IMD

Hiroyuki Kumano a,*, Tetsuo Hirata a, Yosuke Hagiwara b, Fumito Tamura a
a Department of Mechanical Systems Engineering, Shinshu University, 4-17-1 Wakasato, Nagano 380-8553, Japan. b West Japan Railway Co., 2-4-24 Shibata, Kita-ku, Osaka 530-8341, Japan

Abstract:
The effect of storage on flow and heat transfer characteristics of ice slurry was investigated experimentally. After ice slurry had been stored in the storage tank, variations in ice particle size were measured using a microscope, and diameter distribution and average diameter determined. The ice packing factor, Reynolds number and storage time were varied as experimental parameters. The pressure drop and heat transfer coefficient were measured when the ice slurry flowed in the horizontal tube. For laminar flow, the ratios of pipe friction and heat transfer coefficient decreased with storage time. For more than 12 h storage time, the ice slurry could not flow in the tube. The adhesion between ice particles seemed to cause a blockage in the tube. On the other hand, for turbulent flow, the pipe friction and ice slurry heat transfer coefficients were similar to that of the ethanol solution,and the storage effect was insignificant.
© 2011 Elsevier Ltd and IIR. All rights reserved.

1: Introduction
Ice thermal energy storage systems have been identified as effective methods to level electric power loads. In a dynamic type ice thermal energy storage system, ice slurry is used as a phase change material. The ice slurry is fluid and can transport cold thermal energy directly, because it is a mixture of fine ice particles and aqueous solution. Moreover, the ice slurry has a high heat transfer ratio, because the latent heat of fusion of the ice particles can be used and the heat exchange area is very wide. Since ice slurry has as good properties as a second refrigerant, the characteristics of ice slurries in fundamental processes including generation, storage, transportation, and melting have been investigated by many researchers.

In particular, it is important to understand the flow and heat transfer characteristics of ice slurry to design ice thermal energy storage systems. Niezgoda-Zelasko and Zalewski (2006) investigated these flow characteristics experimentally in horizontal tubes. The critical velocity and mass fraction corresponding to a change in the ice slurry character from laminar to turbulent flow were determined. Kitanovski and Poredos (2002) investigated heterogeneous flow in ice slurry analytically. The present authors investigated ice slurry flow characteristics in narrow tubes and found that it can be treated as a pseudoplastic fluid. In addition, they clarified the use of the apparent Reynolds number (Kumano et al, 2010a). Knodel et al (2000) reported the flow and heat transfer characteristics of the ice slurry in a 24mm diameter tube. The slurry velocity was varied from 2,8 to 5,0 m s_1, and a reduction in pressure drop due to flow relaminarisation in the case of a large ice packing factor (IPF) was observed. The flow and melting characteristics of ice slurry have been investigated by many researchers (Guilpart et al, 1999; Lee et al, 2006; Doetsch, 2001; Bellas et al, 2002). The present authors reported the melting heat transfer characteristics of the ice slurry in horizontal tubes (Kumano et al., 2010b). Ayel et al (2003) reviewed the flow and heat transfer behaviours of ice slurries, and Egolf and Kauffeld (2005) reviewed the physical properties of ice slurries.
In a realistic ice thermal energy storage system, ice slurry is produced at night and used for air-conditioning during the day. Therefore, the ice slurry is stored in the storage tank for more than six hours after its production. In the storage process, the size and shape of the ice particles in the ice slurry may vary as a result of various effects, such as Ostwald ripening. Pronk et al (2005) investigated variations in the size and shape of ice particles in slurry during storage, and concluded that Ostwald ripening is the most important mechanism for the increase in average size of the ice crystals. The change in size and shape of the ice particles may influence the flow and heat transfer characteristics. Moreover, agglomeration between ice particles in the slurry may occur and change the slurry flow and heat transfer characteristics. However, the storage effects of the slurry on the flow and heat transfer have not been discussed sufficiently and have therefore been investigated experimentally in this study. Firstly, the size and shape of the ice particles in the ice slurry were observed, and their variation was determined as a result of the storage. Thereafter, the storage time, Reynolds number, Re and IPF were varied as experimental parameters, and the pressure drop and heat transfer coefficients were measured. A comparison between experimental results and theoretical values of the aqueous solution without ice particles was carried out using the pipe friction coefficients and Nusselt number, Nu.

