Yannis Hardalupas

Anomalous heat transfer modes of nanofluids: a review based on statistical analysis

This paper contains the results of a concise statistical review analysis of a large amount of publications regarding the anomalous heat transfer modes of nanofluids. The application of nanofluids as coolants is a novel practise with no established physical foundations explaining the observed anomalous heat transfer. As a consequence, traditional methods of performing a literature review may not be adequate in presenting objectively the results representing the bulk of the available literature. The current literature review analysis aims to resolve the problems faced by researchers in the past by employing an unbiased statistical analysis to present and reveal the current trends and general belief of the scientific community regarding the anomalous heat transfer modes of nanofluids. The thermal performance analysis indicated that statistically there exists a variable enhancement for conduction, convection/mixed heat transfer, pool boiling heat transfer and critical heat flux modes. The most popular proposed mechanisms in the literature to explain heat transfer in nanofluids are revealed, as well as possible trends between nanofluid properties and thermal performance. The review also suggests future experimentation to provide more conclusive answers to the control mechanisms and influential parameters of heat transfer in nanofluids.

Breakup Phenomena in Coaxial Airblast Atomizers

The breakup of a liquid jet with length-to-diameter ratio of 22 surrounded by a coaxial flow of air has been examined by a combination of high-speed photography and phase-Doppler velocimetry. The air-to-liquid momentum and kinetic energy ratios, the Reynolds number of the coaxial water and air jet flows and the exit-plane Weber number have been varied over extensive ranges and the results examined in terms of the breakup length, frequency, droplet size distributions and velocity characteristics. The photographs reveal the deterministic nature of the liquid flow at Reynolds numbers which are sufficient to guarantee turbulent flow, with the formation of a wave-like structure for a short distance followed by the formation of a liquid cluster and subsequent breakup into ligaments and droplets, with the entire process repeated in a periodic manner. Attempts are made to relate the breakup length and the frequency of the process to the air-to-liquid momentum and energy ratios, the exit Weber number and the slip velocity between the two streams at the nozzle exit. The results confirm that the ratio of the frequencies of the wave-like structures and breakup decreased with the slip velocity between the two streams and asymptotically approached a value of around one for values higher than 150 m s-1. The photographs indicate that the droplet sizes in the sprays are due mainly to disintegration of liquid clusters produced after the initial breakup of the liquid jet and the phase Doppler measurements confirm that most of the liquid remained close to the centreline, where the mean diameter reached a maximum and the slip velocity between the droplets and the air flow was low. An atomization model based on the value of the local Weber number on the centreline of the sprays is used to explain the size characteristics of the sprays. The atomization process was affected by the air-to-liquid momentum ratio at the nozzle exit, the annular width of the coaxial atomizer, the liquid-to-air density ratio, the surface tension and the kinematic viscosity and density of the air. The rate of spread of the sprays close to the nozzle reduced with increase of the air and liquid flowrates and was affected by the initial breakup of the liquid jet and the amplitude of the wave-like structure of the liquid jet during breakup rather than by the air flow turbulence.

Velocity and size characteristics of liquid-fuelled flames stabilized by a swirl burner

Velocity and droplet size characteristics of an unconfined quarl burner, of 16 mm quarl inlet diameter, have been measured with a phase-Doppler anemometer at a swirl number of about 0.29: the Reynolds number of the flow was 30000, based on the cold bulk velocity of 30.4 m s-1 and the hydraulic diameter. The atomization was achieved by shear between the swirling air and six radial kerosene jets and the resulting Sauter and arithmetic mean diameters were about 70 and 50 μm respectively after injection: velocity characteristics are presented for three 5 μm-wide size classes, 10, 30 and 60 μm. The flows correspond to no combustion and combustion of natural gas with a heat release of 8 kW supplemented by liquid kerosene flow rates sufficient to generate 21.6 and 37.2 kW : the gas equivalence ratio was 0.45 and atomized kerosene at two flow rates increased the overall ratios to 1.64 and 2.53. In non-reacting flow, droplets 30 μm and smaller are sufficiently small to be entrained by the mean air velocity towards the central part of the flow and into the swirl-induced recirculating air bubble. The 60 μm droplets are able to travel through the bubble uninfluenced by turbulent fluctuations in the air and are ‘centrifuged’ away from the centreline, through acquisition of a mean swirl velocity component, so that a large proportion of the kerosene volume flow rate lies at the edge of the swirling jet. Because larger droplets are centrifuged to the outer part of the flow, whereas the smaller are entrained towards the centreline, the Sauter and arithmetic mean diameters are, by 1.22 quarl exit diameters downstream of the quarl, approximately 65 and 36 μm at the outer part of the flow and 35 and 12 μm near the centreline in the inert flow. In reacting flow, droplets evaporate rapidly in regions of elevated temperatures and hence no droplets are found within the flame brush and recirculation region. The aerodynamic response of each size class to the air velocity is similar to inert flow so that the majority of the kerosene flow is centrifuged away from the flame. On exit from the quarl, the evaporation and burning rates cause the Sauter and arithmetic mean diameters to be about 70 and 50 μm and 60 and 30 μm at the inner and outer edges of the spray respectively. By 1.22 quarl exit-diameters from the exit of the quarl, the air motion entrains droplets smaller than about 30 μm towards the flame, at the inner edge of the spray, so that the Sauter and arithmetic mean diameters are 60 and 40 μm at the outer edge of the jet. There is comparatively little effect of changing the flow rate of kerosene because the combustion is controlled by the low available number of smaller droplets, although the Group combustion number corresponds to ‘cloud’ burning. The relative response of droplets to the mean and turbulent components of air motion, including the ‘centrifuging’ effect, can be scaled to other flows through dimensionless numbers defined in the text.

