756 CHEMICAL ENGINEERING

12.12. PLATE HEAT EXCHANGERS

12.12.1. Gasketed plate heat exchangers
A gasketed plate heat exchanger consists of a stack of closely spaced thin plates clamped
together in a frame. A thin gasket seals the plates round their edges. The plates are
normally between 0.5 and 3 mm thick and the gap between them 1.5 to 5 mm. Plate
surface areas range from 0.03 to 1.5 m2 , with a plate width:length ratio from 2.0 to
3.0. The size of plate heat exchangers can vary from very small, 0.03 m2 , to very large,
1500 m2 . The maximum flow-rate of fluid is limited to around 2500 m3 /h.
The basic layout and flow arrangement for a gasketed plate heat exchanger is shown in
Figure 12.60. Corner ports in the plates direct the flow from plate to plate. The plates are
embossed with a pattern of ridges, which increase the rigidity of the plate and improve
the heat transfer performance.
Plates are available in a wide range of metals and alloys; including stainless steel,
aluminium and titanium. A variety of gasket materials is also used; see Table 12.8.

Selection
The advantages and disadvantages of plate heat exchangers, compared with conventional
shell and tube exchangers are listed below:

Advantages
1. Plates are attractive when material costs are high.
2. Plate heat exchangers are easier to maintain.

Figure 12.60. Gasketed plate heat exchanger

 HEAT-TRANSFER EQUIPMENT 757
Table 12.8. Typical gasket materials for plated heat exchangers
Material Approximate temperature Fluids
limit, ° C
Styrene-butane rubber 85 Aqueous systems
Acrylonitrile-butane rubber 140 Aqueous system, fats,
aliphatic hydrocarbons
Ethylene-propylene rubber 150 Wide range of chemicals
Fluorocarbon rubber 175 Oils
Compressed asbestos 250 General resistance to organic
chemicals

3. Low approach temps can be used, as low as 1 Ž C, compared with 5 to 10 Ž C for

shell and tube exchangers.
4. Plate heat exchangers are more flexible, it is easy to add extra plates.
5. Plate heat exchangers are more suitable for highly viscous materials.
6. The temperature correction factor, Ft , will normally be higher with plate heat
exchangers, as the flow is closer to true counter-current flow.
7. Fouling tends to be significantly less in plate heat exchangers; see Table 12.9.

Disadvantages
1. A plate is not a good shape to resist pressure and plate heat exchangers are not
suitable for pressures greater than about 30 bar.
2. The selection of a suitable gasket is critical; see Table 12.8.
3. The maximum operating temperature is limited to about 250 Ž C, due to the perfor-
mance of the available gasket materials.
Plate heat exchangers are used extensively in the food and beverage industries, as
they can be readily taken apart for cleaning and inspection. Their use in the chemical
industry will depend on the relative cost for the particular application compared with a
conventional shell and tube exchanger; see Parker (1964) and Trom (1990).

Table 12.9. Fouling factors (coefficients), typical values for plate heat exchangers

Fluid Coefficient (W/m2 ° C) Factor (m2 ° C/W)

Process water 30,000 0.00003
Towns water (soft) 15,000 0.00007
Towns water (hard) 6000 0.00017
Cooling water (treated) 8000 0.00012
Sea water 6000 0.00017
Lubricating oil 6000 0.00017
Light organics 10,000 0.0001
Process fluids 5000 20,000 0.0002 0.00005

Plate heat exchanger design

It is not possible to give exact design methods for plate heat exchangers. They are propri-
etary designs, and will normally be specified in consultation with the manufacturers.
Information on the performance of the various patterns of plate used is not generally
758 CHEMICAL ENGINEERING

available. Emerson (1967) gives performance data for some proprietary designs, and
Kumar (1984) and Bond (1980) have published design data for APV chevron patterned
plates.
The approximate method given below can be used to size an exchanger for comparison
with a shell and tube exchanger, and to check performance of an existing exchanger for
new duties. More detailed design methods are given by Hewitt et al. (1994) and Cooper
and Usher (1983).

