porous channels, the conventional Navier-Stokes equations for gas channel, and volume-averaged Navier-Stokes equations for porous media layer are adopted individually in this study. Conservation of mass, momentum and energy equations are solved numerically in a coupled gas and porous media domain in a channel using the vorticity-velocity method with power law scheme. Detailed development of axial velocity, secondary flow and temperature fields at various axial positions in the entrance region are presented. The friction factor and Nusselt number are presented as a function of axial position, and the effects of the size of porous media inside semi-porous channel are also analyzed in the present study.

Commentary by Dr. Valentin Fuster
2003;():649-655. doi:10.1115/HT2003-47103.

The k-ε model are performed to investigate numerically the steady, turbulent, incompressible flow and heat transfer converging radially between two stationary disks, which is as a continuously developing flow problem under the internal boundary layer approximations. The effect of relaminarization was considered. This present study has presented a good agreement with the laminar investigation of Murphy et al [1], where no heat transfer was considered. At large values of the dimensionless radii (>> 1) the velocity profile becomes parabolic and invariant and the friction factor approaches the classic value obtained for fully developed flow between infinite plates, 24/Re0 , where Re0 is an overall Reynolds number based on the volumetric flow rate and the disk spacing and is independent of radius. At radii less than one a typical external boundary layer evolves close to the wall with an approximately uniform core region, the boundary layer thickness decreases from one-half the disk spacing to values proportional to the local radii as the flow accelerates and the friction factor approaches the constant 2.17/Re0 . A local Nusselt number, Nu = 230(r/R)0.650 (1 − r/R)−0.386 , where r is radial coordinate and R the radius of the disk, was estimated. A large overall Reynolds number was imposed and a relaminarization of the flow was observed. It was suggested that these results can be applicable for laminar and turbulent flow under Re0 = 106 .

Commentary by Dr. Valentin Fuster
2003;():657-664. doi:10.1115/HT2003-47108.

Thermal transfers occurring in the vicinity of an air-liquid interface have been studied numerically through a Large Eddy Simulation technique. Results obtained clearly show the unsteady response of the liquid film submitted to such thermal stress. Structures are created at the interface in a very small layer and it has been found that turbulence acts strongly in that layer. Moreover, even if the configuration studied tends to weaken buoyancy, the dissipation was found to obey a −3 power law. This clearly indicates that the buoyancy pilots the way turbulence behaves on the flow field; the latter is mainly characterized by a strong mixing phenomenon taking place in the already identified layer at the interface.

Commentary by Dr. Valentin Fuster
2003;():665-672. doi:10.1115/HT2003-47138.

A three-dimensional computational analysis is employed to investigate the transient flow and heat transport within an annular-core High-Temperature Reactor pebble bed, the blind core, and the surrounding graphite structures including 36 embedded coolant feeding channels. Well established models of pressure loss in the pebble bed, heat transfer between helium and the pebbles, and the heat transport within the porous medium by conduction and radiation (using an effective conductivity) have been implemented into an extended version of CFX-4. The graphite structures of the reflector are simulated closely coupled to the flow simulation as heat conducting solids. The model has been verified for axisymmetric power distributions by comparison with known results from 2D analysis. Various 3D-power distributions are assumed to represent a possible loss of forced cooling incident after unusual plant operation such as asymmetric positioning of absorber rods. It is confirmed, that 2D analysis is not adequate, because the predicted maximum temperature of the fuel elements after fission shutdown may exceed the maximum temperature predicted by 2D analysis.

Commentary by Dr. Valentin Fuster
2003;():673-680. doi:10.1115/HT2003-47139.

Polysilicon growth has increased due to its broad applications and market demand. The traditional method for polysilicon growth is based on the Siemens process. To improve the throughput, a new system with either large growth surface or other mechanism for high deposition rate is necessary. A novel design, using a vertical tubular CVD reactor has been recently proposed, in which an enlarged surface reaction area exits. This study is to investigate the optimal conditions for growth through numerical simulation of heat and mass transfer in the proposed vertical tubular CVD reactor. A complex computational model is developed that is capable of describing multi-component fluid flow, gas/surface chemistry, conjugate heat transfer, thermal radiation, and species transport. Different from the classical Siemens system, the bulk poly-silicon in a vertical tube growth has a complicated geometry. To accurately predict the various parameters covering broad range of scales, a multi-block grid generation system is used. Numerical computation has been conducted under different operating conditions, and in particular the effect of cooling gas flow direction and flow rate on the temperature distribution of the system and the polysilicon deposition rate has been investigated. Numerical results show that cooling from the top of the system is preferred.

