Performance Study and Efficiency Improvement of Ice Slurry Production by Scraped-Surface Method

In this study, the performance of ice slurry production by scraped-surface method was experimentally investigated. Temperature change characteristics, ice packing fraction (IPF) of ice slurry, power consumption of scraping system and coefficient of performance (COP) were measured by varying the concentration of sodium chloride solution, scraping speed, and solution flow rate. The effect of nanosilica on efficiency of ice slurry production was also studied. The results showed that scraping power consumption accounted for only a small proportion (about 5%) of the total power consumption of the system. An increase in the concentration of sodium chloride caused a decrease in the IPF and a decrease in the COP of the system. With the solution flow rate at 1.3 m3/h and scraping speed at 13 rpm, the maximum COP (2.43) was obtained. Furthermore, the addition of nanosilica had a significant effect on improving the system COP.


Introduction
Over the last three decades, interest in using phase-change ice slurry coolants has grown significantly [1,2].Ice slurry refers to a mixture of small ice crystals (typically 0.1 to 1 mm in diameter) and carrier fluid.Ice slurry cooling is a promising technology, because of the high energy storage density associated with the latent heat of phase change and the fast cooling rate due to the large surface area available for heat transfer created by its numerous crystals.Due to these features, ice slurry can be used in many applications, such as comfort cooling of buildings, mine cooling, food cooling and medical cooling [3][4][5].
The major requirement for the widespread use of ice slurry cooling is a reliable, energy efficient and cost-effective production technology.Various methods have been proposed for ice slurry production, including the scraped-surface method [6,7], supercooling method [8], direct-injection or direct heat exchange method [9,10], fluidized-bed method [11], and vacuum method [12,13].Among these methods, the scraped-surface method is the most technologically developed and widely accepted method.The major advantage of existing scraped-surface ice slurry generators over other ice slurry generation technologies is that the mechanical agitation results in exceptionally high heat transfer rates that translate into rapid cooling rates providing an excellent end product.Furthermore, scraped-surface ice slurry generators offer a modular design, which allows for easy expansion with growth in demand.The limitation of scraped-surface ice slurry generators is the minimum concentration of the freezing point depressant that can be used on the ice slurry side to generate the ice.At very low depressant concentrations the continuous accumulation of ice layers on the ice slurry generator walls cannot be prevented which would eventually block the rotation of the scraper-blades and cause the freeze-up of the ice slurry generator.Increasing the additive concentration prevents freeze-up, however, it would reduce heat transfer rates and affect the coefficient of performance (COP) of the system.One additional drawback of existing scraped-surface ice slurry generators is that the rotating scrapers, brushes and orbital rods wear over time and have to be replaced at a given time interval.
Existing studies on the scraped-surface method for making ice slurry mainly focus on the heat and mass transfer characteristics of the ice crystal formation process [14][15][16][17].Qin et al. [14] studied the heat transfer of the scraped-surface heat exchanger via laboratory experiments, and found that the heat transfer coefficient with phase change was about three to five times greater than that without phase change.Martinez et al. [15] investigated the effects of two different scrapers on the heat transfer coefficient and found that the use of adaptable scrapers greatly improved the heat transfer coefficient.However, existing research also shows that the overall energy consumption of this method is higher than that of the current heating, ventilation, air conditioning and refrigeration (HVAC and R) system due to the high energy demands of ice makers, and this is also an important reason why this method has not been widely used [18].Researchers generally believe that the main reason for the high energy consumption of scraped-surface method is that the continuous rotation of the scraping system during the ice-making process requires a large amount of energy [19].In this context, Matsumoto et al. [20] studied the scraping force of ice growing on cooling surface by varying the degree of supercooling, ice formation times, concentrations of the solution, surface temperatures, surface roughness and number of scrapings.The results showed that these factors affected the scraping force to some extent, but the power consumption of the scraping system has not yet been reported.Up to now, few studies have evaluated the scraping power consumption.Moreover, the effects of factors such as scraping speed and solution flow rate on the energy consumption during the scraping process have not been reported.Therefore, it is necessary to study the effects of these factors to increase the efficiency of ice slurry production.
Improving the crystallization and accelerating the formation of ice crystals are important areas of research in recent years.In previous reports, researchers have pointed out that nanoparticles can reduce the supercooling degree of base fluid and enhance thermal conductivity significantly [21][22][23][24][25].The supercooling degree is usually defined as (T-T f ) K, where T is the temperature of the theoretical freezing point and T f is the temperature immediately before supercooling dissolution [26].In this paper, we have followed this definition.Zhang et al. [27] investigated the nucleation of nanofluids and found that the supercooling degree was reduced by adding nanoparticles, such as α-Al 2 O 3 , γ-Al 2 O 3 and SiO 2 , into pure water.Harikrishnan et al. [28] studied the phase transition temperature and latent heat of CuO-oleic acid nanofluids, and concluded that CuO-oleic acid nanofluids can decrease the solidification time and can be recommended as better phase change materials (PCMs) for thermal energy storage applications.Wu et al. [29] investigated the supercooling degree of Al 2 O 3 -H 2 O nanofluids and found that the supercooling degree of 0.2 wt% nanofluids was reduced by 70.9% and the starting time of freezing was advanced by 32.9%.He et al. [30] investigated TiO 2 -BaCl 2 -H 2 O nanofluids and found that the supercooling degree was reduced by 84.92% with volume fraction of 1.13%.Jia et al. [31] investigated the solidification behavior of TiO 2 nanofluid and deionized water.The experimental results showed that the supercooling degree of TiO 2 nanofluids without surfactants was approximately 11.5% lower than that of deionized water, and the values did not change significantly with nanoparticle concentration.Wang et al. [32] reported that the supercooling degree of Cu/pure water nanofluids with 0.1 wt% concentration was reduced by 20.5%.Liu et al. [33,34] researched the supercooling degree and nucleation behavior of graphene oxide nanofluids and reported that many nucleation sites were found on the surface of graphene oxide nanosheets, which could reduce the supercooling degree of deionized water dramatically.However, these experimental results are based on small sample sizes.At present, few investigations and applications for ice slurry production with nanofluids have been reported.It is not clear whether nanofluids can be applied to large-scale production of ice slurry.
In this study, the effects of solution characteristics and system operating parameters on ice-making characteristics and energy consumption were studied.Simultaneously, nanosilica was used as an additive and its effect on enhancing ice crystal formation and improving system energy efficiency was investigated.

