Advances in Chemical

Engineering
Chapter 4

Computer Aided Design of Shell and Tube

Heat Exchangers (Incorporating Most
Recent Developments)
C M Narayanan1*; Sneha Bhadra; Mukul Kanta Das2
1
National Institute of Technology, Durgapur, India.
*
Correspondence to: CM Narayanan, National Institute of Technology, Durgapur, India.
Email: [email protected]

1. Introduction

Heat exchangers are practically omnipresent in all process industries, power plants, heat
recovery units and the like. The feedstock is to be preheated or the product solution (product gas
mixture) is to be cooled down to a specific temperature and for these, heat exchangers become
invariable. As a result, efficient and economical design and operation of heat exchangers
becomes a fundamental parameter that is crucial to the overall economy of the industry.

Among the industrial heat exchangers, exchangers of shell and tube configuration are
one of the most popular ones, particularly for large capacity installations. These exchangers
are composed of a tube bundle (consisting of 50 – 1000 or more tubes) enclosed within a large
diameter shell. The tubes are held at both ends by drilling them into two tubesheets ( fixed
tubesheet construction ). The effective length of each tube (Le) is the length of the tube between
the two tubesheets ( those portions of the tubes that are drilled into the tubesheets are excluded
). Popular values of used in industrial exchangers are 2.5 m, 3.0 m, 3.5 m, 5.0 m and 6.0 m.
Of these, = 5.0 m, 6.0 m are most popular. Tubes are usually either 19 mm OD or 25.4 mm
OD and the tube pitch ( ), which is the center to center distance between adjacent tubes, is
commonly maintained at 1.25 to 1.5 times the tube OD (see tube count Tables 3A to 3F).

2
 Advances in Chemical Engineering

Table 3A: Tube Count Tables (Database – 3)

19 mm OD tubes on 25.4 mm triangular pitch

Fixed tubesheet construction Floating head construction U – tube construction

C M Narayanan

Shell ID ( T.E.M.A. L or M ) ( T.E.M.A. P or S ) ( T.E.M.A. U )

mm Number of passes Number of passes Number of passes

1 2 4 6 1 2 4 6 2 4 6

203.2 42 40 26 24 31 26 16 12 32 24 24

254.0 73 66 52 44 56 48 42 40 52 48 40

304.8 109 102 88 80 88 78 62 68 84 76 74

336.5 136 128 112 102 121 106 94 88 110 100 98

387.3 183 172 146 148 159 148 132 132 152 140 136

438.1 237 228 208 192 208 198 182 180 206 188 182

488.9 295 282 258 248 258 250 228 220 266 248 234

539.7 361 346 318 320 320 314 290 276 330 316 296

590.8 438 416 382 372 400 384 352 336 400 384 356

635.0 507 486 448 440 450 442 400 392 472 440 424

685.8 592 574 536 516 543 530 488 468 554 528 502

736.6 692 668 632 604 645 618 574 556 648 616 588

787.4 796 774 732 708 741 716 666 648 744 716 688

838.2 909 886 836 812 843 826 760 740 852 816 788

889.0 1023 1002 942 920 950 930 878 856 974 932 908

939.8 1155 1124 1058 1032 1070 1052 992 968 1092 1056 1008

990.6 1277 1254 1194 1164 1209 1184 1122 1096 1224 1180 1146

1066.8 1503 1466 1404 1372 1409 1378 1314 1296 1434 1388 1350

1143.0 1726 1690 1622 1588 1635 1608 1536 1504 1652 1604 1560

1219.2 1964 1936 1870 1828 1887 1842 1768 1740 1894 1844 1794

1371.6 2519 2466 2380 2352 2399 2366 2270 2244 2426 2368 2326

1524.0 3095 3058 2954 2928 2981 2940 2832 2800 3006 2944 2884
www.openaccessebooks.com

3
 Advances in Chemical Engineering

Table 3B: Tube count tables (Database – 3)

19 mm OD tubes on 25.4 mm square / rotated square pitch

Floating head construction U – tube construction

Shell
( T.E.M.A. P or S ) ( T.E.M.A. U )
ID
Number of passes Number of passes
mm
1 2 4 6 2 4 6