Procedure
The design procedure is similar to that for shell and tube exchangers.
1. Calculate duty, the rate of heat transfer required.
2. If the specification is incomplete, determine the unknown fluid temperature or fluid
flow-rate from a heat balance.
3. Calculate the log mean temperature difference, TLM .
4. Determine the log mean temperature correction factor, Ft ; see method given
below.
5. Calculate the corrected mean temperature difference Tm D Ft ð TLM .
6. Estimate the overall heat transfer coefficient; see Table 12.1.
7. Calculate the surface area required; equation 12.1.
8. Determine the number of plates required D total surface area/area of one plate.
9. Decide the flow arrangement and number of passes.
10. Calculate the film heat transfer coefficients for each stream; see method given
below.
11. Calculate the overall coefficient, allowing for fouling factors.
12. Compare the calculated with the assumed overall coefficient. If satisfactory, say

0% to C 10% error, proceed. If unsatisfactory return to step 8 and increase or
decrease the number of plates.
13. Check the pressure drop for each stream; see method given below.
This design procedure is illustrated in Example 12.13.

Flow arrangements
The stream flows can be arranged in series or parallel, or a combination of series and
parallel, see Figure 12.61. Each stream can be sub-divided into a number of passes;
analogous to the passes used in shell and tube exchangers.

Estimation of the temperature correction factor

For plate heat exchangers it is convenient to express the log mean temperature difference
correction factor, Ft , as a function of the number of transfer units, NTU, and the flow
arrangement (number of passes); see Figure 12.62. The correction will normally be higher
for a plate heat exchanger than for a shell and tube exchanger operating with the same
temperatures. For rough sizing purposes, the factor can be taken as 0.95 for series flow.
 HEAT-TRANSFER EQUIPMENT 759

(a) Series flow

(b) Looped ( parallel ) flow

(c) 2 Pass / 2 Pass

5 Channels per pass
19 Thermal plates
21 Plates total
Counter-current flow

Figure 12.61. Plate heat-exchanger flow arrangements

The number of transfer units is given by:

NTU D t0
ti /TLM
where ti D stream inlet temperature,Ž C,
t0 D stream outlet temperature,Ž C,
TLM D log mean temperature difference,Ž C.
Typically, the NTU will range from 0.5 to 4.0, and for most applications will lie between
2.0 to 3.0.

Heat transfer coefficient

The equation for forced-convective heat transfer in conduits can be used for plate heat
exchangers; equation 12.10.
The values for the constant C and the indices a,b,c will depend on the particular type
of plate being used. Typical values for turbulent flow are given in the equation below,
760 CHEMICAL ENGINEERING

1.00

0.95
Correction factor, Ft

1.1 4.4

0.90

3.3

0.85
2.2

0.80
0 1 2 3 4 5 6
NTU

Figure 12.62. Log mean temperature correction factor for plate heat exchangers (adapted from Raju and Chand
(1980))

which can be used to make a preliminary estimate of the area required.

hp de
D 0.26Re0.65 Pr 0.4 /w 0.14 12.77
kf

where hp D plate film coefficient,

Gp de up de
Re D Reynold number D D
 
GP D mass flow rate per unit cross-sectional area D w/Af , kgm
2 s
1 ,
w D mass flow rate per channel, kg/s,
Af D cross-sectional area for flow, m2 ,
up D channel velocity, m/s,
de D equivalent (hydraulic) diameter, taken as twice the gap between
the plates, m.
The corrugations on the plates will increase the projected plate area, and reduce the
effective gap between the plates. For rough sizing, where the actual plate design is not
known, this increase can be neglected. The channel width equals the plate pitch minus
the plate thickness.
There is no heat transfer across the end plates, so the number of effective plates will
be the total number of plates less two.
 HEAT-TRANSFER EQUIPMENT 761

Pressure drop
The plate pressure drop can be estimated using a form of the equation for flow in a
conduit; equation 12.18.
up2
Pp D 8jf Lp /de  12.78
2
where LP D the path length and up D Gp /.
The value of the friction factor, jf , will depend on the design of plate used. For
preliminary calculations the following relationship can be used for turbulent flow:
jf D 0.6 Re
0.3
The transition from laminar to turbulent flow will normally occur at a Reynolds number of
100 to 400, depending on the plate design. With some designs, turbulence can be achieved
at very low Reynolds numbers, which makes plate heat exchangers very suitable for use
with viscous fluids.
The pressure drop due the contraction and expansion losses through the ports in the
plates must be added to the friction loss. Kumar (1984) suggests adding 1.3 velocity heads
per pass, based on the velocity through the ports.
2
upt 
Ppt D 1.3 Np 12.79
2
where upt D the velocity through the ports w/Ap , m/s,
w D mass flow through the ports, kg/s,
Ap D area of the port D d2pt /4, m2 ,
dpt D port diameter, m,
Np D number of passes.