Commentary by Dr. Valentin Fuster
2003;():681-687. doi:10.1115/HT2003-47146.

The inverse problem of using temperature measurements to estimate the moisture content and temperature-dependent moisture diffusivity together with the heat and mass transfer coefficients is analyzed in this paper. In the convective drying practice, usually the mass transfer Biot number is very high and the heat transfer Biot number is very small. This leads to a very small temperature sensitivity coefficient with respect to the mass transfer coefficient when compared to the temperature sensitivity coefficient with respect to the heat transfer coefficient. Under these conditions the relative error of the estimated mass transfer coefficient is high. To overcome this problem, in this paper the mass transfer coefficient is related to the heat transfer coefficient through the analogy between the heat and mass transfer processes in the boundary layer. The resulting parameter estimation problem is then solved by using a hybrid constrained optimization algorithm OPTRAN.

Topics: Mass transfer , Drying
Commentary by Dr. Valentin Fuster
2003;():689-697. doi:10.1115/HT2003-47153.

A three-dimensional computational model has been developed to describe the compressible, multi-component, turbulent, reacting plasma jet coupled with the orthogonal injection of carrier gas and particles. This model has been applied to plasma spray process that includes physical phenomena such as heating, melting, accelerating, and evaporation of in-flight particles. The entrained particles, NiCrAlY and ZrO2 , are discretely treated in a Lagrangian coordinate and stochastically generated by sampling from the probability distributions of the particle size and its velocity at the injection nozzle. In this study, special attention has been directed to the effects of carrier gas injection on the characteristics of plasma jet. The computational results show that the injection of carrier gas from the orthogonal injector above the plasma jet introduce the 3-D phenomena of plasma gas flow. The plasma jet is defected and the thermo-fluid flow near the injector is locally deformed.

Commentary by Dr. Valentin Fuster
2003;():699-709. doi:10.1115/HT2003-47158.

Molecular dynamics (MD) simulations of liquid-vapor interfaces were performed to determine mean property variations and property fluctuations in the liquid-vapor interfacial region at various reduced temperatures. The interfacial region typically has a thickness on the order of a few nanometers for systems of practical interest. The system’s initial conditions were specified as a bulk liquid region sandwiched between two bulk vapor regions. Simulations were run using a Lennard-Jones 6-12 potential function between the atoms with appropriate parameters for Argon atoms. As the simulation was performed, interfacial region property data was collected over time. The resulting property data are shown to establish trends similar to those indicated by theoretical and experimental results reported elsewhere. The peak fluctuations of mass density and free energy density were determined to be approximately equal in magnitude when normalized with the difference in their respective bulk values at a given temperature. These fluctuations were found to increase rapidly with temperature. The fluctuations in the interfacial thickness and interfacial position follow a functional dependence on temperature similar to that exhibited by the mean value of interfacial thickness. In addition to exploring fluctuations in the interfacial region, two new methods were developed to determine interfacial tension through methods involving integration of excess free energy density across the interfacial region. These techniques were shown to yield mean results similar to theoretical predictions and those using conventional techniques. In addition, the time required for computation using the new techniques is significantly reduced due to less computational time per step and fewer required steps for convergence to a mean value.

Commentary by Dr. Valentin Fuster
2003;():711-714. doi:10.1115/HT2003-47164.

In typical atomistic simulations of simple liquids, the Lennard-Jones interatomic pair potential is truncated so that algorithms scale as Natoms rather Natoms 2 , which would be the case if an interaction were computed explicitly for all atom pairs. However, it is known that interfacial properties are sensitive to the cutoff radius selected. Corrections for the missing ‘tails’ of the potential can reduce the error, but cannot eliminate it because the liquid and vapor densities are also sensitive to the cutoff radius. In light of this, we have developed and implemented a NlogN particle-particle particle-mesh (P3 M) algorithm to evaluate the 1/r6 dispersive forces between Lennard-Jones fluid molecules without truncation. Statistical expression for the surface tension also scale as N2 if potentials are not truncated, so we also developed a P3 M formulation for computing surface tension. The techniques are demonstrated on a thin liquid film suspended in equilibrium with its own vapor. Simulations at several temperatures between the triple point and the critical point are compared with the available data. The expense of the algorithm is competitive for simple geometries and seems preferable in non-trivial geometries without the possibility of tail corrections.

