- Email: [email protected]
Influence of salt mixture on the heat transfer during spray cooling of hot metals
International Journal of Heat and Mass Transfer 78 (2014) 76–83
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer journal homepage: www.elsevier.com/locate/ijhmt
Influence of salt mixture on the heat transfer during spray cooling of hot metals Khalid H.M. Abdalrahman 1, Sabariman ⇑, Eckehard Specht 2 Institute of Fluid Dynamics and Thermodynamics, Otto von Guericke University, 39106 Magdeburg, Germany
a r t i c l e
i n f o
Article history: Received 30 July 2012 Received in revised form 17 June 2014 Accepted 23 June 2014
Keywords: Dissolved salts Spray cooling Hot metals Electrical conductivity Leidenfrost temperature
a b s t r a c t A hot aluminum alloy AA6082 disc at 560 °C was cooled by a spray nozzle with a constant volumetric spray flux of 3 kg/m2/s. The temperature history during the cooling process was recorded with use of an infrared camera. The influence of salinity in the cooling water was observed by dissolving various concentrations of MgSO4 to the deionized water. Deionized water was used as the reference. Moreover, eight different types of real solution used in the metal processing industry were compared. The heat transfer coefficient a in the film boiling, maximum heat flux, as well as the Leidenfrost and DNB temperature increase with the electrical conductivity (EC) of the real solutions. EC is introduced as a mean to measure the salinity in the cooling water. The change of salinity content in the cooling water must be maintained by product quality control in industries as it can significantly lead to the change of the cooling rate. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction During the cooling processes in the continuous casting of metals and in the hardening of metal alloys, water sprays are often used. At high surface temperatures, the heat is transferred through a vapor film until it reaches the minimum heat flux at which the temperature is called the Leidenfrost temperature. When the surface temperature falls below the Leidenfrost temperature, the heat is transferred through transition and nucleate boiling regime until the surface temperature reaches 100 °C before finally starting the single-phase convection. In the range between the Leidenfrost temperature and 100 °C, the heat transfer can reach maximum values in the range of 5–10 times higher than that in the film boiling. Therefore, the Leidenfrost temperature will have a significant influence on the overall duration of the cooling process. In spray cooling, the primary parameter in the determination of the heat flux is the temperature of the surface. The volumetric spray flux is known as the second important parameter that can also influence the heat flux during the cooling process [4]. Gaugler ⇑ Corresponding author. Address: Institute of Fluid Dynamics and Thermodynamics, Otto von Guericke University Magdeburg, Universitaetsplatz 2, Magdeburg 39106, Germany. E-mail addresses: [email protected] (K.H.M. Abdalrahman), i.sabariman@st. ovgu.de ( Sabariman), [email protected] (E. Specht). 1 Address: Institute of Fluid Dynamics and Thermodynamics, Otto von Guericke University Magdeburg, Universitaetsplatz 2, Magdeburg 39106, Germany. 2 Address: Institute of Fluid Dynamics and Thermodynamics, Otto von Guericke University Magdeburg, Universitaetsplatz 2, G10-134, Magdeburg 39106, Germany. http://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.06.070 0017-9310/Ó 2014 Elsevier Ltd. All rights reserved.
[12] found that the volumetric spray flux has a dominant effect on heat transfer compared to other hydrodynamic properties of the spray. Al-Ahmadi et al. [5] also found that the Leidenfrost temperature, critical temperature, and maximum heat flux are strongly dependent on the local spray mass flux. Other spray parameters such as droplet size, velocity, and the type of nozzle play a less significant role [2,5,18]. Müller et al. [1] presented the change of the heat transfer coefficient (HTC) vs. surface temperature for Nickel. They investigated the HTC variation at different values of volumetric spray flux. Jeschar et al. [3] found that in the range of film boiling, HTC can only be influenced by the volumetric spray flux. Other parameters, such as drop velocity, distance nozzle to surface and nozzle type, do not require to be considered separately, as they are contained in the volumetric spray flux. However, the study of Wendelstorf et al. [4] shows HTC in the film boiling depends not only on the volumetric spray flux but also on the temperature. Puschmann et al. [14] found a similar result where HTC increased with the volumetric spray flux. They also confirmed that with a lower volumetric spray flux, the drop diameter exerts no influence on the HTC obtained. However, when they doubled the droplet velocity, they obtained about a difference of 50 W/m2/K in the HTC. The effect of salinity has been studied as a third relevant parameter that influences the heat treatment process. Huang et al. [7] made an intensive analysis of the influence of salt on the Leidenfrost transition. They explain the reason for the elevated Leidenfrost temperature resulting from the addition of salts. In their analysis, the suppression of bubble coalescence by dissolved salt, salt deposition during the initial contact of the deposition, and earlier
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
collapse of the vapor film due to the increasing salt concentration at the liquid–vapor interface were studied as the mechanisms associated with the Leidenfrost transition. KCl and NaCl were used. As a result, they found that the Leidenfrost temperature was significantly increased by the presence of dissolved salt. Cui et al. [6] conducted an experiment to observe the influence of three different salts (NaCl, Na2SO4, and MgSO4) dissolved in a water spray used for the cooling of a hot surface. Copper heated up to 240 °C was used. It was found that the individual salts influence the heat transfer process in different ways. Some increased the performance of heat transfer in nucleate boiling while others increased both nucleate and transition boiling heat transfer. MgSO4 produced the largest increase in nucleate boiling heat transfer, with Na2SO4 producing a somewhat less increase and NaCl producing the smallest increase. In addition, it was clearly shown that a higher concentration of salts lead to a higher maximum heat flux. Alam et al. [15] also found by using atomized spray that the Leidenfrost temperature increased with higher salt concentrations dissolved in the cooling water. Jeschar et al. [16] measured the influence of salts on the cooling process of a Nickel sphere dipped into a bath of water. NaSO4, NaCl, NaHCO3, CaCl2, and MgSO4 were used as individual salts dissolved up to 1 g/l concentration. The Leidenfrost temperature was increased up to 100 K for all salts. The solution with Mg2+ and SO2 4 content had the strongest effect. Although previous works have reported that salinity has a strong influence on the heat treatment process, there is no extended research published quantifying the effect of salt mixture to the Leidenfrost and DNB temperature, maximum heat flux, and heat transfer coefficient in the film boiling. In this study, those parameters are quantified for eight different real solutions used in the metal processing industry containing different salts. Electrical conductivity (EC), which has often been used as an index of the total dissolved solids (TDS), is introduced as a means of estimating the salinity level of the cooling water since EC depends on the overall ionic concentration in water [9]. The use of a single salt (MgSO4) with different concentrations (0.25–12 g/l) was chosen to show the correlation between the salt concentration and its electrical conductivity. Constant and accurate reproducibility of the spatial volumetric spray flux is the most important factor in order to study the effect of salinity on spray cooling heat transfer.
77
2. Experimental method 2.1. Experimental setup The experimental setup is shown schematically in Fig. 1. Basically, it consists of four main units. First, there is the heating system which is an electrically-heated furnace to heat up the sample to the desired level of temperature. In this work, all samples were heated up to 560 °C. Second, a metal sample is hung up on the rail that can shift the sample cater free from the furnace to the cooling unit direction. A disc of Aluminum Alloy AA6082 with a diameter of 140 mm and a thickness of 3 mm was used in the experiments (detailed information is summarized in Table 1). Third, there is a spray cooling unit by which the quenching process takes place, which consists of a centrifugal pump supplying water at the certain level of flow rate verified by a flow meter and a hydraulic nozzle used as a spraying component producing water droplets to the hot metal sample. A constant volumetric spray flux of 3 kg/m2/s was applied in this experiment. This volumetric spray flux was produced by applying a flow rate of 20 l/h water through a hydraulic nozzle. Fourth, there is the data acquisition system which consists of an infrared (IR) camera connected to a computer system for the image recording process. Prepared samples have to be painted black on the back side before being put into the electrical furnace for the heating process. After that, the hot metal sample is shifted manually along the sliding rail towards the stop position where the water spray will be aimed. The stop position is adjusted in such a way so that the camera focus, center point of metal sample, and center point of spray cone are aligned. Once it reaches the stop position, the spray cover is removed and the IR camera is activated to start recording the temperature history during the cooling process. The sample moving process was verified to not cause a significant temperature drop. Every time this will not cause more than 2 °C temperature difference. 2.2. Determination of emissivity The most important parameter of the IR camera setting is the emissivity of the surface. This is the reason why the sample must
Fig. 1. Schematic diagram of quenching experimental facility (top view).