203.2 28 26 16 12 28 24 12

254.0 52 48 44 24 52 44 32

304.8 80 76 66 56 78 72 70

336.5 104 90 70 80 96 92 90

387.3 136 128 128 114 136 132 120

438.1 181 174 154 160 176 176 160

488.9 222 220 204 198 224 224 224

539.7 289 272 262 260 284 280 274

590.8 345 332 310 308 348 336 328

635.0 398 386 366 344 408 392 378

685.8 477 456 432 424 480 468 460

736.6 554 532 510 496 562 548 530

787.4 637 624 588 576 648 636 620

838.2 730 712 682 668 748 728 718

889.0 828 812 780 760 848 820 816

939.8 937 918 882 872 952 932 918

990.6 1048 1028 996 972 1056 1044 1020

1066.8 1224 1200 1170 1140 1224 1224 1212

1143.0 1421 1394 1350 1336 1436 1408 1398

1219.2 1628 1598 1548 1536 1640 1628 1602

1371.6 2096 2048 2010 1992 2108 2084 2068

1524.0 2585 2552 2512 2476 2614 2584 2558

4
 Advances in Chemical Engineering

Table –3C: Tube count tables (Database – 3)

25.4 mm OD tubes on 31.75 mm square / rotated square pitch

Floating head construction U – tube construction

Shell
( T.E.M.A. P or S ) ( T.E.M.A. U )
ID
Number of passes Number of passes
mm
1 2 4 6 2 4 6

203.2 17 12 8 12 14 8 6

254.0 30 30 16 18 30 24 12

304.8 52 48 42 24 44 40 32

336.5 61 56 52 50 60 48 44

387.3 85 78 62 64 80 72 74

438.1 108 108 104 96 104 100 100

488.9 144 136 130 114 132 132 120

539.7 173 166 154 156 172 168 148

590.8 217 208 194 192 212 204 198

635.0 252 240 230 212 244 240 230

685.8 296 280 270 260 290 284 274

736.6 345 336 310 314 340 336 328

787.4 402 390 366 368 400 384 372

838.2 461 452 432 420 456 444 440

889.0 520 514 494 484 518 504 502

939.8 588 572 562 548 584 576 566

990.6 661 640 624 620 664 644 640

1066.8 776 756 738 724 764 748 750

1143.0 900 882 862 844 902 880 862

1219.2 1029 1016 984 972 1028 1008 1004

1371.6 1310 1296 1268 1256 1320 1296 1284

1524.0 1641 1624 1598 1576 1634 1616 1614

5
 Advances in Chemical Engineering

Table –3D: Tube count tables (Database – 3)

25.4 mm OD tubes on 31.75 mm triangular pitch

Fixed tubesheet construction Floating head construction U – tube construction

Shell
( T.E.M.A. L or M ) ( T.E.M.A. P or S ) ( T.E.M.A. U )
ID
Number of passes Number of passes Number of passes
mm
1 2 4 6 1 2 4 6 2 4 6

203.2 27 26 8 12 18 14 8 12 14 12 6

254.0 42 40 34 24 33 28 16 18 28 24 24

304.8 64 66 52 44 51 48 42 44 52 40 40

336.5 81 74 62 56 73 68 52 44 64 56 52

387.3 106 106 88 92 93 90 78 76 90 80 78

438.1 147 134 124 114 126 122 112 102 122 112 102

488.9 183 176 150 152 159 152 132 136 152 140 136

539.7 226 220 204 186 202 192 182 172 196 180 176

590.8 268 262 236 228 249 238 216 212 242 224 216

635.0 316 302 274 272 291 278 250 240 286 264 246

685.8 375 360 336 324 345 330 298 288 340 320 300

736.6 430 416 390 380 400 388 356 348 400 380 352

787.4 495 482 452 448 459 450 414 400 456 436 414

838.2 579 554 520 504 526 514 484 464 526 504 486

889.0 645 622 586 576 596 584 548 536 596 572 548

939.8 729 712 662 648 672 668 626 608 668 636 614

990.6 808 792 744 732 756 736 704 692 748 728 700

1066.8 947 918 874 868 890 878 834 808 890 856 830

1143.0 1095 1068 1022 1000 1035 1008 966 948 1028 992 972

1219.2 1241 1220 1176 1148 1181 1162 1118 1092 1180 1136 1100

1371.6 1577 1572 1510 1480 1520 1492 1436 1416 1508 1468 1442

1524.0 1964 1940 1882 1832 1884 1858 1800 1764 1886 1840 1794

6
 Advances in Chemical Engineering

Table 3E: Tube count tables (Database – 3)

15.875 mm OD tubes on 20.637 mm square / rotated square pitch

Floating head construction U – tube construction

Shell
( T.E.M.A. P or S ) ( T.E.M.A. U )
ID
Number of passes Number of passes
mm
1 2 4 6 2 4 6

