- Email: [email protected]
Evaporation heat transfer and pressure drop of ammonia in a mixed configuration chevron plate heat exchanger
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Available online at www.sciencedirect.com
ScienceDirect w w w . i i fi i r . o r g
journal homepage: www.elsevier.com/locate/ijrefrig
Evaporation heat transfer and pressure drop of ammonia in a mixed configuration chevron plate heat exchanger Mohammad S. Khan a,*, Tariq S. Khan b, Ming-C. Chyu c, Zahid H. Ayub d a
Department of Mechanical Engineering, Mohammad Ali Jinnah University, Islamabad Campus, Islamabad, Pakistan b Department of Mechanical Engineering, Petroleum Institute, United Arab Emirates c Department of Mechanical Engineering, Texas Tech University, USA d ISOTHERM, Inc., Texas, USA
article info
abstract
Article history:
Ammonia is a naturally occurring environment friendly refrigerant with attractive thermo-
Received 4 September 2013
physical properties. Experimental investigation of heat transfer and pressure drop during
Received in revised form
steady state evaporation of ammonia in a commercial plate heat exchanger has been
23 December 2013
carried out for an un-symmetric 30 /60 chevron plate configuration. Experiments were
Accepted 28 December 2013
conducted for saturation temperatures ranging from 25 C to 2 C. The heat flux was
Available online 8 January 2014
varied between 21 kW m2 and 44 kW m2. Experimental results show significant effect of saturation temperature, heat flux and exit vapor quality on heat transfer coefficient and
Keywords:
pressure drop. Current mixed plate configuration data are compared with previous studies
Plate heat exchanger
on the same heat exchanger with symmetric plate configurations. This comparison high-
Ammonia
lighted importance of optimization in selection of the heat exchangers. Correlations for
Heat transfer
two phase Nusselt number and friction factor for each chevron plate configuration
Evaporation
considered are developed. A Nusselt number correlation generalized for a range of chevron
Pressure drop
angles is also proposed. ª 2014 Elsevier Ltd and IIR. All rights reserved.
Transfert de chaleur d’e´vaporation et chute de pression d’ammoniac dans un e´changeur de chaleur a` plaques Mots cle´s : Echangeur de chaleur a` plaques ; Ammoniac ; Transfert de chaleur ; Evaporation ; Chute de pression
1.
Introduction
In today’s highly industrialized world, energy is one of the major concerns. With rapid consumption of fossil fuels, saving
energy has become an attractive topic for the researchers. Recently, a marked improvement in heat exchanger technology has been observed. With improved technology vivid energy savings are now possible with efficient natural refrigerants in compact heat exchangers, used in various applications.
* Corresponding author. Tel.: þ92 51 111878787; fax: þ92 51 4486705. E-mail addresses: [email protected], [email protected] (M.S. Khan). 0140-7007/$ e see front matter ª 2014 Elsevier Ltd and IIR. All rights reserved. http://dx.doi.org/10.1016/j.ijrefrig.2013.12.015
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Nomenclature
x
Bo Dh f Gr h i k Nu P Pcr P* Pr q00 Re t T U v
Greek symbols b chevron or corrugation angle (deg) D change or difference m dynamic viscosity (kg m1 s1)
boiling number hydraulic diameter (m) fanning friction factor refrigerant mass flux (kg m2 s1) heat transfer coefficient (W m2 K1) enthalpy (kJ kg1) thermal conductivity (W m1 K1) Nusselt number pressure (kPa) or Plate pitch (m) critical pressure (kPa) reduced pressure (P/Pcr) Prandtl number heat flux (kW m2) Reynolds number plate thickness (m) temperature ( C) overall heat transfer coefficient (W m2 K1) specific volume (m3 kg1)
Plate heat exchangers (PHEs) are gaining popularity in recent years. However, data for two phase applications are still scarcely available in open literature, Khan et al. (2009). The gasketed plate heat exchangers are compact in nature, have high area to volume ratio, flexible in design and are easy to operate. Perhaps one of the major advantages of plate heat exchanger is its small size and relatively low energy usage. The flexible design of a plate heat exchanger allows overall reduced operational and maintenance costs. Because of their compact nature, these heat exchangers are not only energy efficient and cost effective but also adaptable for a variety of industrial applications. Ammonia is a naturally available refrigerant and has long been used in the refrigeration industry. It is a relatively low cost refrigerant with attractive thermo-physical properties. Ammonia based systems require relatively low maintenance. On the other hand ammonia is toxic and when mixed in air is flammable under certain conditions. However, being selfalarming has low safety compromise. While chloro-fluorocarbons are threat to ozone depletion and hydro-fluorocarbons add to global warming, natural refrigerants like ammonia are not only environment friendly but have efficient thermo-hydraulic characteristics. Therefore, due to danger of ozone depletion and global warming, interest in the use of natural refrigerants in air-conditioning and refrigeration industry has increased Khan et al. (2012a,b). The PHE used in the present study was configured in a single pass U-arrangement with counter flow setup. Such an arrangement is schematically shown in Fig. 1. It is well known that two phase heat transfer in heat exchangers is a more efficient mode of heat transfer compared to single phase heat transfer. Although some two phase work on compact heat exchangers was reported in the last century, however, major two phase experimentation on plate heat exchangers has been reported in the last decade. It may be interesting to note that two phase flow investigations carried out on plate heat exchangers with natural refrigerants such as
93
exit vapor quality
Subscripts acc acceleration core core ele elevation eq equivalent f liquid g vapor h hot stream m measured port port r refrigerant sp single phase tp two phase
ammonia are comparatively less in number, Khan et al. (2009). Recently Khan et al. (2012a,b) carried out experimental investigations on evaporation of ammonia in a commercial plate heat exchanger with two different symmetric (60 /60 and 30 /30 ) chevron plate configurations. The heat transfer coefficient and pressure drop are reported to increase with an increase in chevron angle and decrease in corrugation depth and mean channel spacing. Some other recent two phase studies using ammonia as refrigerant (Ayub, 2003; Sterner and Sunden, 2006; Djordjevic and Kabelac, 2008; Arima et al., 2010) are also discussed in Khan et al. (2012a,b). The current study uses the same heat exchanger as was employed in Khan et al. (2012a,b) but with an unsymmetric (mixed) chevron plate configuration. The mixed configuration is achieved by combining 30 and 60 chevron angle plates. The present work focuses on experimental investigation of thermal hydraulic characteristics of a 30 /60 un-symmetric commercial chevron plate heat exchanger for two phase ammonia evaporation. The experimental results for the un-symmetric configuration highlights the importance
Stream 2
Stream 1 Fig. 1 e Schematic diagram of single pass U-arrangement for counter flow plate heat exchanger setup (Kakac and Liu, 2002).
94
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Table 1 e Primary experimental parameters and their ranges. Parameter Saturation temperature, Tsat Exit vapor quality, x Equivalent Reynolds number, Reeq Heat flux, q00
Range 25 C < Tsat < 2 C 0.5 < x < 0.9 1225 < Reeq < 3000 21 kW m2 < q00 < 44 kW m2
of optimization in achieving the desired heat transfer with an economical pressure drop penalty. The experimental parameters considered in the present study and their ranges are tabulated in Table 1. Experimentation was carried out at five levels of the saturation temperature between 25 C and 2 C. Experiments were performed in such a way that effects of heat flux, exit vapor quality and saturation temperature could be studied for a selected value of mass flux. The refrigerant properties were evaluated at the saturation temperature. The single phase heat transfer coefficient on hot fluid side was estimated by a single phase heat transfer correlation developed by Khan et al. (2010) on the same plate heat exchanger. The two phase heat transfer coefficient on the refrigerant side was estimated from the overall heat transfer coefficient.
2.
Experimental setup
The experimental setup is schematically shown in Fig. 2. It consists of a refrigerant loop, a plate heat exchanger test loop, water/glycol solution loop, instrumentation and a data
acquisition system. A cooling tower served the cooling requirements of the condenser and the compressor. Ammonia was circulated in the refrigeration and test loop. Primary measurements in the experiments were flow rates, fluid temperatures, system pressure and differential pressure drop for both hot side and refrigerant side, across the plate heat exchanger. A commercially available gasketed plate heat exchanger (TL-90) has been used in this study. Chevron plates made of stainless steel (SS-316) were used. Table 2 provides important geometric details of the plate heat exchanger while the parameters are defined in Fig. 3. Further details of the experimental setup and procedure could be found in Khan et al. (2012a,b).
3.
Data reduction
3.1.
Heat transfer coefficient
In order to determine the heat transfer characteristics of the plate heat exchanger during evaporation of liquid ammonia, the overall heat transfer coefficient, wall thermal resistance and heat transfer coefficient on the hot fluid side must be known. The average heat transfer coefficient on the refrigerant side (htp) was determined from the overall heat transfer coefficient: 1 1 t 1 ¼ htp U k hsp
Fig. 2 e Schematic diagram of the experimental facility.