78
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
Table 1 Geometry and material details. Diameter of the disc Thickness Density Thermal conductivity Specific heat
140 mm 3 mm 2770 kg/m3 170 W/m/K 1050 J/kg/K
be painted black with a graphite coating on the back side in order to get a higher emissivity for a more accurate measurement. For the calibration of emissivity of a graphite coating, a small copper cylinder was machined with an inside bore. A thermocouple was inserted inside the bore so that the temperature of the metal could be measured. This cylinder was painted black with the graphite coating and placed inside a horizontal cylindrical furnace. An IR Camera was then positioned in front of the outlet of this furnace from a distance of 80 cm. The temperature of the furnace was increased in steps of 50 °C from 50 to 700 °C. The temperature of the cylinder measured by thermocouple and that of the IR camera were aligned with each other by adjusting the emissivity value in the settings of the camera. In this way, the emissivity was recorded vs. the temperature of the cylinder surface. This procedure was repeated with both the heating and cooling direction. It was found that emissivity of the coating changed from 0.92 to 0.94. However, the mode value of emissivity was found to be 0.92. Therefore, it can be inferred that a constant value of emissivity of 0.92 should be used for the optimum accurate measurement of the surface temperature. A detailed review was also made in another publication [13]. Moreover, for purposes of accuracy, the thermal image was set to capture and store at the extremely high frame rate of 150 Hz within the range of 240 80 pixels. 2.3. Volumetric spray flux measurement The volumetric spray flux was measured separately by changing the targeted sample with a patternator. A patternator is a device designed for volumetric spray flux measurement. It consists of brass tubes with diameters of 10 mm. Each tube is fixed in such a way an aligned formation is formed. At the bottom end, each tube is connected to the plastic bottle into which the water drops will be collected. Water droplets are sprayed over the patternator tubes and then the mass of the water droplets md collected by the plastic glass is measured after a certain period of time Dt. The area of the opening of the tube is denoted by A. Then, the volumetric spray _ ¼ md /(ADt). The measurement was flux can be derived from m estimated to be accurate within ±4% as several repeating measurements were fluctuating in this range (Fig. 2). Since only one nozzle with a single flow rate set-value is used in this work, it is not necessary to make further verification on the influence of droplet diameter and velocity. 2.4. Preparation of test solution In this experiment, deionized water is used as a solvent of the single-salt solution. This is to minimize the influence of dissolved chemicals and gases. A salt solution was prepared by dissolving the required amount of powdered salt with a purity of P99.5% in deionized water at room temperature and then stirring the solution to ensure complete mixing. In each test, a fresh salt solution was prepared before the experiment. The amount of salt was weighed with a precision balance with an accuracy of ±0.01 g. The temperature of the solution was carefully measured before being applied in the cooling experiment with an accuracy of ±0.5 °C. Real solutions used in different metal processing companies was used no more than 2 days after arrival. All the waters
Fig. 2. Volumetric spray flux distribution for the hydraulic nozzle used in the experiment.
were treated uniformly to have the temperature Tsp of approximately 18 °C. 2.5. Energy balance analysis and accuracy In order to calculate the heat transfer coefficient for a specific position on the metal disc, an energy balance is performed as follows:
q c s
DT ¼ q_ sp þ q_ k þ q_ R þ q_ a Dt
where q, c, and s are the density, specific heat capacity, and the thickness of the metal disc, respectively. The change of the stored enthalpy in the disc caused by temperature decrease is equal to the sum of the heat transfer by the spray q_ sp , radial conduction q_ k , radiation q_ R , and free convection q_ a . The heat flux due to spray has the largest magnitude when compared to all other sources. Heat loss due to radiation and natural convection is less than 5% that of the spray and is therefore negligible. This heat flux q_ sp is defined by the water spray temperature Tsp with the equation
q_ sp ¼ asp T s T sp
where q_ sp ; asp , and Ts are the heat flux, heat transfer coefficient, and surface temperature of the metal disc, respectively. Due to the radial temperature distribution of the spray with the minimum temperature in the center, heat is conducted in the metal disc plane from the outer regions to its center. This heat flux is calculated using the Fourier approach for cylindrical coordinates
q_ k ¼ s k
1 @ @T r r @r @r
where s is the thickness of the disc, k is the conductivity of material, and r is the radius. The gradient oT/or is known from the temperature measurement. The energy balance equation is the basis for a numerical procedure. This procedure calculates the time-dependent heat transfer coefficient for a specific position on the metal disc surface from the varying temperatures at this position. Since the metal disc used is considered a thin sheet, it was verified that a temperature difference of only about 1 °C exists between the front and backside surface. The heat flux was also analyzed with a 2D inverse solution. The results found were in good agreement to those calculated with the simplified energy balance model. A detailed review of this comparison has been made elsewhere in another publication [8]. The specific heat capacity and conductivity have an accuracy of about 10%. The temperature measurement accuracy follows the IR camera specifications. With a certain numerical technique oT/ot
79
can be accurately determined. The percentage of cooling rate error is about 3%. The overall accuracy for the determination of the heat transfer coefficient was 5–10%, and the repeatability was considered good as statistical fluctuation can be controlled within 5% as shown by Fig. 3. For this reason, for every group of the experimental plan, repetitions were made and only the data which has acceptable repeatability will be considered. Since main factors that could influence the measurement were blocked and identified, poor repeatability was assumed to be caused by random error.
Electrical conductivity in µS/cm
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
25000 20000 15000 10000 5000 0 0
3. Results and discussion
2
4
6
8
10
12
Concentration in g/L
3.1. Influence of single salt 3.1.1. Salinity and electrical conductivity (EC) Salinity depends on the dissolved ions. Dissolved ions in the water can be positively charged ions (Na+, Mg2+, etc.) or negatively charged ions (Cl, SO2 4 , etc.), which can come from the dissolved salts in the water. Electrical conductivity (EC) is a measure on the ability of the material to conduct electrical current, which in a solution depends on the presence of charged ions and its concentration. Therefore, EC has often been used as an index of the total dissolved solids (TDS) contained in a solution. The ratio of EC to concentration (TDS) generally ranges from 1.33 to 1.82 [9]. In this work it is proposed to use the measured value of EC as an indirect indicator of the salinity level of the solution. As a basic understanding, solutions with different levels of salinity have been created by dissolving different concentrations of MgSO4 to deionized water. In this work, the concentrations used were 0.25, 1, 1.5L, 2 g/L. Deionized water was used as reference water. All previous studies [6,15,16] have shown that MgSO4 has the strongest effect. For illustration purposes, an extreme condition of 12 g/L was included in certain comparison figures. Fig. 4 shows the correlation between measured EC and MgSO4 concentration. It is obviously seen that up until 5 g/L, the concentration varies almost linearly with the EC with a ratio of 1.53. For concentrations higher than 5 g/L the correlation is inherently nonlinear, and it is
Fig. 4. Measured EC of MgSO4 solution.
suspected that at higher concentrations, interactions among ions can impede their mobility [17]. On the other hand, concentrations higher than 1 g/L are not likely to be adopted in a spray cooling process; they are only used for the clarity of illustration so that the correlation or the influence of the salinity level is more easily observed. However, to investigate the practical aspect, the idea was proposed in this work to also review real solutions used in metal processing industries with respect to different salinity levels. The results will give important information on the fluctuating range of the main characteristics of the cooling process due to the change in the salinity level. The highest EC value from the real solutions is about 2580 lS/cm which is comparable to that of MgSO4 solution with a concentration of 1.5 g/L. 3.1.2. Temperature profile Fig. 5 depicts the cooling profile of different concentrations of MgSO4. It is clearly shown that the higher the concentration, the shorter the cooling duration is entirely. When it is observed further, it is seen that the higher EC, which indicates a higher concentration of salt (MgSO4), shortens the film boiling duration significantly. It took 21.5 s for deionized water to decrease from 560 °C to its Leidenfrost temperature. Below this temperature, abrupt increase of heat flux occurred, and it took only 1.5 s to reach 100 °C. These results were compared to that of the 1 g/L MgSO4 solution, which took 15 s to reach its Leidenfrost temperature and from this point also took 1.5 s to reach 100 °C. By doubling the concentration of MgSO4 in the solution to 2 g/L, the film boiling duration is further significantly reduced to 11 s, but it still took 1.2 s to reach 100 °C. For the extreme case of the 12 g/L MgSO4 solution, it took only 4 s in total to reach 100 °C. The latter case is only used for the clarity of illustration. As explained by Cui et al. [6], a higher concentration creates greater surface roughness, which in turn can break through
600 1. DI Water 2. MgSO4 0.25 g/L 3. MgSO4 1 g/L 4. MgSO4 2 g/L 5. MgSO4 12 g/L
Temperature in C
500 400 300 200 5
4
3
2
1
100 0 0
Fig. 3. Cooling profile repeated three times.