Experimental Apparatus
The surface-scraped ice slurry generation system, as shown in Figure 1, includes an ice slurry generator, a condenser, a compressor, an expansion valve, a rotating blade, a circulating pump, and an ice slurry storage tank.The height of the ice slurry storage tank is 750 mm.The inner diameter is 780 mm and the outer diameter is 800 mm.The ice-making solution stored in the ice slurry storage tank is sent to the annular pipe at the top of the ice slurry generator by the circulation pump and sprayed on the inner wall surface of the scraping tank.The ice slurry generator consists of a circular shell-and-tube type heat exchanger, cooled on its outer shell side by an evaporating refrigerant, and attached with spring loaded rotating blades on its inner side to scrape the ice and prevent any crystal deposits on the cooled surface, as shown in Figure 2. The height of the heat exchanger is 400 mm.The inner diameter of the inner tube is 340 mm and the wall thickness is 10 mm.The inner diameter of the outer shell is 420 mm and the wall thickness is 40 mm.The detail description of scraping system is presented in Table 1.During the experiment, the rotating blades are driven by the motor reducer, and the rotation speed can be adjusted.
an ice slurry storage tank.The height of the ice slurry storage tank is 750 mm.The inner diameter is 780 mm and the outer diameter is 800 mm.The ice-making solution stored in the ice slurry storage tank is sent to the annular pipe at the top of the ice slurry generator by the circulation pump and sprayed on the inner wall surface of the scraping tank.The ice slurry generator consists of a circular shell-and-tube type heat exchanger, cooled on its outer shell side by an evaporating refrigerant, and attached with spring loaded rotating blades on its inner side to scrape the ice and prevent any crystal deposits on the cooled surface, as shown in Figure 2. The height of the heat exchanger is 400 mm.The inner diameter of the inner tube is 340 mm and the wall thickness is 10 mm.The inner diameter of the outer shell is 420 mm and the wall thickness is 40 mm.The detail description of scraping system is presented in Table 1.During the experiment, the rotating blades are driven by the motor reducer, and the rotation speed can be adjusted.
The solution height of the ice storage tank is about 220 mm.There are three Pt100 thermometer probes (Toprie, Shenzhen, Guangdong, China) within the tank, which are set at a height of 40 mm, 120 mm and 200 mm from the bottom respectively.The tank is equipped with a stirring paddle.During the ice making process, the stirring paddle automatically opened for 30 s every 570 s, so that the ice slurry solution was evenly mixed.At this time, the real-time temperature was recorded, and it was found that the display temperatures of the three Pt100 thermometer probes were almost the same.Therefore, the measured temperature was the actual temperature of the ice slurry.The volume flow rate of nanofluids was measured by a turbine flowmeter (LWGYB-15, Fanyang, Shanghai, China).The energy consumption of the total system, and that of the rotational scraping system were measured by two separate digital watt-hour meters (DTS1330-LCD, Zhnqi, Suining, Jiangsu, China).
The reliability of the measurement results is represented through the uncertainty analysis.The uncertainties of direct measurement parameters given by the manufacturers are listed in Table 2, including temperature, volume flow rate, power consumption and weight.The sources of measurement errors were analyzed and the uncertainties of derivation parameters, such as COP and ice packing fraction (IPF), were evaluated [35].The uncertainty of particle size of nanosilica was obtained through the analysis of the TEM images.The values of the uncertainty of these derivation parameters are also given in Table 2.

