See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/285108896

Design method for plate evaporators and condensers

Conference Paper · January 1995

CITATIONS READS

49 6,867

1 author:

Bernard Thonon
CEA
72 PUBLICATIONS 1,160 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

AHEAD advanced heat exchangers View project

ASTECH View project

All content following this page was uploaded by Bernard Thonon on 28 August 2018.

The user has requested enhancement of the downloaded file.

 DESIGN METHOD FOR PLATE EVAPORATORS AND CONDENSERS

B. THONON
CEA-Grenoble, DRN/DTP/STI/GRETh*
17 av des Martyrs, 38054 Grenoble, FRANCE

INTRODUCTION
The first applications for plate heat exchangers were for milk pasteurisation and afterward for heat
recovery in several industrial processes, but essentially in single-phase flows [1] and [2]. Classical
plate heat exchangers are gasketed, and their range of application is limited by pressure or
temperature, but with brazed or welded plate heat exchangers temperature from -180°C to + 500°C
and pressure up to 30 bars can be reached. This allows to use plate heat exchangers in industrial
processes either as evaporators or as condensers [3], [4] and [5], brazed or welded plate heat
exchangers are also used in ammonia cooling units [6] and [7] and in automotive air conditioning
[8] and [9]. Some large units are used in catalytic reforming processes [10] and [11].
These heat exchangers have high heat transfer performances, and allow a significant size reduction.
Therefore, as they are compact, the liquid hold-up is smaller than in conventional heat exchangers.
As plate heat exchangers works in parallel flow (co or countercurrent), the temperature enthalpy
diagram method can be applied to the selection of the optimal number or size of the channel, and to
assess the design method there is a need of a good knowledge of the heat transfer and pressure drop
laws in corrugated channels.

PLATE HEAT EXCHANGER DESCRIPTION

Heat exchanger geometry
A plate heat exchanger consists of a pack of corrugated plates, which creates the channels (Figure
1). The hot fluid and the cold fluid circulate in co or countercurrent in the different channels (Figure
2). Multipass plate heat exchangers exist, but for two-phase flow processes it is recommended to
use single-pass arrangement on the phase change side.

* Groupement ADEME/CEA pour la Recherche sur les Echangeurs Thermiques

Figure 1 : Schematic view of a plate heat exchanger Figure 2 :Countercurrent flow

Channel geometry
The channel is created by assembling two corrugated plates, and the way of packing the plates leads
to the fact that the mean gap between two plates is equal to the height of the corrugation. The plates
are characterised by the length and the width of the heat transfer surface, but also by the corrugation
angle (figure 3).

corrugation
angle

height
pitch
length pitch

width
Figure 3 : Description of the plate geometry

DESIGN METHOD
Introduction
The goal of a design method is to find the geometry which gives the required heat duty for an
acceptable pressure drop, and this with the lowest heat transfer surface. If we follow the evolution of
the heat rate and pressure drop in function of the number of channel, three cases exist (Figure 4) :
- Zone A : The heat exchanger is oversized and there is no pressure drop limitation.
- Zone B : The heat exchanger is undersized and there is no pressure drop limitation.
- Zone C : The heat exchanger is undersized and there is a pressure drop limitation.

Heat rate
Required
heat rate

zone C zone B zone A Number of channels

Pressure drop

Maximum
pressure drop

Number of channels

Figure 4 : Thermal and hydraulic performances in function of the number of channels

The optimal geometry should be located at the border between zone A and zone B.
As the heat exchanger geometry depends not only on the number of channels, but also on the
corrugation angle and on the plate size, a software is required if we want to test all the different
possible geometries.
The design method is based on the thermal checking of a given heat exchanger, and to search, by
changing the heat exchanger geometry, all the configurations which give the required heat duty for
an acceptable pressure drop [12].

