The First Industrial Application of Non Azeotropic - Purdue e-Pubs

Purdue University
Purdue e-Pubs
International Refrigeration and Air Conditioning
Conference
School of Mechanical Engineering
1988
The First Industrial Application of Non Azeotropic
Mixture
J. C. Blaise
Electricite de France
T. Dutto
Electricite de France
J. L. Ambrosino
Institut Francais du Petrole
Follow this and additional works at: http://docs.lib.purdue.edu/iracc
Blaise, J. C.; Dutto, T.; and Ambrosino, J. L., "The First Industrial Application of Non Azeotropic Mixture" (1988). International
Refrigeration and Air Conditioning Conference. Paper 35.
http://docs.lib.purdue.edu/iracc/35
This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact [email protected] for
additional information.
Complete proceedings may be acquired in print and on CD-ROM directly from the Ray W. Herrick Laboratories at https://engineering.purdue.edu/
Herrick/Events/orderlit.html
THE FIRST INDUSTRIAL APPLICATION
OF NON AZEOTROPIC MIXTURE
J. C. Blaise
T. outto
Electricit e de France
Direction des Etudes et Recherches
B. P. N°1 - 77250 Moret S/Loing - FRANCE
J. L. Ambrosino
Institut Francais du Petrole
B. P. 311 - 92506 Rueil Malmaison Cedex - FRANCE
Abstract
Many controls, simulation s and tests have been performed in
laboratori es or test-rigs on non azeotropic mixtures.
Consequen tly,
Electricit e de France (EDF), the Institut Francais du Petrole (IFP),
and Company QUIRI have decided to build together such a unit on an
industrial site, using all their own top technolog ies to optimize
the operation and the control. Such a test in an industrial
environme nt is absolutely necessary to measure the reliabilit y of
this new technology .
Experimen ts were carried out on a heat pump which recovers
energy from a refrigerat ing installati on and warms water from 58 to
68° c.
Character istics of the installati on being;
- 2 open-type reciprocat ing compresso rs powered by
e~ectric-motors of 75 kW; the total swept volume
is 496
m jh at 1450 r.p.m.
- A perfect counter-fl ow condenser.
- Electronic expansion valves to feed the dry-ex
evaporato r.
As the aim of our experimen tation is to compare the cop and the
thermal power of the heat pump, we replace pure fluid Rl2 by a non
azeotropic mixture. So first we determine performanc es of the heat
pump with Rl2 to have a reference, then we replace Rl2 by a ternary
mixture. As the installati on includes 31 sensors (temperatu re,
pressure, flow meter, electric power) connected to a computer, we
can evaluate the influence of a non azeotropic mixture on the
volumetric and isentropic compressio n efficienci es, and on the
overall coefficien t of heat transfer of the condenser. Tests with
the non azeotropic mixture, give the following results (compared
with Rl2).
* +
* +
* * -
20% for the thermal power at the condenser,
1 1 5% for the coefficien t of performan ce,
2% for the volumetric efficiency ,
10% for the overall coefficien t of heat transfer for
the same heat flux.
At the end of the tests we successive ly created leakages
re~pe7tively at the input of the dry ex evaporato r (low pressure,
~l~qu~~ ~nd.gas ~hases) and at the top of the rece~ver where gas
is
~~ e~u~l~br~um w~th liquid.
We took a sample of one kilogramme of
l~qu~d fro~ the receiver before and after each leakage to know the
concentrat 1on of each component of the mixture. The following table
shows mass concentra tion of each component.
ll
At the
beginning
of test
mass
concentration %
After leakage
at the input
of the
evaporato r
After leakage at
the top of the
receiver
1,54
1,25
component A
1,54
component B
71,52
71,53
69,85
component c
26,94
26,93
28,90
Initial load of the heat pump is 300 kg.
Experimen tal results show no variation of mass concentra tion
when a leakage of 30 kg occurs at the input of the evaporato r. The
leakage at the top of the receiver lasts until gas crosses the
Between 20 and 30% of the total load was lost. we
expansion valve.
can therefore conclude that leakages do not modify the compositio n
of this mixture enough to change working conditions of the heat
pump.
