During the 1980s, direct liquid cooling emerged as one of the most promising thermal management techniques for electronic systems. Many observers believe that, later in this decade, direct liquid cooling with inert, dielectric liquids will become the method of choice for the thermal management of advanced computers. A clear understanding of the available empirical data and relevant heat transfer theories is an essential prerequisite for successful application of this technique. Following a brief survey of electronic packaging trends, attention is turned to the many options in direct liquid cooling and to the selection of coolants for electronic systems. Next, single-phase natural- and forced-convection, including impinging and streaming flow, are reviewed and the best available correlations presented. Ebullient heat transfer is discussed in depth, with a detailed exploration of boiling incipience, fully-established boiling, and critical heat flux for these highly-wetting liquids. Closing remarks focus on the prospects for the widespread implementation of direct liquid cooling in the electronic industry.

View full abstract

The destabilizing effect of a rigid wall on the elastic one-dimensional flat plate placed in uniform irrotational flows adjoining a rigid wall was analyzed by means of linear potential theory. The dynamic characteristics of the fluid forces acting upon the disturbed elastic plate were identified, and it was speculated that the destabilizing effect was caused by the increase in the virtual mass of fluid situated on the wall side of the uniform flows partitioned by the elastic plate and by the decrease in the critical stable flow velocity.

View full abstract

Numerical solutions of the Navier-Stokes equations for viscous shear flows past a circular cylinder are obtained for a range of Reynolds numbers, Re, 2≤Re≤40 and a shear parameter, ε, 0.01≤ε≤1.0. The outer boundary values and the initial values of numerical calculations are determined by our previous analytical solutions utilizing Oseen's approximation, and numerical calculations are carried out so as to be single-valued throughout the field of flow in the process of repetition. The flow patterns, drag, lift, moment and pressure distributions on a circular cylinder are compared with the results calculated using the same methods under fixed boundary conditions and the results obtained in other papers. From these comparisons, it is seen that the values of some properties are markedly affected by the outer boundary condition, and also that the present boundary values are favorable in comparison with previous ones.

View full abstract

The steady axisymmetric laminar flow in a poppet valve is analyzed by means of a conventional finite difference method (ψ-ω method) . After obtaining the numerical results for stream functionψ and vorticity ω in the cylindrical coordinate system, a streamline coordinate system (φ, ψ) is applied to compute pressure distributions, because good pressure distributions cannot be obtained using the cylindrical coordinates. The pressure on the valve body surface is experimentally measured. The numerical results for pressure are in good agreement with the experimental ones.

View full abstract

In order to examine the feasibility of direct simulation of bubbly flow, the applicability of the VOF (volume of fluid) method to analyses of a single rising bubble was examined in this study. Calculated bubble shapes and terminal velocities under wide ranges of Eotvos number and Morton number were compared with the experimental data summarized by Grace et al. Except for the cases in which bubble shapes were spherical-cap and skirted, the VOF method could predict them well by assigning only eight cells to the bubble diameter. Hence, it was confirmed that some modification of this method will enable us to simulate bubbly flow directly under a wide range of flow conditions. The relationship between bubble shape and velocity distribution was also examined within the ranges in which the VOF method is valid. It was found that the secondary vortex appearing in wobbling bubbles induces a velocity component normal to the bubble interface, and this velocity is one of the causes of the wobbling shape of the bubble.

View full abstract

The unsteady motion of small particles, which are uniformly distributed at first, in a square chamber with a pulsating inlet velocity is studied numerically. The force between particle and fluid is assumed to be proportional to the relative velocity. The problem considered is governed by four parameters, the Reynolds number, the dimension ratio (which characterizes the particles as coarse or fine), the period and the amplitude of the pulsating inlet velocity (which are made dimensionless). The motion of the particles is classified into three types : the exhaust type, in which the particles are exhausted through the outlet ; the circulation type, in which the particles circulate in the chamber ; and the adhesion type, in which the particles adhere to the wall of the chamber. The effects of the period and the amplitude of the inlet velocity on the ratio of the number of particles of each type to the total number of particles are examined taking the Reynolds number and the dimension ratio as parameters. The suitable ranges of the period and the amplitude of the inlet velocity, in which the number of exhausted particles becomes large, and the number of circulating particles becomes small, are found, taking the dimension ratio as a parameter. The mechanism of the exhausting of the particle is qualitatively clarified on the basis of the relationship between the streamline and the trajectory of the particles.