5
10
15 Time in s
20
25
30
Fig. 5. Temperature profile of four different concentrations of MgSO4 at center point.
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
5
5
Heat Flux in MW/m2
4.5 4 3.5
4
1. DI Water 2. MgSO4 0.25 g/L 3. MgSO4 1 g/L 4. MgSO4 2 g/L 5. MgSO4 12 g/L
3 3
2.5 2
2
1.5
1
1
Maximum heat flux in MW/m2
80
6 5 4 3 2 1 0 0
0.5 0 0
100
200 300 400 500 Surface Temperature in C
500 1000 1500 2000 2500 3000 3500 4000 Electrical conductivity in S/cm
600
Fig. 8. Maximum heat flux of four different concentrations of MgSO4 at center point.
Fig. 6. Heat flux over surface temperature of four different concentrations of MgSO4 at center point.
the vapor layer insulating the heated surface and increase direct liquid–solid contact. Since the film boiling duration dominates the entire duration of the cooling process, an earlier start to the transition boiling will eventually shorten the duration of the cooling process entirely. Corresponding heat flux over surface temperature from different concentrations is shown in Fig. 6. 3.1.3. Leidenfrost and DNB temperature As the Leidenfrost temperature is shifted to higher values, the cooling duration is shortened. This result is in line with the finding explained excellently by Huang et al. [7] and Craig et al. [11]. Bubble merging may be inhibited by dissolved salts as some electrolytes do retard coalescence of bubbles. Given this, a higher surface temperature is required to start the Leidenfrost transition. A higher concentration, i.e. a higher EC, may increase the number of charged ions in the solution. A higher EC means that a higher level of electrolytes is created so that a more intense retardation of bubble coalescence might take place. This also agrees with what was observed by Cui et al. [6], which estimated that the precipitating salt particle from the evaporating droplet would provide a bubble nucleation site to trigger bubble formation and promote the nucleate boiling heat transfer earlier. However, Huang et al. [7] explained their finding that the increase of the salt concentration on the liquid–vapor interface significantly accounts for the increase of Leidenfrost temperature due to the increase of boiling point. In this work, the Leidenfrost temperature was estimated from a relative minimum value of wall heat flux in the film boiling region. While the DNB temperature is estimated from the maximum heat flux point. Specifically, Fig. 7 shows that up to 2 g/L, the Leidenfrost temperature increased in the range of 10–30 °C more than that of deionized water. These results agree with those
found by Jeschar et al. [16], who made this investigation by immersing a nickel ball into a pool of solution. Huang et al. [7] found that the average Leidenfrost temperatures for salt solutions were roughly 10–33 °C higher than that of deionized water. Fig. 7 also shows a higher EC promoting the onset of transition boiling, i.e. the DNB temperature, from about 165 to 190 °C. This result agrees with the result found by Cui et al. [6], which states that the greater surface roughness created by the deposited salt (MgSO4) particle increased the temperature at which the maximum heat flux occurred. 3.1.4. Maximum heat flux By increasing the EC from 6.7 to 4000 lS/cm, the maximum heat flux dramatically changes from 1.6 MW/m2 for deionized water to a maximum value of approximately 3.0 MW/m2 when a 2 g/L solution of MgSO4 was used (Fig. 8). This is a similar trend to that of the result of Cui et al. [6], which also used a volumetric spray flux of 3 kg/m2/s. They found that the surface heat flux increased with salt concentration for all temperatures but only until 24 g/L. Further increase reduces the heat transfer in the nucleate region. 3.1.5. Heat transfer coefficient in film boiling In this investigation, the heat transfer coefficient a in the film boiling is also determined. As depicted by Fig. 9, a increases with an increase in EC. However, it starts showing significant dependency on surface temperature beginning at EC = 2000 lS/cm. This experimental result shows a similar trend to that highlighted by Wendelstorf et al. [4], which found that in film boiling a increases with an increase in volumetric spray flux, but at higher level the dependency of a on the surface temperature is obviously observed. Other than that, these quantitative values in film boiling will
350
3
TLeidenfrost
250 200 150
TDNB
100
2.5 in kW/m2 K
Temperature in C
300
T = 400 C
2 1.5 1
T = 500 C
0.5
50
0
0 0
500 1000 1500 2000 2500 3000 3500 4000 Electrical conductivity in S/cm
Fig. 7. Leidenfrost and DNB temperatures of four different concentrations of MgSO4 at center point.
0
500 1000 1500 2000 2500 3000 3500 4000 Electrical Conductivity in S/cm
Fig. 9. Heat transfer coefficient in the film boiling region of four different concentrations of MgSO4 at center point.
81
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83 Table 2 Salts contained in the real solutions and corresponding measured EC. Parameters
DI Water
AG
K1
BF
A90
AL
AF
C1
C2
Magnesium [mg/l] Calcium [mg/l] Sodium [mg/l] Sulfate [mg/l] Carbonate [mg/l] Chloride [mg/l] EC [lS/cm]
<0.03 <2 <0.03 <5 2.94 3 6.7
14 84 88 <5 152.6 11 530
7.8 38.5 96 <5 140.9 62 692
<2 2 186 62 151 43 841
17 49.7 94.4 36 115.3 121 869
11 138 149 180 111 241 1477
53 150 91 301 126 195 1565
106 248.5 77 867 40.4 170 2190
93 88 305 628 49.1 191 2580
enrich what has been found by Cui et al. [6] since they only used low initial temperatures on a copper sample.
3.2.2. Temperature profiles The temperature profile of the cooling process from three different real solutions at the center point is plotted in Fig. 10, and the corresponding heat flux over surface temperature is shown in Fig. 11. It is evident that the higher the EC value, the shorter the duration of the entire cooling process. The temperature profiles themselves indicate the influence of salinity on the cooling rate. The highest EC among these solutions is from the C2 solution, and the value of its EC is comparable to that of the 1.5 g/L MgSO4 solution. When compared, it is clear that the influence from the salts mixture is stronger than the single salt since it causes a shorter cooling duration. 3.2.3. Leidenfrost and DNB temperature In Fig. 12 the trend of the Leidenfrost and DNB temperatures at the center point is mapped according to the increase in EC values. It is shown that up to 2580 lS/cm, the Leidenfrost temperature increased in the range of 10–100 °C more than that of deionized water. Furthermore, the higher EC also promotes the onset of the transition boiling, i.e. the DNB temperature, from about 165 °C for deionized water to 230 °C for the C2 solution. In this work, temperatures where the maximum heat flux takes place are assumed
Temperature in C
600
400 300 200 4
3
2
1
100 0 0
5
10
15 20 Time in s
25
30
1. DI Water 2. BF 3. AL 4. C2
4
5 3
4 3
2
2
1
1 0 0
100
200 300 400 Surface temperature in C
500
600
Fig. 11. Heat flux over surface temperature of three samples of different real solutions at center point.
