Evaporation heat transfer coefficients of ammonia with a small

content of water in a plate heat exchanger in an absorption

refrigeration machine

Igor Ventura

Under supervision of Prof. Filipe Mendes

Dep. Physics, IST, Lisbon, Portugal

December 6, 2010

Abstract

Evaporation heat transfer coefficients of the mixture ammonia-water in a plate heat exchanger

(PHE) are investigated experimentally in this study. The study is performed in the evaporator of an

absorption refrigeration machine prototype at LSAS. During experiments to characterize the prototype,

data on the evaporator was collected for this study. In the evaporator, the refrigerant is cooled with

water and both fluids circule in counterflow. The experimental parameters in this work include the

refrigerant mass flux with a range from 0,90 to 3,09 kgm−2s−1, heat flux from 616,6 to 2962 W/m2

,

pressure from 3,75 to 6,54 bar and ammonia concentration from 0,93 to 0,99. Correlations proposed

on literature for heat transfer coefficients for evaporation in PHEs doesn’t fit to our data. A correction

for those correlations were proposed based on the fact that the ammonia concentration of the mixture

influences the calculation of heat transfer coefficients.

Keywords: Evaporation; Plate heat exchanger; Heat transfer coefficient; Correla-

tion; Absorption; Ammonia-water; Experiment

1 Introduction

Plate heat exchangers (PHE) have been used since

1930’s in chemical and food processing industries

for single-phase heat transfer from liquid-to-liquid.

In the 70’s they also began to be used in heat trans-

fer processes with phase changing, particularly con-

densers and evaporators in chillers and heat pumps.

1

The use of PHE in refrigeration machines as be-

come more common due to their compactness, heat

transfer area, the facility to create and maintain a

turbulent flow and higher heat transfer coefficients

when compared with other usual heat exchangers.

In the last years, developed countries start to

making some efforts to reduce the energy consump-

tion and carbon dioxide emissions to the atmo-

sphere. In this context, absorption refrigeration

machines powered by solar energy can be seen as a

promising technology as a replacement of conven-

tional air-conditioning systems which represents an

important section of total electric energy consump-

tion in the world.

The work present in this paper fits the develop-

ment of an air-cooled absorption refrigeration ma-

chine prototype of small power at Laboratório de

Sistemas de Arrefécimento Solar (LSAS). This pro-

totype were built to have a refrigeration power of

5kW with a COP of 0,56. The working fluid used

were an ammonia-water mixture.

Particularly, this work study the heat transfer

process in the evaporator . The evaporation pro-

cess in a refrigeration machine is very important on

its performance.

The evaporator is the machine’s component

that receive heat from the environment we pretend

to cool and it is extremely important that its de-

sign fit in the cooling power capacity of the ma-

chine. The heat exchanger used as evaporator on

the prototype is a PHE.

Many studies that have been perform to the

PHEs can be found in open literature. However,

most of them are focused in single-phase heat trans-

fer from liquid-to-liquid, especially for water [1].

Only in the last two decades some work has been

published about two-phase heat transfer on this

exchangers type and therefore, there are limited

data available to the use of PHEs in processes with

phase-changing like evaporation.

Recently, Y.Y Yan and T.F. Lin, D. Donowski

and S.G. Kandlikar, Y.Y. Hsieh and T.F. Lin and

Han and Kim had proposed correlations for heat

transfer coefficients during the evaporation of re-

frigerants in a PHE [2, 3, 4, 5]. Some other

authors had studied evaporation process in PHE

[6, 7, 8, 9, 10, 11].

However the only available work on ammonia-

water mixture evaporation in a PHE was done by

F. Táboas who study the forced boiling of the mix-

ture in a PHE [12, 13].

There are no works where the study of heat

transfer coefficients are done with mixtures with

different concentrations of the refrigeran. In this

work, we try to correlate our data to a heat trans-

fer coefficient correlation and evaluate the influence

of ammonia concentration in the heat transfer cor-

relations.

