Investigation into the fouling & scaling of a

printed circuit heat exchanger

Project ObJectlves Background

To demonstrate that, with proper manage- PCHEs are made using a similar technique
ment of the system, the heat transfer effec- to electronic printed circuit boards, hence
tiveness of small bore Printed Circuit Heat their name, with flat metal plates chemically
Exchangers (PCHEs) is not impaired by: miilec to produce fluid flow passages. The
particulate fouling plates are then diffusion bonded into blocks,
which are in turn welded together to form a
crystallisation fouiing (scaling) heat exchanger core, occupying substan-
Potential Users tially less space than a conventional plate
unit.
Process and reiated industries
Research Organlsatlon The surface finish of components to be
diffusion bonded is vital, as it involves the
Energy Technology Unit
metal surfaces being held together at very
Department of Mechanical Engineering
high temperatures (close to melting point),
Heriot-Watt Universiiv
while interfacial crystal growth occurs. The
Edinburgh EH14 4AS
resulting bond is as strong as the parent
Tel: 031 449 51 11
metal.
Mr P A Kew
Supporting Organlsatlons The passages in a PCHE are typically 0.3 to
ICI Chemicals and Polymers Ltd 1.5 mm deep, leading to prime heat transfer
BP International surface densities, expressed in terms of
effective heat transfer area per unit total
Project Summary core volume, of 500 to 2,500 m2Im3. h i s is
Two programmes of tests were carried out several times higher than prime surface
at the Department of Mechanical Engin- densities in gasketed-plate exchangers and
eering, Heriot-Watt University, to evaluate an order of magnitude higher than normal
the susceptibility of a PCHE to particulate Chemlcally mllled sfalnless steel plate prime surface densities in shell-and-tube
and crystallisation fouling. The test facility as used In PCHEs exchangers. The laminar flow of the fluid
comprised a PCHE and a Double Pipe within the narrow channels leads to high
Heat Exchanger (DPHO, representativeof a heat transfer coefficients, resulting in an
typical industrial shell-and-tube unit, for In the second set of tests, cooling water extremely compact heat exchanger.
comparison of results. The units were sup- was treated specifically to induce scaling. In
plied in parallel by a clean process fluid this case, channels of the PCHE did be- Fouling of Heat Exchangers
loop and an open-system cooling water come blocked, with those adjacent to the Although PCHEs are accepted to be com-
loop. process water inlet, where temperatures pact and thermally effective, there is a
were highest, being affected most freq- perception that the very narrow and tortu-
in the first set of tests, strainers of differing uently. When the system was properly ous channels might be susceptible to
gauges were used to filter the fluid streams. managed, by adequate chemical treatment blockage by fouling deposits if used with
The results indicated that the PCHE was of the cooling water, no blockages any but the cleanest process and cooling
no more susceptible to particulate fouling occurred It was found that where channel fluids. This project was undertaken to
than the DPHE. It was also found that regu- blockage did occur, there was only a small evaluate this propensity and to demon-
lar backwashing of the heat exchanger and degradation of heat transfer performance, strate that, given good system manage-
the strainers maintained specified fluid flow as measured by effectiveness.Furthermore, ment, the heat transfer effectiveness of
rates and prevented blockage of PCHE citric acid was found to be partially effective PCHEs is not impaired by either particu-
channels by relatively large particles. in removing the scale formed. late or crystallisation (scaling) fouling.

%
5 . . . , 6
...given good system management, the heat transfer
... effectiveness of PCHEs is not impaired by either particulate
or crystallisation (scaling)fouling tt
*'
~

D...","." .IYI.D*Y.II
~
 Experimental Programme
Partlculate foullng
Three tests were carried out using strainers
of dinering gauge to filter the streams (Table 1).

The water Row rates for the tests were 15 Vmin

and 30 Vmin for the process fiuid and cool-
ing water respectively.Throughout the tests
the strainers were removed and cleaned
regularly to maintain the specified flow
rates

For the duration of the tests, the cooling

water was treated to inhibit corrosion,
minimise crystallisation and prevent bio-
logical growth, so that only the affects of par-
ticulate fouling were recorded

Prior to Test 1 the 'clean performance' of

the system was monitored and recorded.
The PCHE was backwashed following
Tests 1 and 2 and the performance of the
system monitored again to provide a stalting
point for the following test.