(1)
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
pressure drops due to acceleration of the refrigerant, due to change in elevation, due to inlet/exit manifolds, and due to change of pressure in the main core of the corrugated plate heat exchanger. The core pressure drop may therefore be obtained as (Kakac and Liu, 2002):
Table 2 e Geometric characteristics of chevron plates used in experiments. Geometric characteristic Plate width, Lw (mm) Vertical distance between centers of ports, Lv (mm) Port diameter, Dp (mm) Horizontal distance between centers of ports, Lh (mm) Mean channel spacing, b (mm) Plate thickness, t (mm) Effective area of plate, A (m2) Corrugation pitch, l (mm) Surface enlargement factor, 4 Plate thermal conductivity, k (W m1 K1) Chevron angle a
95
185 565
DPcore ¼ DPm DPacc DPelev DPport 43 125 2.9 0.6 0.095 13.25 and 6.25a 1.117 14 45
(3)
Using this pressure drop the two-phase friction factor for evaporation of liquid ammonia, was obtained by the following relation (Yan and Lin, 1999): ftp ¼
DPcore Dh 2G2r nm Lp
(4)
Pressure drop data reduction procedure has been elaborated in Khan et al. (2012a,b).
For b ¼ 30 and 60 plates, respectively.
4. The following single phase correlation developed by Khan et al. (2010)was used to estimate hsp: hsp ¼ 0:144Re0:78 Pr0:35
kf Dh
m
0:14
mwall
(2)
Further details are mentioned in Khan et al. (2012a,b).
3.2.
For the measurements and results derived from the experiments, the experimental uncertainties were calculated according to the procedure outlined by Moffat (1988). The method for single sample experiments involves estimation of overall uncertainty, wR in the calculated result, R, using the following relationship: "
Pressure drop wR ¼
The measured pressure drop across the plate heat exchanger for upward flow of refrigerant consists of four components;
Experimental uncertainty
n X vR i¼1
vXi
2 #1=2 wi
where R is expressed as: R ¼ (X1, X2, X3, ........Xi, Xn)
Fig. 3 e Basic geometric characteristics of the chevron plate (Kakac and Liu, 2002).
(5)
96
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Table 3 e Experimental parameters and their estimated uncertainties. Parameter Pressure drop, DP (kPa) Temperature, T ( C) Hot fluid mass flux, Gh (kg m2 s1) Refrigerant mass flux, Gr (kg m2 s1) Heat flux, q00 (kW m2) Uncertainty in calculated water/glycol side heat transfer coefficient in single phase study (Khan et al., 2010) Deviation of single phase correlation from single phase heat transfer data (Khan et al., 2009) Refrigerant heat transfer coefficient, htp (W m2 K1) Two phase friction factor, ftp
Uncertainty 0.22% 0.1 C 0.85% < 0.1% <1% 8%
1.8%
10.9% 2.8%
Here Xi are the independent parameters and wi represents the associated uncertainty in the independent parameter. Based on the precision of primary measurements, instrument specifications and fluid property variations, the uncertainties for each set of measurements are calculated in the present experimental study. Finally, the overall uncertainty in Nusselt number came out to be within 10.9%. After accounting for errors in the hot and cold fluid stream data measurements, the maximum errors in the primary measurements of mass flow rate, pressure and temperature were calculated. The uncertainties are tabulated in Table 3.
5.
Results and discussions: heat transfer
5.1. Effects of saturated temperature, heat flux and exit vapor quality Figs. 4 and 5 show experimental heat transfer coefficient versus heat flux and exit vapor quality at various temperatures. For a fixed mass flux, two phase ammonia evaporation data are obtained at five different saturation temperatures (saturation
Fig. 4 e Heat transfer coefficient versus heat flux for several saturation temperatures at a mass flux of 6.5 kg mL2 sL1.
pressures) for the un-symmetric chevron plate configuration considered in this study. Data are shown for saturations temperatures of 2 C, 9 C, 14.5 C, 19.5 C and 25 C. Figs. 4 and 5 show that although the effect of heat flux and exit vapor quality is not substantial, the heat transfer coefficient increases with an increase in the saturation temperature. It may also be observed from Figs. 4 and 5 that heat transfer coefficient follows similar trend with heat flux and exit vapor quality. Several parameters, such as transport properties could be responsible for increase in heat transfer coefficient, however, most importantly the higher temperatures result in smaller size bubbles escaping at higher frequency so the heat transfer coefficient increases with an increase in temperature. This is in agreement with work reported by Chyu et al. (2001). At higher temperatures they reported bubbles to grow, combine and slide over the tube surface with fluid flow, hence enhanced the heat transfer. It may also be observed from Figs. 4 and 5 that heat transfer coefficient increases moderately with an increase in heat flux up to 34 kW m2 and for exit vapor quality between 0.45 and 0.67 for the considered plate configuration. The htp decreases gradually with further increase in heat flux. Recently, Arima et al. (2010) have also shown that for a given saturation pressure and mass flux, the heat transfer coefficient decreases with average heat flux. Generally, before reaching the critical heat flux (dry-out regime), the boiling heat transfer is governed by convection boiling and/or nucleate boiling phenomena. Danilova et al. (1981) claimed that nucleate boiling dominates during evaporation in a PHE and Thonon et al. (1995) concluded that only convective boiling should be considered in PHEs. Sterner and Sunden (2006) reported that the relevant boiling mechanism changes from convective to nucleate boiling at certain specific geometric dimension of the channel. Djordjevic and Kabelac (2008) report that both nucleate and convective boiling mechanisms are present during evaporation of refrigerant in a PHE. Although, not shown here, it was found in preliminary experiments that for 0.1 < x < 0.4 the heat transfer coefficient increased with an
Fig. 5 e Heat transfer coefficient versus exit vapor quality for several saturation temperatures at a mass flux of 6.5 kg mL2 sL1.
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
increase in the heat flux. Hence nucleate boiling phenomenon was dominant in the low vapor quality region. No such major change in heat transfer coefficient is found with increasing heat flux in the current range of exit vapor quality, it can therefore be deduced that for higher exit vapor quality (0.5 > x > 0.8) the evaporation is not governed by nucleate boiling phenomenon. Though, this inference leads to convective boiling dominance, however, further data are required at various mass flux rates to confirm whether the evaporative boiling is dominantly convective in nature or not. The behavior of the heat transfer coefficient shown in Figs. 4 and 5 is in fair agreement with Ayub (2003) and Djordjevic and Kabelac (2008). Ayub (2003) proposed a two phase ammonia evaporation correlation in PHEs. The reported correlation includes effect of saturation pressure. He reported that the heat transfer coefficient increased with an increase in saturation pressure. Ayub (2003) maintains that unlike shell and tube heat exchangers, where bulk of the heat transfer is governed by pool boiling, due to narrow passages in the plate evaporators, the dominant mode of heat transfer could be convective in nature. A dry-out phenomenon has also been observed beyond exit vapor quality of 0.8 in the current study. This is in agreement with Djordjevic and Kabelac (2008) and Arima et al. (2010) who reported presence of dry-out regime at higher heat flux and vapor quality in their studies.
5.2.
Comparison with other studies
Khan et al. (2009) have reported that most of the two phase studies carried out on PHEs during the last decade are based on HFCs and a very few studies are conducted with ammonia as refrigerant (Ayub, 2003; Sterner and Sunden, 2006; Djordjevic and Kabelac, 2008; Arima et al., 2010). Moreover, no data were found on evaporation of ammonia for a mixed plate configuration plate heat exchangers in the literature. As mentioned earlier, Ayub (2003) presented industrial data based correlation for a wide range of chevron angle plate configurations at 8.5 kg m2 s1. Though bulk of the data for present unsymmetrical plate configuration has been taken at a fixed mass flux of 6.5 kg m2 s1, however, for the comparison purpose, some experiments are also carried out at 8.5 kg m2 s1 for a saturation temperature of 2 C. Comparison of present heat transfer data with correlation reported by Ayub (2003) is shown in Fig. 6. Present data are compared only with Ayub (2003) since no other two phase data for ammonia evaporation in the 30 /60 plate configuration has been reported. As could be observed from Fig. 6, Ayub’s (2003) correlation qualitatively follows a pattern similar to that of the current experimental heat transfer data. It is, however, interesting to note here that unlike previous studies on the same PHE with symmetric configurations (Khan et al., 2012a,b), the Ayub’s correlation (2003) under predicts experimental data gathered for the un-symmetric configuration. This difference between previously reported investigations and present data may be attributed to chevron plate geometric differences (surface corrugations). For example, a major difference between present investigation and previous studies on the same PHE (Khan et al., 2012a,b) is that previously plates used in a particular symmetric configuration were similar in all respects. While in the present study, in addition to the chevron angle, the
97
Fig. 6 e Comparison of heat transfer versus exit vapor quality between present study and Ayub (2003) at L2 C. respective corrugation pitch and corrugation depth of both commercial plates were different for the un-symmetric plate configuration considered.
6.