2

2 Experimental set-up and proce-

dure

2.1 System description

This experiment wasn’t dedicated to the study of

ammonia-water evaporation in a PHE. All data col-

lected for this study were measured while testing

the operating conditions of the absorption refriger-

ation machine prototype.

The prototype studied is a single stage

ammonia-water chiller with plate heat exchangers

using water to the heat transfer at the main compo-

nents. The water from the Absorber and Condenser

is air-cooled. There are to boilers to simulate the

water from the solar panels and the charge. The

machine have two internal heat recovery heat ex-

changers as it is usual in this kind of absorption

cycles. The Generator is coupled with a separation

vessel. Also two refining devices were used dur-

ing the experiments. Those were coupled with the

generator vessel and were tested by working with

a contraflow between the vapor generated and the

rich solution.

At the first column an aerosol is created by

a nozzle to promote the mass and heat exchange.

The second column is random packed with Novalox

ceramic saddles.

The beginning of the experiments on the proto-

type were done without any refining system incor-

porated.

All the system were thermally isolated.

During the experiments on the prototype sev-

eral experimental conditions were tested and all the

external inputs were changed: the water flow rates

and temperatures at generator, absorver, condenser

and evaporator; the rich solution and refrigerant

flow rate through the expansion valve.

The amount of ammonia at the solution circuit

was also changed retaining it in the refrigeration

circuit.

The variation of external inputs during the ex-

periments and the use of different systems to treat

vapor produced at the generator (refining columns)

changed the concentration of the mixture. There-

fore, data points obtained for the study of heat

transfer at the evaporator have different values of

ammonia concentration on the mixture.

2.2 Plate heat exchanger

The heat exchanger used as evaporator on the pro-

totype is a nickel brazed plate heat exchanger,

model NB-26-18H manufactured by Alfa Laval.

This heat exchanger have 18 plates forming 17chan-

nels where 9 are intend for water flow and 8 for the

ammonia-water mixture.

The plates of the PHE have a herringbone cor-

rugation type with a chevron angle of 25%.

The water side and refrigerant side on the PHE

circulate in counterflow.

Figure 1 shows a schematic of the PHE geomet-

ric characteristics

3

!"#$%&'("&))* !#+,- (- .(/0 12 "+-!(!'! +3 3+45 (-6$78$-6$-' )++8! &-6 & 6&'& &"94(!('(+- !*!'$%0 :' (-")46$!& 5$35(/$5&-' )++82 ',+ ,&'$5 )++8! ;+-$ 3+5 85$#$&'$5&-6 '#$ &-+'#$5 3+5 '#$ '$!' !$"'(+-<2 &-6 & "+)6 ,&'$5=/)*"+) )++80 :- +56$5 '+ "+-'5+) >&5(+4! '$!' "+-6('(+-!+3 ?7@1AB ;(-")46(-/ (%8+!$6 #$&' C4D2 5$35(/$5&-'%&!! C4D &-6 !*!'$% 85$!!45$< (- '#$ '$!' !$"'(+-2 ,$-$$6 '+ "+-'5+) '#$ '$%8$5&'45$ &-6 C+, 5&'$ (- '#$+'#$5 '#5$$ )++8!0

!"#" $%&'()%'*+, -./ 0..1

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

!"!" 234

E#5$$ "+%%$5"(&) SS7T1U 8)&'$! %&-43&"'45$6 F* '#$V&+5( P$&' E5$&'%$-'2 E&(,&- 3+5% '#$ NPQ 3+5 '#$!&'45&'$6 C+, F+()(-/ '$!'0 E#$ 8)&'$ !453&"$! &5$85$!!$6 '+ F$"+%$ /5++>$6 ,('# & "+554/&'$6 !(-4!+(6&)!#&8$ &-6 UA! +3 "#$>5+- &-/)$2 ,#("# (! '#$ &-/)$ +3 W7/5++>$! '+ '#$ >$5'("&) &D(! +3 '#$ 8)&'$0 E#$ 6$'&()$6"+-K/45&'(+- 3+5 '#$ NPQ "&- F$ !$$- (- .(/0 M0 E#$"+554/&'$6 /5++>$! +- '#$ 5(/#' &-6 )$3' +4'$5 8)&'$!