Results
The DPHE showed no significant changes
during the three tests, either in effective-
ness or in pressure drop across the unit,
indicating that no fouling had occurred.
Visual examination of the inner tube surface

I
corroborated this fact.
I For the PCHE the results were as follows:
far
Schematlc of test
*Testl:
Paltlculata Foullng Test laclllty The pressure drop across the PCHE
increased from 0.67 to 0.88 bar during the
Pariiculate fouling, or silting, occurs when The test facility comprised a hot fluid loop
three week test, indicating that 14% of the
solid particles from the fiuid stream are supplying a watedethylene glycol mixture to
channels had become blocked. However,
deposited on the heat transfer surface. simulate a clean process stream; and a
the effectiveness value of the unit remained
Most fluid streams contain particulate cooling water loop incorporating an open
constant, slowing that there was no
matter originating from a variety of sources. cooling tower.
adverse effect on overall heat transfer,
Silt has low adhesion to the heat transfer
indicating that the distribution of blocked
surface; the thickness of the deposit is The streams were passed in paraiiel
channels was even over the heat exchanger
therefore limited by the rate of removai as through the two test heat exchangers:
cross section. After backwashing the
fluid flow velocity increases through the a PCHE comprising a core of 11 plates, pressure drop returned to clean performance
restricted passage. each with 48 etched channels, arranged values. Examination of the debris removed
Crystalllsatlon Fouling in a counterfiow configuration. A futiher revealed a number of relatively large
10 plates were sandwiched either side of particles including leaf material, seeds,
Crystdiisation or precipitation fouling occurs the core to provide additional structural grass and dead insects.
when the saturation point of a particular strength.
compound within a fiuid is reached.
a DPHE constructed to represent a typi- Test 2:
Crystals will form in the supersaturated
fluid, either on the heat transfer surface cal industrial tube-and-shell heat The greatest rise in pressure drop across
itself, or on other crystals or particles within exchanger. This acted as a standard for the PCHE occurred during Test 2.A max-
the fluid creating a fouling layer. comparative assessment of fouling of the imum value of I . I bar was recorded, indi-
PCHE during the tests. cating 21% of the channels had blockaaes.
Crystals formed on and adhering to the
heat transfer surface often have strong
adhesion, so the rate of removal by fluid
flow through the heat exchanger is low, and
fouling will continue to build up with time.

Generally, the solubility of a compound

increases with temperature; however, there
are several, termed 'inverse solubility
salts', which exhibit a reduction in soiubility
with increasing temperature. Of these com-
pounds, calcium carbonate and calcium
phosphate were identified as being the
most likely to precipitate from the water
used in the tests carried out during this
nrnied.
 The scale was found to be composed prin-
cipally of calcium carbonate and calcium
phosphate. The heat exchanger was
cleaned by injecting citric acid into each
10 channel using a hypodermic syringe; how-
ever, because of the wavy shape of the
channels, the extent cf the blockages was
not known and it is likely that some deposit
0.9 remained in the channels after cleaning.

I Test 5:
0.8 The DPHE showed no significant variation
in pressure drop or effectiveness and no
scaling was apparent.
I
L 0.7 Pressure drop across the PCHE increased
L from 0.75 to 0.86 bar, but there was no
I significant fall in effectiveness. When the
headers were removed after the test, there
The step change in performance at the end was no visible evidence of scaling or
of Test 2 occurred after backwashing. particulate fouling, sc the cause of the
change in pressure drop remained unclear.

Following Test 5 citric acid was circulated

through the PCHE for 30 minutes, and this
procedure was repeated after an interval of
three days. Although some of the channels
remained blocked, the pressure drop at
commencement of Test 6 was slightly
CHE results of tests 1-3
improved over that at the start of Test 5,
suggesting that some descaling had been
achieved
There were improvements in pressure drop
value each time the strainers were cleaned.
1 Tabb2 Tests4-6sunvnary
I