Experimental Procedure
All the experiments were performed following the same procedure.A brine solution or nanofluid was prepared with certain concentrations and then poured into the ice slurry storage tank.Then, the volume flow rate of solution and blade scraping speed were set to certain values.The The solution height of the ice storage tank is about 220 mm.There are three Pt100 thermometer probes (Toprie, Shenzhen, Guangdong, China) within the tank, which are set at a height of 40 mm, 120 mm and 200 mm from the bottom respectively.The tank is equipped with a stirring paddle.During the ice making process, the stirring paddle automatically opened for 30 s every 570 s, so that the ice slurry solution was evenly mixed.At this time, the real-time temperature was recorded, and it was found that the display temperatures of the three Pt100 thermometer probes were almost the same.Therefore, the measured temperature was the actual temperature of the ice slurry.The volume flow rate of nanofluids was measured by a turbine flowmeter (LWGYB-15, Fanyang, Shanghai, China).The energy consumption of the total system, and that of the rotational scraping system were measured by two separate digital watt-hour meters (DTS1330-LCD, Zhnqi, Suining, Jiangsu, China).
The reliability of the measurement results is represented through the uncertainty analysis.The uncertainties of direct measurement parameters given by the manufacturers are listed in Table 2, including temperature, volume flow rate, power consumption and weight.The sources of measurement errors were analyzed and the uncertainties of derivation parameters, such as COP and ice packing fraction (IPF), were evaluated [35].The uncertainty of particle size of nanosilica was obtained through the analysis of the TEM images.The values of the uncertainty of these derivation parameters are also given in Table 2.

Experimental Procedure
All the experiments were performed following the same procedure.A brine solution or nanofluid was prepared with certain concentrations and then poured into the ice slurry storage tank.Then, the volume flow rate of solution and blade scraping speed were set to certain values.The cooling unit was switched on.The solution in the ice storage tank was continuously cooled, and the system power consumption was recorded after the temperature of the solution reached 5 • C.Then, ice began to grow on the inner wall and was scraped.Ice slurry sample was taken every 10 minutes and the ice packing fraction was measured.The coefficient of performance (COP) is defined as the ratio of available cooling by the refrigeration cycle, Q and system power consumption, W, which is measured by digital watt-hour meter: where m IS is the total mass of ice slurry, c p (T) is the average specific heat capacity of the solution, T 1 and T 2 are the initial and final temperatures of the solution, α is the ice packing factor (IPF) of the ice slurry, and L the melting latent heat of ice.The measurement of IPF was based on calorimetry method and the energy conservation principle.It was assumed that there was negligible heat loss during the measurement process.The value of IPF was calculated by measuring the initial and final temperatures and masses of the sample after mixing with 200 g of hot water at 40 • C. The room temperature was controlled at 25 • C by an air conditioner and three adjusting heaters.The fluctuation of the ambient temperature was ±0.5 • C. The IPF of each ice slurry sample for different ice-making time was measured three times and the average value was used for analysis.
In order to verify the reliability of this method for IPF measurement, the solid crushed ice with masses of 10, 20, 30, 40 and 50 g was weighed and added to 150 g of liquid water at 0 • C. The ice slurries with theoretical IPF of 6.25%, 11.76%, 16.67%, 21.05% and 25.0% were obtained, and then the IPF of each ice slurry sample was measured according to the calorimetry method.The calculated results were 6.12%, 11.64%, 16.83%, 21.19% and 24.73%, respectively.Comparing the above data, it can be seen that the error between the experimental value and the theoretical value is small.Therefore, it is considered that the IPF obtained by the calorimetry method is reliable.

Materials
The purity of sodium chloride is over 99.5%.The silica nanofluid with average particle diameter of 25 nm used in the experiments was supplied directly by Shenzhen Jingcai Co., Ltd (Shenzhen, China) at an original concentration of 30 wt%.The base fluid of the silica nanofluid provided by the manufacturer is deionized.We were unable to determine the different surfactants or dispersants the manufacturer employed as additives to stabilize the nanofluids, because this is proprietary information.During the experiment, a certain amount of deionized water and sodium chloride were added to dilute the mother nanofluids to obtain the desired concentration of the silica nanofluids.Each time 1 kg of nanofluid was prepared and the composition is shown in Table 3.The diluted nanofluids were subjected to ultrasonic vibration for 2 h.In order to verify whether agglomerated particles were present in the diluted nanofluids, the nanofluids were examined for particle size distribution under a transmission electron microscope (TEM, TECNAI G2F20, FEI, Hillsboro, OR, USA). Figure 3 shows a TEM image of 0.5 wt% silica nanofluid with an average size of 25 nm.The particle size result for this nanofluid is consistent with the data provided by the manufacturer.We noticed from the TEM image that a majority of nanoparticles fell near this average size, with a few smaller and a few larger particles, which should yield an overall average particle size of 25 nm.From similar TEM images of the other nanofluids with different particle size and different mass concentration, no agglomeration was observed.This was due to: (a) the nanofluid manufacturers have developed successful surfactants or dispersants that are already present in these purchased nanofluids, making them stably suspended and free from agglomeration or coagulation; (b) the sonication of the dilute sample ensured breaking off of agglomerated particles.The main variation parameters involved in the calculation of IPF include the specific heat of aqueous sodium chloride solution and silica nanofluid, which vary with the concentration of the additive.The specific heat of the aqueous sodium chloride solution also changes with temperature.During the process of the measurement of IPF, the average temperature before and after the ice slurry melts is approximately 273.1 K, so the specific heat at 273.1 K is used as the basis for IPF calculation.And the specific heat can be obtained by querying the handbook of physical property parameter.The specific heat of aqueous sodium chloride solution at different concentrations is shown in Table 4.
The specific heat of the nanofluid is evaluated according to the formula [36]: where Cp,nf is the specific heat of the nanofluid, ψ is the mass fraction of the nanosilica, Cp,n is the specific heat of the nanosilica, and Cp,f is the specific heat of the base fluid.This formula has been found appropriate for use with nanofluids as validated by O'Hanley et al. and Ferrouillat et al [36,37].The main variation parameters involved in the calculation of IPF include the specific heat of aqueous sodium chloride solution and silica nanofluid, which vary with the concentration of the additive.The specific heat of the aqueous sodium chloride solution also changes with temperature.During the process of the measurement of IPF, the average temperature before and after the ice slurry melts is approximately 273.1 K, so the specific heat at 273.1 K is used as the basis for IPF calculation.And the specific heat can be obtained by querying the handbook of physical property parameter.The specific heat of aqueous sodium chloride solution at different concentrations is shown in Table 4.The specific heat of the nanofluid is evaluated according to the formula [36]: where C p,nf is the specific heat of the nanofluid, ψ is the mass fraction of the nanosilica, C p,n is the specific heat of the nanosilica, and C p,f is the specific heat of the base fluid.This formula has been found appropriate for use with nanofluids as validated by O'Hanley et al. and Ferrouillat et al [36,37].
In the experiment, 4 wt% sodium chloride aqueous solution was used as the base liquid.Therefore, the calculation conditions of the specific heat of silica nanofluids are the specific heat of 4 wt% aqueous sodium chloride solution (as shown in Table 4) and the specific heat of nanosilica, which equals to 1000 J/kg K.The specific heat of silica nanofluid at different concentrations is shown in Table 5.