Commentary by Dr. Valentin Fuster
2003;():715-719. doi:10.1115/HT2003-47172.

In this work, electronic thermal conduction in thin gold film is modeled via the Boltzmann Transport Equation (BTE). The BTE is solved using a Monte Carlo Method (MCM). Temperature profiles for various film thicknesses are computed. Results show that the electronic thermal transport in gold is still diffusion-like at film thicknesses as small as 100 nm, implying that the Fourier law of conduction can be applied at this scale to predict the steady-state thermal heat transfer without comprising the physics. However, the Fourier law does not predict the temperature profiles accurately if the film thickness is reduced to 10 nm or below.

Topics: Heat conduction
Commentary by Dr. Valentin Fuster
2003;():721-730. doi:10.1115/HT2003-47182.

Numerical simulations have been performed to study the effects of the inner wall rotation on the unsteady multicellular flow of natural convection in the conductive regime (Ra = 8000). We consider a tall air-filled vertical annulus between differentially heated concentric cylinders with the inner cylinder allowed to rotate. The unsteady Boussinesq equations are discretized using a finite volume method with a second order time stepping scheme. The natural convection flow is axisymmetric in this regime, whereas it is known that the mixed convection flow becomes 3D over a range of Reynolds number. We observe the transition in a range of Reynolds number close to the critical Reynolds number of the Taylor-Couette flow. The rotation has a weak influence on the axisymmetric time-periodic natural convection flow before the transition, whereas the flow becomes 3D and chaotic after.

Commentary by Dr. Valentin Fuster
2003;():731-738. doi:10.1115/HT2003-47190.

Three firing schemes for an industrial oxygen-fired glass melting furnace were examined to determine the thermal performance and relative merits of each scheme. A comprehensive computer model was used to investigate the effects of each scheme on the combustion and heat transfer in the furnace. The three-dimensional computer model, suitable for predicting and analyzing fluid flow, combustion and heat transfer has been used to simulate the combustion space of the furnace. The turbulent flow field is obtained by solving the Favre averaged Navier-Stokes equations and using the k-ε model to calculate the turbulent shear stresses and close the equation set. The combustion model consists of a single step, irreversible, infinitely fast reaction. A mixture fraction is used to track the mixing of fuel and oxidant and thus reaction progress in this mixing limited model. An assumed shape PDF method is utilized to account for turbulent fluctuations. Radiative heat transfer in the combustion gases and between surfaces is modeled using the discrete ordinates method coupled with the weighted-sum-of-gray-gases model. The model furnace for all three firing schemes was the same size and shape, was charged from the rear end wall and was pulled from the front wall. The three schemes investigated were: 1) non-interlaced side-wall fired, 2) interlaced side-wall fired, and 3) end fired. The results show that all three arrangements provide similar thermal performance and heat transfer characteristics. However, the flow field for the non-interlaced arrangement is very complex in the region where jets from opposing walls meet at the furnace center line. This type of jet interference can lead to unstable flow, particularly at the centerline of the furnace. Unstable flow conditions can affect the heat transfer characteristics of the furnace and make the furnace difficult to operate. Conversely, the interlaced and end-fired schemes do not exhibit the jet interference seen in the non-interlaced arrangement. While the results indicate that the thermal performance of all three arrangements were similar, the possibility of jet interference suggests that an interlaced or end-fired arrangement is preferable.

Commentary by Dr. Valentin Fuster
2003;():739-748. doi:10.1115/HT2003-47193.

Thermal properties are generally determined through solving inverse problems. Because the temperature is a non-linear function of the properties, the solutions are usually effected by linearizing the equations. The statistics of these linearized estimates are based upon the assumptions that the measurement noise has zero mean and is normally distributed, yielding unbiased and normally distributed parameter estimates. In fact, even for this type of noise, nonlinear functions can lead to bias and nonnormal distributions of estimated properties. We examine these effects and show that for typical thermal systems that while the estimates are unbiased and normal, the confidence limits may be inaccurately defined and the residuals of the fits may not be zero mean and uncorrelated. Characterizing the estimated parameters is critical when nonlinear models are to be used for extrapolation.