Example 12.13
Investigate the use of a gasketed plate heat exchanger for the duty set out in Example 12.1:
cooling methanol using brackish water as the coolant. Titanium plates are to be specified,
to resist corrosion by the saline water.

Summary of Example 12.1

Cool 100,000 kg/h of methanol from 95Ž C to 40Ž C, duty 4340 kW. Cooling water inlet
temperature 25Ž C and outlet temperature 40Ž C. Flow-rates: methanol 27.8 kg/s, water
68.9 kg/s.
Physical properties: Methanol Water
Density, kg/m3 750 995
Viscosity, mN m
2 s 3.4 0.8
Prandtl number 5.1 5.7
Logarithmic mean temperature difference 31Ž C.
762 CHEMICAL ENGINEERING

Solution
NTU, based on the maximum temperature difference
95
40
D D 1.8
31
Try a 1 : 1 pass arrangement.
From Figure 12.62, Ft D 0.96
From Table 12.2 take the overall coefficient, light organic - water, to be 2000 Wm
2Ž C
1 .
4340 ð 103
Then, area required D D 72.92 m2
2000 ð 0.96 ð 31
Select an effective plate area of 0.75 m2 , effective length 1.5 m and width 0.5 m; these
are typical plate dimensions. The actual plate size will be larger to accommodate the
gasket area and ports.

Number of plates D total heat transfer area / effective area of one plate
D 72.92/0.75 D 97
No need to adjust this, 97 will give an even number of channels per pass, allowing for
an end plate.
Number of channels per pass D 97
1/2 D 48
Take plate spacing as 3 mm, a typical value, then:
channel cross-sectional area D 3 ð 10
3 ð 0.5 D 0.0015 m2
and hydraulic mean diameter D 2 ð 3 ð 10
3 D 6 ð 10
3 m

Methanol
27.8 1 1
Channel velocity D ð ð D 0.51 m/s
750 0.0015 48
up de 750 ð 0.51 ð 6 ð 10
3
Re D D D 6750
 0.34 ð 10
3
Nu D 0.2667500.65 ð 5.10.4 D 153.8 12.77
hp D 153.80.19/6 ð 10
3  D 4870 Wm
2Ž C
1

Brackish water
68.9 1 1
Channel velocity D ð ð D 0.96 m/s
995 0.0015 48
955 ð 0.96 ð 6 ð 10
3
Re D D 6876
0.8 ð 10
3
Nu D 0.2668760.65 ð 5.70.4 D 162.8 12.77

3
2Ž
1
hp D 162.80.59/6 ð 10  D 16,009 Wm C
 HEAT-TRANSFER EQUIPMENT 763

Overall coefficient
From Table 12.9, take the fouling factors (coefficients) as: brackish water (seawater)
6000 Wm
2Ž C
1 and methanol (light organic) 10,000 Wm
2Ž C
1 .
Take the plate thickness as 0.75 mm. Thermal conductivity of titanium 21 Wm
1Ž C
1 .
1 1 1 0.75 ð 10
3 1 1
D C C C C
U 4870 10,000 21 16,009 6000
U D 1754 Wm
2Ž C
1 , too low

Increase the number of channels per pass to 60; giving 2 ð 60 C 1 D 121 plates.
Then, methanol channel velocity D 0.51 ð 48/60 D 0.41 m/s, and Re D 5400.
Cooling water channel velocity D 0.96 ð 48/60 D 0.77 m/s, and Re D 5501.
Giving, hp D 4215 Wm
2Ž C
1 for methanol, and 13,846 Wm
2Ž C
1 for water.
Which gives an overall coefficient of 1634 Wm
2Ž C
1 .
Overall coefficient required 2000 ð 48/60 D 1600 Wm
2Ž C
1 , so 60 plates per pass
should be satisfactory.

Pressure drops
Methanol
Jf D 0.605400
0.3 D 0.046
Path length D plate length ð number of passes D 1.5 ð 1 D 1.5 m.