LA PREMIERE APPLICATION INDUSTRIELLE DE MELANGE NON-AZEOTROPIQUE.
RESUME : De nombr.eux essais, controles et verificati ons ont ete menes
en laboratoir e sur des boucles d'essais a propos des melanges de
fluides. E.D.F., IFF et la Societe QUIRI ant pris la decision de
construire ensemble en site industrie l, en utilisant leurs technologies les plus evoluees, une unite mettant en oeuvre ces melanges.
Un tel essai est absolumen t indispensa ble pour mesurer la faisabilit e
de cette nouvelle technolog ie.
Les essais ant ete menes sur une P.A.C. qui recupere l'energle d'une unite de refrigerat ion et qul chauffe de l'eau de 58°C a
68°C. Les caracteris tiques de l'installa tion sont :
- 2 compresseu rs a pistons de type ouvert entraines par
moteur electrique de 75 kW chacun, le volume global
balaye est de 496 m3/h a 1450 tr/rnm;
- un condenseur a centre-cou rant parEait ;
-des detendeurs electroniq ues pour alimenter l'evaporat eur
a detente seche.
Comme le but de notre experlence est de comparer le C.O.P. et
la puissance thermique de la P.A.C., nous remplagons le fluids pur
Rl2 par un melange non azeotropiq ue. Nous determinon s tout d'abord
les performanc es de la P.A.C. avec du Rl2 pour obtenir le point de
reference, puis nous remplagons le Rl2 par un melange ternaire.
Comme l'installa tion comports 31 capteurs (temperatu re, pression,
debit, puissance) connectes a un ordinateu r, nous pouvons evaluer
et isenl'~nfluence du melange sur les rendements volumetriq ues
tropJ.ques globaux et sur les coefflcien ts d'echange_ globaux au consuiresultats
denseur. Les tests menes avec le melange donnent les
vants (compares au Rl2)
+ 20 % pour la puissance therrnlque au condenseur
T
1,5 % pour le C.O.P.
- 2 1 pour le rendement volumetrlq ue
- 10 % pour le coefficien t d'echange global du condenmeme flux.
seur
a
12
A la f1n de ces tests , nous avons cr~~
succe ssivem ent des
fuite s respe ctivem ent a ).'ent ree de
l'~vaporateur a d~tente seche
(bass e press ion, phase liquid e et gaz)
et en parti e haute de la
boute ille oil le gaz est en equil ibre
avec le liqui ds. Nous avons pr~­
lev~ un ~chantillon d'un kilogr
amme de liqui ds de la boute ille avant
et apres chaqu e fuite pour conna itre
la
compo sant dU melan ge. Le table au Su1va conce ntrati on de chaqu e
nt montr e la conce ntrati on
massi que de chaqu e compo sant.
:Conc entra tion
:mass ique %
Au debut du: Apres la fuite a . Apres
la fuite en
test
· l'entr ee de l'e- ; haut de la boute
1lle:
vapor ateur
~Composant
A
1,54
1,54
1,25
~Composant
B
71,52
71,53
69,85
:comp osant
c
26,94
26,93
28,90
La charg e initi ale de la P.A.c . est
de 300 kg.
Les resul tats exp~r1mentaux montr ent
qu'il n'y a pas de varia tion de la conce ntrati on massi que quand
une fuite de 30 kg se produ it
a l'ent ree de l.'~vaporate
ur. La fuite en haut de la
boute ille a dure
jusqu 'a ce que du gaz trave rse le deten
deur. Entre 20 et 30 ~ de la
charg e total e ont ~te perdu s. Nous
pouvo ns done concl ure que les
fuite s ne modi fient pas la comp ositio
n de ce melan ge suffis amme nt
pour chang er les condi tions de foncti
onnem ent de la P.A.C ..