View full abstract

Unsteady aerodynamic responses of a mistuned cascade subjected to incoming wakes from an upstream cascade are investigated in this study, where the extended Nishiyama-Funazaki method is employed. Numerical calculations are made, focussing on the effects of stagger angle mistuning of the cascade. It is accordingly found that there is some possibility of reducing wake-induced unsteady forces by controlling stagger angles without the severe expense of cascade performance deterioration. In addition, intense correlation between unsteady lifts and the corresponding steady lifts in the case of in-phase incoming wake conditions is observed.

View full abstract

As a flow model for so-called "secondary instability" in a three-dimensional boundary layer, a flat plate boundary layer with artificially generated embedded streamwise vortices embedded in it is examined using a hot-wire anemometer and flow visualization in order to determine the onset conditions of the secondary instabiliy and its response to acoustic excitation. The results show that the nondimensional parameter, with which the onset conditions of the secondary instability can be predicted, is the product of critical inflection point height and inverse of maximum shear at critical conditions. Transition seems to start when this value reaches about 0. 2. Acoustic disturbance seems to excite the secondary instability at certain conditions. Direct response to the disturbance frequency was observed within a frequency range of 290Hz-450Hz. Below 280Hz, the spectral peak of velocity fluctuation was held constant at 345Hz.

View full abstract

Instability of boundary layers over a concave wall and a rotating disk which were thought to be essentially different in instability sources, are compared in order to investigate whether or not a single crossflow parameter can be defined. Using a newly defined crossflow parameter, prediction was attempted on a yawed cylinder boundary layer transition. By comparing the calculation with the experiment, it was found that this parameter can document fairly well the onset of crossflow instability.

View full abstract

Effects of the centrifugal force due to wall curvature and the Coriolis force due to system rotation on turbulent boundary layers along a convex surface are studied experimentally using a curved channel of a constant cross-section rotating about an axis perpendicular to the main flow. Time-averaged and turbulent components of the velocities inside the channel were measured using hot wire probes during the channel rotation. Besides wall curvature, channel rotation was found to have a large effect on stabilization or destabilization of the turbulent motion in the boundary layer. Stabilizing effects appear more dominant on the suction side, and the flow tends to become laminarized there. When the Coriolis force acts toward the convex surface resulting in higher pressure on that side, the destabilizing effects due to the force cancel the stabilizing effects of the centrifugal force, and the velocity profiles are almost unchanged in the downstream direction. Comparison was made with the turbulence behavior on the concave surface in the same channel.

View full abstract

In this work, we report on the fractal aspect of an iso-velocity set in a turbulent boundary layer. The iso-velocity set is a one-dimensional point set defined by an instantaneous velocity signal. We use both a box-counting algorithm and a method based on the probability distribution function. The analysis is carried out carefully with clarification of the region of scale similarity. The iso-velocity set clearly indicates fractal features, except for near the local mean velocity, and their dimension is less than 0. 4 in space. Scale similarity extends from the Kolmogorov scale to the largest eddy scale. The instantaneous turbulent energy field also shows fractal features and has a close connection with the iso-velocity set.