The temperature-enthalpy diagram

The flow rates and the two inlet temperatures are fixed by the process, and for a given heat duty, the
outlet conditions can be calculated by an enthalpy balance.
& (h − h
1out ) = M 2 ( h2 out − h2in )
M & (1)
1 1in
The knowledge of the physical properties of each fluids leads to the buildings functions relating the
enthalpy to the temperature.
h = f( T ) T = f −1 ( h ) (2)
As we know the inlet and outlet thermal condition, the energy balance between the inlet and any
point x inside the exchanger is given by :
for a cocurrent heat exchanger
M& (h − h ) = M & (h − h ) (3)
1 1in 1x 2 2x 2 in
and for a countercurrent heat exchanger.
& (h − h
1 out ) = M 2 ( h2 x − h2 in )
M & (4)
1 1x
Since these relations are valid for all points inside the heat exchanger, the enthalpy evolution of one
fluid can be related to the other one. Therefore the temperature enthalpy diagram may be determined
through the f functions (Figure 5).

T1in
Temperature enthalpy of
phase-change
Vapour

Tsat1 Two-phase flow

T2out
Liquid dew
points
bubble
points
T1out
Vapour
Tsat2
Two-phase flow
Liquide
Enthalpy
T2in
Zone1 Zone2 Zone3 Zone4 Zone5

Enthalpy of phase-change

Figure 5 : Temperature/enthalpy diagram

Segmentation of temperature/enthalpy diagram
In single phase flows, the heat transfer coefficients are relatively constant along the heat exchanger,
and the segmentation can be chosen without any flow constraint. But in two-phase flows, there are
some strong variations of the heat transfer coefficient near the dew and bubble points. In
consequence, these particularly points needs to be consider as borders for the segmentation (figure
5).

The checking mode

The evolution curve is segmented in an appropriate number of cells. Then in each cell, limited by
the point x and x', the local energy balance is expressed :
M& (h − h ) = U A ∆T ln (5)
x x' xx'
Where U is the overall heat transfer coefficient and Axx' the heat transfer surface.
The mean log temperature difference is given by :
(T1 x − T2 x ) − (T1 x' − T2 x ' )
∆T ln = (6)
T1 x − T2 x
Ln
T1 x' − T2 x'
The overall heat transfer coefficient can be evaluated from the local heat transfer coefficients of
both fluids.
−1
 1 1 e 
U = + + w + Rf  (7)
 α1 α 2 λ w 

Since these quantities (α1 and α2) are known or can be evaluated from correlations, the required
heat transfer surface Axx' can be calculated. The required heat transfer surface Ar is determined by
the integration of (5) from the inlet to the outlet of the heat exchanger. The actual heat transfer
surface Ac is given by the number of channels and by the length and the width of the plates. Then,
the ratio of the required and actual heat transfer surface characterises the thermal sizing of the heat
exchanger.

Great care is to be taken in the choice of the fouling resistance Rf . It has been shown that plate heat
exchangers are less prone to fouling than conventional shell and tube heat exchanger due to the
higher level of the shear stress [13]. Furthermore, if the TEMA values are applied, the heat
exchanger will be overdesigned, and it may lead to poor operating conditions. It is often
recommended that the fouling margin should not exceed 25% of extra surface.

But to assess the design method there is a need of heat transfer and pressure laws, for the different
flow conditions (single-phase, evaporation and condensation).

HEAT TRANSFER
Single phase flow
Extensive information and literature have been summarised by Raju and Bansal [1] and Shah and
Focke [2]. These previous works have shown that the major geometric parameter is the corrugation
angle. For given flow conditions, the heat transfer coefficient can be 6 times greater than in a
comparable plain channel.
Some correlation have been established in function of the corrugation angle, but they cannot be
generalised to give a high degree of prediction because the shape of the corrugation and the
operating conditions have some influence [14].
The general form of these correlation is :
α dh
Nu = a Reb Pr c = (8)
λ
With 0.1 < a < 2.
0.6 < b < 0.8
0.3 < c < 0.4
The Reynolds and Prandtl numbers are defined by :
m& d h µ cp
Re = Pr =
µ λ
Generally, the plate heat exchanger manufacturers have a good knowledge of the thermal
performances in single phase flow.