Results and discussion: pressure drop
While the thermal analysis of a heat exchanger determines heat transfer characteristics of the exchanger, frictional loss effects are equally important as the pressure drop offered by fluid properties and the heat exchanger geometry provides a measure of pumping power necessary to maintain the desired flow rate. This pumping power requirement adds to the initial and operating costs of the heat exchangers. Thus design and selection of a heat exchanger unit is influenced by both heat transfer characteristics and pressure drop offered. Two phase differential pressure drop data are collected during steady state evaporation of liquid ammonia. For a fixed refrigerant mass flux of 6.5 kg m2 s1, the pressure drop is measured by a differential pressure measurement device installed between inlet and exit ports of the plate heat exchanger. Pressure drop data are collected for the same five saturations temperatures as were in heat transfer experimentation. Unfortunately, no pressure drop data were found for evaporation of ammonia in plate heat exchangers in the published literature. Therefore, only a qualitative comparison is possible with pressure drop data for boiling of HFCs in compact heat exchangers.
6.1. Effect of saturation temperature and exit vapor quality For the given plate configuration, pressure drop is plotted against exit vapor quality for various saturation temperatures in Fig. 7. It may be observed that the pressure drop increased with an increase in saturation temperature for the considered plate configuration in the present study. It may also be observed from Fig. 7 that pressure drop is lesser for low saturation temperatures.
98
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
number, Boeq, and P* which is the reduced pressure defined as P/Pcr. The equivalent Reynolds number and Boiling number are given as:
Reeq ¼
Greq Dh m
(7)
Boeq ¼
q00 Greq ifg
(8)
where "
Greq
Fig. 7 e Experimental pressure drop as a function of exit vapor quality at 6.5 kg mL2 sL1.
It may also be observed that the pressure drop either increased, or decreased with an increase in exit vapor quality, x, for all saturation temperatures considered in the present investigation. It may be noted that the change in pressure drop remained almost insignificant up to x ¼ 0.7, for higher saturation temperatures. However, for higher x values (greater than 0.7), the pressure drop is observed to decrease with exit vapor quality. Pressure drop is observed to decrease more sharply at exit vapor quality greater than 0.7, in general.
7.
Two phase Nusselt number correlation
Several heat transfer applications require two phase flow which results in a more compact plate heat exchanger. However, benefits of better heat transfer characteristics of plate heat exchangers can be utilized effectively in the multi facet industry only when dependable correlations for the estimation of Nusselt number and pressure drop are available. It is known that very few studies on phase change heat transfer and two phase flow have been conducted on plate heat exchangers. Very few data are available for evaporation of ammonia in vertical plate heat exchangers. This lack of data requires an urgent need for correlations that can be used in designing of plate heat exchangers or to estimate their performance. An attempt has been made to develop correlations for heat transfer and pressure drop based on the experimental results of the present study. Applicability of these correlations is verified by comparing them with the experimental data. In the present study, the following type of correlation has been selected to estimate the experimental Nusselt number: m Nutp ¼ C Reeq Boeq ðP Þj
r ¼ Gr 1 x þ x f rg
!#1=2 (9)
Reeq is a function of plate geometry and equivalent mass flux, which in turn, is a function of vapor quality and liquid to vapor density ratio. The mass flux also depends on channel spacing, b. Similarly, Boeq depends on heat flux and latent heat of vaporization. Since the above mentioned correlation pattern includes all important experimental parameters measured and considered in this study, therefore, this correlation pattern has been found appropriate for flow boiling heat transfer of ammonia in the plate heat exchanger. Values of the coefficient, C, and exponents, m and j are determined using the classical regression analysis. The Nusselt number correlation hence developed for the plate configuration considered in this study, based on experimental data is given below: 0:079 0:33 ðP Þ Nutp ¼ 129:26 Reeq Boeq
(10)
Comparison between the experimental Nusselt number data and the Nutp estimated from correlation (Eq. (10)) for the chevron plate configuration considered in the present study is presented in Fig. 8. The maximum deviation of this correlation from experimental data is within 10%.
(6)
The two phase Nusselt number, Nutp, is a function of equivalent Reynolds number, Reeq, equivalent Boiling
Fig. 8 e Comparison of two phase experimental Nutp with correlation estimating the Nutp.
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
It can be observed that the developed correlation represents the experimental data quite well. The error analysis shows that for the considered plate configuration about 74% of the total experimental data fall within 4% error band while only 12% of the data fall between 7% to 10% error band.
7.1. A generalized correlation for the two phase Nusselt number Khan et al. (2012a,b) have presented correlations to estimate Nusselt number for the same commercial PHE that is used in un-symmetric plate configuration in the present study. The correlations proposed for two symmetric plate configurations earlier are as follows: (Khan et al., 2012a,b) 0:085 0:21 ðP Þ ; Nutp ¼ 82:5 Reeq Boeq 0:04 0:52 Nutp ¼ 169 Reeq Boeq ðP Þ ;
60 =60 30 =30
(11) (12)
Comparison of these correlations (Eqs. (11)e(12)) with Eq. (10) shows that the Nutp dependence on saturation pressure of ammonia increases with a decrease in chevron angle. Also the combined effect of equivalent Reynolds and Boiling numbers reduced with decreasing chevron angle. The lead coefficient, exponents of equivalent Reynolds and Boiling numbers product and reduced pressure in the developed correlations (Eqs. (10)e(12)) being different for each corrugated plate configuration considered, may be further related to chevron angle. Therefore, Nusselt number could be represented by the following more general functional form which also includes effect of chevron angle: mfb g jfb g Nutp ¼ Cfb g Reeq Boeq ðP Þ
(13)
Where b* is given as b/bmax and curly brackets represent functional form. bmax is taken equal to 60 which is the maximum chevron angle considered in the present and previous Khan et al. (2012a,b) investigations on the same PHE. In order to determine C{b*}, m{b*} and j{b*}, the C, m and j are individually correlated with chevron angle over the whole range of 30 e60 . It should be noted that the chevron angle for 30 /60 plate configuration is estimated to be equal to 45 . Linear curve fitting of the data provides the following equations for C, m and j as functions of b*. Cfb g ¼ 173:52b þ 257:2
(14)
mfb g ¼ 0:09b þ 0:0005
(15)
jfb g ¼ 0:624b 0:822
(16)
well and is valid for the whole range of chevron plate angles from 30 to 60 .
8.
Two phase friction factor correlation
Fig. 9 shows effect of Reeq on the two phase Fanning friction factor, ftp. It may be observed from Fig. 9 that the substantial reduction of ftp in the low equivalent Reynolds number range becomes gradual with further increase in Reeq. It may also be observed that the friction factor increases with an increase in the saturation temperature. In Fig. 9 although, the general trends for various saturation temperatures are quite similar, however, close examination shows that the overall difference between maximum and minimum values of ftp is large at higher saturation temperature (2 C) compared to low saturation temperature (25 C) over the range of Reeq flux considered in this study. Nevertheless, the friction factor increases almost linearly with increasing saturation temperature, in general. Since the effects of equivalent Reynolds number and saturation temperature on pressure drop are apparent, therefore, the following correlation has been considered to estimate the two phase experimental friction factor for ammonia evaporation in the vertical plate heat exchanger. m ftp ¼ C Reeq ðP Þj
(18)
As mentioned earlier, Reeq is a function of plate geometry, exit vapor quality and refrigerant properties, such as viscosity and liquid to vapor density ratio. The coefficient, C, and exponents, m and j are independent of fluid nature and are determined using the regression analysis approach. Using calculated values of these coefficients and exponents, the experimental friction factor could then be represented by the following correlation: 1:26 0:9 ðP Þ ftp ¼ 305; 590 Reeq
(19)
Comparison between experimental friction factor data and ftp estimated from correlation (Eq. (19)) is shown in Fig. 10. It
The above equations for C{b*}, m{b*} and j{b*} could be plugged in functional form of Nutp correlations given by Eq. (13). The Nutp correlation is then given as: mð0:09b þ0:0005Þ ð0:624b 0:822Þ Nutp ¼ ð173:52b þ257:12Þ Reeq Boeq ðP Þ (17) The above correlation shows similar trends for all chevron angle plate configurations. It represents experimental data
99
Fig. 9 e Variation of friction factor with equivalent Reynolds number for a mass flux of 6.5 kg mL2 sL1.
100
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Fig. 10 e Comparison of experimental friction factor with correlation (Eq. (19)).
can be observed that the friction factor correlation (Eq. (19)) represents the experimental data quite well. Error analysis shows that more than 88% of the total data fall within the error band of 6%, while only about 12% data fall between 7% and 10%. This comparison confirms that the experimental pressure drop data are well estimated by the developed correlation for the entire range of experimental parameters.
9. Comparison of heat transfer and pressure drop results for symmetric and un-symmetric plate configurations The effect of chevron angle on heat transfer and pressure drop for both symmetric and un-symmetric plate configurations is shown in Figs. 11 and 12.
Fig. 11 e Effect of chevron angle on heat transfer coefficient for various plate configurations at L2 C and L25 C.