Shadow Doppler technique for sizing particles of arbitrary shape

The output from a linear diode array is used in a modified laser Doppler velocimeter to measure the size and shape of irregular particles. The sizing accuracy for transparent and opaque particles between 30 and 140 µm is better than 10%. The inaccuracy caused by trajectories that lay at angles of less than 24° to the axis of the array was less than +5%, and a further inaccuracy of +5% was caused by defocusing of the particle from the center of the velocimeter measuring volume by up to ±500 µm. The advantages of the shadow Doppler technique over other techniques for sizing irregular particles, such as amplitude systems with pointer volumes, are that the shadow Doppler technique records shape, the optical arrangement is more robust, less precise alignment is required, and the equipment can be constructed at low cost.

Electromagnetically controlled multi-scale flows

We generate a class of multi-scale quasi-steady laminar flows in the laboratory by controlling a quasi-two-dimensional shallow-layer brine flow by multi-scale Lorentz body forcing. The flows’ multi-scale topology is invariant over a broad range of Reynolds numbers, Re2D from 600 to 9900. The key multi-scale aspects of this flow associated with its multi-scale hyperbolic stagnation-point structure are highlighted. Our multiscale flows are laboratory simulations of quasi-two-dimensional turbulent-like flows, and they have a power-law energy spectrum E(k) ∼ k −p over a range 2π/L < k < 2π/η where p lies between the values 5/3 and 3 which are obtained in a two-dimensional turbulence that is forced at the small scale η or at the large scale L, respectively. In fact, in the present set-up, p + Ds = 3 in agreement with a previously established formula; Ds ≈ 0.5 is the fractal dimension of the set of stagnation points and p ≈ 2.5. The two exponents Ds and p are controlled by the multi-scale electromagnetic forcing over the entire range of scales between L and η for a broad range of Reynolds numbers with separate control over L/η and Reynolds number. The pair dispersion properties of our multi-scale laminar flows are also controlled by their multi-scale hyperbolic stagnation-point topology which generates a sequence of exponential separation processes starting from the smaller-scale hyperbolic points and ending with the larger ones. The average mean square separation ∆ 2 has an approximate power law behaviour ∼t γ with ‘Richardson exponent’ γ ≈ 2.45 in the range of time scales controlled by the hyperbolic stagnation-points. This exponent is itself controlled by the multi-scale quasi-steady hyperbolic stagnation-point topology of the flow.

Spatial distribution of fluorescence intensity within large droplets and its dependence on dye concentration

The dependence of fluorescence intensity distributions within droplets on added dye concentration has been calculated by extension of the geometrical-optics approximation and verified by experimental observations. With rising dye concentration, surface plots of the equatorial fluorescence pattern show decreasing relevance of intensity enhancement at focusing points of internal light rays and increasing effects of linear absorption on the characteristic features of the distribution. For comparison with experimentally obtained images of the fluorescence intensity distribution within droplets, a method for calculating volume-integrated intensity distributions was developed in which image distortion at the fluid-air interface is included. A comparison of the calculated and the experimentally determined fluorescence intensity distributions within a droplet confirmed the accuracy of the geometrical-optics approach at high dye concentrations. However, discrepancies from experimental results are visible at low dye concentrations owing to nonlinear optical effects.