View full abstract

The experimental results of lift and drag acting on a circular cylinder with tangential blowing immersed in a uniform shear flow are presented for seven shear parameters : K=+0. 15, +0. 045, +0. 030, 0, -0. 030, -0. 045, -0. 15 at the Reynolds number 6×10⁴. The effect of the velocity gradient of the uniform shear flow on the characteristic values is discussed for the coefficients of momentum of the blowing ranging from 0 to 0. 4. In this study, we also discuss in depth the contribution of the characteristic values to the lift. These values include the angles and the pressure coefficients of the stagnation point and the separation points as well as the minimum pressure and base pressure. It was found in this study that the shear parameters strongly affected the coefficients of lift and drag, and different starting points for the forced reattachment phenomenon. All the characteristic variables increasing the lift coefficient can be clearly classified into three categories : the first category was influenced only by the shear parameter, the second, only by the location of the tangential jet, and the third, by both parameters. A comparison of various features such as the stagnation point clarifies the contributions of each feature to the lift and drag on the cylinder.

View full abstract

The third-order upwind finite difference scheme is applied to the convective terms in three-dimensional Navier-Stokes equations to directly simulate the free jet flow emitted from a slit nozzle. The calculated flow field ranges from the nozzle exit to the fully developed turbulent region. The calculated results of turbulent characteristics (mean velocity distributions, turbulent intensity distributions, velocity autocorrelations and power spectra) are compared with the experimental data. The mesh size is roughly ten times the Kolmogorov microscale ; however, the calculated turbulent characteristics represent fairly well the experimental ones, except that the data are strongly dependent upon small-scale eddies. This means that the simulation documents well the motion of large-scale eddies which play an important role in the flow-field formation. Neglecting the small-scale eddies in the third-order upwind finite difference scheme does not greatly affect the calculated results of large-scale eddy motions.

View full abstract

This paper describes the detailed study of the flow structure over rows of two-dimensional square ribs on a ground plane for various values of S/D, and the optimum value of S/D to augment the turbulence of the free stream. The pitch between the centers of two adjoining square ribs was varied at S/D=2, 3, 4, 5, 7, 9, 13 and 17. The time-mean velocity, static pressure and the velocity vectors were measured by Pitot and static pressure tubes, and a cylindrical yawmeter of 6mm in diameter having three pressure holes. The turbulence intensities and autocorrelation were obtained using the data processing system and the F. F. T. analyzer connected to the hot-wire anemometer. As a result, it is concluded that the pitch ratio S/D=9 is optimum for augmenting the turbulence intensity.

View full abstract

Detailed experimental analysis has been made on shock reflection over a wedge with surface roughness in a shock tube. The incident shock Mach number M was 1. 43. The experimental parameters were the reflecting wedge angle and the size of gutters. The wave structure is similar amongst model wedges. Detailed measurement into the triple point trajectory shows that it is not a straight line but a slowly varying curve with a kink point where the slope of the tangent to the curve increases discontinuously and that the kink point coincides with the location where the secondary shock, formed by the accumulation of waves reflected from gutters, overtakes the triple point. Three effective values are introduced for experimental analysis of this essentially unsteady phenomenon. Before the kink point the effective angles of incidence and reflection differ greatly from those obtained from the three-shock theory. The angles agree fairly well with the theory beyond this point.

View full abstract

To estimate the blade performance of a Darrieus-type cross-flow water turbine, differences between the characteristics of a Darrieus blade which moves along a circular path and that of a blade at rest in a uniform steady flow are examined quantitatively through the use of experimental data for three types of blades with different cambers or chord lengths. It is found that a unique stall character strongly affects the drag force acting on the Darrieus blade. A model of the changing pattern of the drag coefficient due to dynamic stall is proposed from the results of measurement for estimating the blade characteristics. As a result, the blade characteristics and performance of the Darrieus-type runner are predicted well by the use of aerofoil data in a uniform steady flow, taking into account the effect of a circling motion and the dynamic effect of the unsteady flow around the Darrieus blade.

