CORRELATING EVAPORATION HEAT TRANSFER COEFFICIENT OF REFRIGERANT R134a IN A PLATE HEAT EXCHANGER
Vincent D. Donowski and Satish G. Kandlikar
Mechanical Engineering Department Rochester Institute of Technology Rochester, NY
ABSTRACT
The present study deals with developing a predictive methodology for evaporation heat transfer coefficient of refrigerant R134a flowing in a plate heat exchanger. Correlation schemes proposed by Yan and Lin (1999b) for modeling the heat transfer coefficient in both a single phase watertowater experiment and a twophase evaporation experiment with refrigerant R 134a in a plate heat exchanger are found to result in large discrepancies with their own experimental data. In the present study Yan and Lin’s plate heat exchanger data is used to develop improved correlations for singlephase heat transfer coefficient of water as well as two phase evaporation heat transfer coefficients of refrigerant R134a. Two schemes are presented for correlating the twophase data, and they are compared with the correlating schemes available in literature (Yan and Lin, 1999b; and Kandlikar, 1991). It is found that the new correlation schemes developed in this work provide an accurate fit to the experimental data, and a final correlation sche me is proposed for future use in modeling the evaporation heat transfer coefficient of refrigerant R134a in a plate heat exchanger configuration.
1
NOMENCLATURE
Bo 
Boiling number, eq. (18) 
Co 
Convective number, eq. (17) 
c _{p} 
specific heat, J/kgK 
D _{h} 
hydraulic diameter, m 
D _{i} 
tube inside diameter, m 
f 
friction factor 
F fl 
fluiddependent parameter 
G 
mass flux, kg/m ^{2} s 
h _{L}_{G} 
enthalpy of vaporization, J/kg 
Nu 
Nusselt number, dimensionless, eq. (5) 
Pr 
Prandtl number, dimensionless 
q ” _{w} 
imposed wall heat flux, W/m ^{2} 
Re 
Reynolds number, dimensionless 
x 
vapor quality 
Greek Symbols
a 
heat transfer coefficient, W/m ^{2} K 
h 
viscosity, N s/m ^{2} 
l 
thermal conductivity, W/m K 
r 
density, kg/m ^{3} 
Subscripts 

eq 
equivalent all liquid number 
G 
vapor phase 
L 
liquid phase 
m 
average value for the two phase mixture or between the inlet and exit 
r 
refrigerant, twophase 
TP 
twophase 
w 
water 
wall 
wall/fluid near the wall 
2
1.
INTRODUCTION
Plate heat exchangers have been a popular choice for use in chemical and food processing applications since the 1930s (Hewitt et al., 1994). Within a compact design, plate heat exchangers offer low fabrication costs, ease of cleaning, adaptability, and high thermal efficiency. Introduction of plate heat exchangers as evaporators and condensers in refrigeration and air conditioning applications has gained most of its interest in the last 20 years (Yan and Lin, 1999b). To date, most studies about plate heat exchanger applications in published literature focus on the singlephase liquidtoliquid heat transfer. As a result, limited data resources are available for plate heat exchangers used as evaporators in twophase flow boiling applications. With public awareness of environmental safety concerns in recent years, degradation of the ozone layer in the atmosphere through the use of CFC refrigerants has been of great concern. In response to these issues, new HFC refrigerants such as R134a, R143a, and R125 have been developed as substitute chemicals for use in refrigeration and air conditioning systems. An excellent example of this conversion to new refrige rants is the introduction of refrigerant R134a into automobile air conditioning systems, in place of the traditional CFC refrigerant system. In order to improve the design of heat exchange systems utilizing these new refrigerants, it is necessary to first gain an understanding of the fluid heat transfer characteristics through experimentation and data collection. For these new refrigerants, the majority of published works have focused on singlephase liquidto liquid heat transfer. One recent study (Yan and Lin, 1999b), involves twophase evaporation heat transfer of the new refrigerant R134a in a plate heat exchanger. The present work analyzes the experimental data of Yan and Lin (1999b) and proposes two new correlation schemes for the evaporation heat transfer coefficient of refrigerant R134a in a plate heat exchanger. These new correlation schemes are compared with the original experimental data and other correlation schemes available in literature.
2. REVIEW OF LITERATURE
In this section previously published studies by other investigators are briefly reviewed to provide relevant information to the present work.
3
2.1
Early Correlation Schemes
Many early literature studies focus on the singlephase heat transfer of fluids flowing in a smooth tube to develop correlation schemes to model the heat transfer coefficients in other geometries. The most widely used singlephase correlation in a pipe is the wellknown Dittus Boelter equation (Dittus and Boelter, 1930) and is given by
Nu = 0.023 Re
0.8
Pr
0.4
(1)
where Re represents the fluid Reynolds number, Pr is the Prandtl number, and Nu is the Nusselt number. Other noteworthy equations are the SiederTate equation (Incropera and Dewitt, 1981),
Nu = 0.023 Re ^{0}^{.}^{8} Pr ^{1}^{/}^{3} (h _{m} /h _{w}_{a}_{l}_{l} ) ^{0}^{.}^{1}^{4}
and the Gnielinski correlation (Gnielinski, 1976),
Nu
=
( 
f 
/ 2 )( Re 
 ) 1000 Pr 