Characteristics of the spray from a diesel injector

Abstract Spatial and temporal profiles of the velocity of the entrained air and 60 and 30 μm droplets, together with the associated fluxes, from a 5-hole diesel spray exhausting into atmosphere at a repetition rate of 10 Hz have been measured with a phase-Doppler anemometer. The nozzle diameters, fuel charge per hole, injection duration and the area-averaged spray velocity during this duration were 0.18 mm, 2.35 mm3, 0.7 ms and U0 = 132 m/s, respectively. The Sauter mean diameter of the fuel droplets decreased from a maximum centreline value of around 80 μm at 100 diameters from the nozzle to 38 μm at 780 diameters, and a similar decrease was observed between 1 and 2 ms after the start of injection at the upstream location. The flux carried by the 30 μm droplets was up to twice that associated with the 60 μm droplets, 2 ms after injection, although the velocities of the larger droplets were consistently higher than those of the smaller droplets. The maximum measured ensemble-averaged relative velocity was 0.45 U0 for 60 μm droplets just after the arrival of the spray at 550 diameters from the nozzle. The magnitudes of the Weber number imply that droplet breakup was always confined to the leading edge of the spray and was limited to, at most, the initial 1 2 ms of the passage of the spray past a given point. Breakup was mostly complete by 550 diameters from the nozzle. Thus, the measured decrease in the mean diameter was due to small droplets, generated by breakup at the leading edge of the spray, losing velocity due to aerodynamic drag and falling behind the leading edge. Droplets generated late in the injection schedule were likely to overtake those generated earlier and together with the fan-spreading effect, which arises from the combination of the root mean square (RMS) droplet radial velocity and the radial profile of the ensemble-averaged droplet axial velocity, led to RMS velocities in the axial component of the droplets that were not associated with the transfer of turbulent motion from the air.

On the measurement of particle concentration near a stagnation point

Conclusions(1)This work has evaluated the particle number density measured by a single particle counting instrument, based on either the particle mean velocity or on the particle residence time in the measuring volume.(2a)In regions where the mean velocity of a size class is near zero the number density should be based on the residence time of the particle, Eq. (3).(2b)Equation (3) can be used for two-dimensional flow and removes the need to measure the magnitude of the velocity vector, as is the case with the definition of Eq. (1), provided that V(di) — and hence A(di) — is known. A twochannel laser-Doppler anemometer, however, permits the direct, “on-line” measurement of V(di) and A(di).(3)In regions where the mean velocity of all size classes is non-zero, there is little difference in the values of the Sauter mean diameter returned by the two equations.(4)In instruments which do not have the facility for measuring the residence time, it is suggested that the Sauter mean diameter should be evaluated directly from the measured value of ni.(5)For instruments based on laser-Doppler anemometry, the correction for the effect of frequency shifting on the cross-sectional area, A(di) and volume, V(di), of the anemometer is of the order of 25% for small particles and for Nf/N0 = 1.4.(6)Further work is required to establish the theoretical foundation of Eq. (3) in relation to the work of Buchhave et al. (1979). The accuracy with which C(di) can be measured is determined by the tolerances on A(di), V(di) and zp.Further experimental work is also required to determine the accuracy with which these quantities are known.

Experiments with Disk Stabilized Kerosene-Fuelled Flames

Abstract Kerosene-fueled flames with the spray emanating from a commercial atomizer positioned on the axis of stabilizing disks with blockage ratios of 0.74, 0.56 and 0.39 without and with a quarl diffuser are characterised in terms of flammability limits, droplet velocities according to their size, mean droplet sizes,liquid fluxes and local temperatures. Detailed measurements were obtained with a disk blockage ratio of 0.74, for an overall equivalence ratio of 0.26 and a Reynolds number of 53700 based on the area averaged velocity of the air in the annulus upstream of the bluff body and the outer diameter of the pipe. The velocities and liquid fluxes are presented for three 5 µm size classes, namely 10-15, 30-35 and 50-55 µlm, to emphasize the behaviour of small, medium and large droplet sizes in the region of flame stabilization and isothermal-flow results are presented in the absence of the quarl to allow comparison of the near disk flow with and without combustion. The results show that decrease of th...

Investigating the use of nanofluids to improve high heat flux cooling systems

Abstract The thermal performance of high heat flux components in a fusion reactor could be enhanced significantly by the use of nanofluid coolants, suspensions of a liquid with low concentrations of solid nanoparticles. However, before they are considered viable for fusion, the long-term behaviour of nanofluids must be investigated. This paper reports an experiment which is being prepared to provide data on nanofluid stability, settling and erosion in a HyperVapotron device. Procedures are demonstrated for nanofluid synthesis and quality assessment, and the fluid sample analysis methods are described. The end results from this long-running experiment are expected to allow an initial assessment of the suitability of nanofluids as coolants in a fusion reactor.