Sie sind auf Seite 1von 8

CHAPTER 2

Basic Principles
The accumulation of deposits on the surfaces of a heat exchanger increases the
overall resistance to heat flow. Fig. 2.1 illustrates how the temperature distribution
is affected by the presence of the individual fouling layers.

FIGURE 2.1. Temperature distribution across fouled heat exchanger surfaces

T1 and T6 represent the temperatures of the bulk hot and cold fluids respectively.
Under turbulent flow conditions these temperatures extend almost to the boundary
layer in the respective fluids since there is good mixing and the heat is carded
physically rather than by conduction as in solids or slow moving fluids. The
boundary layers (the regions between the deposit and the fluid), because of their
near stagnant conditions offer a resistance to heat flow. In general the thermal
conductivity of foulants is low unlike that of metals which are relatively high. For
these reasons, in order to drive the heat through the deposits relatively large
temperature differences are required, whereas the temperature difference across the
metal wall is comparatively low.
Fouling of Heat Exchangers

Thermal conductivities of some common foulant-like materials are given in


Table 2.1 that also includes data for common construction materials. The effects
of even thin layers of foulant may be readily appreciated.

TABLE 2.1

Some thermal conductivities of foulants and metals

Material Thermal conductivity


W/mK
Alumina 0.42
Biofilm (effectively water) 0.6
Carbon 1.6
Calcium sulphate 0.74
Calcium carbonate 2.19
Magnesium carbonate 0.43
Titanium oxide 8
Wax 0.24

Copper 400
Brass 114
Monel 23
Titanium 21
Mild steel 27.6

The resistance to heat flow across a solid surface is given as

x
-~- (2.1)

where x is the solid thickness

and 2 is the thermal conductivity of the particular solid.

Referring to the diagram (Fig. 2.1) the resistances of the solids to heat flow are:

For Deposit 1 x---Lwhere 2~ is the thermal conductivity of Deposit 1


21

For Deposit 2 xz where 22 is the thermal conductivity of Deposit 2


Basic Principles

and for the metal wall xm where ,;1,mis the thermal conductivity of the metal

For steady state conditions the heat flux q

q = T2-T3= T3-T4= T4-T5 (2.2)

Also q= a , ( ~ - g ) = a , ( ~ - ~) (2.3)

where a~ and tz2 are the heat transfer coefficients for the hot and cold fluids
respectively.

Equation (2.3) can be rewritten as

~-~_~-~
(2.4)

1 1
and represent the resistance to heat flow of the hot and cold fluids
al a 2
respectively.

The total resistance to heat flow will be the sum of the individual resistances, i.e.
R r the total thermal resistance will be given by

Rr = ( x ~ ) (+x ~ 2 2 ) (+ x ~ . ) + ~ 41 1
- (2.5)
a I a 2

The overall temperature driving force to accomplish the heat transfer between the
hot and cold fluids is the sum of the individual temperature differences

i.e. (~- r~) + ( ~ - ~ ) + ( ~ - ~ ) + (r, - ~ ) + ( ~ - ~)

or T I - T 6

.'.q- T~-T6 (2.6)


ev
10 Fouling o f Heat Exchangers

If the heat exchange area required for the required heat transfer is A, then the rate
of heat transfer Q

(2.7)
ev

In general the design of heat exchangers involves the determination of the


required area A. The necessary heat transfer, the temperatures and the fluids are
generally known from the process specification, the individual heat transfer
coefficients of the fluids may be calculated, and values of the fouling resistances on
either side of the heat exchanger would have to be estimated. It is the latter that
can be difficult and if the resistances are incorrectly estimated difficulties in
subsequent operation may be manifest.
At first sight it may be thought possible to calculate the fouling resistance, i.e.

x:12:
where x:is the deposit thickness

and 2r is the foulant thermal conductivity

The difficulty is however, that this involves a knowledge of the likely thickness
of the deposit laid down on the heat exchanger surfaces and the corresponding
thermal conductivity. In general these data are not available. It is therefore
necessary to assign values for the fouling resistance in order that the heat
exchanger may be designed.
An alternative way of writing Equation 2.6 for clean conditions when the heat
transfer surfaces are clean, is:

q= Uc(T~ - T6) (2.8)

where Uc represents the overall heat transfer coefficient for clean conditions, i.e.