1 
+ 
12.7 
f / 2 
( Pr 
2/3 
 
1 
) 
[
1
+
(
D
i
/
L
)
2 / 3
](
Pr/ Pr
w
)
0.11
(2) 

. 
(3) 
The concept of modeling the twophase heat transfer coefficient can be traced back to the 1960s. Studies of twophase phenomenon can likely be traced back earlier; however, Bergles and Rohsenow (1964) made a significant contribution to this area by suggesting an additive model for convective and nucleate boiling components. Chen (1966) is also noted for suggesting an additive model for saturated boiling, using a nucleate boiling term based on a pool boiling equation by Foster and Zuber (1955), and a convective contribution utilizing the DittusBoelter (1930) equation noted earlier. Both of these models present an ideal that flow boiling can be separated into a nucleate boiling term as characterized by still fluid being boiled by heat applied directly at nucleation sights, and a convective boiling term to characterize heat transferred to the moving fluid. When these terms are brought together in the additive model, the dominant
4
contributing term is decided based on the liquid Reynolds number, a dimensionless number characterized by liquid mass flow and viscosity.
2.2 Experimentation With Smooth Tubes
Kattan et al. (1998a, 1998b, 1998c) conducted smooth tube experimentation for flow boiling of several refrigerants, including refrigerant R134a. Data from Kattan et al.’s work shows that the heat transfer coefficient will continue to increase with vapor quality, but will show a sudden drop at a qua lity of approximately 0.85. For very low mass flux the heat transfer coefficient shows little change as vapor quality increases. Yan and Lin (1998) show similar results to Kattan et al. (1998b) for the evaporation of refrigerant R134a in a small pipe. Little change is seen in the heat transfer coefficient as vapor quality increases for low values of heat and mass flux. Higher heat transfer values are seen for higher mass flux, or lower saturation temperature. For higher heat flux, higher values of heat transfer are seen; however, the heat transfer coefficient decreases as vapor quality increases. Gungor and Winterton (1986) perform an extensive review of a large flow boiling data base to propose a correlation for twophase flow boiling based on the two component model suggested by Chen (1966). The flow boiling correlation by Kandlikar (1990) was developed with over 10,000 data points and provides an accurate representation of parametric trends for qualities up to 0.70.8. It has been tested with R 134a data as well. Since the augmentation techniques employed in tubes, compact evaporators, and other enhanced geometries delay the transition to dryout beyond a quality of 0.8, the present work focuses on improving the correlation in this range.
2.3 Augmented Surfaces
Augmented surface tubes are designed to provide better heat transfer than smooth tubes through the use of enhanced surfaces. Some examples of enhanced surfaces include fins, coil springs, twisted tape, and helical ribs. These specialized surfaces serve to increase the heat transfer area in fluid flow, thereby adding significant enhancement to the liquid heat exchange capabilities, as well as improvements resulting from an increase in the local heat transfer coefficient.
5
For refrigerant R134a, Singh et al. (1996) provide an experimental study of the flow boiling heat transfer coefficient in a microfin tube. Results show that the heat transfer coefficient is higher for higher heat flux and for higher mass flux. For low mass flux little change is seen in the heat transfer coefficient as vapor quality increases. Kandlikar (1991) proposes a model for correlating flow boiling heat transfer in augmented tubes based upon previous work (Kandlikar, 1990) with flow boiling in smooth tubes. Kand likar uses the additive model suggested by Bergles and Rohsenow (1964) as a basis for developing twophase correlating schemes.
2.4 Compact Evaporators and Plate Heat Exchangers
Kandlikar (1991) suggested that heat exchangers, serving as compact evaporators in this particular study, can be treated similarly to augmented tubes for analysis. From previous work in correlating heat transfer coefficients in smooth tubes (Kandlikar, 1990), Kandlikar introduced augmentation factors to account for surface enhancements in augmented tubes and compact evaporators. Results from the Kandlikar (1991) study show the new model was able to predict the heat transfer coefficients within 1015 percent. For evaporation heat transfer of refrigerant R134a in a plate heat exchanger, Yan and Lin (1999b) provide the only detailed published data in a usable form by other researchers. Yan and Lin first experimented with liquidonly, watertowater heat transfer to analyze the uncertainty in experimentation, and to develop a singlephase correlation which can be adapted to the twophase situation. The twophase experiment is then employed for refrigerant R134a to investigate the effects of mean vapor quality, mass flux, heat flux, and pressure on the evaporation heat transfer coefficient. A correlation scheme is introduced by Yan and Lin (1999b) for modeling the evaporation heat transfer coefficient of refrigerant R134a in the plate heat exchanger.
2.5 Summary of Experimental Data and Related Correlation Work
Table 1 presents a brief summary of some of the more relevant studies discussed above. For the twophase heat transfer of refrigerant R134a, published studies have been focused
6
toward smooth and augmented surface tubes. The experimental ranges of the study by Yan and Lin (1999b) are also presented in this table.
2.6 Objectives of the Present Work
The objectives of the present work are as follows:
1. Develop a correlation scheme for predicting the evaporation heat transfer coefficient of refrigerant R134a in a plate heat exchanger based on experimental data from the literature and the additive model concept for convective and nucleate boiling contribution terms.
2. 
Compare the newly developed correlation scheme to other models proposed in the literature for similar situations. 
3. 
Propose the correlation scheme best suited for use in modeling the evaporation heat transfer coefficient of refrigerant R134a in a plate heat exchanger. 
3. 
DATA ANALYSIS 
Yan and Lin (1999b) provide experimental data for evaporation heat transfe r of refrigerant R134a in a plate heat exchanger, which serves as the data source for the present
work. Three plates of commercial SS316 stainless steel formed the plate heat exchanger. Each plate surface had a corrugated sine shape and 60 degree of chevron angle to the plate vertical axis. The plates measured 500.0 mm in length by 120.0 mm in width, with corners rounded at a
25.0 
mm radius. The thickness of the plates was 0.4 mm, with 3.3 mm pitch between plates and 
10.0 
mm pitch of the corrugation. Other dimensions show the connection ports to be spaced at 
70.0 
mm horizontally and 450.0 mm vertically, each port measuring 30.0 mm in diameter. When 
the three plates were pressed together, two counter flow channels were formed. Contrary V shape corrugations on adjoining plates formed a lattice grid which forced flow to move along the groove. This lattice formation in plate heat exchangers causes a notably high Reynolds number and turbulent flow, which enhance heat transfer capabilities. For further details, authors are referred to Yan and Lin (1999b).
4. RESULTS AND DISCUSSION
7
The singlephase data is first compared with the correlation developed by Yan and Lin (1999b). An improved singlephase correlation, Correlation I, is then developed and compared against the data. In the twophase region, the correlation developed by Yan and Lin is first compared with their own experimental data. The first correlation scheme, Correlation II, developed here is based on the model by Kandlikar (1991). In the second correlation scheme, Correlation III, this model is further modified to provide a better fit especially in the higher quality range.
4.1 Yan and Lin SinglePhase Correlation
Yan and Lin (1999b) proposed the following singlephase correlation us ing waterto water data from their plate heat exchanger tests:
Nu = 0.2121 Re ^{0}^{.}^{7}^{8} Pr ^{1}^{/}^{3} (h _{m} /h _{w}_{a}_{l}_{l} ) ^{0}^{.}^{1}^{4} 
. 
(4) 