For better readability, experimental data of only extreme saturation temperatures (2 C and 25 C) considered in the current study have been plotted. The heat transfer enhancement with chevron angle is clearly evident from Fig. 11. These results are also in agreement with almost all previously conducted studies on PHEs with different chevron angle plate configurations except Djordjevic and Kabelac (2008). They used two plate configurations in their investigations; soft plate configuration (27 /27 ) and high pressure drop (63 /63 ) configuration. They have reported that the soft plate configuration provides higher heat transfer coefficient compared to their high pressure drop configuration. This is contrary to findings of all previous investigations referenced in this work. The enhancement in heat transfer is generally attributed to the plate surface corrugations that generate swirl flows and periodic disruptions. Additionally, the smaller corrugation pitch and channel spacing also help enhance the heat transfer. Therefore, chevron plate corrugations encourage early turbulence which in turn helps increase the heat transfer. Although, overall trends of the heat transfer data shown in Fig. 11 for all three plate configurations are similar to a larger extent, however, some variability exists with respect to effect of chevron angle. Fig. 11 shows that the dry-out effect weakens from chevron angle plate configuration (60 /60 ) to chevron angle plate configuration (30 /30 ). For instance, it is observed that for the two phase experiments on 30 /30 chevron plate configuration, the dry-out effect is less significant compared to the other two configurations. Fig. 12 shows effect of chevron angle on pressure drop for the three configurations. Pressure drop is observed to be the maximum for 60 /60 plate configuration and minimum for 30 /30 plate configuration. This was expected due to geometric difference between three chevron plate configurations. It may also be noted from Fig. 12 that the effect of exit vapor quality on pressure drop is either insignificant or very small up to about 0.7 for all configurations. The two types of plates used in present experimentation are distinct from one another, mainly, in three aspects; the chevron angle, the corrugation pitch, l and the mean channel spacing, b. The 60
Fig. 12 e Effect of chevron angle on pressure drop for various plate configurations at L2 C and L25 C.
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
plate offers higher pressure drop due to its higher chevron angle, its smaller mean channel spacing (b ¼ 2.2 mm) and corrugation pitch (l ¼ 6.25 mm) compared to b ¼ 3.6 mm and l ¼ 13.5 mm for 30 plate. In the current un-symmetric configuration, the corrugation pitch and mean channel spacing are taken as average of the two plates. Fig. 12 shows that for 30 /60 and 30 /30 plate configurations; the pressure drop reduces with an increase in exit vapor quality for all saturation temperatures considered. On the contrary, the pressure drop increases with an increase in exit vapor quality for 60 /60 plate configuration. It may also be noted that pressure drop decreases with an increase in saturation temperature for 60 /60 plate configuration only. However, the pressure drop trends remained similar regardless of change in saturation temperature. In comparison to symmetric 60 /60 plate configuration the 30 plate geometry seems to govern the similar pressure drop trends for soft and mixed configurations. Although, the pressure drop for unsymmetric plate configuration (30 /60 ) is higher than that of 30 /30 plate configuration, it is interesting to note that the difference in pressure drop offered by mixed configuration and 30 /30 configuration is small compared to hard plate configuration. It is because of the fact that at low refrigerant mass flux, the resistance offered by 60 plate used in mixed configuration with 30 plate is compensated by 30 plate that has larger mean channel spacing and corrugation pitch. Since the current two phase experiments are carried out at rather low flow rates, the channel geometry and refrigerant properties appear to play major role in determining the pressure drop across the channel. These trends demonstrate strong influence of mean channel spacing, corrugation pitch and corrugation depth on pressure drop. Referring to the heat transfer results shown in Fig. 11, it may be observed that increase in heat transfer coefficient was more prominent than that in the pressure drop with chevron angle. For instance, in comparison to 30 /30 plate configuration at 25 C, on average, the heat transfer coefficient increases by about 2 times for 30 /60 plate configuration and by about 4 times for 60 /60 plate configuration. In comparison, the pressure drop for the same saturation temperature for 30 / 60 plate configuration is almost same as that of 30 /30 plate configuration but pressure increases about 3 times for 60 /60 plate configuration compared to 30 /30 plate configuration (Fig. 12). For the highest saturation temperature (2 C) considered, the average increase in heat transfer coefficient with chevron angle is about 2 and 2.5 times that of the 30 /30 plate configuration for 30 /60 and 60 /60 plate configurations respectively. Pressure loss penalty for 30 /60 plate configuration remained almost the same, however, decreased for 60 / 60 plate configuration by more than 1.5 times that of 30 /30 plate configuration. This comparison of results demonstrates a strong need for optimization of heat exchanger design and selection for their economic utilization in the industry.
10.
Conclusions
Steady state evaporation heat transfer and pressure drop of ammonia in a commercial plate heat exchanger with unsymmetric 30 /60 plate configuration has been investigated.
101
Experiments were conducted for equivalent Reynolds number range of 1225 to 3000, while the saturated temperature of ammonia varied from 25 C to 2 C. Results are reported for exit vapor quality of 0.5e0.9. Experimental results show a strong influence of saturation temperature and other operational conditions on heat transfer and pressure drop in the PHE. For the exit vapor quality greater than 0.5, convective boiling is found to dominate the heat transfer with heat flux and vapor quality having cumulative effect on heat transfer coefficient. Heat transfer coefficient and pressure drop are found to increase with an increase in saturation temperature. The effect of exit vapor quality on the two phase pressure drop is not significant except at higher values. The two phase Fanning friction factor is found to decrease with equivalent Reynolds number and increased with an increase in saturation temperature. Uncertainty analysis has shown a maximum error of about 10.9% in the experimental Nusselt number data and 2.8% in experimental friction factor data. Based on the experimental data, an empirical correlation for two phase Nusselt number is developed as a function of reduced pressure, equivalent Reynolds number and Boiling number for the 30 /60 plate configuration. Bulk of the heat transfer data are predicted within 7% of the error band. The two phase friction factor correlation is also proposed as a function of equivalent Reynolds number and reduced pressure for the above mentioned chevron plate configuration. This correlation represents experimental data within an error band of 10%. Heat transfer and pressure drop data of the present study are compared with some previous ammonia evaporation studies on the same PHE with symmetric plate configurations. This comparison necessitates a strong need for optimization in design and selection of plate heat exchangers to minimize energy requirements. Based on this comparison, a Nusselt number correlation generalized for all chevron angle plate configurations employed on this PHE is also proposed.
Acknowledgments The authors gratefully acknowledge the financial support provided by American Society of Heating Refrigerating and Air-Conditioning Engineers (ASHRAE) for this work under 1352-RP. Also thanks are due to Isotherm, Inc. of USA and Thermowave of Germany for donating the plate heat exchanger and spare heat exchanger plates with gaskets.
references
Arima, H., Kim, H.J., Okamoto, A., Ikegami, Y., 2010. Local heat transfer characteristics of ammonia in a vertical plate evaporator. Int. J. Refrig. 33, 359e370. Ayub, Z.H., 2003. Plate heat exchanger literature survey and new heat transfer and pressure drop correlations for refrigerant evaporators. Heat. Trans. Eng. 24 (5), 3e16. Chyu, M.-C., Zheng, J., Jin, G., 2001. Evaporation of Ammonia Outside Smooth and Enhanced Tubes with Miscible Oil. ASHRAE Report, RP-977. Djordjevic, E., Kabelac, S., 2008. Flow boiling of R134a and ammonia in a plate heat exchanger. Int. J. Heat. Mass Trans. 51, 6235e6242.
102
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Danilova, D.N., et al., 1981. Heat transfer in different plate geometries. Kholod Tek. 4, 25e31. Kakac, S., Liu, S., 2002. Heat Exchangers Selection, Rating, and Thermal Design. CRC. Khan, T.S., Khan, M.S., Chyu, M.-C., Ayub, Z.H., 2012a. Experimental investigation of evaporation heat transfer and pressure drop of ammonia in a 60 chevron plate heat exchanger. Int. J. Refrig. 35 (2), 336e348. Khan, M.S., Khan, T.S., Chyu, M.-C., Ayub, Z.H., 2012b. Experimental investigation of evaporation heat transfer and pressure drop of ammonia in a 30 chevron plate heat exchanger. Int. J. Refrig. 35 (6), 1757e1765. Khan, T.S., Khan, M.S., Chyu, M.-C., Ayub, Z.H., Javed, A., Chattha, 2009. Review of heat transfer and pressure drop correlations for evaporation of fluid flow in plate heat exchangers (RP1352). HVAC&R Res. 15 (2), 169e188.
Khan, T.S., Khan, M.S., Chyu, M.-C., Ayub, Z.H., 2010. Experimental investigation of single phase convective heat transfer coefficient in a corrugated plate heat exchanger for multiple plate configurations. Appl. Therm. Eng. 30 (8e9), 1058e1065. Moffat, R.J., 1988. Describing the uncertainties in experimental results. Exp. Therm. Fluid Sci. 1, 3e17. Sterner, D., Sunden, B., 2006. Performance of plate heat exchangers for evaporation of ammonia. J. Heat. Trans. Eng. 27 (5), 45e55. Thonon, B., Vidil, R., Marvillet, C., 1995. Recent research and developments in plate heat exchangers. J. Enhanc. Heat. Transf. 2 (1e2), 149e155. Yan, Y.Y., Lin, T.F., 1999. Evaporation heat transfer and pressure drop of refrigerant R-134a in a plate heat exchanger. J. Heat. Trans. 12 (1), 118e127 (Transaction of the ASME).