(2.9)
gc +~-t -
al a2

and allowing for the fouling resistances on either side of the heat transfer surface

1 x~ x2 x,. + ~ + ~ (2.10)
uo +

Rewriting Equation 2.8 for fouled conditions, to give the heat flux
Basic Principles 11

q = Uo ( T~ - T6) (2.11)

Because the temperature driving force across the heat exchanger usually varies
along the length of the heat exchanger, it is necessary to employ some mean value
of the temperature difference in using Equations 2.8 and 2.11.
If ATt and AT2 are the temperature difference between the hot and cold fluids at
either end of the heat exchanger then the temperature difference may be taken as
the arithmetic mean, i.e.

AT,,, = ATe- AT2 (2.12)


2

but more usually the log mean temperature difference is used, i.e.

AT,,, = AT~- AT2 (2.13)


ln(A~

For more background to the use of mean temperature differences in the design
of heat exchangers the reader is referred to such texts as Hewitt, Shires and Bott
[1994].
The mean temperature difference may be substituted in Equation 2.11 to give
the heat flux

q = U o A T~ (2.14)

and if the total available heat transfer area A is taken into account

Q=UoAAT ~ (2.15)

In the design of heat exchangers A is usually unknown, so rearranging Equation


2.15 provides a means of estimating the required heat transfer area, i.e.

A= Q (2.16)

The choice of the individual fouling resistances for the calculation of Uo can
have a marked influence on the size of the heat exchanger and hence the capital
cost.
For a heat exchanger transferring heat from one liquid to another with the
individual liquid heat transfer coefficients of 2150 and 2940 W/mZK and fouling
12 Fouling of Heat Exchangers

resistances of 0.00015 and 0.0002 m2/WK on the surfaces of the heat exchanger,
the total resistance to heat flow is

1 1
0.00015 + 0.0002 m2/WK
2150 2940

= 0.00047 + 0.00034 + 0.00015 + 0.0002 m2/WK

= 0.00116 m2K/W

For the given design conditions, i.e. thermal load and temperature difference
this fouling resistance represents an increase in the required heat exchanger area
over and above the clean area requirements of

0.00035
x 100% = 43.2%
0.00081

i.e. the cost of the heat exchanger will be increased considerably due to the
presence of the fouling on the heat exchanger.

If the same fouling resistances are applied to a heat exchanger transferring heat
between two gases where the individual heat transfer coefficients are much lower
due to the low thermal conductivity of gases, say 32.1 and 79.2 W/m2K, the
situation is quite different.

Under these conditions the total thermal resistance is

1 1
+, + 0.00035 m2K/W
32.1 79.2

= 0.0312 + 0.0126 + 0.00035 m2K/W

= 0.0442 m2K/W

In these circumstances the increase in required area in comparison to the clean


conditions is

0.00035
x 100% = 0.8%
0.0438

For the liquid/liquid exchanger the choice of fouling resistances represents a


considerable increase in the required surface in comparison within the clean
Basic Principles 13