Influence of Concentration of Sodium Chloride Solution
The temperature change characteristics in the ice slurry storage tank were measured by varying the sodium chloride concentration in the range of 1 wt%, 2 wt%, 3 wt%, 4 wt%, 5 wt%, and 6 wt% under the condition of the fixed rotary scraping speed (n = 29 rpm) and flow rate of the solution (q = 1.0 m 3 /h).When the temperature of the aqueous solution dropped to 5 • C, the temperature change began to be recorded.Taking 3 wt% as an example, the relationship between temperature and time is shown in Figure 4.The temperature of the solution kept falling over time.When the temperature dropped to the freezing point temperature (−1.8 • C), it was found that the ice crystals became continuously formed and the temperature was basically unchanged.

Influence of Concentration of Sodium Chloride Solution
The temperature change characteristics in the ice slurry storage tank were measured by varying the sodium chloride concentration in the range of 1 wt%, 2 wt%, 3 wt%, 4 wt%, 5 wt%, and 6 wt% under the condition of the fixed rotary scraping speed (n = 29 rpm) and flow rate of the solution (q = 1.0 m 3 /h).When the temperature of the aqueous solution dropped to 5 °C, the temperature change began to be recorded.Taking 3 wt% as an example, the relationship between temperature and time is shown in Figure 4.The temperature of the solution kept falling over time.When the temperature dropped to the freezing point temperature (−1.8°C), it was found that the ice crystals became continuously formed and the temperature was basically unchanged.temperature decreased.The sodium chloride concentrations of 1 wt%, 2 wt%, 3 wt%, 4 wt%, 5 wt%, and 6 wt% corresponded to the ice slurry formation temperatures of −0.6, −1.2, −1.8, −2.4,−3.0, and −3.6 • C, and the theoretical freezing point temperatures of −0.6, v1.2, v1.8, v2.4,−3.0, and −3.6 • C, respectively.Therefore, the ice slurry formation temperatures were consistent with the theoretical freezing point temperatures.This indicates that the supercooling degree of this ice-making method is almost nonexistent.Figure 6 displays the relationship between the IPF of the ice slurry and ice-making time.At the same ice-making time, higher initial concentration of aqueous sodium chloride solution resulted in lower IPF of ice slurry, suggesting that the formation of ice slurry became slower.This is mainly because the increase in sodium chloride concentration lowered the freezing point of the solution and increased the system cooling load.The total power consumption of the ice-making system and the power consumption of the scraping system were monitored.Taking 3 wt% sodium chloride solution as an example, the experimental results are shown in Figure 7.When the IFP of the ice slurry reached 13.9%, the total power consumption was 1.9 kWh.The power consumption of the scraping system was 0.1 kWh, which accounted for only 5.26% of the total power consumption.Experimental results at other sodium chloride concentrations also indicated that the scraping power consumption accounted for about 5% of the total power consumption of the system.Thus, the results suggest that the scraping system does consume a certain amount of energy, but it does not account for a large proportion.Compared with the supercooling method, the scraping method has the advantages of simple equipment, small floor space, and easy temperature control, so this method is a more promising method for ice slurry production.
The COPs invariably declined as the concentration of sodium chloride increased, as shown in Figure 8.The COP reduced because the increase in solution concentration caused a corresponding decrease in the freezing point of the solution and the required evaporation temperature of the Figure 6 displays the relationship between the IPF of the ice slurry and ice-making time.At the same ice-making time, higher initial concentration of aqueous sodium chloride solution resulted in lower IPF of ice slurry, suggesting that the formation of ice slurry became slower.This is mainly because the increase in sodium chloride concentration lowered the freezing point of the solution and increased the system cooling load.Figure 6 displays the relationship between the IPF of the ice slurry and ice-making time.At the same ice-making time, higher initial concentration of aqueous sodium chloride solution resulted in lower IPF of ice slurry, suggesting that the formation of ice slurry became slower.This is mainly because the increase in sodium chloride concentration lowered the freezing point of the solution and increased the system cooling load.The total power consumption of the ice-making system and the power consumption of the scraping system were monitored.Taking 3 wt% sodium chloride solution as an example, the experimental results are shown in Figure 7.When the IFP of the ice slurry reached 13.9%, the total power consumption was 1.9 kWh.The power consumption of the scraping system was 0.1 kWh, which accounted for only 5.26% of the total power consumption.Experimental results at other sodium chloride concentrations also indicated that the scraping power consumption accounted for about 5% of the total power consumption of the system.Thus, the results suggest that the scraping system does consume a certain amount of energy, but it does not account for a large proportion.Compared with the supercooling method, the scraping method has the advantages of simple equipment, small floor space, and easy temperature control, so this method is a more promising method for ice slurry production.
The COPs invariably declined as the concentration of sodium chloride increased, as shown in Figure 8.The COP reduced because the increase in solution concentration caused a corresponding decrease in the freezing point of the solution and the required evaporation temperature of the refrigerant.This conclusion is consistent with previous research results [38].The total power consumption of the ice-making system and the power consumption of the scraping system were monitored.Taking 3 wt% sodium chloride solution as an example, the experimental results are shown in Figure 7.When the IFP of the ice slurry reached 13.9%, the total power consumption was 1.9 kWh.The power consumption of the scraping system was 0.1 kWh, which accounted for only 5.26% of the total power consumption.Experimental results at other sodium chloride concentrations also indicated that the scraping power consumption accounted for about 5% of the total power consumption of the system.Thus, the results suggest that the scraping system does consume a certain amount of energy, but it does not account for a large proportion.Compared with the supercooling method, the scraping method has the advantages of simple equipment, small floor space, and easy temperature control, so this method is a more promising method for ice slurry production.