Evaporation
In such corrugated channels, annular flow pattern dominates in a large range of applications, unless
the vapour quality is low [8] and [15]. This flow pattern, in evaporation, allows a high heat transfer
coefficient. As in tube, two phenomena occurs : nucleate boiling (nb) and convective boiling (cv).
Ohara [8] and Marvillet [16] have observed that nucleate boiling have no influence, unless for high
heat flux or low mass flux.
In nucleate boiling, the heat transfer coefficient is function of the heat flux and of fluid properties.
The Cooper correlation [17] is used to estimated the nucleate boiling term :
q&
α nb = 55 M −0.5 p * 0.12 q& 2 / 3 ( − log10 p*) −0.55 =
~
(9)
Tw − Tsat
To estimate the convective term, we use an enhancement factor F, which characterises the heat
transfer enhancement due to the liquid vapour interaction.
α cv = F α L (10)
Where the liquid heat transfer coefficient is calculated with an appropriate correlation for corrugated
channels.
c
 m& (1 − x ) d h   µ L c pL  λ L
b
αL =a    (11)
 µL   λ L  dh
The enhancement factor F is calculated with the Chen correlation, which was originally developed
for plain tubes.
18.
F = 1+ (12)
Xtt 0.79
Where Xtt is Lockhart-Martinelli number :
0.9 0.5 0.1
 1− x   ρG   µL 
Xtt =       (13)
 x   ρL   µG 

The two phase heat transfer coefficient is taken as the greater value of the nucleate boiling term and
the convective boiling term.
α = Max(α nb ,α cv )

The critical dryness quality, above the one the heat transfer coefficient drops, is at lower values than
in plain tubes, typically xc ≈ 0.7. For a pure vapour, the heat transfer coefficient can be estimated
from single phase correlation, thus the two phase heat transfer coefficient is estimated by
interpolation between the value at α xc =0 .7 and the value of a pure vapour αG0 .

α ( x > 0.7) = α ( x = 0.7 ) −

x−x
1− x
c α
c
(
( x = 0.7) − α GO ) (14)
With
c
 m& d h   µ G c pG  λ G
b
α G0 = a     (15)
 µ G   λ G  dh

This approach was originally used by Cornwell [18], and it seams to give estimation of our data.
Table 1 summarises the results for several type of plate heat exchangers.

Geometry Fluid Heat flux Mass velocity Validity

kW/m2 kg/m2 s % in ± 30%
Plain [16] R 22 10-60 200-600 77 %
Corrugated [14] R 22 5-60 200-900 85 %
Enhanced R 113 3-16 70-600 70 %
Table 1 : Comparison with experimental results

Condensation
Some local measurements of the heat transfer coefficient [4] and [19] have clearly shown the
existence of two regimes of condensation depending on the mass flux :
- At low mass flux ( Re L < 300 − 1000 ) : a gravity controlled regime, where the liquid flow may be
separate from vapour flow.
- At high mass flux : a shear controlled regime, where a turbulent flow model must be applied

For the gravity controlled regime, the Nusselt theory underpredicts by a factor 2 to 3 the heat
transfer coefficient [4] and [20]. This difference may comes from the fact that the liquid film is
drained in the furrow of the corrugations.

In the shear controlled regime, several correlations exist based on heat-momentum analogy, and one
of the most common is the Boyko-Kruzilihn correlation. The heat transfer coefficient in
condensation is reported to heat transfer coefficient of the liquid.
0.5
 ρ 
α = α L0 1 + x  L − 1  (16)
  ρG  
With
c
 m& d h   µ L c pL  λ L
b
α L0 = a     (17)
 µ L   λ L  dh

This correlation has been tested on several heat exchangers and seems to give some good results for
condensation of water steam [21], but for refrigeration application it was found to be less reliable.
The heat transfer coefficient is underestimated for low corrugation angle, but for high corrugation
angle the heat transfer coefficient is overestimated.
PRESSURE DROP
Single phase flow
The pressure drop, in corrugated channels, can be increased by a factor one hundred compared to a
equivalent plain tube, and this comes from the higher level of turbulence in the channel. The
pressure drop is calculated from :
m& 2 L
∆p = 4 f (18)
2 ρ dh
Where f is the channel friction factor, which is a function of the Reynolds number and which
depends on the corrugation angle. For a Reynolds number of 1000., the increased of the friction
factor is : f 30° / f plain = 9.2 and f 60° / f plain = 83
It must be noted that for a Reynolds number of 1000, the flow is in laminar regime in a plain
channel, while it is in turbulent regime for corrugated channels.