Using the definition of the Nusselt number, Nu, as 

Nu = a D _{h} /l _{L} 
, 
(5) 
equation 4 can be rearranged to provide the following equation for the heat transfer coefficient.
a _{w} = 0.2121 Re ^{0}^{.}^{7}^{8} Pr ^{1}^{/}^{3} (h _{m} /h _{w}_{a}_{l}_{l} ) ^{0}^{.}^{1}^{4} l _{L} /D _{h}
.
(6)
It should be noted that the singlephase correlation proposed by Yan and Lin (1999b) shows striking similarity to the SiederTate equation (Incropera and Dewitt, 1981) provided in eq. 2, which is known to provide reliable results in modeling singlephase liquid heat transfer in smooth tubes. Here the singlephase model is applied to the plate heat exchanger, with an increased multiplier to account for surface enhancement through the use of corrugated plates. Two assumptions must be made in plotting the proposed correlation. Firstly, it is assumed that
8
the bulk fluid temperatures are 95 C for the hot side inlet and 15 C for the cold side inlet. Second, it is assumed that the viscosity ratio raised to such a low power can be taken as 1.0. When Yan and Lin’s correlation is plotted on the same plot with the experimental data, as shown in Fig. 1, large deviations are noted. This was extensively investigated by contacting the authors, but the error could not be resolved. Their correlation seems to be off from the data by a fixed multiplier. To correct this situation, an improved correlation is developed in the present work as described in the following section.
4.2 Correlation Scheme I – Improved SinglePhase Correlation
Noting the similarity between the Yan and Lin (1999b) proposed correlation and the SiederTate equation, it seems logical to begin the development of a new correlation using smooth tube equations as the basis. Starting with the wellknown DittusBoelter (eq. 1) and SiederTate (eq. 2) equations, the improved correlation begins to take the form
a _{w} = E _{1} Re ^{E}^{2} Pr ^{E}^{3} l _{L} /D _{h}
(7)
Here the variables E _{1} , E _{2} , and E _{3} can be varied to fit the data. According to the DittusBoelter and SiederTate equations, and eq. (6), the variable E _{2} becomes 0.78 or 0.8, and the variable E _{3} takes on the value of 1/3 or 0.4. Combinations of these variables are used, and the multiplier E1 varied using a spreadsheet solver to develop an improved correlation to best fit the data. The new Correlation Scheme I results in
a _{w} = 0.2875 Re ^{0}^{.}^{7}^{8} Pr ^{1}^{/}^{3} l _{L} /D _{h}
(8)
A comparison of the singlephase Correlation I (eq. 8) to the Yan and Lin’s correlation (eq. 6) and the experimental data is provided in Fig. 1. As seen from this figure, the improved correlation provides an excellent fit to the experimental data for the singlephase watertowater heat transfer in the plate heat exchanger. This equation will be used in the next phase of the present work for developing an improved twophase correlation scheme.
9
4.3
Yan and Lin’s TwoPhase Correlation Scheme
Yan and Lin (1999b) proposed the following correlation in modeling their own twophase data for evaporation heat transfer of refrigerant R134a,
a _{r} = 1.926 Re _{e}_{q} Pr ^{1}^{/}^{3} Re ^{}^{0}^{.}^{5} Bo _{e}_{q} l _{L} / D _{h} 
(9) 

where, 

Re _{e}_{q} = G _{e}_{q} D _{h} / h _{L} 
(10) 

Bo _{e}_{q} = q” _{w} / (G _{e}_{q} h _{L}_{G} ) 
(11) 