View full abstract

A new enhancing technique was developed for natural convection adjacent to a vertical, heated plate. In order to enhance the heat transfer, V-shaped plates of which the edges faced upstream, were attached to a vertical plate in a staggered layout. These plates work not only as an extended surface but also as a heat transfer promoter. The overall heat transfer coefficients of the above enhanced surface were measured and compared with a nontreated, flat surface and a conventional finned surface. The results show that the highest heat transfer performance achieved is for the experimental surface and that the heat transfer coefficient is 40% higher than that of a conventional finned surface with the same total surface area and fin height.

View full abstract

The proposed cold storage method using heat transfers of direct-contact liquid-vapor phase change between the heat transfer medium and water in a closed vessel, whose effectiveness for a small-size vessel was confirmed in the previous report, was adopted and expanded for the actual size of a storage vessel. HCFC142b and water or water solution of a surface active agent were used as the heat transfer medium and cold storage material, respectively. The chemical reaction (hydration) heat of the HCFC142b and water was utilized. As a result, it was confirmed by the experiment using large-scale apparatus that efficient cold storage and release was achieved. In particular, for the water solution of a surface active agent, a high solid packing factor and high rate of cold storage were maintained even for the thickest storage material layer.

View full abstract

Transient laminar free convection in a vertical annulus is investigated numerically. The induced flow is due to a step change in the outer wall temperature, while the inner wall is adiabatic. The range of Grashof number considered is 4lesGγ*les50 000, while the Prandtl number is 0. 7. Results are presented showing the variations of the temperature and velocity profiles with time and location. Also, variations with time of the induced flow, the heat absorbed by the fluid, and the pressure distribution along the annulus are given.

View full abstract

A new reflectometer system is developed to measure the radiation characteristics of solid materials from a heat transfer science point of view. This system employs an ellipsoidal mirror in a simple optical alignment, and measures the spectrum of the normal-incident hemispherical reflectance in the wavelength region of 0. 35 to 1. 10μm. A single spectral scanning gives a hemispherical reflectance spectrum which completely covers both the specular and diffuse components of reflection. The reflectance is directly interpreted in terms of the absorptance for the normal incidence. The system also readily measures the angular characteristics of reflection. The system is applied to a systematic study of the reflection characteristics of rough metal surfaces. The predominant contribution of the specular component to the angular characteristics is demonstrated.

View full abstract

The definition of turbulence of in-cylinder flow fields has been an important issue for researchers working on in-cylinder gas motion. This study proposes a definition of the "turbulence component" in both the calculation and measurement. Also, a new method to make a precise comparison between the calculation and measurement is presented theoretically. This method implies changing the coefficient, C_D, which is included in expressions for turbulence energy, its dissipation rate and turbulence length scale. Employing this method demonstrates the possibility of reducing the ambiguity of the cut-off frequency method to determine turbulence experimentally and also of improving the reliability of the calculation code.

View full abstract

Combined Brayton-Rankine cycles for power generation are calculated using the modulated empirical formulas of gas and steam tables in a computer programme with accurate assumptions. Three computer programmes were created to calculate the thermodynamic properties of the working fluids and to evaluate the performance of combined Brayton-Rankine cycles. Also, many parameters were varied to study the effect of their variation on the peformance of the combined Brayton-Rankine cycle power plants. The computational results for existent power plants were compared with the actual data of those power plants, and an excellent agreement was obtained between them.

View full abstract

In recent years, outdoor units of air conditioners have tended to be set on veranda space on each floor of high-rise buildings. In this situation, it is probable that the intake outlet flows into the outdoor units of air conditioners directly because of atmospheric wind. This is called a short cycle between the inlet and outlet. Atmospheric wind problems around a high-rise building and inlet-outlet flows of outdoor units are solved numerically to investigate the influence of atmospheric wind upon the short cycle. The test building is only a short distance from neighboring buildings. Atmospheric wind is assumed to flow with a constant velocity in a constant direction. The results are compared with experimental data on real buildings to discuss differences between the analytical wind model and natural wind.

View full abstract