CME 341

Heat Transfer
Lecture 11_1
Design of double pipe
and
shell and tube heat exchanger

Dr. Hameed Muhamad

Chemical Engineering Department

College of Engineering

Abu Dhabi University

Analysis and Design of Double-Pipe Exchangers
Use of the Seider–Tate or Hausen equations requires the calculation of the viscosity correction factor, Ø=(µ/µw)0.14

The detailed thermal analysis of a double-pipe heat exchanger involves the calculation of the overall heat-transfer
coefficient according to Equation (3.9). The inside coefficient, hi, can be computed using the equations for flow in
circular pipes presented in Chapter 2, i.e., Equation (2.34), (2.37), (2.38), or (2.39), depending on the flow regime.
The configuration is indicated schematically in Figure 3.16, in which Tave and tave are the average temperatures of
the fluids in the outer and inner pipes, respectively, and Tw is the average wall temperature.
An energy balance gives:

Solving for Tw yields:

Substituting A =πDL for the areas and rearranging, we obtain:

(3.18)
Use of Equation (3.18) involves an iterative procedure since Tw is required for the
calculation of hi and ho, and vice versa. However, a single iteration is usually sufficient.
The values of hi and ho are first computed by assuming that Øi and Øo are both unity.

These preliminary values are used in Equation (3.18) to compute Tw.

The values of µw for the two fluids are then obtained, and

the viscosity correction factors are calculated.

These factors are then multiplied by the preliminary values of hi and ho to obtain the
final values of the film coefficients.

The thermal analysis of a double-pipe exchanger is illustrated in the following example.

Example 3.3
10,000 lb/h of benzene is to be heated from 60F to 120F by heat exchange with an aniline stream that will be cooled from
150F to 100F. A number of 16ft hairpins consisting of 2-in. by 1.25-in. schedule 40 stainless steel pipe (type 316, k = 9.4
Btu/h . ft . F) are available and will be used for this service. Assuming that the benzene flows in the inner pipe, how many
hairpins will be required?
Solution
(a) Fluid properties at the average stream temperatures are obtained from Figures A.l and A.2, and Table A.15.
Note: The correction factor for entrance effects can usually be omitted from the Seider–Tate and Hausen
equations when working with industrial heat-transfer equipment.

For the inner pipe, the effect of the return bends on the heat transfer is minor, so it is reasonable (and
conservative) to use the entire length of the flow path in the correction term. Although this length is unknown in
the present instance, the exchanger will have at least one hairpin containing 32 ft of pipe. Hence,
 Therefore, entrance effects are negligible in this case.

Neglecting entrance effects,

Note: In the annulus the flow is disrupted at the return bends, so it is appropriate to use the length of pipe in one leg of
a hairpin to estimate entrance effects. Thus,
Hence, entrance effects are negligible. (Including the correction term in the Hausen equation gives
ho = 283 Btu/h . ft2 . F, and this does not alter the solution.)

(f) Calculate the pipe-wall temperature.

(h) Obtain fouling factors. A check of Table 3.3 shows no appropriate listings. However, these organic process
chemicals are expected to exhibit low fouling tendencies. Hence, we take

(i) Compute the overall heat-transfer coefficient.

(j) Calculate the required surface area and

number of hairpins.

From Table B.2, the external surface area per foot of 1.25-in. schedule 40 pipe is 0.435 ft2/ft. Therefore,
Important Design Criteria for fluid placement of Double-Pipe Exchangers
The preceding example constitutes a design problem in which all of the design parameters are specified except one,
namely, the total length of the heat exchanger.

Design problems frequently include specifications of maximum allowable pressure drops on the two streams.

In that case, pressure drops for both streams would have to be calculated in order to determine the suitability of the
exchanger.

In the above example, benzene was specified as the tube-side fluid. Some guidelines for positioning the fluids are
given in Table 3.4. It should be understood that these general guidelines, while often valid, are not ironclad rules,
and optimal fluid placement depends on many factors that are service specific.
For the example problem, neither fluid is corrosive to stainless steel or highly fouling, so the first applicable
criterion on the list for tube-side fluid is viscosity.

Benzene has a significantly lower viscosity than aniline, indicating that it should be placed in the inner pipe.
However, this criterion has relatively low priority and is actually relevant to finned-pipe or shell-and-tube
exchangers where the fins or tubes act as turbulence generators for the shell-side fluid.

The next criterion is stream pressure, which was not specified in the example.

The last criterion for tube-side fluid is the hotter fluid, which is aniline. Neither criterion for the shell-side fluid
is applicable.

The upshot is that either fluid could be placed in the inner pipe, and both options should be investigated to
determine which results in the better design.
Design of Shell-and-Tube Exchangers
The complete thermal design of a shell-and-tube heat exchanger is a complex and lengthy process, and is usually
performed with the aid of a computer program .

Frequently, however, one requires only a rough approximation of the heat-transfer area for the purpose of making a
preliminary cost estimate of the exchanger.

Tabulations of overall heat-transfer coefficients such as the one presented in Table 3.5 are used for this purpose.

One simply estimates a value for the overall coefficient based on the tabulated values and then computes the
required heat-transfer area from Equation (3.1).