Influence of Solution Flow Rate
Ice slurry formation characteristics of 4 wt% sodium chloride solution were experimentally measured at flow rates of 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 m 3 /h while maintaining the fixed rotary scraping speed (n = 29 rpm).The measurement results are shown in Figure 9.The ice slurry formation temperature was basically the same under different flow rates, which indicated that the ice slurry formation temperature mainly depended on the sodium chloride concentration of the solution.When the flow rate was 1.3 m 3 /h, the ice crystals were generated slightly earlier than with other flow rates.The COPs invariably declined as the concentration of sodium chloride increased, as shown in Figure 8.The COP reduced because the increase in solution concentration caused a corresponding decrease in the freezing point of the solution and the required evaporation temperature of the refrigerant.This conclusion is consistent with previous research results [38].

Influence of Solution Flow Rate
Ice slurry formation characteristics of 4 wt% sodium chloride solution were experimentally measured at flow rates of 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 m 3 /h while maintaining the fixed rotary scraping speed (n = 29 rpm).The measurement results are shown in Figure 9.The ice slurry formation temperature was basically the same under different flow rates, which indicated that the ice slurry formation temperature mainly depended on the sodium chloride concentration of the solution.When the flow rate was 1.3 m 3 /h, the ice crystals were generated slightly earlier than with other flow rates.

Influence of Solution Flow Rate
Ice slurry formation characteristics of 4 wt% sodium chloride solution were experimentally measured at flow rates of 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 m 3 /h while maintaining the fixed rotary scraping speed (n = 29 rpm).The measurement results are shown in Figure 9.The ice slurry formation temperature was basically the same under different flow rates, which indicated that the ice slurry formation temperature mainly depended on the sodium chloride concentration of the solution.When the flow rate was 1.3 m 3 /h, the ice crystals were generated slightly earlier than with other flow rates.
The relationship between IPF and ice-making time at different flow rates is shown in Figure 10.At the same ice-making time, as the flow rate of the solution increased, the IPF of the ice slurry first increased and then decreased.When the flow rate was 1.3 m 3 /h, the IPF of the ice slurry reaches the highest value.It was also found that when the flow rate reached 1.3 m 3 /h, the COP of the system reached the maximum value, as shown in Figure 11.The ice-making solution flowed out of the annular shower tube and flowed from the top to the bottom along the inner wall surface.When the flow rate was low, the amount of cooling obtained per unit of time was small, which resulted in a low IPF.When the flow rate was increased from 1.3 to 1.4 m 3 /h, the ice slurry generation time was extended from 1080 s to 1160 s, an increase of 7.41%.During the experiment, we observed that when the flow rate increased to 1.4 m 3 /h, the ice-making solution flowed out of the annular shower tube with a serious splash and part of the solution did not flow down the inner wall surface.Therefore, the solution actually participating in the effective heat exchange through the inner wall surface was reduced and the amount of cold obtained per unit time of the ice-making solution was also decreased, resulting in the increase of ice slurry generation time and the decrease of the system COP.The results shown in Figures 9-11 indicated that the optimal flow rate of solution for this ice-making system was 1.3 m 3 /h.