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

!"5" 6*,%' 0..1 &.' ,%7, 7%8,(.+

E#$ ,&'$5 )++8 (- '#$ $D8$5(%$-'&) !*!'$% 3+5 "(57"4)&'(-/ '#$ #+' ,&'$5 '#5+4/# '#$ '$!' !$"'(+- "+-'&(-! &MA ) ,&'$5 '#$5%+!'&' ,('# & A0R JZ #$&'$52 &-6 & ,&'$584%8 (! 4!$6 '+ 65(>$ '#$ #+' ,&'$5 &' & !8$"(K$6 ,&'$5

.(/0 M0 S"#$%&'(" 6(&/5&% +3 NPQ0

1ATU 9"9" 37(%:; <"=" >(+ ? @+,%'+*,(.+*0 A.B'+*0 .& 3%*, *+C D*77 <'*+7&%' EF G!HH!I #H55J#HEE

Figure 1: Schematic diagram of PHE

2.3 Measuring equipment and data ac-

quisition

Four types of measurements were made to charac-

terize the performance of the machine and of the

Evaporator: temperature; pressure; volume flow

rates; ammonia concentration.

The temperatures were measured at the in-

let and outlet of each side of each component.

These temperatures were measured with four wired

PT100 sensors, calibrated to ensure a maximum

absolute error of 0,1K.

The high pressure and the low pressure of the

machine were measured at the Generator vessel and

at the absorber outlet. Also, another pressure sen-

sor was present at the entrance of the evaporator,

measuring the difference between that point and

the absorver outlet. The pressure sensors used en-

sured a maximum absolute error of 0,125 bar for the

high pressure and 0,05 bar for the low pressure.

The mass flow rate of the refrigerant at the PHE

were calculated from the mass conservation equa-

tions at the separation vessel at the generator.

The volume flow rate of the rich and poor solu-

tions were measured with an EM flow meter which

has a maximum relative error of 0,4 %.

The water flow rate at the Evaporator was mea-

sured with a turbine flow meter with an maximum

error of 1%.

All the other parameters that characterize the

working conditions of the machine, and particu-

larly the evaporator, were calculated based on these

measurements, using one external function Engi-

neering Equation Solver (EES) that gives several

thermodinamic proprieties of ammonia-water mix-

ture when given three known properties of the Mix-

ture. The correlations used by this function can be

found at [20]. The relative error in the calculation

of the vapour concentration due to the measure-

ments errors is lower than 0,5%.

Each data point corresponds to averaged mea-

surements, taken each 10 seconds, over a period of

15 minutes of steady state running.

4

3 Data reduction

In order to proceed with calculations of heat trans-

fer coefficients of ammonia-water mixture from the

experimental data some data reduction analysis is

needed.

Shah and Wanniarachchi [17] suggested that

the hydraulic diameter of the PHE channels can

be expressed as two times the mean channel spac-

ing if the channel width is much larger than the

channel spacing,

De = 2 b (1)

The overall heat transfer coefficient in the evap-

orator, U , can be expressed as

U =Q

A∆Tln(2)

where Q is the heat transfer rate from the hot

fluid to the refrigerant, A is the heat transfer area

of the PHE and ∆Tln is the logarithmic mean tem-

perature difference which can be calculated by the

expression

∆Tln =�Tin −�Tout

ln�

�Tin�Tout

� (3)

with

�Tin = Tw, in − Tr, out (4)

�Tout = Tw, out − Tr, in (5)

where the temperatures differences are repre-

sented at Figure 2.

Figure 2: Temperature profile on a heat

exchanger.

The use of the concept of logarithmic mean

temperature difference is only valid for counter-

current and co-current single-phase heat exchanger

configurations. However, accordingly with J.

Claesson, Logarithmic Mean Temperature Difer-

ence approach can be used in two-phase heat trans-

fer like the evaporation as an approximation to the

real mean temperature difference[14].