G _{e}_{q} = G [(1 
– x _{m} ) + x _{m} (r _{L} / r _{G} ) ^{0}^{.}^{5} ] 
. 
(12) 
The terms Re _{e}_{q} and Bo _{e}_{q} are, respectively, the equivalent Reynolds number and equivalent Boiling number as proposed by Akers (1958). The Reynolds number, Re, and Prandtl number, Pr, are defined for all liquid flow:
Re = G D _{h} / h _{L} 
(13) 
Pr = h _{L} c _{p} / l _{L} . 
(14) 
Fluid properties used in testing the correlation are taken from the REFPROP (NIST, 1999) refrigerant property computer program. Results of using the Yan and Lin correlation vs. experimental data can be found in Fig. 2 through Fig. 11. Again for the twophase scenario, significant discrepancy is found between the Yan and Lin’s correlation and their experimental data. Extensive checks were performed to resolve the discrepancy but it was clear that there is a major error in their correlation as published in the literature. For purposes of the present work it is desirable to accept that a discrepancy exists in their correlation, and to develop an improved model to fit the experimental data set.
4.4 Other TwoPhase Correlation Schemes
10
Twophase correlation schemes were proposed in the works of Kattan et. al. (1998c), Gungor and Winterton (1986), and Kandlikar (1991). A brief review of these works was presented in section 2 of this paper, and in Table 1. Here the importance of these studies to the present work is evaluated for development of an improved model. Kattan et al. (1998c) developed a twophase model for various refrigerants based on data from previous works (Kattan et al., 1998a; and Kattan et al., 1998b). The proposed correlation scheme utilizes the study of boiling flow pattern as a basis for the proposed model. While the Kattan flow pattern model provides excellent results in the referenced work, the proposed model is complex and limited at this point to flow in horizontal tubes. Based on an early general correlation proposed by Chen (1966), Gungor and Winterton (1986) developed an additive model for twophase boiling in tubes and annuli. This model takes into account the contributions of pool boiling and convective boiling, which are brought together in determining the heat transfer coefficient. The results provided in the literature for the Gungor and Winterton model show an uncertainty of +40 %, which is larger than desired for this study. An additive model was also suggested by Kandlikar (1991) for flow boiling applications. The Kandlikar work is based on an additive model suggested by Bergles and Rohsenow (1964), consisting of convective and nucleate boiling components. Kandlikar takes a previous model (Kandlikar, 1990) for flow boiling in smooth tubes and extends it to augmented tubes and compact evaporators. Results in the Kandlikar (1991) study show the Kandlikar correlation scheme to provide a reliable fit to data, with an uncertainty of +15 %.
4.5 Kandlikar Two Phase Correlation Scheme
For the present work, it is desirable to find a best fit correlation for the evaporation heat transfer coefficient of refrigerant R134a. The Kandlikar (1991) correlation noted above has been shown in the literature to be adaptable to augmented surface systems while maintaining reasonable accuracy in results. Kandlikar suggests the following correlation scheme to model flow boiling for augmented tubes:
a _{T}_{P}_{,}_{N}_{B}_{D} = 0.6683 Co ^{}^{0}^{.}^{2} (1 – x) ^{0}^{.}^{8} a _{L} E _{C}_{B}
+ 1058.0 Bo ^{0}^{.}^{7} (1 – x) ^{0}^{.}^{8} F _{f}_{l} a _{L} E _{N}_{B}
11
(15)
and