350
1. TLe of real solutions 2. TLe of MgSO4
300
1 TLeidenfrost
2
250 TDNB
3
200 150
4
3. TDNB of real solutions 4. TDNB of MgSO4
100 0
1000
2000
3000
4000
Electrical conductivity in S/cm Fig. 12. Leidenfrost and DNB temperature of eight different real solutions at center point.
to be the DNB temperatures. In comparison with the effect of a single salt, both the Leidenfrost and the DNB temperature from the salts mixture were more strongly affected. 3.2.4. Maximum heat flux Subsequently, the trend of the maximum heat flux at the center point of the real solutions is plotted in Fig. 13. It is shown that by increasing the EC from 6.7 to 2580 lS/cm, the maximum heat flux dramatically changes from 1.6 MW/m2 for deionized water to a maximum value of approximately 5.0 MW/m2 when the C2 solution was used. When compared to the single salt, the rate of the increase on the maximum heat flux values from the salts mixture is in the range of 1–2.5 MW/m2 higher.
1. DI Water 2. BF 3. AL 4. C2
500
Heat flux in MW/m2
3.2.1. Overview of real solutions In the case of salt mixtures, eight different real solutions used in some industrial plants during the casting process were investigated. In Table 2, the analysis of these solutions is summarized. The analysis of deionized water, which is used as reference water, is also included. All solutions are ranked from left to right from smallest to largest measured EC. In general, the basic characteristics of the cooling process with the real solutions tend to be similar to that of a single-salt solution as reviewed in sections below:
Temperature in C
3.2. Influence of salt mixture
6
35
Fig. 10. Profile of temperature over time of three samples of different real solutions at center point.
3.2.5. Heat transfer coefficient in the film boiling In order to compare with the single salt effect, the heat transfer coefficient a in the film boiling region of the real solutions at the center point is also reviewed. As shown by Fig. 14, a increases with an increase in EC. In addition, a similar trend as found in the
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
Maximum heat flux in MW/m 2
82
5. Concluding remarks
6 5 4 3 2 1 0 0
1000
2000
3000
Electrical conductivity in S/cm Fig. 13. Maximum heat flux of eight different real solutions at center point.
4
in kW/m2K
3.5 T = 400 C
3 2.5 2
T = 500 C
1.5 1 0.5 0 0
500
1000
1500
2000
2500
3000
The experimental work for studying the effect of different salinity levels contained in the cooling water is established. Two different groups of solution were tested. The first is a group of solutions created from varying concentrations of a single salt MgSO4. And the second is a group of solutions taken from the real solution used in different metal processing industries. A disc of AA6082 with a diameter of 140 mm and a thickness of 3 mm is heated up to 560 °C. Then it is quenched at the front side, while the temperature at the back side is recorded with use of an infrared camera. Repeatability of the measured temperature and volumetric spray flux are also verified. In this work, it is proposed to use EC instead of concentration for the sake of practicability when measuring the salinity level in the cooling water. Until 6 g/L a linear correlation with a certain factor was found. From the measured temperature, the Leidenfrost and DNB temperature, maximum heat flux, and heat transfer coefficient in film boiling are estimated by using a 1D simplified energy balance model. From both groups of solutions used, it is observed that these values increase with the increase in EC. However, it is found that the effect resulted from a salts mixture is stronger than that of a single salt. It is believed that the interaction between ions might change properties in the solutions that ultimately change the effect found from the individual salt on the cooling process. In order to verify the results, further work is still required especially that of observing the quenching behavior of a solution created from a binary salt mixture.
Electrical conductivity in S/cm Fig. 14. Heat transfer coefficient in the film boiling region of eight different real solutions at center point.
investigation with a single salt, that a significant dependency of a on the surface temperature is also observed. However, from the comparable EC values at the same surface temperature, it is observed that a difference of 0.5–1 kW/m2K exists.
4. Other considerations It is interesting to review the fact that the influence of a single salt is not exactly the same as those found from a salt mixture especially on the specific parameters compared above. The salt mixture has a stronger effect than the single salt at comparable EC values. It is believed that the interaction between ions might change properties in the solutions, e.g. surface tension, saturation temperature, and liquid density, which ultimately change the effect found from the individual salt on the cooling process. Craig [10] highlighted on his review on bubble coalescent and specificion effects that ion combinations separated on the interface result in a reduction in the mobility of the interface. The reduced mobility leads to an increase in the stability of the film. When a sufficient concentration of ions in bulk is reached, the surface mobility is suppressed to an extent that bubble coalescence does not take place within the life time of a collision. The types of ions contained in the sample of real solutions (shown in Table 2), Mg2+, Ca2+, Na+, SO2 4 , NO3 , and Cl , are categorized in Craig’s review [10] as the ions that when combined with each other will inhibit bubble coalescence. This situation agrees with the observation found in Huang et al., which is that inhibition of bubble coalescence will result in the Leidenfrost point becoming unachievable at the corresponding surface temperature, and therefore a higher surface temperature is required to start the Leidenfrost transition. However, further work is still required to confirm this suspected cause especially that of observing the quenching behavior of a solution created from a binary salt mixture.
Conflict of interest None declared. References [1] H.R. Müller, R. Jeschar, Heat transfer during water spray cooling of nonferrous metals, Metallkunde 74 (1983) 257–264. [2] M. Ciofalo, A. Caronia, M. Di Liberto, S. Puleo, The Nukiyama curve in water spray cooling: its derivation from temperature–time histories and its dependence on the quantities that characterize drop impact, Int. J. Heat Mass Transfer 50 (2007) 4948–4966. [3] R. Jeschar, E. Specht, C. Köhler, Heat transfer during cooling of heated metal objects with evaporating liquids, in: H. Tensi, B. Liscic, W. Luty (Eds.), Theory and Technology of Quenching, Berlin, Springer-Verlag, 1992, pp. 73–92. [4] J. Wendelstorf, K.H. Spitzer, R. Wendelstorf, Spray water cooling heat transfer at high temperatures and liquid mass fluxes, Int. J. Heat Mass Transfer 51 (2008) 4902–4910. [5] H.M. Al-Ahmadi, S.C. Yao, Spray cooling of high temperature metals using high mass flux industrial nozzles, Exp Heat Transfer 2 (2008) 38–54. [6] Q. Cui, S. Chandra, S. McCahan, The effect of dissolving salts in water sprays used for quenching a hot surface: Part 2 – spray cooling, ASME J. Heat Transfer 125 (2003) 333–338. [7] C.K. Huang, V.P. Carey, The effect of dissolved salt on the Leidenfrost transition, Int. J. Heat Mass Transfer 50 (2006) 269–282. [8] A. Nallathambi, E. Specht, Estimation of heat flux in array of jets quenching using experimental and inverse finite element method, J. Mater. Process. Technol. 209 (2009) 5325–5332. [9] R. Das, N.R. Samal, P.K. Roy, D. Mitra, Role of electrical conductivity as indicator of pollution in shallow lakes, Asian J. Water Environ. Pollut. 3 (1) (2005) 143–146. [10] V.S.J. Craig, Bubble coalescence and specific-ion effects, Curr. Opin. Colloid Interface Sci. 9 (2004) 178–184. [11] V.S.J. Craig, B.W. Ninham, R.M. Pashley, The effect of electrolytes on bubble coalescence in water, J. Phys. Chem. 97 (39) (1993) 10192–10197. [12] R.E. Gaugler, An experimental study of spray cooling of high temperature surfaces (Ph.D. dissertation), Department of Mechanical Engineering, Carnegie Institute of technology, Pittsburg, 1966. [13] U. Alam, J. Krol, E. Specht, J. Schmidt, Enhancement and local regulation of metal quenching using atomized spray, J. ASTM Int. 5 (10) (2008). [14] F. Puschmann, E. Specht, J. Schmidt, Local distribution of the heat transfer in water spray quenching, in: International Conference on Continuous Casting of Non-Ferrous Metals. Proceedings of the DGM Frankfurt, 2000, pp. 101–107. [15] U. Alam, Experimental study of local heat transfer during quenching of metals by spray and multiple jets (Ph.D. thesis), Otto-von-Güricke Universität Magdeburg, 2011.
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83 [16] R. Jeschar, H. Kraushaar, H. Griebel, Influence of gases dissolved in cooling water on heat transfer during stable film boiling, Steel Res. 67 (1996) 227–234. [17] R.D. More, G. Richards, A. Story, Electrical conductivity as an indicator of water chemistry of hydrologic process, Streamline Watershed Manage. Bull. 11 (2008) 25–29.
83
[18] J. Schmidt, H. Boye, Influence of velocity and size of droplets on the heat transfer in spray cooling, Chem. Eng. Technol. 24 (2001) 255–260.