The heat transfer rate in the PHE can be ob-

tained by a energy balance on the water side

Qw = mw cp∆Tw (6)

where mw is the mass flow rate of water on the

PHE measured with the flowmeter, cp is the water

specific heat capacity and ∆Tw is the water tem-

perature difference between the inlet and outlet of

the PHE.

5

3.1 Heat transfer coefficients of water

For calculating the heat transfer coefficients of

ammonia-water mixture we need to know the heat

transfer coefficients of water. As we already dis-

cussed, there are many studies that focus on single

phase heat transfer in the literature. Since many

of these studies are also related to the heat trans-

fer coefficient of water, we used one correlation for

the heat transfer coefficients of water that can be

applied to our data.

In our experimental measurements the

Reynold’s number on the water side ranged

from 174 to 1128 while Prandtl’s number ranged

from 7,05 to 9,74 where Reynolds’s number and

Prandtl’s number are defined as

Re =GDh

µ(7)

Pr =µ cpk

(8)

Kumar [1] had proposed a correlation for heat

transfer coefficients of water in a plate heat ex-

changer that match with our data

Nu = C1RemPr0,33�µm

µw

�0,17

(9)

where, for chevron angles equal or below 30º

and Reynold’s numbers bigger than 10, the param-

eters C1 and m are 0,348 and 0,663, respectively.

The Nusselt number, Nu, is defined as

Nu =hDh

k(10)

3.2 Ammonia-water heat transfer coef-

ficients

The overall heat transfer coefficient can be derived

from the local heat transfer coefficients assuming

no fouling resistances

1

U=

1

hw+

1

hr+

∆x

k(11)

where ∆x is the plate wall thickness, is k ther-

mal conductivity of the plates and hr, hw are the

heat transfer coefficient from the refrigerant side

and water side, respectively.

By knowing the heat transfer coefficient from

the water side we can calculate the heat transfer

coefficient of ammonia-water mixture.

4 Analysis of experimental re-

sults

In the three different series of experimental mea-

surements 150 data points were obtained in which

66 are from the set-up with no refination, 42 are

from the set-up with the spray column and the

other 42 from the packed column.

Tables 1 and 2 show the minimum and maxi-

mum values of some measured and calculated pa-

rameters from all data collected during the exper-

iments with the prototype for the water side and

refrigerant side, respectively.

6

Measures Values

Min Max

Tw, in (ºC) 8,3 21,1

Tw, out (ºC) 62 19,6

ms (l/s) 0,09 0,58

Rew 174,4 1128,0

Qw (W/m2) 616,6 2962,0

Table 1: Minimum and maximum values

measured and calculated for water side.

Measure Values

Min Max

Tr in (ºC) -2,3 13,7

Tfr out (ºC) 1,0 18,0

x 0,93 0,99

P (bar) 3,75 6,54

m (g s−1) 1,44 4,95

Rer 19,5 73,4

Table 2: Minimum and maximum values

measured and calculated for refrigerant side.

4.1 Comparison with available correla-

tions

Correlations for heat transfer coefficients in the

evaporation of ammonia-water mixture in a plate

heat exchangers are important to the design of

evaporators used in air-conditioning and refriger-

ation systems.

In the last years some correlations have been

proposed to the evaporation of several refriger-

ants in plate heat exchangers. Meanwhile, only

F. Táboas [12] have proposed one correlation for

the forced boiling on a desorber of an ammonia-

water mixture with 42% of ammonia concentra-

tion. There are no proposed correlations or data

available in literature about the heat transfer co-

efficients of ammonia-water mixture at with high

concentrations of ammonia and therefore we tried

to apply known correlations types to our data.

The values of the physic proprieties used in the

correlations were calculated by using the mean tem-

perature of water and refrigerant inside the PHE.