I Test 6
The effectiveness of the PCHE decreased
TWjlnMbitor IDlyusanl Bkrc i TWI The effectiveness of the DPHE declined
siightiy during the test, indicating the
susceptibility of the PCHE to biockages by very siightiy and there were random
large particles and collections of particles. changes in pressure drop due to difficuities
After backwashing the performance of the in setiing the cooling water flow rate. A thin
PCHE returned t o clean performance iayer of slimy grey deposit was found at the
values. Amongst the debris removed by cooling water outiet, which dried to a white
backwashing was a single mat, approximately 36 I 120 powder.
15 mm diameter, comprising interwoven
insect wings and legs, seeds and dead 56 0-75 Pressure drop across the PCHE increased
frcm 0.73 to 0.89 bar but was accompanied
nematodes (microscopicworms). J 07 76292 by only a small decrease in effectiveness.
Test 3: When the headers were removed a Small
amount of particuiate fouiing was found at
The PCHE performed consistently during Res.. rs the iniet and a thin layer of slimy grey
Test 3 . There was no significant change in Test 4 deposit, like that in the DPHE, was deposit-
either pressure drop or effectiveness, ed in the region of the process fiuid iniet. All
indicating that little or no fouling had Tne ZIP-E sncv.ec nc s gn i c a i t ,ar a1 01
I3ress"re oroc 3' effect veiess an0 i c visible deposit was rerncved by cleaning, as
occurred. The addition of a stainless Steel described for Test 5.
mesh to the strainer improved the sca i g idasapozent
effectiveness of the strainer at removing Press-re orop across !ne PCnE nc'eisec Concluslons
fibrous particles and, by preventing large lrom 0 E8 :o 1.8 38' t:r'eSc9)Po 13 :o
particles from reaching the strainer basket, 25% cianne c x n a c e noi.e.er 1W.e
reduced the amount of times that the ,.as on, a srna roarcl o r n e'fec: ueness
strainer had to be cleaned.
Ate' me lesi :re net a i c 3.1 e: ieaoers The performance of the PCHE in terms of
Scallng ne'e reioLec ~ C T:le .rl : ? a c.1 a s-.. thermal effectiveness remained fairly
a nspeclon 10 De cave0 c-I -?ere ..as
Three tests were carried out with the
cooling water chemically treated tc induce
no t s D e sca ns a: !ne zcc nc .\ aier e1 - constant throughout ail six tests, indicating
that any scaling was evenly distributed
atic,gn a sma amo-1'. 3 par c + ale IC- - across the heat exchanger cross section.
scaling. It was not possible to control the r g .\as re.eaeo AI me C - I ~ I e-c !-e
water chemistry precisely throughout the core aro l i e qeace' %'lace .\ere 38'1, The first set of tests highlighted that the
test programme, but as the tests were coterec ob naro ,e C A r o o L e s eacr PCHE was no more susceptible to particulate
comparing the affects on the PCHE and approxna:e, Zrnrr n o a i e l ~a n.+e fouling than the DPHE providing that the
those on the DPHE this was considered to cf cnarie s here c c:.(e3 0 , !ne oeocs I follcwing system management was carried
be acceptable. Daily dose rates were as pan c- a', r !ne 'e5 01ci !ne crxess I .3 out:
foilows: inlet where the temperature was highest.
There were no deposits in the pvc pipe correct filtration of the fluids at inlet to
Nalflcc 8301D (Dispersant) 5ml the PCHE to prevent the entrance of large
connected to the outiet header, but it was
Nalfloc 7356 (Inhibitor) 15ml not clear if this was due to the material, or (relative to the channel Size) Particles Or
to its distance from the heat exchanger fibres which could block individual chan-
Biocide Hypo 200-500ml core. nels;
 When the cooling water was treated with
both inhibitor and dispersant to prevent
crystallisation of the solutes present, scale
was not formed. Careful management of
the water quality can prevent the channels
of a PCHE from becoming blocked.
Wlder Appllcatlon Potential
The development of the PCHE has been
driven by its potential to heip reduce the
size of chemical piant while still achieving a
given production output (process intensi-
fication) with all the attendant cost benefits
that brings.

The capital and installation cost of a com-

pact heat exchanger can be significantly
lower than that of a comparable sheil-
and-tube heat exchanger. This factor will
increase the number of situations where
heat recover) is considered cost effective,
and therefore implemented.

PCHEs have been used mainly in refrigera-

tion systems, for such tasks as water and
brine chillers, condensers, iiquid/suction
exchangers, oil coolers, economisers, water/
water exchangers, discharge gas desuper-
heaters and absorption cycle exchangers.

There are now a number of PCHE

regular backwashing of the PCHE and The resuits of the second set of tests indi-
installations in the UK chemicals industry,
cleaning of the strainers and steel mesh Of cate that scaling occurs most frequently including a 1 0 MW nitric acid piant
used): where temperatures are highest. Therefore, economiser with a core surface area of
when determining the water quality required 1,000 m2 and a mass of 10 tonnes. Other
These simple procedures maintain constant to avoid scaling and eventual blockage of installations at Associated Octel and ICI are
fluid flow rates through the heat exchanger, PCHE channels, the maximum water tem-
featured in Good Practice Case Studies 22
and prevent blockage of the PCHE perature in the heat exchanger must be and 109 respectively.
channels. considered, because this is of greater sig-
nificance than the bulk outlet temperature.
Where blocking by large particles occurs. it Chemical cleaning can remove scaling, so
can be removed successfully by back- long as channels have not become totally
washing. blocked, and should be carried out regular-
ly to avoid the build up of deposits.

For further copies of this publlcatlon or other Best Practice programme literature please contact BRECSU or ETSU.

For bulldlngs-related prolects: Enquiries Bureau, Building Research Energy Consetvation Support Unit (BRECSU), Building
Research Establishment, Garston, Watford WD2 7JR. Tel No: 0923 664258. Fax No: 0923 684787.

For industrial prolacts: Energy Efficiency Enquiries Bureau, ETSU, Harweli, Oxfordshire OX11 ORA. Tel No: 0235 436747.
Telex: 83135. Fax No: 0235 432923.

Information on particlpatlon in the Best Practice programme and on energy efficiency generally is also available from your
Regional Energy Efflciency Gffice.