Available online at www.sciencedirect.com
ScienceDirect w w w . i i fi i r . o r g
journal homepage: www.elsevier.com/locate/ijrefrig
Evaporation heat transfer and pressure drop of ammonia in a mixed configuration chevron plate heat exchanger Mohammad S. Khan a,*, Tariq S. Khan b, Ming-C. Chyu c, Zahid H. Ayub d a
Department of Mechanical Engineering, Mohammad Ali Jinnah University, Islamabad Campus, Islamabad, Pakistan b Department of Mechanical Engineering, Petroleum Institute, United Arab Emirates c Department of Mechanical Engineering, Texas Tech University, USA d ISOTHERM, Inc., Texas, USA
article info
abstract
Article history:
Ammonia is a naturally occurring environment friendly refrigerant with attractive thermo-
Received 4 September 2013
physical properties. Experimental investigation of heat transfer and pressure drop during
Received in revised form
steady state evaporation of ammonia in a commercial plate heat exchanger has been
23 December 2013
carried out for an un-symmetric 30 /60 chevron plate configuration. Experiments were
Accepted 28 December 2013
conducted for saturation temperatures ranging from 25 C to 2 C. The heat flux was
Available online 8 January 2014
varied between 21 kW m2 and 44 kW m2. Experimental results show significant effect of saturation temperature, heat flux and exit vapor quality on heat transfer coefficient and
Keywords:
pressure drop. Current mixed plate configuration data are compared with previous studies
Plate heat exchanger
on the same heat exchanger with symmetric plate configurations. This comparison high-
Ammonia
lighted importance of optimization in selection of the heat exchangers. Correlations for
Heat transfer
two phase Nusselt number and friction factor for each chevron plate configuration
Evaporation
considered are developed. A Nusselt number correlation generalized for a range of chevron
Pressure drop
angles is also proposed. ª 2014 Elsevier Ltd and IIR. All rights reserved.
Transfert de chaleur d’e´vaporation et chute de pression d’ammoniac dans un e´changeur de chaleur a` plaques Mots cle´s : Echangeur de chaleur a` plaques ; Ammoniac ; Transfert de chaleur ; Evaporation ; Chute de pression
1.
Introduction
In today’s highly industrialized world, energy is one of the major concerns. With rapid consumption of fossil fuels, saving
energy has become an attractive topic for the researchers. Recently, a marked improvement in heat exchanger technology has been observed. With improved technology vivid energy savings are now possible with efficient natural refrigerants in compact heat exchangers, used in various applications.
* Corresponding author. Tel.: þ92 51 111878787; fax: þ92 51 4486705. E-mail addresses: [email protected], [email protected] (M.S. Khan). 0140-7007/$ e see front matter ª 2014 Elsevier Ltd and IIR. All rights reserved. http://dx.doi.org/10.1016/j.ijrefrig.2013.12.015
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Nomenclature
x
Bo Dh f Gr h i k Nu P Pcr P* Pr q00 Re t T U v
Greek symbols b chevron or corrugation angle (deg) D change or difference m dynamic viscosity (kg m1 s1)
boiling number hydraulic diameter (m) fanning friction factor refrigerant mass flux (kg m2 s1) heat transfer coefficient (W m2 K1) enthalpy (kJ kg1) thermal conductivity (W m1 K1) Nusselt number pressure (kPa) or Plate pitch (m) critical pressure (kPa) reduced pressure (P/Pcr) Prandtl number heat flux (kW m2) Reynolds number plate thickness (m) temperature ( C) overall heat transfer coefficient (W m2 K1) specific volume (m3 kg1)
Plate heat exchangers (PHEs) are gaining popularity in recent years. However, data for two phase applications are still scarcely available in open literature, Khan et al. (2009). The gasketed plate heat exchangers are compact in nature, have high area to volume ratio, flexible in design and are easy to operate. Perhaps one of the major advantages of plate heat exchanger is its small size and relatively low energy usage. The flexible design of a plate heat exchanger allows overall reduced operational and maintenance costs. Because of their compact nature, these heat exchangers are not only energy efficient and cost effective but also adaptable for a variety of industrial applications. Ammonia is a naturally available refrigerant and has long been used in the refrigeration industry. It is a relatively low cost refrigerant with attractive thermo-physical properties. Ammonia based systems require relatively low maintenance. On the other hand ammonia is toxic and when mixed in air is flammable under certain conditions. However, being selfalarming has low safety compromise. While chloro-fluorocarbons are threat to ozone depletion and hydro-fluorocarbons add to global warming, natural refrigerants like ammonia are not only environment friendly but have efficient thermo-hydraulic characteristics. Therefore, due to danger of ozone depletion and global warming, interest in the use of natural refrigerants in air-conditioning and refrigeration industry has increased Khan et al. (2012a,b). The PHE used in the present study was configured in a single pass U-arrangement with counter flow setup. Such an arrangement is schematically shown in Fig. 1. It is well known that two phase heat transfer in heat exchangers is a more efficient mode of heat transfer compared to single phase heat transfer. Although some two phase work on compact heat exchangers was reported in the last century, however, major two phase experimentation on plate heat exchangers has been reported in the last decade. It may be interesting to note that two phase flow investigations carried out on plate heat exchangers with natural refrigerants such as
93
exit vapor quality
Subscripts acc acceleration core core ele elevation eq equivalent f liquid g vapor h hot stream m measured port port r refrigerant sp single phase tp two phase
ammonia are comparatively less in number, Khan et al. (2009). Recently Khan et al. (2012a,b) carried out experimental investigations on evaporation of ammonia in a commercial plate heat exchanger with two different symmetric (60 /60 and 30 /30 ) chevron plate configurations. The heat transfer coefficient and pressure drop are reported to increase with an increase in chevron angle and decrease in corrugation depth and mean channel spacing. Some other recent two phase studies using ammonia as refrigerant (Ayub, 2003; Sterner and Sunden, 2006; Djordjevic and Kabelac, 2008; Arima et al., 2010) are also discussed in Khan et al. (2012a,b). The current study uses the same heat exchanger as was employed in Khan et al. (2012a,b) but with an unsymmetric (mixed) chevron plate configuration. The mixed configuration is achieved by combining 30 and 60 chevron angle plates. The present work focuses on experimental investigation of thermal hydraulic characteristics of a 30 /60 un-symmetric commercial chevron plate heat exchanger for two phase ammonia evaporation. The experimental results for the un-symmetric configuration highlights the importance
Stream 2
Stream 1 Fig. 1 e Schematic diagram of single pass U-arrangement for counter flow plate heat exchanger setup (Kakac and Liu, 2002).
94
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Table 1 e Primary experimental parameters and their ranges. Parameter Saturation temperature, Tsat Exit vapor quality, x Equivalent Reynolds number, Reeq Heat flux, q00
Range 25 C < Tsat < 2 C 0.5 < x < 0.9 1225 < Reeq < 3000 21 kW m2 < q00 < 44 kW m2
of optimization in achieving the desired heat transfer with an economical pressure drop penalty. The experimental parameters considered in the present study and their ranges are tabulated in Table 1. Experimentation was carried out at five levels of the saturation temperature between 25 C and 2 C. Experiments were performed in such a way that effects of heat flux, exit vapor quality and saturation temperature could be studied for a selected value of mass flux. The refrigerant properties were evaluated at the saturation temperature. The single phase heat transfer coefficient on hot fluid side was estimated by a single phase heat transfer correlation developed by Khan et al. (2010) on the same plate heat exchanger. The two phase heat transfer coefficient on the refrigerant side was estimated from the overall heat transfer coefficient.
2.
Experimental setup
The experimental setup is schematically shown in Fig. 2. It consists of a refrigerant loop, a plate heat exchanger test loop, water/glycol solution loop, instrumentation and a data
acquisition system. A cooling tower served the cooling requirements of the condenser and the compressor. Ammonia was circulated in the refrigeration and test loop. Primary measurements in the experiments were flow rates, fluid temperatures, system pressure and differential pressure drop for both hot side and refrigerant side, across the plate heat exchanger. A commercially available gasketed plate heat exchanger (TL-90) has been used in this study. Chevron plates made of stainless steel (SS-316) were used. Table 2 provides important geometric details of the plate heat exchanger while the parameters are defined in Fig. 3. Further details of the experimental setup and procedure could be found in Khan et al. (2012a,b).
3.
Data reduction
3.1.
Heat transfer coefficient
In order to determine the heat transfer characteristics of the plate heat exchanger during evaporation of liquid ammonia, the overall heat transfer coefficient, wall thermal resistance and heat transfer coefficient on the hot fluid side must be known. The average heat transfer coefficient on the refrigerant side (htp) was determined from the overall heat transfer coefficient: 1 1 t 1 ¼ htp U k hsp
Fig. 2 e Schematic diagram of the experimental facility.