Commentary by Dr. Valentin Fuster
2003;():749-756. doi:10.1115/HT2003-47246.

This paper deals with the problem of heat transfer in square cavities partially filled with porous material. Local flow and energy equations are integrated in a representative elementary volume in order to obtain a set of equations valid in both the clear flow region and in the porous matrix. A unique set of equations is discretized with the control volume method and solved with the SIMPLE algorithm. Enhancement of convective currents within the porous substrate is detected as the Rayleigh number increases. Thin boundary layers along the cavity vertical walls and stratification of the thermal field are observed for Ra > 109 .

Commentary by Dr. Valentin Fuster
2003;():757-763. doi:10.1115/HT2003-47253.

This work presents a numerical investigation of turbulent flow past a porous structure in a channel using linear and non-linear eddy viscosity macroscopic models. Parameters such as porosity and permeability of the porous material are varied in order to analyze their effects on the flow pattern, particularly on the damping of the recirculating bubble after the entrance and exit regions. The numerical technique employed for discretizing the governing equations is the control-volume method. The SIMPLE algorithm is used to correct the pressure field. The classical wall function is utilized in order to handle flow calculation near the wall. A discussion on the use of this technique for simulating the flow in question is presented. Comparisons of results simulated with both linear and non-linear turbulence models are shown.

Commentary by Dr. Valentin Fuster
2003;():765-770. doi:10.1115/HT2003-47255.

The present work investigates the efficiency of the multigrid method when applied to solve laminar flow in a two-dimensional tank filled with a porous material. The numerical method includes finite volume discretization with the flux blended deferred correction scheme on structure orthogonal regular meshes. Performance of the correction storage (CS) multigrid algorithm is compared for different numbers of sweeps in each grid level. Up to four grids, for both multigrid V- and W-cycles, are considered. Effects of medium permeability on converged rates are presented. Results indicate that W-cycles perform better in reducing the required computational effort and that the lower the permeability, faster solutions are obtained.

Commentary by Dr. Valentin Fuster
2003;():771-776. doi:10.1115/HT2003-47276.

Modeling of electron field emission has not advanced significantly since Fowler and Nordheim described the phenomenon eighty years ago. While their approach provides remarkable agreement with experiments for a large number of cases, the theory is strictly valid for planar geometries and low temperatures. Carbon nanotubes have been considered for field emission energy conversion devices. Under high-temperature conditions and significant field enhancement, the approximations used in the Fowler-Nordheim formalism become invalid. The present work predicts electron current densities emitted from carbon nanotubes using Airy functions to predict transmission and temperature dependent supply functions. Results indicate that Fowler-Nordheim compares favorably with the Airy function model for materials with large work function (φ ≈ 5eV, in the present study) at room temperatures. However, for materials with smaller work functions, the difference between the Airy function model and Fowler-Nordheim can be significant.

Commentary by Dr. Valentin Fuster
2003;():777-783. doi:10.1115/HT2003-47280.

A nanofluid is a fluid containing suspended solid particles, with sizes of the order of nanometers. Normally the fluid has a low thermal conductivity compared to the suspended particles. Therefore introduction of these particles into the fluid increases the effective thermal conductivity of the system. It is of interest to predict the effective thermal conductivity of such a nanofluid under different conditions like varying particle volume fraction, varying particle size, changing fluid conductivity or changing fluid viscosity, especially since only limited experimental data are available. Also, some controversy exists about the role of Brownian motion in enhancing the nanofluid’s thermal conductivity. We have developed a novel technique to compute the effective thermal conductivity of a nanofluid using Brownian dynamics simulation, which has the advantage of being computationally less expensive than molecular dynamics. We obtain the contribution of the nanoparticles towards the effective thermal conductivity using the equilibrium Green-Kubo method. Then we combine that with the thermal conductivity of the base fluid to obtain the effective thermal conductivity of the nanofluid, and thus are able to show that the Brownian motion contributes greatly to the thermal conductivity.

Commentary by Dr. Valentin Fuster
2003;():785-792. doi:10.1115/HT2003-47294.