a _{T}_{P}_{,}_{C}_{B}_{D} = 1.1360 Co ^{}^{0}^{.}^{9} (1 – x) ^{0}^{.}^{8} a _{L} E _{C}_{B}
+ 667.2 Bo ^{0}^{.}^{7} (1 – x) ^{0}^{.}^{8} F _{f}_{l} a _{L} E _{N}_{B}
.
(16)
Equations 15 and 16 solve for the heat transfer coefficient for the nucleate boiling dominant, and
convective boiling dominant cases, respectively. The actual heat transfer coefficient is the larger
of the two values. The factors E _{C}_{B} and E _{N}_{B} are the enhancement factors to be applied to the
convective boiling contribution and nucleate boiling contribution, respectively. The convective number, Co, and the boiling number, Bo, are defined as
Co = (r _{G} / r _{L} ) ^{0}^{.}^{5} ((1 – x) / x) ^{0}^{.}^{8} 
(17) 
Bo = q” / (G h _{L}_{G} ) 
(18) 
A fluiddependent parameter, F _{f}_{l} , is used in the correlation to account for varying fluid
characteristics. For stainless steel surfaces, Kandlikar (1991) suggests that this fluiddependent parameter is equal to 1 for all fluids. The term a _{L} is the all liquid heat transfer coefficient for the
singlephase situation. Correlation I (eq. 8) was introduced earlier as an improved fit to the singlephase experimental data, and can be employed here as the all liquid heat transfer coefficient for refrigerant R134a. For the plate heat exchanger, fluid flow is turbulent even at very low Reynolds numbers. As a result, it is assumed that the heat transfer will rely dominantly on the convective boiling contribution. Hence, eq. 16 will be used for comparison with the experimental data. Fluid properties are again taken from the REFPROP program. Employment of the Kandlikar correlation here predicts the data with an uncertainty of +20 % for mean vapor quality between 0.2 and 0.8. Fitting the Kandlikar correlation to the data to minimize absolute error results in the correlation values E _{C}_{B} = 0.659 and E _{N}_{B} = 0.006 for the augmentation effects of the heat exchanger, giving the final equation as
a _{T}_{P}_{,}_{C}_{B}_{D} = 1.1360 Co ^{}^{0}^{.}^{9} (1 – x) ^{0}^{.}^{8} a _{L} E _{C}_{B}
where,
+ 667.2 Bo ^{0}^{.}^{7} (1 – x) ^{0}^{.}^{8} F _{f}_{l} a _{L} E _{N}_{B}
12
,
(19)
E _{C}_{B} = 0.659 E _{N}_{B} = 0.006 .
Although the above scheme based on the Kandlikar correlation provides a reasonable fit to the data, this scheme results in a curve that does not match the shape of the data, especially in the high quality region as seen in Fig. 2 through Fig. 11 for various mass flux, heat flux and, pressure conditions. Therefore a new model development is undertaken to match the experimental data and the trends of heat transfer coefficient in the plate heat exchanger.
4.6 Improved Correlation II – Two Phase With Augmentation Factors
With reasonable accuracy achieved by eq. (19) above, it is logical to develop an improved correlation scheme based on the same additive model. Using the same exact form as the Kandlikar (1991) correlation, the model is represented by
a _{r} = [y1 Co ^{y}^{2} E _{C}_{B} + y3 Bo ^{y}^{4} F _{f}_{l} E _{N}_{B} ] (1 – x) ^{y}^{5} a _{L} ^{y}^{6}
.
(20)
Variables y _{1} through y _{6} , and the augmentation factors, E _{C}_{B} and E _{N}_{B} , are solved for in fitting the correlation to the experimental data. The fluiddependent parameter, F _{f}_{l} , is again taken to be 1.0 for the stainless steel plates. In fitting the proposed correlation to the heat exchanger data using a spreadsheet solver, the following values are obtained for the constants in eq. (20):
y _{1} = 2.312, 
y _{2} = 0.3, 
y _{3} = 667.3, 
y _{4} = 2.8, 
y _{5} = 0.003, 
y _{6} = 1, 
E _{C}_{B} = 0.512, and E _{N}_{B} = 0.338 . 