(1)
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
pressure drops due to acceleration of the refrigerant, due to change in elevation, due to inlet/exit manifolds, and due to change of pressure in the main core of the corrugated plate heat exchanger. The core pressure drop may therefore be obtained as (Kakac and Liu, 2002):
Table 2 e Geometric characteristics of chevron plates used in experiments. Geometric characteristic Plate width, Lw (mm) Vertical distance between centers of ports, Lv (mm) Port diameter, Dp (mm) Horizontal distance between centers of ports, Lh (mm) Mean channel spacing, b (mm) Plate thickness, t (mm) Effective area of plate, A (m2) Corrugation pitch, l (mm) Surface enlargement factor, 4 Plate thermal conductivity, k (W m1 K1) Chevron angle a
95
185 565
DPcore ¼ DPm DPacc DPelev DPport 43 125 2.9 0.6 0.095 13.25 and 6.25a 1.117 14 45
(3)
Using this pressure drop the two-phase friction factor for evaporation of liquid ammonia, was obtained by the following relation (Yan and Lin, 1999): ftp ¼
DPcore Dh 2G2r nm Lp
(4)
Pressure drop data reduction procedure has been elaborated in Khan et al. (2012a,b).
For b ¼ 30 and 60 plates, respectively.
4. The following single phase correlation developed by Khan et al. (2010)was used to estimate hsp: hsp ¼ 0:144Re0:78 Pr0:35
kf Dh
m
0:14
mwall
(2)
Further details are mentioned in Khan et al. (2012a,b).
3.2.
For the measurements and results derived from the experiments, the experimental uncertainties were calculated according to the procedure outlined by Moffat (1988). The method for single sample experiments involves estimation of overall uncertainty, wR in the calculated result, R, using the following relationship: "
Pressure drop wR ¼
The measured pressure drop across the plate heat exchanger for upward flow of refrigerant consists of four components;
Experimental uncertainty
n X vR i¼1
vXi
2 #1=2 wi
where R is expressed as: R ¼ (X1, X2, X3, ........Xi, Xn)
Fig. 3 e Basic geometric characteristics of the chevron plate (Kakac and Liu, 2002).
(5)
96
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Table 3 e Experimental parameters and their estimated uncertainties. Parameter Pressure drop, DP (kPa) Temperature, T ( C) Hot fluid mass flux, Gh (kg m2 s1) Refrigerant mass flux, Gr (kg m2 s1) Heat flux, q00 (kW m2) Uncertainty in calculated water/glycol side heat transfer coefficient in single phase study (Khan et al., 2010) Deviation of single phase correlation from single phase heat transfer data (Khan et al., 2009) Refrigerant heat transfer coefficient, htp (W m2 K1) Two phase friction factor, ftp
Uncertainty 0.22% 0.1 C 0.85% < 0.1% <1% 8%
1.8%
10.9% 2.8%
Here Xi are the independent parameters and wi represents the associated uncertainty in the independent parameter. Based on the precision of primary measurements, instrument specifications and fluid property variations, the uncertainties for each set of measurements are calculated in the present experimental study. Finally, the overall uncertainty in Nusselt number came out to be within 10.9%. After accounting for errors in the hot and cold fluid stream data measurements, the maximum errors in the primary measurements of mass flow rate, pressure and temperature were calculated. The uncertainties are tabulated in Table 3.
5.
Results and discussions: heat transfer
5.1. Effects of saturated temperature, heat flux and exit vapor quality Figs. 4 and 5 show experimental heat transfer coefficient versus heat flux and exit vapor quality at various temperatures. For a fixed mass flux, two phase ammonia evaporation data are obtained at five different saturation temperatures (saturation
Fig. 4 e Heat transfer coefficient versus heat flux for several saturation temperatures at a mass flux of 6.5 kg mL2 sL1.
pressures) for the un-symmetric chevron plate configuration considered in this study. Data are shown for saturations temperatures of 2 C, 9 C, 14.5 C, 19.5 C and 25 C. Figs. 4 and 5 show that although the effect of heat flux and exit vapor quality is not substantial, the heat transfer coefficient increases with an increase in the saturation temperature. It may also be observed from Figs. 4 and 5 that heat transfer coefficient follows similar trend with heat flux and exit vapor quality. Several parameters, such as transport properties could be responsible for increase in heat transfer coefficient, however, most importantly the higher temperatures result in smaller size bubbles escaping at higher frequency so the heat transfer coefficient increases with an increase in temperature. This is in agreement with work reported by Chyu et al. (2001). At higher temperatures they reported bubbles to grow, combine and slide over the tube surface with fluid flow, hence enhanced the heat transfer. It may also be observed from Figs. 4 and 5 that heat transfer coefficient increases moderately with an increase in heat flux up to 34 kW m2 and for exit vapor quality between 0.45 and 0.67 for the considered plate configuration. The htp decreases gradually with further increase in heat flux. Recently, Arima et al. (2010) have also shown that for a given saturation pressure and mass flux, the heat transfer coefficient decreases with average heat flux. Generally, before reaching the critical heat flux (dry-out regime), the boiling heat transfer is governed by convection boiling and/or nucleate boiling phenomena. Danilova et al. (1981) claimed that nucleate boiling dominates during evaporation in a PHE and Thonon et al. (1995) concluded that only convective boiling should be considered in PHEs. Sterner and Sunden (2006) reported that the relevant boiling mechanism changes from convective to nucleate boiling at certain specific geometric dimension of the channel. Djordjevic and Kabelac (2008) report that both nucleate and convective boiling mechanisms are present during evaporation of refrigerant in a PHE. Although, not shown here, it was found in preliminary experiments that for 0.1 < x < 0.4 the heat transfer coefficient increased with an
Fig. 5 e Heat transfer coefficient versus exit vapor quality for several saturation temperatures at a mass flux of 6.5 kg mL2 sL1.
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
increase in the heat flux. Hence nucleate boiling phenomenon was dominant in the low vapor quality region. No such major change in heat transfer coefficient is found with increasing heat flux in the current range of exit vapor quality, it can therefore be deduced that for higher exit vapor quality (0.5 > x > 0.8) the evaporation is not governed by nucleate boiling phenomenon. Though, this inference leads to convective boiling dominance, however, further data are required at various mass flux rates to confirm whether the evaporative boiling is dominantly convective in nature or not. The behavior of the heat transfer coefficient shown in Figs. 4 and 5 is in fair agreement with Ayub (2003) and Djordjevic and Kabelac (2008). Ayub (2003) proposed a two phase ammonia evaporation correlation in PHEs. The reported correlation includes effect of saturation pressure. He reported that the heat transfer coefficient increased with an increase in saturation pressure. Ayub (2003) maintains that unlike shell and tube heat exchangers, where bulk of the heat transfer is governed by pool boiling, due to narrow passages in the plate evaporators, the dominant mode of heat transfer could be convective in nature. A dry-out phenomenon has also been observed beyond exit vapor quality of 0.8 in the current study. This is in agreement with Djordjevic and Kabelac (2008) and Arima et al. (2010) who reported presence of dry-out regime at higher heat flux and vapor quality in their studies.
5.2.
Comparison with other studies
Khan et al. (2009) have reported that most of the two phase studies carried out on PHEs during the last decade are based on HFCs and a very few studies are conducted with ammonia as refrigerant (Ayub, 2003; Sterner and Sunden, 2006; Djordjevic and Kabelac, 2008; Arima et al., 2010). Moreover, no data were found on evaporation of ammonia for a mixed plate configuration plate heat exchangers in the literature. As mentioned earlier, Ayub (2003) presented industrial data based correlation for a wide range of chevron angle plate configurations at 8.5 kg m2 s1. Though bulk of the data for present unsymmetrical plate configuration has been taken at a fixed mass flux of 6.5 kg m2 s1, however, for the comparison purpose, some experiments are also carried out at 8.5 kg m2 s1 for a saturation temperature of 2 C. Comparison of present heat transfer data with correlation reported by Ayub (2003) is shown in Fig. 6. Present data are compared only with Ayub (2003) since no other two phase data for ammonia evaporation in the 30 /60 plate configuration has been reported. As could be observed from Fig. 6, Ayub’s (2003) correlation qualitatively follows a pattern similar to that of the current experimental heat transfer data. It is, however, interesting to note here that unlike previous studies on the same PHE with symmetric configurations (Khan et al., 2012a,b), the Ayub’s correlation (2003) under predicts experimental data gathered for the un-symmetric configuration. This difference between previously reported investigations and present data may be attributed to chevron plate geometric differences (surface corrugations). For example, a major difference between present investigation and previous studies on the same PHE (Khan et al., 2012a,b) is that previously plates used in a particular symmetric configuration were similar in all respects. While in the present study, in addition to the chevron angle, the
97
Fig. 6 e Comparison of heat transfer versus exit vapor quality between present study and Ayub (2003) at L2 C. respective corrugation pitch and corrugation depth of both commercial plates were different for the un-symmetric plate configuration considered.
6.
