|
Article Excerpt
INTRODUCTION
A nanofluid is a new type of heat transfer fluid that operates via nanoparticles suspended in a conventional host liquid (Choi 1995), nanorefrigerant is one kind of nanofluid, and the host fluid of nanorefrigerant is refrigerant (Wang et al. 2005). A nanorefrigerant has a higher heat transfer coefficient than the host refrigerant, and it can be used to improve the performance of refrigeration systems (Wang et al. 2005, Bi et al. 2008). The thermal conductivity of a nanorefrigerant should be known in order to obtain the nanorefrigerant's heat transfer properties and to find the proper ratio of nanoparticles to refrigerant.
Experiments on the thermal conductivity of nanofluids show that a nanofluid has a much higher thermal conductivity than a pure host fluid (Choi 1995; Eastman et al. 1996, 2001; Das et al. 2003; Patel et al. 2003; Chon et al. 2005; Hong et al. 2005, 2006; Liu et al. 2005; Li and Peterson 2006). The host fluids in previous research almost always were limited to water or ethylene glycol (EG), and there are no published papers on the thermal conductivity of nanorefrigerants. As the thermal conductivities of most refrigerants are about one order of magnitude lower than that of water or EG, the characteristics of nanorefrigerants may be different from those of water- or EG-based nanofluids, and experiments on the thermal conductivities of nanorefrigerants featuring a variety of nanoparticles are necessary to determine the characteristics of nanorefrigerants.
Models for predicting nanorefrigerant thermal conductivity also are useful. A nanorefrigerant is one kind of solid-liquid mixture, but experiments show that the classical models for a solid-liquid mixture's thermal conductivity, such as the Maxwell model, Bruggeman model, CSM model, Rayleigh model, Monecke model, and Cichocki-Felderhof model, are not applicable for nanofluids (Wang et al. 2003), so they cannot be used to determine the thermal conductivity of a nanorefrigerant. In recent years, some models on the thermal conductivity of nanofluids have been proposed (Prasher et al. 2005; Jang and Choi 2004; Xuan et al. 2003; Yu and Choi 2003; Wang et al. 2003). These models proved effective on some water- and EG-based nanofluids with suitable empirical constants but have not been validated by experiments on nanorefrigerants.
The objective of this research is to investigate nanorefrigerant thermal conductivity through experimentation and to develop a model for predicting the thermal conductivity of nanofluids, particularly that of nanorefrigerants. First, the experimental setup, principle, and procedure are introduced, and the results are discussed. Then, typical models for the thermal conductivity of nanofluids are compared with experimental results. These models are analyzed on the basis of the above comparison. These analyses help suggest a new model, which will be proposed and validated by the results of the thermal conductivity of nanorefrigerants and other nanofluid experiments.
EXPERIMENTS ON NANOREFRIGERANT THERMAL CONDUCTIVITY
Experimental Setup and Principle
The experimental setup is schematically shown in Figure 1. The nanorefrigerant is put in a vessel, which is placed in a constant temperature bath. A probe covered by a layer of heat resistance is vertically immersed in the nanorefrigerant. The probe is not only a heat producer for the nanorefrigerant but also an electric resistance thermometer. The layer of heat resistance is not only an electrical insulator but also a protection layer for the probe. While the probe is electrified, its electric resistance is measured by a thermal constants analyzer (Hot Disk 2008). Then the measured data are analyzed by a computer and the nanorefrigerant's thermal conductivity is calculated. A thermometer is immersed in the nanorefrigerant to measure its temperature.
[FIGURE 1 OMITTED]
The measuring principle of the thermal constants analyzer is based on the transient plane source method (Gustafsson et al. 1986). There exists a definite relationship between the fluid's thermal conductivity and the temperature increase of the probe if no natural convection occurs in the fluid. The fluid's thermal conductivity, [k.sub.nf], can be calculated by Equation 1,
[k.sub.nf] = [W/[[[pi].sup.1.5][r.sub.prp][DELTA]T(t)]][square root of ([at/[r.sub.prb.sup.2]], (1)
where [DELTA]T(t) is the temperature rise of the fluid near the probe. Because the thickness of the layer is very small, the temperature difference between the fluid and probe can be ignored and [DELTA]T(t) is equal to the temperature rise of the probe, as shown in Equation 2:
[DELTA]T(t) = [1/[alpha]][[[[R.sub.prb](t)]/[R.sub.0]] - 1] (2)
If natural convection occurs in the fluid, the value of calculated [k.sub.nf] based on Equation 1 will vary with time, and the result may be wrong. In this case, the thermal constants analyzer can give an alarm automatically to avoid using an unreliable result.
In order to avoid this natural convection, the parameters of the analyzer should be properly controlled. The parameters of the analyzer in the experiments in this research are shown in Table 1.
Table 1. Parameters of the Thermal Constants Analyzer Power Measurement Sensor Temperature Disk Time Radius Coefficient Type of Resistance 0.01 W 5 s 2.001 mm 0.471/ K Kapton (0.034 Btu/h) (0.006565 ft) (0.262/[degrees]F)
R-113 was used as the host fluid of the nanorefrigerant in the tests because it is in the liquid state under room temperature and normal atmosphere, while more commonly used refrigerants, such as R-134a and R-410A, are in the gaseous state under the same conditions. Nanorefrigerants based on R-113 are also easier to prepare than more commonly used refrigerants. The thermal conductivity deviation of R-113 from other refrigerants is much smaller than that of refrigerants made from water or EG. Therefore, the experimental results of an R-113-based nanorefrigerant still may be useful in identifying the characteristics of nanorefrigerants.
The thermal constants analyzer was calibrated by the comparison between the measured thermal conductivities of pure R-113 at standard atmosphere and those calculated by refrigerant property software, REFPROP 8.0...
|