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer journal homepage: www.elsevier.com/locate/ijhmt
Influence of salt mixture on the heat transfer during spray cooling of hot metals Khalid H.M. Abdalrahman 1, Sabariman ⇑, Eckehard Specht 2 Institute of Fluid Dynamics and Thermodynamics, Otto von Guericke University, 39106 Magdeburg, Germany
a r t i c l e
i n f o
Article history: Received 30 July 2012 Received in revised form 17 June 2014 Accepted 23 June 2014
Keywords: Dissolved salts Spray cooling Hot metals Electrical conductivity Leidenfrost temperature
a b s t r a c t A hot aluminum alloy AA6082 disc at 560 °C was cooled by a spray nozzle with a constant volumetric spray flux of 3 kg/m2/s. The temperature history during the cooling process was recorded with use of an infrared camera. The influence of salinity in the cooling water was observed by dissolving various concentrations of MgSO4 to the deionized water. Deionized water was used as the reference. Moreover, eight different types of real solution used in the metal processing industry were compared. The heat transfer coefficient a in the film boiling, maximum heat flux, as well as the Leidenfrost and DNB temperature increase with the electrical conductivity (EC) of the real solutions. EC is introduced as a mean to measure the salinity in the cooling water. The change of salinity content in the cooling water must be maintained by product quality control in industries as it can significantly lead to the change of the cooling rate. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction During the cooling processes in the continuous casting of metals and in the hardening of metal alloys, water sprays are often used. At high surface temperatures, the heat is transferred through a vapor film until it reaches the minimum heat flux at which the temperature is called the Leidenfrost temperature. When the surface temperature falls below the Leidenfrost temperature, the heat is transferred through transition and nucleate boiling regime until the surface temperature reaches 100 °C before finally starting the single-phase convection. In the range between the Leidenfrost temperature and 100 °C, the heat transfer can reach maximum values in the range of 5–10 times higher than that in the film boiling. Therefore, the Leidenfrost temperature will have a significant influence on the overall duration of the cooling process. In spray cooling, the primary parameter in the determination of the heat flux is the temperature of the surface. The volumetric spray flux is known as the second important parameter that can also influence the heat flux during the cooling process [4]. Gaugler ⇑ Corresponding author. Address: Institute of Fluid Dynamics and Thermodynamics, Otto von Guericke University Magdeburg, Universitaetsplatz 2, Magdeburg 39106, Germany. E-mail addresses: [email protected] (K.H.M. Abdalrahman), i.sabariman@st. ovgu.de ( Sabariman), [email protected] (E. Specht). 1 Address: Institute of Fluid Dynamics and Thermodynamics, Otto von Guericke University Magdeburg, Universitaetsplatz 2, Magdeburg 39106, Germany. 2 Address: Institute of Fluid Dynamics and Thermodynamics, Otto von Guericke University Magdeburg, Universitaetsplatz 2, G10-134, Magdeburg 39106, Germany. http://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.06.070 0017-9310/Ó 2014 Elsevier Ltd. All rights reserved.
[12] found that the volumetric spray flux has a dominant effect on heat transfer compared to other hydrodynamic properties of the spray. Al-Ahmadi et al. [5] also found that the Leidenfrost temperature, critical temperature, and maximum heat flux are strongly dependent on the local spray mass flux. Other spray parameters such as droplet size, velocity, and the type of nozzle play a less significant role [2,5,18]. Müller et al. [1] presented the change of the heat transfer coefficient (HTC) vs. surface temperature for Nickel. They investigated the HTC variation at different values of volumetric spray flux. Jeschar et al. [3] found that in the range of film boiling, HTC can only be influenced by the volumetric spray flux. Other parameters, such as drop velocity, distance nozzle to surface and nozzle type, do not require to be considered separately, as they are contained in the volumetric spray flux. However, the study of Wendelstorf et al. [4] shows HTC in the film boiling depends not only on the volumetric spray flux but also on the temperature. Puschmann et al. [14] found a similar result where HTC increased with the volumetric spray flux. They also confirmed that with a lower volumetric spray flux, the drop diameter exerts no influence on the HTC obtained. However, when they doubled the droplet velocity, they obtained about a difference of 50 W/m2/K in the HTC. The effect of salinity has been studied as a third relevant parameter that influences the heat treatment process. Huang et al. [7] made an intensive analysis of the influence of salt on the Leidenfrost transition. They explain the reason for the elevated Leidenfrost temperature resulting from the addition of salts. In their analysis, the suppression of bubble coalescence by dissolved salt, salt deposition during the initial contact of the deposition, and earlier
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
collapse of the vapor film due to the increasing salt concentration at the liquid–vapor interface were studied as the mechanisms associated with the Leidenfrost transition. KCl and NaCl were used. As a result, they found that the Leidenfrost temperature was significantly increased by the presence of dissolved salt. Cui et al. [6] conducted an experiment to observe the influence of three different salts (NaCl, Na2SO4, and MgSO4) dissolved in a water spray used for the cooling of a hot surface. Copper heated up to 240 °C was used. It was found that the individual salts influence the heat transfer process in different ways. Some increased the performance of heat transfer in nucleate boiling while others increased both nucleate and transition boiling heat transfer. MgSO4 produced the largest increase in nucleate boiling heat transfer, with Na2SO4 producing a somewhat less increase and NaCl producing the smallest increase. In addition, it was clearly shown that a higher concentration of salts lead to a higher maximum heat flux. Alam et al. [15] also found by using atomized spray that the Leidenfrost temperature increased with higher salt concentrations dissolved in the cooling water. Jeschar et al. [16] measured the influence of salts on the cooling process of a Nickel sphere dipped into a bath of water. NaSO4, NaCl, NaHCO3, CaCl2, and MgSO4 were used as individual salts dissolved up to 1 g/l concentration. The Leidenfrost temperature was increased up to 100 K for all salts. The solution with Mg2+ and SO2 4 content had the strongest effect. Although previous works have reported that salinity has a strong influence on the heat treatment process, there is no extended research published quantifying the effect of salt mixture to the Leidenfrost and DNB temperature, maximum heat flux, and heat transfer coefficient in the film boiling. In this study, those parameters are quantified for eight different real solutions used in the metal processing industry containing different salts. Electrical conductivity (EC), which has often been used as an index of the total dissolved solids (TDS), is introduced as a means of estimating the salinity level of the cooling water since EC depends on the overall ionic concentration in water [9]. The use of a single salt (MgSO4) with different concentrations (0.25–12 g/l) was chosen to show the correlation between the salt concentration and its electrical conductivity. Constant and accurate reproducibility of the spatial volumetric spray flux is the most important factor in order to study the effect of salinity on spray cooling heat transfer.
77
2. Experimental method 2.1. Experimental setup The experimental setup is shown schematically in Fig. 1. Basically, it consists of four main units. First, there is the heating system which is an electrically-heated furnace to heat up the sample to the desired level of temperature. In this work, all samples were heated up to 560 °C. Second, a metal sample is hung up on the rail that can shift the sample cater free from the furnace to the cooling unit direction. A disc of Aluminum Alloy AA6082 with a diameter of 140 mm and a thickness of 3 mm was used in the experiments (detailed information is summarized in Table 1). Third, there is a spray cooling unit by which the quenching process takes place, which consists of a centrifugal pump supplying water at the certain level of flow rate verified by a flow meter and a hydraulic nozzle used as a spraying component producing water droplets to the hot metal sample. A constant volumetric spray flux of 3 kg/m2/s was applied in this experiment. This volumetric spray flux was produced by applying a flow rate of 20 l/h water through a hydraulic nozzle. Fourth, there is the data acquisition system which consists of an infrared (IR) camera connected to a computer system for the image recording process. Prepared samples have to be painted black on the back side before being put into the electrical furnace for the heating process. After that, the hot metal sample is shifted manually along the sliding rail towards the stop position where the water spray will be aimed. The stop position is adjusted in such a way so that the camera focus, center point of metal sample, and center point of spray cone are aligned. Once it reaches the stop position, the spray cover is removed and the IR camera is activated to start recording the temperature history during the cooling process. The sample moving process was verified to not cause a significant temperature drop. Every time this will not cause more than 2 °C temperature difference. 2.2. Determination of emissivity The most important parameter of the IR camera setting is the emissivity of the surface. This is the reason why the sample must
Fig. 1. Schematic diagram of quenching experimental facility (top view).