Results and discussion: pressure drop
While the thermal analysis of a heat exchanger determines heat transfer characteristics of the exchanger, frictional loss effects are equally important as the pressure drop offered by fluid properties and the heat exchanger geometry provides a measure of pumping power necessary to maintain the desired flow rate. This pumping power requirement adds to the initial and operating costs of the heat exchangers. Thus design and selection of a heat exchanger unit is influenced by both heat transfer characteristics and pressure drop offered. Two phase differential pressure drop data are collected during steady state evaporation of liquid ammonia. For a fixed refrigerant mass flux of 6.5 kg m2 s1, the pressure drop is measured by a differential pressure measurement device installed between inlet and exit ports of the plate heat exchanger. Pressure drop data are collected for the same five saturations temperatures as were in heat transfer experimentation. Unfortunately, no pressure drop data were found for evaporation of ammonia in plate heat exchangers in the published literature. Therefore, only a qualitative comparison is possible with pressure drop data for boiling of HFCs in compact heat exchangers.
6.1. Effect of saturation temperature and exit vapor quality For the given plate configuration, pressure drop is plotted against exit vapor quality for various saturation temperatures in Fig. 7. It may be observed that the pressure drop increased with an increase in saturation temperature for the considered plate configuration in the present study. It may also be observed from Fig. 7 that pressure drop is lesser for low saturation temperatures.
98
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
number, Boeq, and P* which is the reduced pressure defined as P/Pcr. The equivalent Reynolds number and Boiling number are given as:
Reeq ¼
Greq Dh m
(7)
Boeq ¼
q00 Greq ifg
(8)
where "
Greq
Fig. 7 e Experimental pressure drop as a function of exit vapor quality at 6.5 kg mL2 sL1.
It may also be observed that the pressure drop either increased, or decreased with an increase in exit vapor quality, x, for all saturation temperatures considered in the present investigation. It may be noted that the change in pressure drop remained almost insignificant up to x ¼ 0.7, for higher saturation temperatures. However, for higher x values (greater than 0.7), the pressure drop is observed to decrease with exit vapor quality. Pressure drop is observed to decrease more sharply at exit vapor quality greater than 0.7, in general.
7.
Two phase Nusselt number correlation
Several heat transfer applications require two phase flow which results in a more compact plate heat exchanger. However, benefits of better heat transfer characteristics of plate heat exchangers can be utilized effectively in the multi facet industry only when dependable correlations for the estimation of Nusselt number and pressure drop are available. It is known that very few studies on phase change heat transfer and two phase flow have been conducted on plate heat exchangers. Very few data are available for evaporation of ammonia in vertical plate heat exchangers. This lack of data requires an urgent need for correlations that can be used in designing of plate heat exchangers or to estimate their performance. An attempt has been made to develop correlations for heat transfer and pressure drop based on the experimental results of the present study. Applicability of these correlations is verified by comparing them with the experimental data. In the present study, the following type of correlation has been selected to estimate the experimental Nusselt number: m Nutp ¼ C Reeq Boeq ðP Þj
r ¼ Gr 1 x þ x f rg
!#1=2 (9)
Reeq is a function of plate geometry and equivalent mass flux, which in turn, is a function of vapor quality and liquid to vapor density ratio. The mass flux also depends on channel spacing, b. Similarly, Boeq depends on heat flux and latent heat of vaporization. Since the above mentioned correlation pattern includes all important experimental parameters measured and considered in this study, therefore, this correlation pattern has been found appropriate for flow boiling heat transfer of ammonia in the plate heat exchanger. Values of the coefficient, C, and exponents, m and j are determined using the classical regression analysis. The Nusselt number correlation hence developed for the plate configuration considered in this study, based on experimental data is given below: 0:079 0:33 ðP Þ Nutp ¼ 129:26 Reeq Boeq
(10)
Comparison between the experimental Nusselt number data and the Nutp estimated from correlation (Eq. (10)) for the chevron plate configuration considered in the present study is presented in Fig. 8. The maximum deviation of this correlation from experimental data is within 10%.
(6)
The two phase Nusselt number, Nutp, is a function of equivalent Reynolds number, Reeq, equivalent Boiling
Fig. 8 e Comparison of two phase experimental Nutp with correlation estimating the Nutp.
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
It can be observed that the developed correlation represents the experimental data quite well. The error analysis shows that for the considered plate configuration about 74% of the total experimental data fall within 4% error band while only 12% of the data fall between 7% to 10% error band.
7.1. A generalized correlation for the two phase Nusselt number Khan et al. (2012a,b) have presented correlations to estimate Nusselt number for the same commercial PHE that is used in un-symmetric plate configuration in the present study. The correlations proposed for two symmetric plate configurations earlier are as follows: (Khan et al., 2012a,b) 0:085 0:21 ðP Þ ; Nutp ¼ 82:5 Reeq Boeq 0:04 0:52 Nutp ¼ 169 Reeq Boeq ðP Þ ;
60 =60 30 =30
(11) (12)
Comparison of these correlations (Eqs. (11)e(12)) with Eq. (10) shows that the Nutp dependence on saturation pressure of ammonia increases with a decrease in chevron angle. Also the combined effect of equivalent Reynolds and Boiling numbers reduced with decreasing chevron angle. The lead coefficient, exponents of equivalent Reynolds and Boiling numbers product and reduced pressure in the developed correlations (Eqs. (10)e(12)) being different for each corrugated plate configuration considered, may be further related to chevron angle. Therefore, Nusselt number could be represented by the following more general functional form which also includes effect of chevron angle: mfb g jfb g Nutp ¼ Cfb g Reeq Boeq ðP Þ
(13)
Where b* is given as b/bmax and curly brackets represent functional form. bmax is taken equal to 60 which is the maximum chevron angle considered in the present and previous Khan et al. (2012a,b) investigations on the same PHE. In order to determine C{b*}, m{b*} and j{b*}, the C, m and j are individually correlated with chevron angle over the whole range of 30 e60 . It should be noted that the chevron angle for 30 /60 plate configuration is estimated to be equal to 45 . Linear curve fitting of the data provides the following equations for C, m and j as functions of b*. Cfb g ¼ 173:52b þ 257:2
(14)
mfb g ¼ 0:09b þ 0:0005
(15)
jfb g ¼ 0:624b 0:822
(16)
well and is valid for the whole range of chevron plate angles from 30 to 60 .
8.
Two phase friction factor correlation
Fig. 9 shows effect of Reeq on the two phase Fanning friction factor, ftp. It may be observed from Fig. 9 that the substantial reduction of ftp in the low equivalent Reynolds number range becomes gradual with further increase in Reeq. It may also be observed that the friction factor increases with an increase in the saturation temperature. In Fig. 9 although, the general trends for various saturation temperatures are quite similar, however, close examination shows that the overall difference between maximum and minimum values of ftp is large at higher saturation temperature (2 C) compared to low saturation temperature (25 C) over the range of Reeq flux considered in this study. Nevertheless, the friction factor increases almost linearly with increasing saturation temperature, in general. Since the effects of equivalent Reynolds number and saturation temperature on pressure drop are apparent, therefore, the following correlation has been considered to estimate the two phase experimental friction factor for ammonia evaporation in the vertical plate heat exchanger. m ftp ¼ C Reeq ðP Þj
(18)
As mentioned earlier, Reeq is a function of plate geometry, exit vapor quality and refrigerant properties, such as viscosity and liquid to vapor density ratio. The coefficient, C, and exponents, m and j are independent of fluid nature and are determined using the regression analysis approach. Using calculated values of these coefficients and exponents, the experimental friction factor could then be represented by the following correlation: 1:26 0:9 ðP Þ ftp ¼ 305; 590 Reeq
(19)
Comparison between experimental friction factor data and ftp estimated from correlation (Eq. (19)) is shown in Fig. 10. It
The above equations for C{b*}, m{b*} and j{b*} could be plugged in functional form of Nutp correlations given by Eq. (13). The Nutp correlation is then given as: mð0:09b þ0:0005Þ ð0:624b 0:822Þ Nutp ¼ ð173:52b þ257:12Þ Reeq Boeq ðP Þ (17) The above correlation shows similar trends for all chevron angle plate configurations. It represents experimental data
99
Fig. 9 e Variation of friction factor with equivalent Reynolds number for a mass flux of 6.5 kg mL2 sL1.
100
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Fig. 10 e Comparison of experimental friction factor with correlation (Eq. (19)).
can be observed that the friction factor correlation (Eq. (19)) represents the experimental data quite well. Error analysis shows that more than 88% of the total data fall within the error band of 6%, while only about 12% data fall between 7% and 10%. This comparison confirms that the experimental pressure drop data are well estimated by the developed correlation for the entire range of experimental parameters.
9. Comparison of heat transfer and pressure drop results for symmetric and un-symmetric plate configurations The effect of chevron angle on heat transfer and pressure drop for both symmetric and un-symmetric plate configurations is shown in Figs. 11 and 12.
Fig. 11 e Effect of chevron angle on heat transfer coefficient for various plate configurations at L2 C and L25 C.