Hsieh and Lin correlations

In the study of flow boiling heat transfer of refriger-

ant R-140a in a PHE, Hsieh and Lin proposed a cor-

relation for heat transfer coefficients. The expres-

sion used to correlate the heat transfer coefficients

is the term of nucleate boiling present in the first

proposed correlation for flow boiling heat transfer

coefficients in a compact heat exchanger by Kan-

dlikar [18]. These authors also included one cor-

recting factor, firstly proposed by Seider and Tate

[19], that have in account the velocity differences

between the bulk fluid and the fluid close to the

walls of the exchanger. The correlation type used

was

hf = ak

DeReb Pr

13Boc

�µm

µw

�0,14

(12)

where a, b and c are the constants that corre-

7

late this expression to their data and Boeq is the

equivalent Boiling number and is defined by

Bo =q��

Gilg(13)

Yan and Lin correlation

Y.Y. Yan and T.F. Lin (1999) proposed one cor-

relation for heat transfer coefficients of refrigerant

R-134a during his evaporation on a plate heat ex-

changer. The expression used to correlate their

data is simillar with the correlation used by Hsieh.

In their correlation, they didn’t use the correction

due to fluid velocity diferences, but they used a

correction for the mass flux of a two-phase fluid.

Their experimental data were correleted with

the expression

hf = ak

DhReb Pr

13BoceqG

∗ (14)

where Boeq is the equivelent Boiling number

calculated with a equivelent mass flux

Bo =q��

Geq ilg(15)

where Geq is the equivalent mass flux firstly pro-

posed by Akers [15]

Geq = GG∗ (16)

with

G∗ =

�(1−Qum) +Qum

�ρlρg

�0,5�

(17)

Han and Kim correlation

Han and Kim proposed an correlation to predict

the evaporation heat transfer coefficients for refrig-

erant R-410A in a brazed plate heat exchanger .

The correlation proposed includes the equivalent

mass flux in the Reynold’s number

hf = ak

DhRebeq Pr

13Boceq (18)

Donowski and Kandlikar correlation

Donowski and Kandlikar (2000) proposed a new im-

proved correlation for the refrigerant R-134a based

on Yan and Lin data. The proposed correlation is

similar to the correlation proposed by Kandlikar in

1990 where they include a convective heat transfer

term on the correlation

hf =k

DeReb Pr

13 (aBoc + dCoe)Quf (19)

Correlation comparison

Parameters a, b, c, d, e and f from the previous

correlations are constants used to adjust those cor-

relations to our data. Table 3 shows the results of

the fitting.

In Figure 3 can be seen the comparison of ex-

perimental data with the correlations obtainded.

The results show that proposed correlations in

literature don’t adjust quite well to our data. In

the plot in Figure 3(c) seems that the correlation

can’t predict the higher and lower heat transfer co-

efficients at the same time.

8

Correlation Parameter Standard deviation δ%

a b c d e f (W m−2K)

Hsieh 0,607 0,765 0,091 - - - 385 35%

Han and Lin 0,0024 1,072 -0,23 - - - 333 28%

Han and Kim 0,00398 0,94 -0,24 - - - 333 28%

Donowski 0,930 0,771 0,034 -0,726 0,930 -0,258 343 29%

Table 3: Obtained parameters for correlate our data with previous correlations.

+30%

- 30%

0 500 1000 1500 2000 2500 30000

500

1000

1500

2000

2500

3000

hf experimental

h fCa

lcul

ado

hf � a Reb kDe

Pr13 Boeq

c G�

a�

0,00

236;

b�

1,07

2;c��

0,23

3

(a) Yan and Lin correlation

+30%

-30%

0 500 1000 1500 2000 2500 30000

500

1000

1500

2000

2500

3000

hf experimental

h fCalculado

hf � a Reeqb kDePr

13 Boeqc

a�0,00398;b�0,942;c��0,244

(b) Han and Kim correlation

+30%

-30%

0 500 1000 1500 2000 2500 30000

500

1000

1500

2000

2500

3000

hf experimental

h fCalculado

hf � a Reb kDePr

13 Boc � ΜmΜw �0,14

a�0,607;b�0,765;c�0,091

(c) Hsieh correlation

+30%

-30%

0 500 1000 1500 2000 2500 30000

500

1000

1500

2000

2500

3000

hf experimental

h fCalculado

hf � kDeReb Pr

13 �a Boc� d Coe�Qu f

a�0,726;b�0,771;c�0,0345;d��0,726;e�0,192;f��0,258

(d) Donowski correlation

Figure 3: Comparision of experimental heat transfer coefficients with the proposed correlations for

ammonia-water mixture during his evaporation on a PHE.