78
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
Table 1 Geometry and material details. Diameter of the disc Thickness Density Thermal conductivity Specific heat
140 mm 3 mm 2770 kg/m3 170 W/m/K 1050 J/kg/K
be painted black with a graphite coating on the back side in order to get a higher emissivity for a more accurate measurement. For the calibration of emissivity of a graphite coating, a small copper cylinder was machined with an inside bore. A thermocouple was inserted inside the bore so that the temperature of the metal could be measured. This cylinder was painted black with the graphite coating and placed inside a horizontal cylindrical furnace. An IR Camera was then positioned in front of the outlet of this furnace from a distance of 80 cm. The temperature of the furnace was increased in steps of 50 °C from 50 to 700 °C. The temperature of the cylinder measured by thermocouple and that of the IR camera were aligned with each other by adjusting the emissivity value in the settings of the camera. In this way, the emissivity was recorded vs. the temperature of the cylinder surface. This procedure was repeated with both the heating and cooling direction. It was found that emissivity of the coating changed from 0.92 to 0.94. However, the mode value of emissivity was found to be 0.92. Therefore, it can be inferred that a constant value of emissivity of 0.92 should be used for the optimum accurate measurement of the surface temperature. A detailed review was also made in another publication [13]. Moreover, for purposes of accuracy, the thermal image was set to capture and store at the extremely high frame rate of 150 Hz within the range of 240 80 pixels. 2.3. Volumetric spray flux measurement The volumetric spray flux was measured separately by changing the targeted sample with a patternator. A patternator is a device designed for volumetric spray flux measurement. It consists of brass tubes with diameters of 10 mm. Each tube is fixed in such a way an aligned formation is formed. At the bottom end, each tube is connected to the plastic bottle into which the water drops will be collected. Water droplets are sprayed over the patternator tubes and then the mass of the water droplets md collected by the plastic glass is measured after a certain period of time Dt. The area of the opening of the tube is denoted by A. Then, the volumetric spray _ ¼ md /(ADt). The measurement was flux can be derived from m estimated to be accurate within ±4% as several repeating measurements were fluctuating in this range (Fig. 2). Since only one nozzle with a single flow rate set-value is used in this work, it is not necessary to make further verification on the influence of droplet diameter and velocity. 2.4. Preparation of test solution In this experiment, deionized water is used as a solvent of the single-salt solution. This is to minimize the influence of dissolved chemicals and gases. A salt solution was prepared by dissolving the required amount of powdered salt with a purity of P99.5% in deionized water at room temperature and then stirring the solution to ensure complete mixing. In each test, a fresh salt solution was prepared before the experiment. The amount of salt was weighed with a precision balance with an accuracy of ±0.01 g. The temperature of the solution was carefully measured before being applied in the cooling experiment with an accuracy of ±0.5 °C. Real solutions used in different metal processing companies was used no more than 2 days after arrival. All the waters
Fig. 2. Volumetric spray flux distribution for the hydraulic nozzle used in the experiment.
were treated uniformly to have the temperature Tsp of approximately 18 °C. 2.5. Energy balance analysis and accuracy In order to calculate the heat transfer coefficient for a specific position on the metal disc, an energy balance is performed as follows:
q c s
DT ¼ q_ sp þ q_ k þ q_ R þ q_ a Dt
where q, c, and s are the density, specific heat capacity, and the thickness of the metal disc, respectively. The change of the stored enthalpy in the disc caused by temperature decrease is equal to the sum of the heat transfer by the spray q_ sp , radial conduction q_ k , radiation q_ R , and free convection q_ a . The heat flux due to spray has the largest magnitude when compared to all other sources. Heat loss due to radiation and natural convection is less than 5% that of the spray and is therefore negligible. This heat flux q_ sp is defined by the water spray temperature Tsp with the equation
q_ sp ¼ asp T s T sp
where q_ sp ; asp , and Ts are the heat flux, heat transfer coefficient, and surface temperature of the metal disc, respectively. Due to the radial temperature distribution of the spray with the minimum temperature in the center, heat is conducted in the metal disc plane from the outer regions to its center. This heat flux is calculated using the Fourier approach for cylindrical coordinates
q_ k ¼ s k
1 @ @T r r @r @r
where s is the thickness of the disc, k is the conductivity of material, and r is the radius. The gradient oT/or is known from the temperature measurement. The energy balance equation is the basis for a numerical procedure. This procedure calculates the time-dependent heat transfer coefficient for a specific position on the metal disc surface from the varying temperatures at this position. Since the metal disc used is considered a thin sheet, it was verified that a temperature difference of only about 1 °C exists between the front and backside surface. The heat flux was also analyzed with a 2D inverse solution. The results found were in good agreement to those calculated with the simplified energy balance model. A detailed review of this comparison has been made elsewhere in another publication [8]. The specific heat capacity and conductivity have an accuracy of about 10%. The temperature measurement accuracy follows the IR camera specifications. With a certain numerical technique oT/ot
79
can be accurately determined. The percentage of cooling rate error is about 3%. The overall accuracy for the determination of the heat transfer coefficient was 5–10%, and the repeatability was considered good as statistical fluctuation can be controlled within 5% as shown by Fig. 3. For this reason, for every group of the experimental plan, repetitions were made and only the data which has acceptable repeatability will be considered. Since main factors that could influence the measurement were blocked and identified, poor repeatability was assumed to be caused by random error.
Electrical conductivity in µS/cm
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
25000 20000 15000 10000 5000 0 0
3. Results and discussion
2
4
6
8
10
12
Concentration in g/L
3.1. Influence of single salt 3.1.1. Salinity and electrical conductivity (EC) Salinity depends on the dissolved ions. Dissolved ions in the water can be positively charged ions (Na+, Mg2+, etc.) or negatively charged ions (Cl, SO2 4 , etc.), which can come from the dissolved salts in the water. Electrical conductivity (EC) is a measure on the ability of the material to conduct electrical current, which in a solution depends on the presence of charged ions and its concentration. Therefore, EC has often been used as an index of the total dissolved solids (TDS) contained in a solution. The ratio of EC to concentration (TDS) generally ranges from 1.33 to 1.82 [9]. In this work it is proposed to use the measured value of EC as an indirect indicator of the salinity level of the solution. As a basic understanding, solutions with different levels of salinity have been created by dissolving different concentrations of MgSO4 to deionized water. In this work, the concentrations used were 0.25, 1, 1.5L, 2 g/L. Deionized water was used as reference water. All previous studies [6,15,16] have shown that MgSO4 has the strongest effect. For illustration purposes, an extreme condition of 12 g/L was included in certain comparison figures. Fig. 4 shows the correlation between measured EC and MgSO4 concentration. It is obviously seen that up until 5 g/L, the concentration varies almost linearly with the EC with a ratio of 1.53. For concentrations higher than 5 g/L the correlation is inherently nonlinear, and it is
Fig. 4. Measured EC of MgSO4 solution.
suspected that at higher concentrations, interactions among ions can impede their mobility [17]. On the other hand, concentrations higher than 1 g/L are not likely to be adopted in a spray cooling process; they are only used for the clarity of illustration so that the correlation or the influence of the salinity level is more easily observed. However, to investigate the practical aspect, the idea was proposed in this work to also review real solutions used in metal processing industries with respect to different salinity levels. The results will give important information on the fluctuating range of the main characteristics of the cooling process due to the change in the salinity level. The highest EC value from the real solutions is about 2580 lS/cm which is comparable to that of MgSO4 solution with a concentration of 1.5 g/L. 3.1.2. Temperature profile Fig. 5 depicts the cooling profile of different concentrations of MgSO4. It is clearly shown that the higher the concentration, the shorter the cooling duration is entirely. When it is observed further, it is seen that the higher EC, which indicates a higher concentration of salt (MgSO4), shortens the film boiling duration significantly. It took 21.5 s for deionized water to decrease from 560 °C to its Leidenfrost temperature. Below this temperature, abrupt increase of heat flux occurred, and it took only 1.5 s to reach 100 °C. These results were compared to that of the 1 g/L MgSO4 solution, which took 15 s to reach its Leidenfrost temperature and from this point also took 1.5 s to reach 100 °C. By doubling the concentration of MgSO4 in the solution to 2 g/L, the film boiling duration is further significantly reduced to 11 s, but it still took 1.2 s to reach 100 °C. For the extreme case of the 12 g/L MgSO4 solution, it took only 4 s in total to reach 100 °C. The latter case is only used for the clarity of illustration. As explained by Cui et al. [6], a higher concentration creates greater surface roughness, which in turn can break through
600 1. DI Water 2. MgSO4 0.25 g/L 3. MgSO4 1 g/L 4. MgSO4 2 g/L 5. MgSO4 12 g/L
Temperature in C
500 400 300 200 5
4
3
2
1
100 0 0
Fig. 3. Cooling profile repeated three times.