13
URE
TION OF NON AZEOTROPIC MIXT
THE FIRST INDUSTRIAL APP ICA
J.C. BLAI SE- T. OUTTO
DES ETUDES ~T RECHERCHES
ELECTRICITE DE FRANCE, DIRECTION
- FRANCE
B. P. N" l - 7 250 MORET SILOING
J.L. AMBROSINO
INSTIT T FRANCAIS DU PETROL[
Cedex - FRANCE
B.P. 311 - 92506 RUEIL-MALMAISON
1 - INTRODUCTION
labo rator ies or
d tests have been perfo rmed in
Many contr ols, simu latio ns a
te de Franc e
trici
Elec
tly,
equen
mixtu res. Cons
test- rigs on non azeo tropi c
and the company QUIRI have deciIFP)
(
le
Petro
du
;ais
Fran<
(EDF), the Insb tut
all their own top
n an indu stria l site, using
ded to build toget her a unit
a test in an
Such
ol.
contr
the
and
perat ion
techn ologi es to optim ize the
bilit y of
relia
the
re
measu
to
utely neces sary
indu stria l envir onme nt is abso
.
this new techn ology
Z - TEST INSTALLA TlON
energ y from a
on a heat pump which recov ers
Expe rimen ts were carri ed o t
6B'C. As we can see on
to
5B°C
from
water
arms
refri gerat 1ng insta llatio n and
th 1nst allat ion are :
figur e l, char acter istic s of
by elec tric motor s of
compressor~ powe red
- 2 open type recip roca ti g
r.p.m .,
1450
at
/h
m
496
is
lume
75 kW ; the total swep t v
ser,
- a perfe ct coun ter-fl ow c nden
to feed the dry-e x evap orato r.
- Elect ronic expan sion val es
cold sourc e
s on heat sourc e (amm onia), on
The heat pump inclu des 31 ensor
eters and
flowm
The
re).
mixtu
c
tropi
or non azeo
(wate r) and on refri gera nt (Rl
to a comp uter. The insta llacted
conne
are
uges
g
ure
the temp eratu re and press
with more deta ils in [1].
tion has alrea dy been descr ibe
14
3 - INVESTlGA TlONS METHOD
When we examined the measuremen ts obtained we were faced with all the
problems of getting precise measuremen ts in an industrial environmen
t (meat
salting factory) and particular ly the problem of stability. The instability
of
the heat pump in operation is mostly due to the variations of
the ammonia
condensing temperatur e in the refrigeran t evaporator . As a matter of
fact, the
activities of the salting vary during the day and this brings about
important
variations of the needs of refrigerati on and upsets the stability
of the heat
pump.
So, among all the obtained measuremen ts, we looked for operating
zones in
which the variations of the ammonia pressure are small. Then we could
study the
performanc es of the heat pump versus the temperatur e of the heat source.
On a stable operating source, we determined an average operating
point in
order to calculate all the performanc es.
To evaluate the influence of a non azeotropic mixture on the performanc
es
of a heat pump compared with RlZ we determined the following efficienci
es for
each fluid.
- Volumetric efficiency : ratio of the flow-rate of refrigeran t at the
suction point to the swept volume of the compressor .
- Isentropic efficiency : ratio of the power absorbed by an isentropic
compression of the refrigeran t to the power determined fl'Om the characteristics of the refrigeran t at the suction and the discharge point.
- Coefficien t of performanc e (COP) : ratio of the power Q obtained
at the
cold source to the electric power used by motors.
15
/
/.
as
- Overall coefficie nt of heat transfer of the condenser h defined
DTL
2
external surface of the condenser (m )
(kW)
condenser
the
by
rejected
power
thermal
from
Logarithm ic mean temperatu re differenc e (°C). It is obtained
ating
the logarithm ic mean temperatu re differenc es of the desuperhe
and condensin g area [2].
4 - TESTS RESULTS
versus ammoThe curves for the average COP and thermal power of heat pump
water are plotted on
nia pressure for the sam~ input and output temperatu res of
figure 2 and J.
of fact, simuThese results correspon d to the expected values. As a matter
increase the thermal
lation showed that the use of a ternary mixture would
equal to 12,5 bars.
power by 21 :l with the same COP under a ammonia pressure
drop on heat source,
The COP does not change because there is no temperatu re
r.
therefore no decrease of irreversi bilities on the evaporato
TABLE l - INCREASE OF THERMAL POWER AND COP COMPARED TO Rl2
Thermal power
c:n
COP (1;)
+ 19,1
+
11,5
11
10,5
PNH 3 (Bar)
1,5
+ 21,3
+ 20
+ 1,6
+
1,4
12
+ 23
+ 1,5
warm source on
We can al<o ob;erve the ;;mall influence of variation ;; of the
is a fluid
mixture
a
So
mixture.
c
azeotropi
non
with
obtained
the advantage s
working condition s
made to measure, but it keeps its advantage s even when the
differ from the initial ones.