Liquid cooled exhaust manifolds are used in turbo charged diesel and gas engines in the marine and various industrial applications. Performance of the manifold has a significant impact on the engine efficiency. Modifying manifold design and changing operational parameters are ways to improve its performance. With the rapid advance of computer technology and numerical methods, Computational Fluid Dynamics (CFD) has become a powerful tool that can provide useful information for manifold optimization. In this study, commercial CFD software (FLUENT® ) was used to analyze liquid cooled exhaust manifolds. Detailed information of flow property distribution and heat transfer were obtained in order to provide a fundamental understanding of the manifold operation. Experimental data was compared with the CFD results to validate the numerical simulation. Computations were performed to investigate the parametric effects of operating conditions (engine rotational speed, coolant flow rate, coolant inlet temperature, exhaust gas inlet temperature, surface roughness of the manifold’s material) on the performance of the manifold. Results were consistent with the experimental observations. Suggestions were made to improve the manifold design and performance.

Commentary by Dr. Valentin Fuster
2003;():793-802. doi:10.1115/HT2003-47307.

A novel algorithm has been developed for the non-destructive determination of the shape of the interface between a melt and a refractory material wall in smelter furnaces. This method uses measurements of temperature and heat flux at a number of points on the outer surface of the furnace and assumes that the inner (guessed) surface of the furnace wall is isothermal. The temperature field is then predicted in the entire furnace wall material by numerically solving a steady state heat conduction equation subject to the measured temperature values on the external surface and the isothermal melt material solidus temperature on the inner surface of the wall. The byproduct of this analysis is the computed heat flux on the external surface. The shape determination method then uses the difference between the measured and the computed heat fluxes on the outer surface of the furnace as a forcing function in an elastic membrane motion concept for the determination of the inner (melt-refractory) surface motion. The inverse determination of the melt-refractory interface shape can be achieved by utilizing this algorithm and any available analysis software for temperature field in the refractory wall. The initial guess of the wall inner shape can be significantly different from the final (unknown) wall shape. The entire wall shape determination procedure requires typically 5–15 temperature field analysis in the furnace wall material.

Commentary by Dr. Valentin Fuster
2003;():803-805. doi:10.1115/HT2003-47312.

Modeling heat generation at nanometer scales in silicon is of great interest and particularly relevant to the heating and reliability of nanoscale and thin-film transistors. Joule heating is usually simulated as the dot product of the macroscopic electric field and current density [1]. This approach does not account for the microscopic non-locality of the phonon emission near a strongly peaked electric field region. It also does not differentiate between electron energy exchange with the various phonon branches and does not give any information regarding the types of phonons emitted. The present work addresses both of these issues: we use a detailed Monte Carlo (MC) simulation to compute sub-continuum and phonon mode-specific heat generation rates, with applications at nanometer length scales.

Commentary by Dr. Valentin Fuster
2003;():807-815. doi:10.1115/HT2003-47331.

This paper deals with the temperature rise due to material damping in rotating pultruded composite shafts. Pultrusion is a special fabrication process for manufacturing composite structural components whose cross section does not change along its length. Rotating shafts carrying gears or pulleys are examples. These shafts are subjected to side loads as they rotate. This causes them to bend back and forth in a cyclic manner. As a result, every element of the shaft generates heat due to material damping. The heat generation per cycle of motion is known for many materials and so a mathematical analysis is possible. Mathematical treatment of thermal problems involving pultruded composites can be simplified using two-dimensional Cartesian or Axisymmetric models. An explicit unsteady finite difference scheme with heat generation is used to study rotating shafts. The method is straightforward and capable of handling unsteady heat conduction in anisotropic components having single or multiple fiber types. The scheme can handle nonzero initial conditions and general convection type boundary conditions. Examples are used to illustrate the method and to investigate effects of the fiber-matrix thermal conductivity ratios, shaft rpm, and material damping on the temperature history in the material. The results show that, depending on the stiffness and thermal conductivity ratios in the principal directions, the temperature rise can be significant.

Commentary by Dr. Valentin Fuster
2003;():817-826. doi:10.1115/HT2003-47343.