A somewhat better procedure is to estimate the individual film coefficients, hi and ho, and use them to compute the
overall coefficient by Equation (3.9).
Example 3.4
In a petroleum refinery, it is required to cool 30,000 lb/h of kerosene from 400 F to 250 F by heat exchange with 75,000
lb/h of gas oil, which is at 110 F. A shell-and-tube exchanger will be used, and the following data are available:

For the purpose of making a preliminary cost estimate, determine the required heat-transfer area of the exchanger.

Solution
(a) Calculate the heat load and outlet oil temperature by energy balances on the two streams.

(b) Calculate the LMTD.

 (c) Calculate the LMTD correction factor.

For the purpose of this calculation, assume that the

kerosene will flow in the shell. This assumption will
not affect the result since the value of F is the same,
regardless of which fluid is in the shell.
400 F

182 F

110 F
250 F

(d) Estimate UD.

From Table 3.5, a kerosene-oil exchanger should have an
overall coefficient in the range 20 to 35 Btu/h.ft2 .F. Therefore,
take

(e) Calculate the required area.

Rating a Shell-and-Tube Exchanger
The thermal analysis of a shell-and-tube heat exchanger is similar to the analysis of a double-pipe exchanger in that an
overall heat transfer coefficient is computed from individual film coefficients, hi and ho. However, since the flow patterns
in a shell-and-tube exchanger differ from those in a double-pipe exchanger, the procedures for calculating the film
coefficients also differ.

The shell-side coefficient presents the greatest difficulty due to the very complex nature of the flow in the shell.

In addition, if the exchanger employs multiple tube passes, then the LMTD correction factor must be used in calculating
the mean temperature difference in the exchanger.

In computing the tube-side coefficient, hi, it is assumed that all tubes in the exchanger are exposed to the same thermal
and hydraulic conditions.
The value of hi is then the same for all tubes, and the calculation can be made for a single tube. Equation (2.34), (2.37),
(2.38), or (2.39) is used, depending on the flow regime.
In computing the Reynolds number, however, the mass flow rate per tube must be used, where

(3.19)

Equation (3.19) simply states that the flow rate in a single tube is the total flow rate divided by the number of fluid circuits,
which is nt/np. With the exception of this minor modification, the calculation of hi is the same as for a double-pipe exchanger.
The method utilizes the graph of modified Colburn factor, jH, versus shell-side Reynolds number shown in Figure 3.17.
All symbols used on the graph are defined in the key.

The graph is valid for single segmental baffles with a 20% cut, which is within the range typically used in practice. It is
also based on TEMA standards for tube-to-baffle and baffle-to-shell clearances.

To use Figure 3.17, one simply reads jH from the graph and then computes ho from:

(3.20)

An approximate curve fit to Figure 3.17, which is convenient for implementation on a programmable calculator or
computer, is:

(3.21)
It should again be noted that the rating procedure given in Figure 3.18 includes only the thermal analysis of the
exchanger. A complete rating procedure must include a hydraulic analysis as well. That is, the pressure drops of both
streams must be computed and compared with the specified maximum allowable pressure drops. The tube-side
pressure drop is readily computed by the friction factor method for pipe flow with appropriate allowances for the
additional losses in the headers and nozzles.
Example 3.5
30,000 lb/h of kerosene are to be cooled from 400F to 250F by heat exchange with 75,000 lb/h of gas oil which is at 110F.
Available for this duty is a shell-and-tube exchanger having 156 tubes in a 21 1/4 -in ID shell. The tubes are 1-in. OD, 14
BWG, 16 ft long on a 1 1/4 -in. square pitch. There is one pass on the shell side and six passes on the tube side. The
baffles are 20% cut segmental type and are spaced at 5-in. intervals. Both the shell and tubes are carbon steel having k =
26 Btu/h.ft . F. Fluid properties are given in Example 3.4. Will the exchanger be thermally suitable for this service?

Solution
Neither fluid is corrosive, but the oil stream may cause
fouling problems so it should be placed in the tubes for
ease of cleaning. Also, the kerosene should be placed in
the shell due to its large ΔT.

Step 1: Calculate Ureq.

From Example 3.4, we have:

The surface area is obtained from the dimensions of the exchanger:

Thus,
External surface area from the table
0.2618
 Step 2: Calculate the clean overall coefficient, UC. We
need hi and ho
(a) Calculate the tube-side Reynolds
number.

(b) Calculate hi

In this calculation, the viscosity correction factor was assumed to be unity since no data were given for the temperature
dependence of the oil viscosity.

(c) Calculate the shell-side Reynolds number.

ds= 21 1/4 -in ID shell Baffles spaced B= 5-in.

Intervals
PT = tube pitch in =1.25
(d) Calculate ho.

(e) Calculate UC.

Step 3: Obtain the required fouling factors.

In the absence of other information, Table 3.3 indicates fouling factors of 0.002–0.003 h . ft2 .F/Btu for kerosene and
0.002– 0.005 h . ft2 . F/Btu for gas oil. Taking 0.0025 for kerosene and 0.0035 for gas oil gives:
Step 4: Calculate UD.

Since this value is greater than the required value of 25.3 Btu/h . ft2 . F,

the exchanger is thermally suitable. In fact, a smaller exchanger would be adequate.

In the above example, the average tube-wall temperature was not required because the variation of fluid
properties with temperature was ignored.

In general, however, the wall temperature is required and is calculated using Equation (3.18) in the same manner
as illustrated for a double-pipe exchanger in Example 3.3.