9

The parameter c in Donowski is negative which

has no physical meaning but can point out that

convective heat transfer isn’t important on this ex-

periment.

According with B. Thonon [16], the heat trans-

fer dependence of nucleate boiling can be determi-

nated by a simple criterion based on the Boiling

number and the Lockhart-Martinelli parameter

Boχ > 0, 15× 10−3 (20)

where χ is the Lockhart-Martinelli parameter

defined as

χ =

�1−Qu

Qu

�0,9�ρgρl

�0,5�µg

µl

�0,1

(21)

Accordingly with this criterion, most of our

data points indicates that heat transfer is governed

by nucleate boiling.

4.2 New proposed correlations

Since previous correlations didn’t fit in our data we

tried to change them in order to get better results.

During the experiments the ammonia concen-

tration on the mixture changed. That change could

affect profile temperatures of the mixture inside of

the PHE in ways that ∆Tln can no longer be a

good approximation to the calculation of global

heat transfer coefficient. Also some thermody-

namic characteristics of the mixture change with

the variation of ammonia concentration, like vis-

cosity, and those proposed correlations may not be

able to predict those variations.

Therefore, we propose to add one parameter

that is function of ammonia concentration to the

previous correlations tested. This way we add a

new multiplier to those correlations in the form of

xg where g is the parameter to be adjusted to the

correlation.

The new proposed correlations we try to test

with our data are then

hf = ak

DeReb Pr

13Boc

�µm

µw

�0,14

xg (22)

hf = ak

DeReb Pr

13Boceq G

∗xg (23)

hf = ak

DeRebeq Pr

13Boceq x

g (24)

hf =k

DeReb Pr

13 (aBoc + dCoe)Qufxg (25)

Table 4 indicates the values obtained for free

parameters of correlations and their relative devia-

tion.

Figure 4 also shows the comparison between

experimental and new predicted heat transfer co-

efficients for the ammonia-water evaporation in a

PHE.

The correction parameter added produced very

satisfactory results. The correlation with a better

result was the equation 25 but according with the

bigger number of free parameters used to fit this

correlation in comparison with others we opted to

consider correlation 24 as the best obtained. Also

the value of fitted parameters in this correlation

seems to be in concordance with others proposed

10

Correlation Parameter Standard deviation δ%

a b c d e f g (W m−2K)

Hsieh 7,04 0,948 0,680 - - - -23,6 186 16%

Yan and Lin 1,02 1,05 0,563 - - - -22,3 187 17%

Han and Kim 1,92 0,885 0,536 - - - -22,3 186 16%

Donowski 73,8 0,960 1,09 0,018 0,028 -0,091 -23,0 185 15%

Table 4: Parameters obtained to correlate our data with the new corrected correlations.

+30%

-30%

0 500 1000 1500 2000 2500 30000

500

1000

1500

2000

2500

3000

hf experimental

h fcalculado

hf � a Reb kDePr

13 Boeqc G� xg

a�1,020;b�1,051;c�0,563;g��22,341

(a) Yan and Lin

+30%

-30%

0 500 1000 1500 2000 2500 30000

500

1000

1500

2000

2500

3000

hf experimental

h fcalculado

hf � a Reeqb kDePr

13 Boeqc xg

a�1,924;b�0,885;c�0,536;d��22,304

(b) Han and Kim

+30%

-30%

0 500 1000 1500 2000 2500 30000

500

1000

1500

2000

2500

3000

hf experimental

h fcalculado

hf � a Reb kDePr

13 Boc � ΜmΜw �0,14 xg

a�7,037;b�0,948;c�0,680;g��23,617

(c) Hsieh

+30%

-30%

0 500 1000 1500 2000 2500 30000

500

1000

1500

2000

2500

3000

hf experimental

h fcalculado

hf � kDeReb Pr

13 �a Boc� d Coe�Qu f xg

a�73,794;b�0,959;c�1,095;d�0,0182;e�0,0284;f��0,0912;g��23,008

(d) Donowski

Figure 4: Comparision of experimental heat transfer coefficients with the new corrected correlations for

ammonia-water mixture during his evaporation on a PHE.