- Volumetri c efficienc y
versus pressure
Figure 4 shows the variation s of volumetri c efficienc y
to the values
dose
are
obtained
values
the
that
notice
can
we
ratio. First
gap of pressure betusually given for this type of compresso r according to the
c efficiem; y
ween suction and discharge . For the same pressure ratio, volumetri
pure refriof
instead
mixture
c
azeotropi
non
use
we
when
~~
2
decreases about
may be due to a vagerant Rl2. This small decrease of volumetri c efficienc y
c mixture.
riation of the ratio Cp/Cv which is higher for the non azeotr.opi
- Isentropi c efficienc y
of the presThe variation s of isentropi c efficienc y for different values
that the
conclude
just
can
We
5.
figure
from
noticed
sure ratios can not be
and 72 ~: for ternary
mean isentropi c efficienc ies are 70 ~: for pure refrigera nt
non azeotropi c mixture.
16
So w~ can conclu de that the use of
a non a~eotropic mlxtu re does not
reduc e
isentr opic and volum etric efflcl encie
5 of an open type recip rocat ing
compr essor.
- Overa ll coeff icien t of heat trans
fer
Exper iment al resul ts (figur e 7) show
a small decre ase of the overa ll coefficien t of heat trans fer of the conde
nser when we use a non azeot ropic
mlxtu re.
On the heat pump equipp ed with a coiled
conde nser with bare pipes , the decre
ase
can be equal to 10 1~ for the same
densi ty of energ y.
But we can also compa re the overa
ll coeff icien t of heat trans fer with
the
same tempe rature of the water at
the input and the outpu t of the conde
nser. To
obtain the same tempe rature drop we
make the water flow- rate of the cold
sourc e
vary in propo rtion to the therm al
power rejec ted by refrig erant . In
this esse,
the overa ll coeff icien t of heat
trans fer is about the same with
R12 or non
azeot ropic mixtu re. For the examp
le shown on figure 6 the overa ll
coeff icien t
of heat trans fer is 916 W/m 2 °C for
2
Rl2 and 933 W/m °C for non azeot ropic
mixture.
5 - INFLUENCE OF lEAKAGES
After having evalu ated the perfor
mance s of a heat pump using a non
azeotropic mixtu re we studie d the influe
nce of refrig erant leakag e on the
mixtu re
comp ositio n.
For this purpo se, we create d two
leaka ges : one at the input of the
dry ex
evapo rator (low press ure, liquid
and gas phasi s) and the other at
the safety
valve of the receiv er where gas is
in equili brium with liqu•d (high
press ure).
In order to determ ine the varia tions
of conce ntrati on of all the compo
nents, we
took a sampl e of one kilogr am of liquid
from the receiv er before and after
each
leakag e.
A leakag e of JO kg occur red at
the input of the evapo rator, the
initia l
load was 300 kg and the leakag e lasted
4 hours .
The leakag e at the top of the receiv
er lasted 4 hours until gas crosse
d the
expan sion valve becau se reFrig erant
load of the heat pump became insuf
flcien t
for a good "or king. The refrig erant
loss is compr i5ed betwe en 55 and 90
kg, i.e
20 to 30 :' of the initia l load 300
kg.
The follow ing table (table 2) shows
the mass conce ntrati on of each compo
nent of the terna ry mixtu re at the
end of each leakag e.