The flow and heat transfer characteristics in the cooling of a heated surface by impinging confined jets have been investigated numerically through the steady state solution of laminar two-dimensional Navier-Stokes and energy equations. The principal objective of this study is to investigate the effect of buoyancy on the associated heat transfer process. Numerical computations are done for vertically downward directed two-dimensional confined slot jets impinging on a hot isothermal surface at the bottom. The computed flow patterns and isotherms for various domain aspect ratios and for a range of jet exit Reynolds numbers (100–500) and Richardson numbers (0–10) are analyzed to understand the heat transfer phenomena. The local and average Nusselt numbers at the hot surface for various conditions are compared. It is observed that for a given domain aspect ratio and Richardson number, the average Nusselt number at the hot surface increases with increasing jet exit Reynolds number. On the other hand, for a given aspect ratio and Reynolds number the average Nusselt number does not change significantly with Richardson number indicating that the buoyancy effects are not very significant on the overall heat transfer process for the range of jet Reynolds number considered in this study.

Commentary by Dr. Valentin Fuster
2003;():827-834. doi:10.1115/HT2003-47362.

In the recent scenario of braking system for automobiles, disc brake takes up a wide range of applications, because of its simplicity in construction, operation and not self energizing as in the case of drum brakes. Since the disc brakes takes up a wide range of application, it is essential to ensure the reliable function of the braking system under varied operating conditions. The reliable function of the disc brake system is purely depends on the system based design. In this work, a linear regression technique is used for the optimal design of the disc brake rotor for varied operating conditions. Various forces involved during braking, energy generated during braking and the corresponding effective stopping distances were also calculated using appropriate governing relations and equations. In the varied operating conditions, the heat energy generated during braking should be driven away form the working surfaces of the components. To analyze this thermal loading and cooling phenomenon, a conventional convective heat transfer approach was also formulated and developed in this work. The analytical findings of the above approaches are demonstrated at the end and it is found to be quite satisfactory.

Commentary by Dr. Valentin Fuster
2003;():835-843. doi:10.1115/HT2003-47364.

In the wide range of braking speed, the disc brakes are subjected to temperature variation and thermal loading. Different modelling approaches ranging from a simple lumped parameter model to complex three-dimensional models are available for the thermal analysis of the disc brakes. Based on the review of the above models, a model has been developed and formulated for the analysis of thermal loading of disc brake. The developed model is proposed to couple with a model for the thermal distortion of disc brake. It is also proposed to conduct the necessary experiments and thermal analysis to validate the results obtained from the synthesized analysis.

Commentary by Dr. Valentin Fuster
2003;():845-856. doi:10.1115/HT2003-47372.

Numerical simulations coupled with LDV experiments were carried out to investigate a slot jet issued into a cross flow, which is relevant in the film cooling of gas turbine combustors. The film cooling fluid injection from slots or holes into a cross-flow produces highly complicated flow fields. In this paper, the time-averaged Navier-Stokes equations were solved on a collocated body-fitted grid system with the V2F turbulence model. The fluid flow and turbulent Reynolds stress fields were compared with the LDV experiments for three jet angles, namely, 30-deg, 60-deg, and 90-deg, and the jet blowing ratio is ranging from 2 to 9. Good agreement was obtained. Therefore, the present solution procedure was also adopted to calculations of 15-deg and 40-deg jets. In addition, the temperature fields, which were difficult to measure by experimental methods, were also computed with a simple eddy diffusivity model to obtain the film cooling effectiveness which was used for evaluation of the various jet-cross-flow arrangements. The results show that a recirculation bubble downstream the jet exists for jet angles larger than 40-deg, but it vanishes when the angle is less than 30-deg, which is in good accordance with the experiments. The blowing ratio has a large effect on the size of the recirculation bubble, and consequently on the film cooling effectiveness. In addition, the influence of boundary conditions for the jet and cross-flow are also addressed in the paper.

Topics: Cooling
Commentary by Dr. Valentin Fuster
2003;():857-864. doi:10.1115/HT2003-47377.

The results of this research show that as the horizontal spacing between the cylinders in a single horizontal row is reduced; the natural convection heat transfer coefficient and thus the rate of heat transfer from each cylinder could increase substantially beyond those for natural convection from a single cylinder before rapidly decreasing. The rate of heat transfer from the array can therefore be significantly increased because of two factors; the enhanced heat transfer coefficient and the ability to add more cylinders in a given volume. The rapid reduction in heat transfer at very small spacing also points to the existence of an optimum tube spacing. This phenomenon does not appear to have been reported before and may lead to significant reductions in the size of many heat exchangers and may prove useful in cooling of electronic components.