For better readability, experimental data of only extreme saturation temperatures (2 C and 25 C) considered in the current study have been plotted. The heat transfer enhancement with chevron angle is clearly evident from Fig. 11. These results are also in agreement with almost all previously conducted studies on PHEs with different chevron angle plate configurations except Djordjevic and Kabelac (2008). They used two plate configurations in their investigations; soft plate configuration (27 /27 ) and high pressure drop (63 /63 ) configuration. They have reported that the soft plate configuration provides higher heat transfer coefficient compared to their high pressure drop configuration. This is contrary to findings of all previous investigations referenced in this work. The enhancement in heat transfer is generally attributed to the plate surface corrugations that generate swirl flows and periodic disruptions. Additionally, the smaller corrugation pitch and channel spacing also help enhance the heat transfer. Therefore, chevron plate corrugations encourage early turbulence which in turn helps increase the heat transfer. Although, overall trends of the heat transfer data shown in Fig. 11 for all three plate configurations are similar to a larger extent, however, some variability exists with respect to effect of chevron angle. Fig. 11 shows that the dry-out effect weakens from chevron angle plate configuration (60 /60 ) to chevron angle plate configuration (30 /30 ). For instance, it is observed that for the two phase experiments on 30 /30 chevron plate configuration, the dry-out effect is less significant compared to the other two configurations. Fig. 12 shows effect of chevron angle on pressure drop for the three configurations. Pressure drop is observed to be the maximum for 60 /60 plate configuration and minimum for 30 /30 plate configuration. This was expected due to geometric difference between three chevron plate configurations. It may also be noted from Fig. 12 that the effect of exit vapor quality on pressure drop is either insignificant or very small up to about 0.7 for all configurations. The two types of plates used in present experimentation are distinct from one another, mainly, in three aspects; the chevron angle, the corrugation pitch, l and the mean channel spacing, b. The 60
Fig. 12 e Effect of chevron angle on pressure drop for various plate configurations at L2 C and L25 C.
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
plate offers higher pressure drop due to its higher chevron angle, its smaller mean channel spacing (b ¼ 2.2 mm) and corrugation pitch (l ¼ 6.25 mm) compared to b ¼ 3.6 mm and l ¼ 13.5 mm for 30 plate. In the current un-symmetric configuration, the corrugation pitch and mean channel spacing are taken as average of the two plates. Fig. 12 shows that for 30 /60 and 30 /30 plate configurations; the pressure drop reduces with an increase in exit vapor quality for all saturation temperatures considered. On the contrary, the pressure drop increases with an increase in exit vapor quality for 60 /60 plate configuration. It may also be noted that pressure drop decreases with an increase in saturation temperature for 60 /60 plate configuration only. However, the pressure drop trends remained similar regardless of change in saturation temperature. In comparison to symmetric 60 /60 plate configuration the 30 plate geometry seems to govern the similar pressure drop trends for soft and mixed configurations. Although, the pressure drop for unsymmetric plate configuration (30 /60 ) is higher than that of 30 /30 plate configuration, it is interesting to note that the difference in pressure drop offered by mixed configuration and 30 /30 configuration is small compared to hard plate configuration. It is because of the fact that at low refrigerant mass flux, the resistance offered by 60 plate used in mixed configuration with 30 plate is compensated by 30 plate that has larger mean channel spacing and corrugation pitch. Since the current two phase experiments are carried out at rather low flow rates, the channel geometry and refrigerant properties appear to play major role in determining the pressure drop across the channel. These trends demonstrate strong influence of mean channel spacing, corrugation pitch and corrugation depth on pressure drop. Referring to the heat transfer results shown in Fig. 11, it may be observed that increase in heat transfer coefficient was more prominent than that in the pressure drop with chevron angle. For instance, in comparison to 30 /30 plate configuration at 25 C, on average, the heat transfer coefficient increases by about 2 times for 30 /60 plate configuration and by about 4 times for 60 /60 plate configuration. In comparison, the pressure drop for the same saturation temperature for 30 / 60 plate configuration is almost same as that of 30 /30 plate configuration but pressure increases about 3 times for 60 /60 plate configuration compared to 30 /30 plate configuration (Fig. 12). For the highest saturation temperature (2 C) considered, the average increase in heat transfer coefficient with chevron angle is about 2 and 2.5 times that of the 30 /30 plate configuration for 30 /60 and 60 /60 plate configurations respectively. Pressure loss penalty for 30 /60 plate configuration remained almost the same, however, decreased for 60 / 60 plate configuration by more than 1.5 times that of 30 /30 plate configuration. This comparison of results demonstrates a strong need for optimization of heat exchanger design and selection for their economic utilization in the industry.
10.
Conclusions
Steady state evaporation heat transfer and pressure drop of ammonia in a commercial plate heat exchanger with unsymmetric 30 /60 plate configuration has been investigated.
101
Experiments were conducted for equivalent Reynolds number range of 1225 to 3000, while the saturated temperature of ammonia varied from 25 C to 2 C. Results are reported for exit vapor quality of 0.5e0.9. Experimental results show a strong influence of saturation temperature and other operational conditions on heat transfer and pressure drop in the PHE. For the exit vapor quality greater than 0.5, convective boiling is found to dominate the heat transfer with heat flux and vapor quality having cumulative effect on heat transfer coefficient. Heat transfer coefficient and pressure drop are found to increase with an increase in saturation temperature. The effect of exit vapor quality on the two phase pressure drop is not significant except at higher values. The two phase Fanning friction factor is found to decrease with equivalent Reynolds number and increased with an increase in saturation temperature. Uncertainty analysis has shown a maximum error of about 10.9% in the experimental Nusselt number data and 2.8% in experimental friction factor data. Based on the experimental data, an empirical correlation for two phase Nusselt number is developed as a function of reduced pressure, equivalent Reynolds number and Boiling number for the 30 /60 plate configuration. Bulk of the heat transfer data are predicted within 7% of the error band. The two phase friction factor correlation is also proposed as a function of equivalent Reynolds number and reduced pressure for the above mentioned chevron plate configuration. This correlation represents experimental data within an error band of 10%. Heat transfer and pressure drop data of the present study are compared with some previous ammonia evaporation studies on the same PHE with symmetric plate configurations. This comparison necessitates a strong need for optimization in design and selection of plate heat exchangers to minimize energy requirements. Based on this comparison, a Nusselt number correlation generalized for all chevron angle plate configurations employed on this PHE is also proposed.
Acknowledgments The authors gratefully acknowledge the financial support provided by American Society of Heating Refrigerating and Air-Conditioning Engineers (ASHRAE) for this work under 1352-RP. Also thanks are due to Isotherm, Inc. of USA and Thermowave of Germany for donating the plate heat exchanger and spare heat exchanger plates with gaskets.
references
Arima, H., Kim, H.J., Okamoto, A., Ikegami, Y., 2010. Local heat transfer characteristics of ammonia in a vertical plate evaporator. Int. J. Refrig. 33, 359e370. Ayub, Z.H., 2003. Plate heat exchanger literature survey and new heat transfer and pressure drop correlations for refrigerant evaporators. Heat. Trans. Eng. 24 (5), 3e16. Chyu, M.-C., Zheng, J., Jin, G., 2001. Evaporation of Ammonia Outside Smooth and Enhanced Tubes with Miscible Oil. ASHRAE Report, RP-977. Djordjevic, E., Kabelac, S., 2008. Flow boiling of R134a and ammonia in a plate heat exchanger. Int. J. Heat. Mass Trans. 51, 6235e6242.
102
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n 4 1 ( 2 0 1 4 ) 9 2 e1 0 2
Danilova, D.N., et al., 1981. Heat transfer in different plate geometries. Kholod Tek. 4, 25e31. Kakac, S., Liu, S., 2002. Heat Exchangers Selection, Rating, and Thermal Design. CRC. Khan, T.S., Khan, M.S., Chyu, M.-C., Ayub, Z.H., 2012a. Experimental investigation of evaporation heat transfer and pressure drop of ammonia in a 60 chevron plate heat exchanger. Int. J. Refrig. 35 (2), 336e348. Khan, M.S., Khan, T.S., Chyu, M.-C., Ayub, Z.H., 2012b. Experimental investigation of evaporation heat transfer and pressure drop of ammonia in a 30 chevron plate heat exchanger. Int. J. Refrig. 35 (6), 1757e1765. Khan, T.S., Khan, M.S., Chyu, M.-C., Ayub, Z.H., Javed, A., Chattha, 2009. Review of heat transfer and pressure drop correlations for evaporation of fluid flow in plate heat exchangers (RP1352). HVAC&R Res. 15 (2), 169e188.
Khan, T.S., Khan, M.S., Chyu, M.-C., Ayub, Z.H., 2010. Experimental investigation of single phase convective heat transfer coefficient in a corrugated plate heat exchanger for multiple plate configurations. Appl. Therm. Eng. 30 (8e9), 1058e1065. Moffat, R.J., 1988. Describing the uncertainties in experimental results. Exp. Therm. Fluid Sci. 1, 3e17. Sterner, D., Sunden, B., 2006. Performance of plate heat exchangers for evaporation of ammonia. J. Heat. Trans. Eng. 27 (5), 45e55. Thonon, B., Vidil, R., Marvillet, C., 1995. Recent research and developments in plate heat exchangers. J. Enhanc. Heat. Transf. 2 (1e2), 149e155. Yan, Y.Y., Lin, T.F., 1999. Evaporation heat transfer and pressure drop of refrigerant R-134a in a plate heat exchanger. J. Heat. Trans. 12 (1), 118e127 (Transaction of the ASME).