conditions. Using the same fouling resistances for a gas/gas heat exchanger
represents negligible additional capital cost.
The traditional method of designing heat exchangers is to consider the potential
fouling problem and assign a suitable fouling resistance to correspond. In order to
assist with this selection, organisations such as the Tubular Exchanger
Manufacturers Association (TEMA) issue tables of fouling resistances for special
applications. The first edition appeared in 1941. From time to time these data are
reviewed and revised. A review was carried out in 1988 and made available
[Chenoweth 1990].
The principal difficulty in this approach to design is the problem of choice. At
best the tables of fouling resistances give a range of mean fouling resistances, but
in general there is no information on the conditions at which these values apply.
For instance there is generally no information of fluid velocity, temperature or
nature and concentration of the foulant. As will be seen later these factors
amongst others, can have a pronounced effect on the development of fouling
resistance. Probably the largest amount of information contained in the tables is
concerned with water. Table 2.2 presents the relevant data published by TEMA
based on a careful review and the application of sound engineering acumen by a
group of knowledgeable engineers, involved in the design and operation of shell
and tube heat exchangers.

TABLE 2.2

Fouling resistances in water systems

Water type Fouling resistance


104 m2K/W
Sea water (43~ maximum outlet) 1.75-3.5
Brackish water (43~ maximum outlet) 3.5 - 5.3
Treated cooling tower water 1.75 - 3.5
(49~ maximum outlet)
Artificial spray pond (49~ maximum outlet) 1.75 - 3.5
Closed loop treated water 1.75
River water 3.5- 5.3
Engine jacket water 1.75
Distilled water or closed cycle condensate 0.9- 1.75
Treated boiler feedwater 0.9
Boiler blowdown water 3.5 - 5.3

Chenoweth [ 1990] gives the assumptions underlying the data contained in Table
2.2. For tubeside, the velocity of the stream is at least 1.22 m/s (4fl/s) for tubes of
non-ferrous alloy and 1.83 m/s (6 ft/s) for tubes fabricated from carbon steel and
other ferrous alloys. For shell-side flow the velocity is at least 0.61 m/s (2ft/s). In
14 Fouling of Heat Exchangers

respect of temperature it is assumed that the temperature of the surface on which


deposition is taking place does not exceed 71~ (160~ It is also assumed that
the water is suitably treated so that corrosion fouling and fouling due to biological
activity do not contribute significantly to the overall fouling. Chenoweth [1990]
comments on the further restrictions of these data. He observes that fouling by
treated water is known to be a function of the prevailing velocity, the surface
temperature, and the pH and is often characterised by reaching an asymptote (see
Fig. 1.2). Although asymptotic values could be identified in the tables, the typical
values listed for design, reflect a reasonable cleaning cycle and heat exchanger
operation without operating upset. The severe limitations imposed by the
assumptions will be readily appreciated.
It has to be said however, that without other information, these published data
are of value in making an assessment of the potential fouling resistance. At the
same time data on fouling resistances have to be treated with caution, they can only
be regarded as a guide. A further limitation is that these values only apply to shell
and tube heat exchangers. Conditions in plate heat exchangers for instance, could
be quite different.
A fundamental flaw in the use of fixed fouling resistances as suggested by the
TEMA tables is that they impose a static condition to the dynamic nature of
fouling. In fundamental terms the use of Equation 2.5 in conjunction with the
tables are not sound unless steady state has been reached. Fig. 1.2 shows that it is
only after the lapse of time that a steady fouling resistance is obtained. In other
words the heat exchanger does not suddenly become fouled when it is put on
stream. For a period of time the heat exchanger will over perform because the
overall resistance to heat flow is lower than that used in the design. To allow for
this overdesign the heat exchanger operator may adjust conditions that in
themselves could exacerbate the fouling problem. For instance, the velocity may
be reduced, in turn this could accelerate the rate of deposition. It is possible that
the imposed conditions could lead to fouling resistances that are subsequently,
greater than those used in the design with attendant operating difficulties. Effects
of this kind will be discussed in more detail later.

REFERENCES

Chenoweth, J., 1990, Final report of the HTRI/TEMA joint committee to review
the fouling section of TEMA standards. Heat. Trans. Eng. 11, No. 1, 73.
Hewitt, G.F., Shires, G.L. and Bott, T.R., 1994, Process Heat Transfer. CRC
Press, Boca Raton.

Das könnte Ihnen auch gefallen