Two phase flow

The knowledge of the pressure drop, during evaporation or condensation, cannot be dissociated
from the heat transfer. On one hand, the pressure drop will give the mechanical energy needed to
flow the process fluid. On the other hand, the saturation temperature depends on the actual pressure.
As the pressure drop in corrugated channels can be important, the saturation temperature can varies
of several degrees between the inlet and the outlet, and affect the mean temperature difference
between the cold and hot fluid.
In two-phase flows, as in single phase flow, there a significant increase of the pressure drop
compared to a plain tube. The two-phase flow pressure drop is estimated from the pressure drops of
the liquid and gas flows.
∆p = ∆p L + C ∆p L ∆pG + ∆pG (19)
Where the liquid and gas pressure drop are calculated from the single-phase flow correlations:
m& 2 (1 − x ) 2 L
∆p L = 4 f L (20)
2 ρL dh
m& 2 x 2 L
∆pG = 4 f G (21)
2 ρ G dh
For plain tubes, the C constant depends on the flow conditions and varies from 5 to 20. For
corrugated channel the value C=8 allows a good estimation of the two-phase flow pressure drop
[14] (for a tube bundle C=8).
CONCLUSION
In single-phase flows, the overall performances of plate heat exchangers can be estimated with a
good accuracy by the proposed model. But in some critical working conditions, complex pass
arrangements, some more sophisticated models needs to be used to take into account flow
maldistribution or edge effects [22].
In two-phase flow applications (evaporation or condensation), the proposed model can be used for
the design of plate heat exchangers, but the accuracy will depend on reliability of the heat transfer
and pressure drop correlations. Some more work is needed to generalise the heat transfer coefficient
correlations in function of the geometrical and operating parameters, this is particularly true in
condensation [23]. Furthermore, a perfect flow distribution is assumed in the model, but in
evaporation processes the liquid/vapour mixture is not uniformly distributed between the channels,
and this will affect the overall performances.
Industrial process fluids are often mixtures and the temperature-enthalpy design method is still valid
as far as the boiling or condensation curve can be established. But some experimental work is
needed to develop and validate models to estimate the heat transfer coefficients for boiling and
condensation of mixtures in complex geometry.
NOMENCLATURE
A heat transfer surface m2 subscripts
C constant (eq. 17) c actual surface
cp heat transfer capacity J/kg K cr critical
dh hydraulic diameter m cv convective boiling
e plate thickness m G gas
F enhancement factor G0 gas only
f friction factor in inlet
h enthalpy J/kg L liquid
L length m L0 liquid only
M& mass flow rate kg/s nb nucleate boiling
~ molar mass g/mol out outlet
M
m& mass flux kg/s m2 r required surface
p pressure Pa sat saturation
p* reduce pressure Pa/Pa w wall
q& heat flux W/m2 1 hot fluid
Rf fouling resistance m2 K/W 2 cold fluid
T temperature K
U overall heat transfer coef. W/m2 K Non dimensional numbers
x quality Nu Nusselt number
α local heat transfer coef. W/m2 K Pr Prandtl number
λ thermal conductivity W/m K Re Reynolds number
µ dynamic viscosity Pa s Xtt Lockhart-Martinelli number
ρ density kg/m3