Commentary by Dr. Valentin Fuster
2003;():865-873. doi:10.1115/HT2003-47450.

A fast, accurate and efficient multi-level boundary element method is developed to solve general boundary value problems. Here we concentrate on problems of two-dimensional steady potential flow and present a fast direct boundary element formulation. This novel method extends the pioneering work of Brandt and Lubrecht on multi-level multi-integration (MLMI) in several important ways to address problems with mixed boundary conditions. We utilize bi-conjugate gradient methods (BCGM) and implement the MLMI approach for fast matrix and matrix transpose multiplication for every iteration loop. Furthermore, by introducing a C-cycle multigrid algorithm, we find that the number of iterations for the bi-conjugate gradient methods is independent of the boundary element mesh discretization for problems of steady-state heat diffusion considered in this paper. As a result, the computational complexity of the proposed method is proportional to only N · log(N), where N is the number of degrees of freedom.

Commentary by Dr. Valentin Fuster
2003;():875-886. doi:10.1115/HT2003-47452.

Higher-order boundary element methods (BEM) are presented for time-dependent convective diffusion in two dimensions. The time-dependent convective diffusion free-space fundamental solutions originally proposed by Carslaw and Jaeger are used to obtain the boundary integral formulation. Boundary element method solutions up to the Peclet number 106 are obtained for an example problem of unsteady convection-diffusion that possesses an exact solution. We investigate the convergence rate and accuracy of the higher-order boundary element formulations. An extremely high accuracy of the BEM solutions for highly convective flows is demonstrated. Moreover, it is shown that the use of time-dependent convective kernels provides an automatic upwinding for the entire range of Peclet numbers and also leads to very efficient algorithms as the Peclet number increases.

Commentary by Dr. Valentin Fuster
2003;():887-897. doi:10.1115/HT2003-47462.

The need to shorten the development time for new engine and vehicles is leading to the increasing use of computational design and simulation methods in the automotive industry. In the last years 3D computational models have been used successfully in vehicle and engine development. It is clear that in this kind of simulation, the input complexity, the output data management and the computational time increase. On the other hand 3D simulations increase the details of the results and their link with the analyzed geometry. During a vehicle design several numerical techniques can be used (finite difference, finite volume, spectral methods, boundary elements, etc.) Often, in a complex simulation, that involves several different physical phenomena such as fluid flow and heat transfer only the use of different simulation techniques allows to obtain good results in a acceptable time. In several industrial applications the use of coupled codes, with different features (1D, 3D or different numerical schemes) could provide an optimal solution for the simulation approach. In this paper an example of a complex simulation of an Under Hood Cooling (UHC) of a vehicle is carried out using two different 3D codes with different numerical approaches with the objective to reduce the simulation time [1].

Topics: Cooling , Modeling
Commentary by Dr. Valentin Fuster
2003;():899-905. doi:10.1115/HT2003-47474.

In the present work, gas-phase reactions between opposing streams of mixtures of hydrogen (H2 ) and methane (CH4 ) in the presence of volumetric energy input were simulated. The goal of the simulations is to estimate the concentrations of precursors responsible for the formation of carbon nanotubes (CNTs). These estimates are expected to help in understanding fundamental mechanisms of CNT formation and in controlling the synthesis process through parameters such as inlet composition and temperature, reactor pressure and absorbed energy. The simulation employs gas-phase kinetics of the GRI-2.11 mechanism with only reactions involving molecules that contain C and H atoms. The results indicate that the concentrations of H radicals, C2 H2 and C atoms increase significantly with increases in volumetric energy deposition rate beyond a threshold.

Commentary by Dr. Valentin Fuster
2003;():907-910. doi:10.1115/HT2003-47498.

Proper positioning of temperature sensors is required to minimize error in measurements of thermal diffusivity. Calculating the proper position of these sensors a priori, however, requires knowing the very thermophysical properties of the material one is testing. Presented here are calculations illustrating how IR video might be used to optimize a biaxial, pulsed thermal diffusivity measurement system. Results will be shown for representative elastomers.