5
10
15 Time in s
20
25
30
Fig. 5. Temperature profile of four different concentrations of MgSO4 at center point.
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
5
5
Heat Flux in MW/m2
4.5 4 3.5
4
1. DI Water 2. MgSO4 0.25 g/L 3. MgSO4 1 g/L 4. MgSO4 2 g/L 5. MgSO4 12 g/L
3 3
2.5 2
2
1.5
1
1
Maximum heat flux in MW/m2
80
6 5 4 3 2 1 0 0
0.5 0 0
100
200 300 400 500 Surface Temperature in C
500 1000 1500 2000 2500 3000 3500 4000 Electrical conductivity in S/cm
600
Fig. 8. Maximum heat flux of four different concentrations of MgSO4 at center point.
Fig. 6. Heat flux over surface temperature of four different concentrations of MgSO4 at center point.
the vapor layer insulating the heated surface and increase direct liquid–solid contact. Since the film boiling duration dominates the entire duration of the cooling process, an earlier start to the transition boiling will eventually shorten the duration of the cooling process entirely. Corresponding heat flux over surface temperature from different concentrations is shown in Fig. 6. 3.1.3. Leidenfrost and DNB temperature As the Leidenfrost temperature is shifted to higher values, the cooling duration is shortened. This result is in line with the finding explained excellently by Huang et al. [7] and Craig et al. [11]. Bubble merging may be inhibited by dissolved salts as some electrolytes do retard coalescence of bubbles. Given this, a higher surface temperature is required to start the Leidenfrost transition. A higher concentration, i.e. a higher EC, may increase the number of charged ions in the solution. A higher EC means that a higher level of electrolytes is created so that a more intense retardation of bubble coalescence might take place. This also agrees with what was observed by Cui et al. [6], which estimated that the precipitating salt particle from the evaporating droplet would provide a bubble nucleation site to trigger bubble formation and promote the nucleate boiling heat transfer earlier. However, Huang et al. [7] explained their finding that the increase of the salt concentration on the liquid–vapor interface significantly accounts for the increase of Leidenfrost temperature due to the increase of boiling point. In this work, the Leidenfrost temperature was estimated from a relative minimum value of wall heat flux in the film boiling region. While the DNB temperature is estimated from the maximum heat flux point. Specifically, Fig. 7 shows that up to 2 g/L, the Leidenfrost temperature increased in the range of 10–30 °C more than that of deionized water. These results agree with those
found by Jeschar et al. [16], who made this investigation by immersing a nickel ball into a pool of solution. Huang et al. [7] found that the average Leidenfrost temperatures for salt solutions were roughly 10–33 °C higher than that of deionized water. Fig. 7 also shows a higher EC promoting the onset of transition boiling, i.e. the DNB temperature, from about 165 to 190 °C. This result agrees with the result found by Cui et al. [6], which states that the greater surface roughness created by the deposited salt (MgSO4) particle increased the temperature at which the maximum heat flux occurred. 3.1.4. Maximum heat flux By increasing the EC from 6.7 to 4000 lS/cm, the maximum heat flux dramatically changes from 1.6 MW/m2 for deionized water to a maximum value of approximately 3.0 MW/m2 when a 2 g/L solution of MgSO4 was used (Fig. 8). This is a similar trend to that of the result of Cui et al. [6], which also used a volumetric spray flux of 3 kg/m2/s. They found that the surface heat flux increased with salt concentration for all temperatures but only until 24 g/L. Further increase reduces the heat transfer in the nucleate region. 3.1.5. Heat transfer coefficient in film boiling In this investigation, the heat transfer coefficient a in the film boiling is also determined. As depicted by Fig. 9, a increases with an increase in EC. However, it starts showing significant dependency on surface temperature beginning at EC = 2000 lS/cm. This experimental result shows a similar trend to that highlighted by Wendelstorf et al. [4], which found that in film boiling a increases with an increase in volumetric spray flux, but at higher level the dependency of a on the surface temperature is obviously observed. Other than that, these quantitative values in film boiling will
350
3
TLeidenfrost
250 200 150
TDNB
100
2.5 in kW/m2 K
Temperature in C
300
T = 400 C
2 1.5 1
T = 500 C
0.5
50
0
0 0
500 1000 1500 2000 2500 3000 3500 4000 Electrical conductivity in S/cm
Fig. 7. Leidenfrost and DNB temperatures of four different concentrations of MgSO4 at center point.
0
500 1000 1500 2000 2500 3000 3500 4000 Electrical Conductivity in S/cm
Fig. 9. Heat transfer coefficient in the film boiling region of four different concentrations of MgSO4 at center point.
81
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83 Table 2 Salts contained in the real solutions and corresponding measured EC. Parameters
DI Water
AG
K1
BF
A90
AL
AF
C1
C2
Magnesium [mg/l] Calcium [mg/l] Sodium [mg/l] Sulfate [mg/l] Carbonate [mg/l] Chloride [mg/l] EC [lS/cm]
<0.03 <2 <0.03 <5 2.94 3 6.7
14 84 88 <5 152.6 11 530
7.8 38.5 96 <5 140.9 62 692
<2 2 186 62 151 43 841
17 49.7 94.4 36 115.3 121 869
11 138 149 180 111 241 1477
53 150 91 301 126 195 1565
106 248.5 77 867 40.4 170 2190
93 88 305 628 49.1 191 2580
enrich what has been found by Cui et al. [6] since they only used low initial temperatures on a copper sample.
3.2.2. Temperature profiles The temperature profile of the cooling process from three different real solutions at the center point is plotted in Fig. 10, and the corresponding heat flux over surface temperature is shown in Fig. 11. It is evident that the higher the EC value, the shorter the duration of the entire cooling process. The temperature profiles themselves indicate the influence of salinity on the cooling rate. The highest EC among these solutions is from the C2 solution, and the value of its EC is comparable to that of the 1.5 g/L MgSO4 solution. When compared, it is clear that the influence from the salts mixture is stronger than the single salt since it causes a shorter cooling duration. 3.2.3. Leidenfrost and DNB temperature In Fig. 12 the trend of the Leidenfrost and DNB temperatures at the center point is mapped according to the increase in EC values. It is shown that up to 2580 lS/cm, the Leidenfrost temperature increased in the range of 10–100 °C more than that of deionized water. Furthermore, the higher EC also promotes the onset of the transition boiling, i.e. the DNB temperature, from about 165 °C for deionized water to 230 °C for the C2 solution. In this work, temperatures where the maximum heat flux takes place are assumed
Temperature in C
600
400 300 200 4
3
2
1
100 0 0
5
10
15 20 Time in s
25
30
1. DI Water 2. BF 3. AL 4. C2
4
5 3
4 3
2
2
1
1 0 0
100
200 300 400 Surface temperature in C
500
600
Fig. 11. Heat flux over surface temperature of three samples of different real solutions at center point.
350
1. TLe of real solutions 2. TLe of MgSO4
300
1 TLeidenfrost
2
250 TDNB
3
200 150
4
3. TDNB of real solutions 4. TDNB of MgSO4
100 0
1000
2000
3000
4000
Electrical conductivity in S/cm Fig. 12. Leidenfrost and DNB temperature of eight different real solutions at center point.
to be the DNB temperatures. In comparison with the effect of a single salt, both the Leidenfrost and the DNB temperature from the salts mixture were more strongly affected. 3.2.4. Maximum heat flux Subsequently, the trend of the maximum heat flux at the center point of the real solutions is plotted in Fig. 13. It is shown that by increasing the EC from 6.7 to 2580 lS/cm, the maximum heat flux dramatically changes from 1.6 MW/m2 for deionized water to a maximum value of approximately 5.0 MW/m2 when the C2 solution was used. When compared to the single salt, the rate of the increase on the maximum heat flux values from the salts mixture is in the range of 1–2.5 MW/m2 higher.