2: Experimental apparatus and procedure
2.1: Variation of ice particle size due to storage
The ice slurry was produced from a 5 weight % ethanol solution. The ethanol solution was cooled down to supercooled state, the supercooled state was then released and the ice slurry was produced. After that, only the ethanol solution was extracted from the ice slurry storage tank, and cooled again to supercooled state. These processes continued until a particular IPF was obtained, after which the ice slurry was used in the experiments. Details of the production unit were described in a previous study (Kumano et al, 2010a), and an explanation of the unit is omitted in this paper. The ice slurry was produced in the production unit and thereafter was stored in the storage tank. Then, the ice particles in the ice slurry were observed using a microscope, and their sizes were measured. The storage time was varied from zero to 24 hours after production of the ice slurry. The ice slurry was stirred using a stirrer during the storage, and the temperature of the ice slurry was maintained at the melting point of the aqueous solution in the ice slurry.
Figure 1 shows photographs of the ice particles in the ice slurry at each storage time. It was found that the size of the ice particles increases gradually with storage time. The ice particle diameter was measured from the photographs and defined as the average of longest and shortest dimensions, with approximately 100 samples being used. Then, all ice particles in one photograph were used as the samples, not to be a factitious measurement. The diameter distribution for each storage time is shown in figure 2. The IPF was set at 15%. The ice particle diameters were distributed between 0,1 and 0,4mm and the percentage of large ice particles increased with storage time. Figure three shows the relationship between average diameter and storage time. The IPF was set at 5, 10 and 15 % and maintained constant through the storage. The average diameter increases with storage time, and the effects of IPF were not significant.

2.2: Experimental apparatus
Figure 4 shows a schematic diagram of the experimental apparatus. Since the details of the apparatus were described in the previous study (Kumano et al, 2010a, b), it is only briefly explained here. The experimental apparatus consisted of a circulation unit for the ice slurry and a measurement unit. The circulation unit consisted of a storage tank, gear pump, entrance section, test section and coriolis-type mass flow meter. The entrance and test sections were both one-metre in length, and the diameter of the tube was 7,5mm. The pressure drop and heat transfer coefficients were measured in the test section. The flow rate and density of the ice slurry were measured by the coriolis-type mass flow meter. The flow rate was used to calculate Re for the ice slurry, and the density of the ice slurry was used to calculate the IPF. The test section consisted of a stainless steel tube with nichrome foil heater wrapped around the tube. The tube was heated at constant heat flux. The heating section was 0,8m into the test section. T-type thermocouples, 0,1mm in diameter, were inserted in the stainless steel tube to measure the temperature of the inner surface of the tube. The thermocouples were inserted 0,1; 0,4 and 0,7m from the beginning of the heating section at upper, middle and lower positions. The temperature of the ice slurry was measured using T-type thermocouples at the front of the entrance section and the rear of the test section to determine the coefficient of kinematic viscosity and the density of the ethanol solution in the ice slurry. The heat transfer coefficients were determined from the ice slurry temperature and the inner surface of the tube, and the heat flux was given by the heater as follows.


2.3: Experimental conditions and procedure
Table 1 shows the experimental conditions, Re, IPF and storage time, used in the study. Re was varied from 1 000 to 2 000 for laminar flow and from 4 500 to 7 500 for turbulent flow and calculated using the kinematic viscosity of the ethanol solution. The kinematic viscosity of the ethanol solution was measured in our previous study (Kumano et al, 2010a) and calculated from the temperature of the ice slurry. IPF was varied from zero to 20 percent, and the heat flux, which was maintained constant through the experiments, was 5 000W m_2. The IPF was decided from the densities of the ice slurry, ethanol solution and ice. The storage time was varied from zero to six hours. In particular, the storage time was varied from zero to 12 hours to clarify its effect, as the IPF was maintained at 15 percent. In the experiments, the pressure drop and heat transfer coefficients of the ice slurry were measured as follows. After production of the ice slurry, it was stored while stirring in the storage tank for the prescribed storage time.




Thereafter, the ice slurry was circulated to the test section, and the temperature of the inner surface in the test section, density of the ice slurry and flow rate were measured. These measurements were carried out at five-second intervals. In this study, the IPF decreased during measurement of the heat transfer coefficients since heat flowed from the heater and pump. Therefore, the viscosity of the ethanol solution in the ice slurry varied owing to the variation in IPF. Thus, the flow rate was controlled so that Re was constant throughout the measurement. It seems the size of the ice particles in the ice slurry decreases with decrease in IPF. Therefore, the initial IPF was set at 10 and 20 percent, and when it had decreased by about 10 percent from the initial value, the measurements were ended to exclude the effect of variation in size of the ice particles in the ice slurry.