Influence of Solution Flow Rate
Ice slurry formation characteristics of 4 wt% sodium chloride solution were experimentally measured at flow rates of 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 m 3 /h while maintaining the fixed rotary scraping speed (n = 29 rpm).The measurement results are shown in Figure 9.The ice slurry formation temperature was basically the same under different flow rates, which indicated that the ice slurry formation temperature mainly depended on the sodium chloride concentration of the solution.When the flow rate was 1.3 m 3 /h, the ice crystals were generated slightly earlier than with other flow rates.The relationship between IPF and ice-making time at different flow rates is shown in Figure 10.At the same ice-making time, as the flow rate of the solution increased, the IPF of the ice slurry first increased and then decreased.When the flow rate was 1.3 m 3 /h, the IPF of the ice slurry reaches the highest value.It was also found that when the flow rate reached 1.3 m 3 /h, the COP of the system reached the maximum value, as shown in Figure 11.The ice-making solution flowed out of the annular shower tube and flowed from the top to the bottom along the inner wall surface.When the flow rate was low, the amount of cooling obtained per unit of time was small, which resulted in a low IPF.When the flow rate was increased from 1.3 to 1.4 m 3 /h, the ice slurry generation time was extended from 1080 s to 1160 s, an increase of 7.41%.During the experiment, we observed that when the flow rate increased to 1.4 m 3 /h, the ice-making solution flowed out of the annular shower tube with a serious splash and part of the solution did not flow down the inner wall surface.Therefore, the solution actually participating in the effective heat exchange through the inner wall surface was reduced and the amount of cold obtained per unit time of the ice-making solution was also decreased, resulting in the increase of ice slurry generation time and the decrease of the system COP.The results shown in Figure 9, Figure 10 and Figure 11 indicated that the optimal flow rate of solution for this ice-making system was 1.3 m 3 /h.

Influence of scraping speed on the system performance
Ice slurry formation characteristics of 4 wt% sodium chloride solution were experimentally measured at different scraping speeds while maintaining the fixed solution flow rate (q=1.0 m 3 /h).Figure 12 shows that as the scraping speed increased, the temperature of ice slurry formation remained unchanged and the ice slurry generation time was delayed.When the scraping speed was 13 rpm, the ice crystals began to appear after cooling for 1030 s.When the scraping speed increased to 29 rpm, the time required increased to 1180 s.The relationship between IPF and ice-making time The relationship between IPF and ice-making time at different flow rates is shown in Figure 10.At the same ice-making time, as the flow rate of the solution increased, the IPF of the ice slurry first increased and then decreased.When the flow rate was 1.3 m 3 /h, the IPF of the ice slurry reaches the highest value.It was also found that when the flow rate reached 1.3 m 3 /h, the COP of the system reached the maximum value, as shown in Figure 11.The ice-making solution flowed out of the annular shower tube and flowed from the top to the bottom along the inner wall surface.When the flow rate was low, the amount of cooling obtained per unit of time was small, which resulted in a low IPF.When the flow rate was increased from 1.3 to 1.4 m 3 /h, the ice slurry generation time was extended from 1080 s to 1160 s, an increase of 7.41%.During the experiment, we observed that when the flow rate increased to 1.4 m 3 /h, the ice-making solution flowed out of the annular shower tube with a serious splash and part of the solution did not flow down the inner wall surface.Therefore, the solution actually participating in the effective heat exchange through the inner wall surface was reduced and the amount of cold obtained per unit time of the ice-making solution was also decreased, resulting in the increase of ice slurry generation time and the decrease of the system COP.The results shown in Figure 9, Figure 10 and Figure 11 indicated that the optimal flow rate of solution for this ice-making system was 1.3 m 3 /h.

Influence of scraping speed on the system performance
Ice slurry formation characteristics of 4 wt% sodium chloride solution were experimentally measured at different scraping speeds while maintaining the fixed solution flow rate (q=1.0 m 3 /h).Figure 12 shows that as the scraping speed increased, the temperature of ice slurry formation remained unchanged and the ice slurry generation time was delayed.When the scraping speed was 13 rpm, the ice crystals began to appear after cooling for 1030 s.When the scraping speed increased to 29 rpm, the time required increased to 1180 s.The relationship between IPF and ice-making time