1. DI Water 2. BF 3. AL 4. C2
500
Heat flux in MW/m2
3.2.1. Overview of real solutions In the case of salt mixtures, eight different real solutions used in some industrial plants during the casting process were investigated. In Table 2, the analysis of these solutions is summarized. The analysis of deionized water, which is used as reference water, is also included. All solutions are ranked from left to right from smallest to largest measured EC. In general, the basic characteristics of the cooling process with the real solutions tend to be similar to that of a single-salt solution as reviewed in sections below:
Temperature in C
3.2. Influence of salt mixture
6
35
Fig. 10. Profile of temperature over time of three samples of different real solutions at center point.
3.2.5. Heat transfer coefficient in the film boiling In order to compare with the single salt effect, the heat transfer coefficient a in the film boiling region of the real solutions at the center point is also reviewed. As shown by Fig. 14, a increases with an increase in EC. In addition, a similar trend as found in the
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83
Maximum heat flux in MW/m 2
82
5. Concluding remarks
6 5 4 3 2 1 0 0
1000
2000
3000
Electrical conductivity in S/cm Fig. 13. Maximum heat flux of eight different real solutions at center point.
4
in kW/m2K
3.5 T = 400 C
3 2.5 2
T = 500 C
1.5 1 0.5 0 0
500
1000
1500
2000
2500
3000
The experimental work for studying the effect of different salinity levels contained in the cooling water is established. Two different groups of solution were tested. The first is a group of solutions created from varying concentrations of a single salt MgSO4. And the second is a group of solutions taken from the real solution used in different metal processing industries. A disc of AA6082 with a diameter of 140 mm and a thickness of 3 mm is heated up to 560 °C. Then it is quenched at the front side, while the temperature at the back side is recorded with use of an infrared camera. Repeatability of the measured temperature and volumetric spray flux are also verified. In this work, it is proposed to use EC instead of concentration for the sake of practicability when measuring the salinity level in the cooling water. Until 6 g/L a linear correlation with a certain factor was found. From the measured temperature, the Leidenfrost and DNB temperature, maximum heat flux, and heat transfer coefficient in film boiling are estimated by using a 1D simplified energy balance model. From both groups of solutions used, it is observed that these values increase with the increase in EC. However, it is found that the effect resulted from a salts mixture is stronger than that of a single salt. It is believed that the interaction between ions might change properties in the solutions that ultimately change the effect found from the individual salt on the cooling process. In order to verify the results, further work is still required especially that of observing the quenching behavior of a solution created from a binary salt mixture.
Electrical conductivity in S/cm Fig. 14. Heat transfer coefficient in the film boiling region of eight different real solutions at center point.
investigation with a single salt, that a significant dependency of a on the surface temperature is also observed. However, from the comparable EC values at the same surface temperature, it is observed that a difference of 0.5–1 kW/m2K exists.
4. Other considerations It is interesting to review the fact that the influence of a single salt is not exactly the same as those found from a salt mixture especially on the specific parameters compared above. The salt mixture has a stronger effect than the single salt at comparable EC values. It is believed that the interaction between ions might change properties in the solutions, e.g. surface tension, saturation temperature, and liquid density, which ultimately change the effect found from the individual salt on the cooling process. Craig [10] highlighted on his review on bubble coalescent and specificion effects that ion combinations separated on the interface result in a reduction in the mobility of the interface. The reduced mobility leads to an increase in the stability of the film. When a sufficient concentration of ions in bulk is reached, the surface mobility is suppressed to an extent that bubble coalescence does not take place within the life time of a collision. The types of ions contained in the sample of real solutions (shown in Table 2), Mg2+, Ca2+, Na+, SO2 4 , NO3 , and Cl , are categorized in Craig’s review [10] as the ions that when combined with each other will inhibit bubble coalescence. This situation agrees with the observation found in Huang et al., which is that inhibition of bubble coalescence will result in the Leidenfrost point becoming unachievable at the corresponding surface temperature, and therefore a higher surface temperature is required to start the Leidenfrost transition. However, further work is still required to confirm this suspected cause especially that of observing the quenching behavior of a solution created from a binary salt mixture.
Conflict of interest None declared. References [1] H.R. Müller, R. Jeschar, Heat transfer during water spray cooling of nonferrous metals, Metallkunde 74 (1983) 257–264. [2] M. Ciofalo, A. Caronia, M. Di Liberto, S. Puleo, The Nukiyama curve in water spray cooling: its derivation from temperature–time histories and its dependence on the quantities that characterize drop impact, Int. J. Heat Mass Transfer 50 (2007) 4948–4966. [3] R. Jeschar, E. Specht, C. Köhler, Heat transfer during cooling of heated metal objects with evaporating liquids, in: H. Tensi, B. Liscic, W. Luty (Eds.), Theory and Technology of Quenching, Berlin, Springer-Verlag, 1992, pp. 73–92. [4] J. Wendelstorf, K.H. Spitzer, R. Wendelstorf, Spray water cooling heat transfer at high temperatures and liquid mass fluxes, Int. J. Heat Mass Transfer 51 (2008) 4902–4910. [5] H.M. Al-Ahmadi, S.C. Yao, Spray cooling of high temperature metals using high mass flux industrial nozzles, Exp Heat Transfer 2 (2008) 38–54. [6] Q. Cui, S. Chandra, S. McCahan, The effect of dissolving salts in water sprays used for quenching a hot surface: Part 2 – spray cooling, ASME J. Heat Transfer 125 (2003) 333–338. [7] C.K. Huang, V.P. Carey, The effect of dissolved salt on the Leidenfrost transition, Int. J. Heat Mass Transfer 50 (2006) 269–282. [8] A. Nallathambi, E. Specht, Estimation of heat flux in array of jets quenching using experimental and inverse finite element method, J. Mater. Process. Technol. 209 (2009) 5325–5332. [9] R. Das, N.R. Samal, P.K. Roy, D. Mitra, Role of electrical conductivity as indicator of pollution in shallow lakes, Asian J. Water Environ. Pollut. 3 (1) (2005) 143–146. [10] V.S.J. Craig, Bubble coalescence and specific-ion effects, Curr. Opin. Colloid Interface Sci. 9 (2004) 178–184. [11] V.S.J. Craig, B.W. Ninham, R.M. Pashley, The effect of electrolytes on bubble coalescence in water, J. Phys. Chem. 97 (39) (1993) 10192–10197. [12] R.E. Gaugler, An experimental study of spray cooling of high temperature surfaces (Ph.D. dissertation), Department of Mechanical Engineering, Carnegie Institute of technology, Pittsburg, 1966. [13] U. Alam, J. Krol, E. Specht, J. Schmidt, Enhancement and local regulation of metal quenching using atomized spray, J. ASTM Int. 5 (10) (2008). [14] F. Puschmann, E. Specht, J. Schmidt, Local distribution of the heat transfer in water spray quenching, in: International Conference on Continuous Casting of Non-Ferrous Metals. Proceedings of the DGM Frankfurt, 2000, pp. 101–107. [15] U. Alam, Experimental study of local heat transfer during quenching of metals by spray and multiple jets (Ph.D. thesis), Otto-von-Güricke Universität Magdeburg, 2011.
K.H.M. Abdalrahman et al. / International Journal of Heat and Mass Transfer 78 (2014) 76–83 [16] R. Jeschar, H. Kraushaar, H. Griebel, Influence of gases dissolved in cooling water on heat transfer during stable film boiling, Steel Res. 67 (1996) 227–234. [17] R.D. More, G. Richards, A. Story, Electrical conductivity as an indicator of water chemistry of hydrologic process, Streamline Watershed Manage. Bull. 11 (2008) 25–29.
83
[18] J. Schmidt, H. Boye, Influence of velocity and size of droplets on the heat transfer in spray cooling, Chem. Eng. Technol. 24 (2001) 255–260.