11

correlations in literature for this type of heat trans-

fer.

The correction factor xg varies from 1,24 to 4,93

where the heat transfer coefficients with lowest am-

monia concentration are the ones with the bigger

correction. This can be seen as a correction to the

temperature profile in the PHE for lowest values

of ammonia concentration where the temperatures

of refrigeran tends to increase at the outlet of the

PHE.

5 Conclusions

In order to preform a more precise study on the

heat transfer of ammonia-water mixture evapora-

tion in a PHE a dedicated experiment was needed.

The collected data during the experiments covered

an extense variation of external conditions. How-

ever more data points were needed to study the in-

fluence of some physical quantities (like heat flux,

mass flux, vapor quality, etc) in the heat transfer

coefficients on the ammonia-water evaporation in a

PHE.

To get more accurate heat transfer coefficients

for ammonia-water mixture during his evapora-

tion a new correlation for the heat transfer coef-

ficients from the water side should be find, since

the PHE geometric characteristics influences those

coefficients.

The proposed correlations for heat transfer co-

efficients during the evaporation of a liquid in a

PHE found on literature doesn’t adapt to our data

due to the change of ammonia concentration dur-

ing the experiments on the refrigeration absorption

machine prototype.

A correction to those correlations were pro-

posed as a multiplicative factor function of ammo-

nia concentration which improved the correlations

those correlations to a satisfatory level. The re-

sults and tests to those correlations show that the

introduced correction factor was related with a cor-

rection to the logarithmic mean temperature difer-

ence method used in the calculation of global heat

transfer coefficient.

References

[1] Zahid H. Ayub, 2003, “Plate Heat Exchan-

ger Literature Survey and New Heat Transfer

and Pressure Drop Correlations for Refrige-

rant Evaporators”, Heat Transfer Engineering,

Vol. 24, Nº 5, p. 3-16.

[2] Y.Y. Yan and T.F. Lin, 1999, “Evaporation

Heat Transfer and Pressure Drop of Refrige-

rant R-134a in a Plate Heat Exchanger”, Tran-

sactions of the ASME, Vol. 121, p. 118-127.

[3] V.D. Donowski and S.G. Kandlikar, 2000,

“Correlating Evaporation Heat Transfer Coef-

ficient of Refrigerant R-134a in a Plate Heat

Exchanger”, Engineering Foundation Confe-

rence on Pool and Flow Boiling, Alasca.

[4] Y.Y. Hsieh and T.F. Lin, 2002, “Saturated flow

boiling heat transfer and and pressure drop

12

of refrigerant R-410a in a vertical plate heat

exchanger”, International Journal of Heat and

Mass Transfer, Vol. 44, p. 1033-1044.

[5] D.H. Han, K.J. Lee and Y.H. Kim, 2003, “Ex-

periments on characteristics of evaporation of

R410a in brazed plate heat exchangers with

different geometric configurations”, Applied

Thermal Engineering, Vol. 23, p. 1209–1225.

[6] Y.Y. Hsieh et al, “Subcooled flow boiling heat

transfer of R-134a and the associated bubble

characteristics in a vertical plate heat exchan-

ger”, International Journal of Heat and Mass

Transfer, Vol. 45, p. 1791-1806.

[7] G.A. Longo, A. Gasparella and R. Sartori,

2004, “Experimental heat transfer coefficient

during refrigerant vaporisation and conden-

sation inside herringbone-type plate heat ex-

changers with enhanced surfaces”, Internatio-

nal Journal of Heat and Mass Transfer, Vol.

47, p. 4124–4136.

[8] D. Sterner and B. Sundén, 2006, “Performance

of plate heat exchangers for evaporation of am-

monia”, Heat Transfer Engineering, Vol. 27, p.