The final equation for the Correlation Scheme II is given by 

a _{r} = [2.312 Co ^{}^{0}^{.}^{3} E _{C}_{B} + 667.3 Bo ^{2}^{.}^{8} F _{f}_{l} E _{N}_{B} ] (1 – x) ^{0}^{.}^{0}^{0}^{3} a _{L} 
(21) 
where
E _{C}_{B} = 0.512,
13
E _{N}_{B} = 0.338.
Figures 2 through 11 show that Correlation II (eq. 21) provides a better fit than the original
Kandlikar (1991) correlation, with a maximum deviatio n of 17 % for all data points. It can be
seen that the experimental data trend is well matched by Correlation II in the entire range of
quality. 

4.7 
Improved Correlation III – TwoPhase Without Augmentation Factors 
A second improved correlation has been developed in the present work by leaving out the
augmentation factors and the fluiddependent parameter. This correlation then takes on the form
a _{r} = [z _{1} Co ^{z}^{2} + z _{3} Bo ^{z}^{4} ] x _{m} ^{z}^{5} a _{L} ^{z}^{6} z _{7}
.
(22)
A multiplier z _{7} has been added to the correlation scheme. Here only the variables z _{1} through z _{7}
need to solved for in fitting the correlation to the experimental data. For the best fit curve for
evaporation of refrigerant R134a in the plate heat exchanger, the resultant values are
z 
_{1} = 1.056, 
z _{2} = 
0.4, 
z _{3} = 1.02, 
z _{4} = 0.9 

_{5} = 0.12, z 
z _{6} = 0.98, and 
z _{7} = 1.055 
. 
With these constants, Correlation III takes the following form.
a _{r} = 1.055 [1.056 Co ^{}^{0}^{.}^{4} + 1.02 Bo ^{0}^{.}^{9} ] x _{m} ^{}^{0}^{.}^{1}^{2} a _{L} ^{0}^{.}^{9}^{8}
.
(23)
Correlation III as suggested here provides an excellent fit with a maximum deviation of 16 %, as
seen in Fig. 2 through Fig. 11. As can be seen, the data is well represented in the entire range of
quality. 

4.7 
Discussion of Results 
14
Figures 2 through 11 show the results of the twophase correlation schemes discussed in this study. Alternate figures show the predicted heat transfer coefficient plotted against the experimental heat transfer coefficient with 20% error bounds. It is easily seen from the figures that the Yan and Lin correlation (eq. 9) yields a large discrepancy, and the Kandlikar correlation (eq. 16) yields a different curve form the experimental data. The newly developed correlations, Correlation II (eq. 21) and Correlation III (eq. 23) both provide an excellent fit to the data set. Correlation III however is developed specifically for the present data set. It lacks the theoretical basis, and the improvement over Correlation II is quite small, about 1 percent. Therefore Correlation II represented by eq. (21) is recommended as the final correlation in the present work.
5.
CONCLUSIONS
The experimental data for evaporation of refrigerant R134a in a plate heat exchanger obtained by Yan and Lin (1999b) is used in developing twophase flow boiling correlation in this study. As a result of the present work, the following observations can be made:

A singlephase correlation, Correlation I (eq. 8), is developed for the singlephase heat 

transfer in the 60 degree chevron plate heat exchanger with D _{h} =6.6 mm. 
This correlation is 


able to represent the experimental singlephase data to within 3 percent. The Kandlikar (1991) scheme for augmented tubes and compact evaporators resulted in a reasonable overall accuracy, but the trend in the high quality region was not correctly represented. The original correlation was intended for qualities below 0.8. 


Two modified correlaions, Correlations II and III were developed to represent the twophase data for R134a in plate heat exchanger obtained by Yan and Lin (1999b). In Correlation II, the basic form of the Kandlikar (1991) correlation is retained, but the exponents for Bo, Co, and (1x), and the constants in the correlation are reevaluated from the best fit to the experimental data. Correlation II is able to correlate the data within 17 percent mean error. It is also able to represent the variation of heat transfer coefficient with quality accurately in the entire range of quality for all mass flux, heat flux and pressure conditions. 
15

Correlation III is developed as a simplified form as given by eq. (23). It is obtained as the 
best fit to the data, resulting in a mean error of 16 percent. However, since the correlation 

lacks a theoretical basis, and the improvement over Correlation II is not significant, it is not 

recommended. 


Correlation II, given by eq. (21) , is recommended for correlating the flow boiling heat 
transfer coefficient for R134a in the 60 degree chevron plate heat exchanger investigated by
Yan and Lin (1999b).
6. REFERENCES
Akers, W. W., Deans, H. A., and Crosser, O. K., 1958, “Condensation heat transfer within horizontal tubes,” Chemical Engineering Progress, Vol. 54, pp. 8990. Bergles, A. E., and Rohsenow, W. M., 1964, “The Determination of Forced Convection, Surface Boiling Heat Transfer,” ASME Journal of Heat Transfer, Vol. 86, pp. 365372. Chen, J. C., 1966, “Correlation for Boiling Heat Transfer to Saturated Fluids in Convective Flow,” Ind. Engng Chem. Proc. Des. Dev., Vol. 5, pp. 322329.
Dittus, F. W. and Boelter, L. M. K., 1930, “Heat Transfer in Automobile Radiators of the Tubular Type,” Publications in Engineering, Vol. 2, p. 443, University of California, Berkeley. Foster, H. K., and Zuber, N., 1955, “Dynamics of Vapour Bubbles and Boiling Heat Transfer,"”A. I. Ch. E. Jl, Vol. 1, pp. 531535. Gnielinski, V., 1976, “New Equations for Heat and Mass Transfer in Turbulent Pipe and Channel Flow,” International Chemical Engineering, Vol. 16(2), pp. 359367. Gungor, K. E., and Winterton, R. H. S., 1986, “A General Correlation for Flow Boiling in Tubes and Annuli,” International Journal of Heat and Mass Transfer, Vol. 29 No. 3, pp. 351
358.
Hewitt, G. F., Shires, G. L., and Bott, T. R., 1994, Process Heat Transfer, CRC Press, Florida, pp. 201203, 327363. Incropera and Dewitt, 1981, Fundamentals of Heat Transfer, John Wiley and Sons, New York, pp. 399407. Incropera, F. P., and Dewitt, D. P., 1990, Fundamentals of Heat and Mass Transfer, Third Edition, John Wiley and Sons, New York, pp. 504505. Kandlikar, S. G., 1990, “A General Correlation for Saturated TwoPhase Flow Boiling Heat Transfer Inside Horizontal and Vertical Tubes,” ASME Journal of Heat Transfer, Vol. 112, pp. 219228. Kandlikar, S. G., 1991, “A Model for Correlating Flow Boiling Heat Transfer in Augmented Tubes and Compact Evaporators,” ASME Journal of Heat Transfer, Vol. 113, pp.
966972.
Kattan, N., Thome, J. R., and Favrat, D., 1998a, “Flow Boiling in Horizontal Tubes: Part 1Development of a Diabatic TwoPhase Flow Pattern Map,” ASME Journal of Heat Transfer, Vol.120, pp. 140147.
16
Kattan, N., Thome, J. R., and Favrat, D., 1998b, “Flow Boiling in Horizontal Tubes: Part 2New Heat Transfer Data for Five Refrigerants,” ASME Journal of Heat Transfer, Vol.120, pp.
148155.
Kattan, N., Thome, J. R., and Favrat, D., 1998c, “Flow Boiling in Horizontal Tubes: Part 3Development of a New Heat Transfer Model Based on Flow Pattern,” ASME Journal of Heat Transfer, Vol.120, pp. 156165. Singh, A., Ohadi, M. M., and Dessiatoun, S., 1996, “Flow Boiling Heat Transfer Coefficients of R134a in a Microfin Tube,” Journal of Heat Transfer, Vol. 118, pp. 497499. Yan, Y.Y., and Lin, T.F., 1998, “Evaporation Heat Transfer and Pressure Drop of Refrigerant R134a in a Small Pipe,” International Journal of Heat and Mass Transfer, Vol. 41, pp. 41834194. Yan, Y.Y., and Lin, T.F., 1999a, “Condensation Heat Transfer and Pressure Drop of Refrigerant R134a in a Small Pipe,” International Journal of Heat and Mass Transfer, Vol. 42, pp. 697708. Yan, Y.Y., and Lin, T.F., 1999b, “Evaporation Heat Transfer and Pressure Drop of Refrigerant R134a in a Plate Heat Exchanger,” ASME Journal of Heat Transfer, Vol. 121, pp.
118127.
17
Table I  Summary of the Experimental Data and Rele vant Correlations
Investigator 
Experimental 
Test Fluid/ 
Experimental 
Results 