Influence of Scraping Speed on the System Performance
Ice slurry formation characteristics of 4 wt% sodium chloride solution were experimentally measured at different scraping speeds while maintaining the fixed solution flow rate (q = 1.0 m 3 /h).
Figure 12 shows that as the scraping speed increased, the temperature of ice slurry formation remained unchanged and the ice slurry generation time was delayed.When the scraping speed was 13 rpm, the ice crystals began to appear after cooling for 1030 s.When the scraping speed increased to 29 rpm, the time required increased to 1180 s.The relationship between IPF and ice-making time is shown in Figure 13.It was found that, at the same ice-making time, the lower the scraping speed, the higher the IPF of ice slurry.Comparing the COPs at different scraping speeds, as shown in Figure 14, it is clear that the COP at the scraping speed of 13 rpm was the highest, which increased by 8.39% with respect to the scraping speed of 29 rpm.This was because when the scraping speed increased, the solution was easily taken away by the scraper, resulting in shorter residence time of the solution on the inner wall.Consequently, the solution could not exchange heat with the refrigerant on the other side sufficiently, which lowered both the cooling volume actually obtained in the system unit time and the system COP.Therefore, the decrease in scraping speed was beneficial to lowering the power consumption of the system and the optimal scraping speed was 13 rpm.
Appl.Sci.2018, 8, x FOR PEER REVIEW 11 of 15 is shown in Figure 13.It was found that, at the same ice-making time, the lower the scraping speed, the higher the IPF of ice slurry.Comparing the COPs at different scraping speeds, as shown in Figure 14, it is clear that the COP at the scraping speed of 13 rpm was the highest, which increased by 8.39% with respect to the scraping speed of 29 rpm.This was because when the scraping speed increased, the solution was easily taken away by the scraper, resulting in shorter residence time of the solution on the inner wall.Consequently, the solution could not exchange heat with the refrigerant on the other side sufficiently, which lowered both the cooling volume actually obtained in the system unit time and the system COP.Therefore, the decrease in scraping speed was beneficial to lowering the power consumption of the system and the optimal scraping speed was 13 rpm.    is shown in Figure 13.It was found that, at the same ice-making time, the lower the scraping speed, the higher the IPF of ice slurry.Comparing the COPs at different scraping speeds, as shown in Figure 14, it is clear that the COP at the scraping speed of 13 rpm was the highest, which increased by 8.39% with respect to the scraping speed of 29 rpm.This was because when the scraping speed increased, the solution was easily taken away by the scraper, resulting in shorter residence time of the solution on the inner wall.Consequently, the solution could not exchange heat with the refrigerant on the other side sufficiently, which lowered both the cooling volume actually obtained in the system unit time and the system COP.Therefore, the decrease in scraping speed was beneficial to lowering the power consumption of the system and the optimal scraping speed was 13 rpm.

Influence of Nanoparticles on the System Performance
Different concentrations of silica nanofluid were prepared by diluting 30 wt% silica nanofluid in sodium chloride solution (4 wt%).During the experiment, the scraping rate was maintained at 29 rpm and the flow rate was maintained at 1.0 m 3 /h.
As shown in Figure 15, as the concentration of nanosilica increased, the time required for the formation of ice slurry decreased from 1030 s to 840 s.Based on the heterogeneous nucleation theory [39], if there are solid particles in parent phase, crystal nucleus will be formed firstly on the solid particle surface, which contributes to reducing nucleation work.As we know, heterogeneous nucleation is the main form of nucleation, and the use of nanosilica as a nucleating agent induced heterogeneous nucleation and advance the freezing time.However, the formation temperature of ice slurry fluctuated between −2.3 and −2.4 • C under different nanosilica concentrations, which indicates that nanosilica has little effect on increasing the freezing point temperature.The effect of nanosilica on the COP of the system is shown in Figure 16.Compared with COP of a 4 wt% sodium chloride solution, the COP was significantly increased by 10.47%, 10.76%, 12.65%, 14.77%, 15.30%, and 17.38%, with addition of nanosilica at concentrations of 0.05 wt%, 0.1 wt%, 0.2 wt%, 0.5 wt%, 0.75 wt%, and 1 wt%, respectively.Obviously, as the amount of nanosilica added increased, the COP of the system also increased.Existing researches have shown that the addition of nanoparticles to solutions can significantly increase the thermal conductivity of solutions and improve the heat transfer performance [21][22][23][24][40][41][42].Therefore, we speculate that the addition of nanosilica also improves the heat transfer performance of the fluid, thereby achieving a higher COP.
However, there is no uniform explanation for the mechanism of the increase in thermal conductivity [43,44].The effect of nanosilica on the COP of the system is shown in Figure 16.Compared with COP of a 4 wt% sodium chloride solution, the COP was significantly increased by 10.47%, 10.76%, 12.65%, 14.77%, 15.30%, and 17.38%, with addition of nanosilica at concentrations of 0.05 wt%, 0.1 wt%, 0.2 wt%, 0.5 wt%, 0.75 wt%, and 1 wt%, respectively.Obviously, as the amount of nanosilica added increased, the COP of the system also increased.Existing researches have shown that the addition of nanoparticles to solutions can significantly increase the thermal conductivity of solutions and improve the heat transfer performance [21][22][23][24][40][41][42].Therefore, we speculate that the addition of nanosilica also improves the heat transfer performance of the fluid, thereby achieving a higher COP.However, there is no uniform explanation for the mechanism of the increase in thermal conductivity [43,44].

Figure 2 .
Figure 2. Schematic of a scraped surface ice slurry generator.

Figure 2 .
Figure 2. Schematic of a scraped surface ice slurry generator.

Figure 3 .
Figure 3. TEM image of silica nanofluid with an average size of 25 nm.

Figure 3 .
Figure 3. TEM image of silica nanofluid with an average size of 25 nm.

Figure 4 .
Figure 4. Variations in temperature during ice slurry production at different initial NaCl concentrations.