1.5 0.412
Pp D 8 ð 0.046 ð 750 ð D 5799 N/m2 12.78
6 ð 10
3 2
Port pressure loss, take port diameter as 100 mm, area D 0.00785 m2 .
Velocity through port D 27.8/750/0.00785 D 4.72 m/s.
750 ð 4.722
Ppt D 1.3 ð D 10,860 N/m2 12.79
2
Total pressure drop D 5799 C 10,860 D 16,659 N/m2 , 0.16 bar.

Water
Jf D 0.65501
0.3 D 0.045
Path length D plate length ð number of passes D 1.5 ð 1 D 1.5 m.


1.5 0.772
Pp D 8 ð 0.045 ð ð 995 ð D 26,547 N/m2 12.78
6 ð 10
3 2
Velocity through port D 68.9/995/0.0078 D 8.88 m/s (rather high)
995 ð 8.88
Ppt D 1.3 ð D 50,999 N/m2 12.79
2
764 CHEMICAL ENGINEERING

Total pressure drop D 26,547 C 50,999 D 77,546 N/m2 , 0.78 bar

Could increase the port diameter to reduce the pressure drop.
The trial design should be satisfactory, so a plate heat exchanger could be considered
for this duty.

12.12.2. Welded plate

Welded plate heat exchangers use plates similar to those in gasketed plate exchangers but
the plate edges are sealed by welding. This increases the pressure and temperature rating
to up to 80 bar and temperatures in excess of 500Ž C. They retain the advantages of plate
heat exchangers (compact size and good rates of heat transfer) whilst giving security
against leakage. An obvious disadvantage is that the exchangers cannot be dismantled
for cleaning. So, their use is restricted to specialised applications where fouling is not a
problem. The plates are fabricated in a variety of materials.
A combination of gasketed and welded plate construction is also used. An aggressive
process fluid flowing between welded plates and a benign process stream, or service
stream, between gasketed plates.

12.12.3. Plate-fin
Plate-fin exchangers consist essentially of plates separated by corrugated sheets, which
form the fins. They are made up in a block and are often referred to as matrix exchangers;
see Figure 12.63. They are usually constructed of aluminium and joined and sealed by
brazing. The main application of plate-fin exchangers has been in the cryogenics indus-
tries, such as air separation plants, where large heat transfer surface areas are needed.
They are now finding wider applications in the chemical processes industry, where large
surface area, compact, exchangers are required. Their compact size and low weight have
lead to some use in off-shore applications. The brazed aluminium construction is limited
to pressures up to around 60 bar and temperatures up to 150Ž C. The units cannot be
mechanically cleaned, so their use is restricted to clean process and service steams. The

Figure 12.63. Plate-fin exchanger

 HEAT-TRANSFER EQUIPMENT 765
construction and design of plate-fin exchangers and their applications are discussed by
Saunders (1988) and Burley (1991), and their use in cryogenic service by Lowe (1987).

12.12.4. Spiral heat exchangers

A spiral heat exchanger can be considered as a plate heat exchanger in which the plates
are formed into a spiral. The fluids flow through the channels formed between the plates.
The exchanger is made up from long sheets, between 150 to 1800 mm wide, formed into
a pair of concentric spiral channels. The channels are closed by gasketed end-plates bolted
to an outer case. Inlet and outlet nozzles are fitted to the case and connect to the channels,
see Figure 12.64. The gap between the sheets varies between 4 to 20 mm; depending on
the size of the exchanger and the application. They can be fabricated in any material that
can be cold-worked and welded.

Figure 12.64. Spiral heat exchanger

Spiral heat exchangers are compact units: a unit with around 250 m2 area occupying
a volume of approximately 10 m3 . The maximum operating pressure is limited to 20 bar
and the temperature to 400Ž C.
For a given duty, the pressure drop over a spiral heat exchanger will usually be lower
than that for the equivalent shell-and-tube exchanger. Spiral heat exchangers give true
counter-current flow and can be used where the temperature correction factor Ft for a
shell-and-tube exchanger would be too low; see Section 12.6. Because they are easily
cleaned and the turbulence in the channels is high, spiral heat exchangers can be used for
very dirty process fluids and slurries.
The correlations for flow in conduits can be used to estimate the heat transfer coefficient
and pressure drop in the channels; using the hydraulic mean diameter as the characteristic
dimension.
The design of spiral heat exchangers is discussed by Minton (1970)