Commentary by Dr. Valentin Fuster
2003;():911-916. doi:10.1115/HT2003-47515.

Thermally-induced natural convection heat transfer in the annulus between horizontal concentric cylinders has been studied using the commercial code Fluent. The boundary layers are meshed all the way to the wall because forced convection wall functions are not appropriate. Various oneand two-equation turbulence models have been considered. Overall and local heat transfer rates are compared with existing experimental data.

Commentary by Dr. Valentin Fuster
2003;():917-922. doi:10.1115/HT2003-47541.

In the present study, a numerical investigation has been carried out into the fundamental problem of airflow past and heat transfer from a circular finned cylindrical tube, placed in a duct. The simulation is carried out using a finite volume method, based on laminar calculation of the transport quantities and employs an unsteady, 3-D, second order upwind scheme. As the work has importance in applications of air-cooled heat exchangers, practical values have been chosen for air velocity, air temperature, fin spacing and clearance between fin outer diameter and duct wall. In experimental determination of the performance of a finned-tube bundle, only overall average values such as drag coefficient and overall Nusselt number are possible. Local measurements are well nigh impossible, as any measurement instrument introduced into the narrow fin space will immediately change the flow field. This work gives an insight into variations of shear and heat transfer that will help the designer to optimize the fin spacing. The validity of the results for instantaneous velocity profiles and Nusselt number distribution comes from their physical plausibility. The agreement of the logical behavior of the studied variables when the fin space or fin clearance is modeled confirms the adequacy of the numerical simulation. The strong viscous effects caused by decreasing the fin space result in an increase in Cd and change in its frequency. Vortices generated on the rear section of the tube are damped, but the flow still shows small oscillations downstream of the tube, which indicates that vortex generation still exists, but has changed its location. Vortices augment the Nusselt number locally with increasing fin distance. Simultaneously, the opposite effect of converging boundary layers and therefore accelerated core flow in the fin space yields a maximum in the average Nusselt number as a function of fin space. The impact of decrease in the clearance between the tube and the duct wall is not as important as the effects of fin space.

Commentary by Dr. Valentin Fuster
2003;():923-932. doi:10.1115/HT2003-47553.

In this paper, we develop a domain decomposition, or the artificial sub-sectioning technique, along with a region-by-region iteration algorithm particularly tailored for parallel computation to address storage and memory issues arising from large-scale boundary element models. A coarse surface grid solution coupled with an efficient physically-based procedure provides an effective initial guess for a fine surface grid model. The process converges very efficiently offering substantial savings in memory. We discuss the implementation of the iterative domain decomposition approach for parallel computation on a modest Windows XP Pentium P4 PC-cluster running under MPI. Results from 3-D BEM heat conduction models including models of upwards of 85,000 nodes demonstrate that the BEM can practically be used to solve large-scale linear- and non-linear heat conduction problems using this algorithm.

Commentary by Dr. Valentin Fuster
2003;():933-947. doi:10.1115/HT2003-47582.

Numerical calculations of the aqueous humor dynamics in the anterior chamber of a rabbit’s eye are presented to delineate the basic flow mechanisms. The calculations are based on a geometrical model of the eye, which represents the Trabecular mesh (TM) as a multi-layered porous zone of specified pore sizes and void fraction. The outer surface of the cornea is assumed to be at a fixed temperature (corresponding to the ambient temperature), while the iris surface is assumed to be at the core body temperature. Results are obtained for both the horizontal upward-facing orientation of the eye, and the vertical orientation of the eye. Parameters varied include: the pore size in the TM to understand how TM blockage influences the flow pattern and the intra-ocular pressure (IOP) distribution; the temperature difference between the iris and the cornea to underscore the important role of buoyancy in driving the aqueous humor flow; and, the pupil size reflecting different levels of ambient light. Buoyancy is observed to be the dominant driving mechanism for the convective motion in both orientations of the eye. Reducing the TM pore size does not appear to have a significant influence on the IOP until the pore size drops below 1 micron beyond which a significant increase in IOP is observed. Variations in the pupil size appears to have little influence on the IOP or flow distributions in view of the dominant role of buoyancy in controlling the flow motion.

Commentary by Dr. Valentin Fuster

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In