System 
Parameters 
Uncertainties 

Gungor and 
Vertical and 
Data taken from 28 authors for 7 fluids for saturated and subcooled boiling 
Used +40 % for experimental uncertainty 
Developed an 

Winterton, 
horizontal tubes 
additive 

(1986) 
and annuli 
model based 

on Chen 

(1966) 

Kandlikar 
Augmented 
Data taken from literature for refrigerants R11, R22, R113 
+15 % used for experimental uncertainty 
Developed an 

(1991) 
tubes and 
additive 

compact 
model to 

evaporators 
account for 

surface 

enhancements 

Kattan, 
Horizontal tubes, Smooth surface, Copper, D _{i} : 10.92 – 12.00 mm 
R134a, R123, 
Accuracy; T: +0.10 ^{o} C G: +0.2 % P: +0.5 % 
Developed 

Thome, and 
R402A, R404A, 
twophase 

Favrat 
R502; 
model based 

(1998c) 
G: 100 – 
on flow 

500 
kg/s m ^{2} 
h: +7.2 % 
pattern 

T _{s}_{a}_{t} : 1.3 – 10.3 ^{o} C 

X: 0.04 – 1.0 q”: 0.44 – 

36.54 kW/m ^{2} 

Yan and 
Small pipe, Smooth surface, D _{i} = 2.0 mm 
R134a; 
Uncertainties; A: +1 % T: +1 ^{o} C P: +0.001 MPa G: +2 % q: +0.5 % X: +0.03 
Developed a 

Lin, (1998) 
T _{s}_{a}_{t} : 5 – 31 ^{o} C q”: 5 – 20 kW/m ^{2} G: 50 – 
twophase 

additive heat 

transfer 

200 
kg/s m ^{2} 
model, similar 

X: ~0.05 – 0.95 
to that 

proposed by 

h 
_{r} : +6 % 
Kandlikar 

(1990) 

Yan and 
Plate heat exchanger, 60 deg chevron, D _{h} = 6.6 mm, Single pass, Three plates, Stainless steel 
R134a; 
Accuracy; A: +7e05 m ^{2} T: +0.2 K P: +0.002 MPa G: +2 % 
Developed a 

Lin, (1999b) 
X: 0.1 – 0.9 P: 0.675 – 0.8 MPa q”: 11 – 
correlation to 

model two 

phase 

evaporation 

15 
kW/m ^{2} 
X: +0.03 
heat transfer 

G: 55 – 
h 
_{r} : +15 % 
coefficient of 

70 
kg/s m ^{2} 
refrigerant 

T _{s}_{a}_{t} : 25.5 – 31 ^{o} C 
R134a 
18
Viel mehr als nur Dokumente.
Entdecken, was Scribd alles zu bieten hat, inklusive Bücher und Hörbücher von großen Verlagen.
Jederzeit kündbar.