REFERENCES
[1] Raju K.S.N. et Bansal J.C. "Plate heat exchanger design", in Low Reynolds Number Flow
Heat Exchanger, Hemisphere Publishing Corporation, 1983
[2] Shah R.K. et Focke W.W. "Plate heat exchanger and their design theory", in Heat Transfer
Equipment Design, Hemisphere Publishing Corporation, 1988
[3] Syed A. "The use of plate heat exchangers as evaporators and condensers in process
refrigeration", Heat Exchange Engineering, E.A. Foumeny et P.J. Heggs ,Ellis Horwood Series in
Chemical Engineering, Volume 1 Design of Heat Exchangers, Chapter 10, 1992
[4] Kumar H. "The design of plate heat exchangers for refrigerants", Proceedings of the Institute
of Refrigeration 1991-92, 5-1 to 5.5, 1992
[5] Panchal C.B., Hillis D.L. et Thomas A. "Convective boiling of ammonia and freon R22 in
plate heat exchanger", ASME/JSME Thermal Engineering Conference Honolulu, pp 261-268, 1983
[6] Wand S. "Practical design tips for plate heat exchangers in ammonia refrigerants systems",
Proceedings of the IIAR 16th Annual Meeting, St Louis, Missouri, 1994
[7] Young M. "Plate heat exchangers as liquid cooling evaporators in ammonia refrigeration
systems", Proceedings of the IIAR 16th Annual Meeting, St Louis, 1994
[8] Ohara T., Yamamoto T. and Fujita H. "Heat transfer and pressure drop of boiling flow in
cross-ribbed flat channels", Int Com of Heat and Mass Transfer, Vol 17, n°5, pp 556-557, 1990
[9] Marvillet C. "Recent developments in heat exchanger for automotive applications", in
Recent developments in finned tube heat exchanger,DTI, Energy Technology, Denmark, 1993
[10] Fauconier J.C., Gorenflo F. and Thomas R., "A shell and plate heat exchanger for the
catalytic reforming unit at OMW refinery Karlsruhe", in Heat Exchanger Technology : Recent
Developments, EETI, France, 1994
[11] Grillot J.M. "Liquid side fouling of compact heat exchangers : first and second periodic
reports.", EEC Joule contract, JOUE-CT92-0033, 1994
 [12] Vidil R., Ratel G. et Grillot J.M. "Thermal hydraulic performance of plate and frame heat
exchangers - The CEPAJ Software", in Design and Operation of Heat Exchangers, Springer Verlag,
1992
[13] Thonon B., Grillot J.M. and Vidil R., "Liquid side fouling of plate heat exchangers",
proceedings of Fouling mitigation of industrial heat exchangers, June 1995, Shell Beach, California,
1995
[14] Thonon B., Vidil R., and Marvillet C. "Recent research and developments in plate heat
exchangers", Journal of Enhanced Heat Transfer, Vol 2, n° 1-2, pp 149-155, 1995
[15] Cohen M. and Carey V.P. " A comparison of the flow boiling performance characteristics of
partially heated cross-ribbed channels with different rib geometries", Int J of Heat & Mass Transfer,
Vol 32, n°12, pp 2459-2474, 1989
[16] Marvillet C. "Welded plate heat exchangers as refrigerants dry-ex evaporators", in Design
and Operation of Heat Exchangers , Springer-Verlag, 1992
[17] Cooper M.G. "Saturation nucleate pool boiling - A simple correlation", IChemE Symposium
Series, n° 86, pp 785-793, 1984
[18] Cornwell K. and Scoones D.S. "Analysis of low quality boiling on plain and low finned
tube", Proceedings of the 2nd UK Heat Transfer Conference, IMechE-IChemE, volume 1, 1988
[19] Navarro J.M. and Bailly A. "Compact brazed plate heat exchangers", in Heat Exchanger
Technology : Recent Developments, EETI, France, 1994
[20] Nakaoka T et Uehara H. "Performance test of a shell and plate type condenser for OTEC",
Experimental Thermal and Fluid Science, Vol 1, pp 275-281, 1988
[21] Tovazhnyanski L.L. et Kapustenko P.A. "Intensification of heat and mass transfer in
channels of plate condensers", Chem Eng Communication, Vol 31, pp 351-366, 1984
[22] Thonon B., Mercier P. and Feidt M. "Flow distribution in plate heat exchangers and
consequences on thermal and hydraulic performances", in Design and Operation of Heat
Exchangers, Springer Verlag, 1992
[23] Thonon B. "Echangeurs à plaques : dix ans de recherche au GRETh", Revue Générale de
Thermique, n°397, pp77-90, 1995

View publication stats