45–55.

[9] G.A. Longo and A. Gasparella, 2007, “Refri-

gerant R134a vaporisation heat transfer and

pressure drop inside a small brazed plate heat

exchanger”, International Journal of Refrige-

ration, Vol. 30, p. 821-830.

[10] E. Djordjevic and S. Kabelac, 2008, “Flow boi-

ling of R134a and ammonia in a plate heat

exchanger, International Journal of Heat and

Mass Transfer, Vol. 51, p. 6235-6242.

[11] H. Arima, J.H. Kim, A. Okamoto and Y. Ike-

gami, 2010, “Local boiling heat transfer cha-

racteristics of ammonia in a vertical plate eva-

porator, International Journal of Refrigera-

tion, Vol. 33, p. 359-370.

[12] Francisco Táboas Touceda, 2006, PhD “Estu-

dio del proceso de ebullición forzada de la mez-

cla amoiaco/agua en intercambiadores de pla-

cas para equipos de refrigeración por absor-

ción”, Universidad Rovira i Virgili, Tarragona.

[13] F. Táboas, M. Vallès et al, 2010, “Flow boi-

ling heat transfer of ammonia/water mixture

in plate heat exchanger”, International Journal

of Refrigeration, Vol. 33, p. 695-705

[14] J. Claesson, 2005, “Correction of logarithmic

mean temperature difference in a compacted

brazed plate evaporator assuming heat flux go-

verned flow boiling heat transfer”, Internatio-

nal Journal of Refrigeration, Vol.28, p. 573-578

[15] W.W. Akers, H.A. Deans and O.K. Crosser,

1958, “Condensation heat trasfer within hori-

zontal tubes”, Chemical Engineering Progress,

Vol. 45, p. 89-90.

[16] B. Thonon, A. Feldman, L. Margat and C.

Marvillet, 1997, “Transition from nucleate boi-

ling to convective boiling in compact heat ex-

13

changers”, International Journal of Refrigera-

tion, Vol. 20, p.592-597.

[17] R.K. Shah and A.S. Wanniarachchi, 1992,

“Plate heat exchanger design theory in indus-

try heat exchanger”, Von Karman Institute for

Fluid Dynamics, Bélgica.

[18] S.G. Kandlikar, 1991, “A Model for Flow

Boiling Heat Transfer in Augmented Rubes

and Compact Evaporators”, ASME Journal of

Heat Transfer, Vol. 113, p. 966-972.

[19] Incropera, Dewitt, Bergman and Lavine, 2007,

“Fundamentals of Heat and Mass Transfer”,

John Wiley & Sons.

[20] O.M. Ibrahim, S.A. Klein, 1993, “Termodina-

mic Properties if Ammonia-water Mixtures”,

ASHRAE Transactions, Vol. 99, p. 1495-1502.

14

Nomenclature

A Heat transfer area of the plate [m2]

b Channel spacing [m]

Bo Boiling number

cp specific heat [J kg−1K−1]

Co Convection number

De Equivalent hydraulic diameter [m]

Dh Diâmetro hidráulico

G Mass flux [kgm−2s−1]

Geq Equivalent mass flux [kgm−2s−1]

h Heat transfer coefficient [W m−2K−1]

ilg enthalpy of vaporization [Jkg−1]

k Thermal conductivity [W m−1K−1]

Nu Nulsset number

P Pressure [bar]

Pr Prandlt number

q�� Heat flux [W m−2]

Qu Vapor quality

Re Reynolds number

U Overall heat trasfer coefficient [W m−2K−1]

x ammonia concentration

COP Coefficient of performance

m Mass flow [kgm−2]

Q Heat transfer rate [W ]

Greek symbols

χ Lockhart-Martinelli parameter

∆T Temperature difference [K]

∆Tln Logarithimic mean temperature difference

[K]

∆x Plate wall thickness [m]

δ% Relative mean deviation

µ Viscosity [km−1s−1]

ρ density [kgm−3]

Subscripts

g gas

in inlet

l liquid

m mean

out outlet

r refrigerant

w wall

w water

15