TABLE 2 - MASS CONCENTRATION AFTER
EACH LEAKAGE
Mass
conce ntrati on
"
At the
begin ning
of test
After leakag e at
the input of the
evapo rator
After leakag e
at the top of
the receiv er
compo nent A
compo nent B
compo nent c
1,54
71,52
26,94
1,54
71,53
26,93
1,25
69,85
28,90
17
of the dry ex evapor ator does
We can notice that the leakage at the input
ents, These results obtaicompon
nt
differe
the
of
tration
concen
the
not modify
obtaine d in laborat o,ry
results
the
with
agree
ation
ned in an indust rial install
on a pipe, where a diphes is
(3]. So we can conclud e that the leakage s created
of a non azeotro pic mixture .
mixture flows, does not modify the compos ition
a bigger decreas e of the most
The leakage at the top of the receive r causes
volatil e compon ents A and B.
before and after the leakage s,
We know the concen tration of each compon ent
rant. Althoug h it is difrefrige
lost
of
but we do not know the exact quantit y
each compon ent in the lost refrige of
tration
concen
exact
the
know
to
ficult
to the concen tration of the mixture
rant, we can suppose that it is very close
refrige rant, with initial concen of
y
quantit
itself. Then if we add the lost
tration is about the same as the
tration , to the remaini ng quantit y, the concen
load.
initial
not modify the compos ition of
We can therefo re conclud e that leakage s do
ons of the heat pump.
conditi
working
the
change
to
enough
mixture
the
6 - CONCLUSION
series of measure ments obtaine d
In this publica tion we have present ed the
a:;:eotr opic mixture of refrige non
a
using
on the first indust rial heat pump
rants.
a ternary non a>eotro pic mixThe tests carried out on the heat pump with
us to verify the followi ng :
allowed
Rl2,
ture and then with a pure fluid, the
power and of COP with the
- The agreem ent of the increas e of the thermal
expecte d values.
c to the use of non azeotro pic
- The absence of mainten ance problem s specifi
mixtur e5.
- The very small influen ce of the
compos ition.
~:efrigerant
leakage s on the
mixture
a heat pump using a non azeoThese tests have showed the reliab ility of
ment.
environ
rial
tropic mixture in an indust
for RlZ,
This type of mixture can be a substit ute
layer.
zone
the
on
ce
influen
bad
its
of
because
contro lled
the use of which is
REFERENCES
NO ( IFP) - TORRE!LLES (QUIRI)
[ l) BLAISE J. C. - DUTTO T. (EO F) - CHERON AMBROSI
mixture
pic
An indust rial applica tion of non azeotro
VIENNE.
EZ
sion
Commis
IIR
1987
Graw Hill Book Company Inc 1950.
[2] Donald KERN - Process heat transfe r - Mac
results obtaine d with non
[3] BLAISE J.C- DUTTO T. - Some practic al
temper ature heat pump high
in
rants
azeotro pic mixture of refrige
PURDUE.
E2
sion
Commis
IIR
1986
18
COP
'Thermal power (kW)
600
..
500
400
:
/
• R12
R12
• Ternary mjx lura
Ternary mixture
300
10
11
12
13
14
10
11
12
13
14
P NH
Fi9Ure 2 -Thermal power versus ammonia pres!lt,J.rc
!!~
Volumetric eff1C1ency
3
(bars)
- COP versus ammon1a pressure
lsentroplc eHIC!CTlCY
100
100
75
75
+
" X XX
:tx'l£
X
X
><
X
*
X Nol"l azemropic milltl.ll'e
X Non azeotropiC m1xture
+ R12
+ R12
50
3
t+
3.4
50 L------------,,-----------~---
3.S
3.4
3,8
CompressiQI"l rm!o
CompreSSIOn
F1gure 4-
Volum('tru~::
aHicJency.
110
Overall coefficient of heat
!kW/m 2 .0 cl
Variation of the t;ernpcrilture
in the eondenser (°C)
transfe~
Pure rafngerant R12
100
1000
900
II
kW
: ·-==:r
ir----: o=-
Tern~ry mixtur~
= 507
R 12
'"''-II
60
A12
j
Ototal =413 kW
Non azeo~ropl!;; mixture ototal
90
800
ra~io
..-------- ----
w.,.,
..-----'_........---_..........-.......----
SOL-------.------,,------,--
----~~~
4
~gure
0 25
7-
Ov~::rall
0.60
0.76
Variation of the thermal
power rejected in Ihe
condenser Q/Q total
ccefficien't of heat tran!;:fcr of the condenser.
Figure 6
19
y
Variation of the "tem~:~erature of Fl12 and non azeo"tropic
m1Xture with the same temperatures of water at the
input and the output of !he condenser.