Figure 4 .
Figure 4. Variations in temperature during ice slurry production at different initial NaCl concentrations.The formation temperature and generation time of ice crystals at different concentrations were compared, as shown in Figure5.It can be seen that as the concentration of the solution increased, the time required for the appearance of ice crystals increased and the freezing point

Figure 5 .
Figure 5. Ice slurry formation characteristics at different initial NaCl concentrations.

Figure 6 .
Figure 6.Relationship between ice packing fraction (IPF) and ice-making time at different initial NaCl concentrations.

Figure 5 .
Figure 5. Ice slurry formation characteristics at different initial NaCl concentrations.

15 Figure 5 .
Figure 5. Ice slurry formation characteristics at different initial NaCl concentrations.

Figure 6 .
Figure 6.Relationship between ice packing fraction (IPF) and ice-making time at different initial NaCl concentrations.

Figure 6 .
Figure 6.Relationship between ice packing fraction (IPF) and ice-making time at different initial NaCl concentrations.

Figure 8 .
Figure 8. Effects of initial NaCl concentration on coefficient of performance (COP).

Figure 9 .
Figure 9. Ice slurry formation characteristics at different flow rates.

Figure 8 .
Figure 8. Effects of initial NaCl concentration on coefficient of performance (COP).

Figure 9 .
Figure 9. Ice slurry formation characteristics at different flow rates.

Figure 8 .
Figure 8. Effects of initial NaCl concentration on coefficient of performance (COP).

Figure 8 .
Figure 8. Effects of initial NaCl concentration on coefficient of performance (COP).

Figure 9 .
Figure 9. Ice slurry formation characteristics at different flow rates.Figure 9. Ice slurry formation characteristics at different flow rates.

Figure 9 .
Figure 9. Ice slurry formation characteristics at different flow rates.Figure 9. Ice slurry formation characteristics at different flow rates.

Figure 10 .
Figure 10.Relationship between IPF and ice-making time at different flow rates.

Figure 11 .
Figure 11.Effects of flow rates on COP.

Figure 10 .
Figure 10.Relationship between IPF and ice-making time at different flow rates.

Figure 10 .
Figure 10.Relationship between IPF and ice-making time at different flow rates.

Figure 11 .
Figure 11.Effects of flow rates on COP.

Figure 11 .
Figure 11.Effects of flow rates on COP.

Figure 12 .
Figure 12.Ice slurry formation characteristics at different scraping speeds.

Figure 13 .
Figure 13.Relationship between IPF and ice-making time at different scraping speeds.

Figure 14 .
Figure 14.Effects of scraping speed on COP.

Figure 12 .
Figure 12.Ice slurry formation characteristics at different scraping speeds.

Figure 12 .
Figure 12.Ice slurry formation characteristics at different scraping speeds.

Figure 13 .
Figure 13.Relationship between IPF and ice-making time at different scraping speeds.

Figure 14 .
Figure 14.Effects of scraping speed on COP.

Figure 13 .
Figure 13.Relationship between IPF and ice-making time at different scraping speeds.

Figure 13 .
Figure 13.Relationship between IPF and ice-making time at different scraping speeds.

Figure 14 .
Figure 14.Effects of scraping speed on COP.Figure 14.Effects of scraping speed on COP.

Figure 14 .
Figure 14.Effects of scraping speed on COP.Figure 14.Effects of scraping speed on COP.

15 3. 4 .
Appl.Sci.2018, 8, x FOR PEER REVIEW 12 of Influence of Nanoparticles on the System Performance Different concentrations of silica nanofluid were prepared by diluting 30 wt% silica nanofluid in sodium chloride solution (4 wt%).During the experiment, the scraping rate was maintained at 29 rpm and the flow rate was maintained at 1.0 m 3 /h.As shown in Figure 15, as the concentration of nanosilica increased, the time required for the formation of ice slurry decreased from 1030 seconds to 840 seconds.Based on the heterogeneous nucleation theory [39], if there are solid particles in parent phase, crystal nucleus will be formed firstly on the solid particle surface, which contributes to reducing nucleation work.As we know, heterogeneous nucleation is the main form of nucleation, and the use of nanosilica as a nucleating agent induced heterogeneous nucleation and advance the freezing time.However, the formation temperature of ice slurry fluctuated between −2.3 and −2.4 °C under different nanosilica concentrations, which indicates that nanosilica has little effect on increasing the freezing point temperature.

Figure 15 .
Figure 15.Ice slurry formation characteristics at different nanosilica concentrations.Figure 15.Ice slurry formation characteristics at different nanosilica concentrations.

Figure 15 .
Figure 15.Ice slurry formation characteristics at different nanosilica concentrations.Figure 15.Ice slurry formation characteristics at different nanosilica concentrations.

Table 1 .
The detail description of scraping system.

Table 1 .
The detail description of scraping system.

Table 2 .
Value of uncertainties.

Table 2 .
Value of uncertainties.

Table 3 .
The composition of different concentrations of silica nanofluids.

Table 3 .
The composition of different concentrations of silica nanofluids.

Table 4 .
The specific heat of aqueous sodium chloride solution.

Table 5 .
The specific heat of silica nanofluid.

Table 4 .
The specific heat of aqueous sodium chloride solution.

Table 5 .
The specific heat of silica nanofluid.