203.2 55 48 34 24 52 40 32

254.0 88 78 62 56 90 80 74

304.8 140 138 112 100 140 128 108

336.5 178 172 146 136 180 164 148

387.3 245 232 208 192 246 232 216

438.1 320 308 274 260 330 312 292

488.9 405 392 352 336 420 388 368

539.7 502 484 442 424 510 488 460

590.8 610 584 536 508 626 596 562

635.0 700 676 618 600 728 692 644

685.8 843 812 742 716 856 816 780

736.6 970 942 868 840 998 956 920

787.4 1127 1096 1014 984 1148 1108 1060

838.2 1288 1250 1172 1148 1318 1268 1222

889.0 1479 1438 1330 1308 1492 1436 1388

939.8 1647 1604 1520 1480 1684 1620 1568

990.6 1840 1794 1700 1664 1882 1816 1754

1066.8 2157 2112 2004 1968 2196 2136 2068

1143.0 2511 2458 2326 2288 2530 2464 2402

1219.2 2856 2808 2686 2656 2908 2832 2764

1371.6 3656 3600 3462 3404 3712 3624 3556

1524.0 4538 4472 4310 4256 4608 4508 4426

7
 Advances in Chemical Engineering

Table 3F: Tube count tables (Database – 3)

19 mm OD tubes on 23.8 mm triangular pitch

Fixed tubesheet construction Floating head construction U – tube construction

Shell
( T.E.M.A. L or M ) ( T.E.M.A. P or S ) ( T.E.M.A. U )
ID
Number of passes Number of passes Number of passes
mm
1 2 4 6 1 2 4 6 2 4 6

203.2 64 48 34 24 34 32 16 18 32 24 24

254.0 85 72 52 50 60 62 52 44 64 52 52

304.8 122 114 94 96 109 98 78 68 98 88 78

336.5 151 142 124 112 126 120 106 100 126 116 108

387.3 204 192 166 168 183 168 146 136 180 160 148

438.1 264 254 228 220 237 228 202 192 238 224 204

488.9 332 326 290 280 297 286 258 248 298 280 262

539.7 417 396 364 348 372 356 324 316 370 352 334

590.8 495 478 430 420 450 430 392 376 456 428 408

635.0 579 554 512 488 518 498 456 444 534 500 474

685.8 676 648 602 584 618 602 548 532 628 600 570

736.6 785 762 704 688 729 708 650 624 736 696 668

787.4 909 878 814 792 843 812 744 732 846 812 780

838.2 1035 1002 944 920 962 934 868 840 978 928 904

889.0 1164 1132 1062 1036 1090 1064 990 972 1100 1060 1008

939.8 1304 1270 1200 1168 1233 1196 1132 1100 1238 1200 1152

990.6 1460 1422 1338 1320 1365 1346 1266 1244 1390 1336 1290

1066.8 1703 1664 1578 1552 1611 1580 1498 1464 1632 1568 1524

1143.0 1960 1918 1830 1800 1875 1834 1736 1708 1882 1820 1770

1219.2 2242 2196 2106 2060 2132 2100 1998 1964 2152 2092 2044

1371.6 2861 2804 2682 2660 2730 2684 2574 2536 2748 2680 2628

1524.0 3527 3476 3360 3300 3395 3346 3228 3196 3420 3340 3286

8
 Advances in Chemical Engineering

1.1. Tube Sheet Layout

Tubes may be laid on the tubesheet using a square pitch arrangement, in which the tubes
are aligned in line (Figure 1) or using a rotated square or triangular layout, in which cases
a staggered tube arrangement is employed (Figures 2, 3). When the arrangement of tubes is
staggered, the flow of shellside fluid ( which flows over the tubes ) becomes more tortuous,
there shall be more intimate contacting between fluid elements and consequently, the shellside
heat transfer coefficient gets enhanced. It has been observed that the magnitude of shellside
heat transfer coefficient ( ) attained with a triangular pitch layout is often 1.25 to 1.30 times
that obtained with a square pitch layout. However, increased totuosity of flow path causes
increased resistance to flow of shellside fluid and this would demand higher pumping power
requirement and higher operating cost.

Figure 1: Square pitch layout

A square pitch arrangement, therefore, though provides the lowest shellside heat transfer
coefficient among the three, causes the lowest pressure drop as well and thereby brings down
the operating cost. In an arrangement like this, the tubes remain easily accessible for external
cleaning. Thus, if large scale fouling is anticipated on the outer surface of tubes, a square pitch
arrangement is to be preferred. The number of tubes that can be accommodated within a given
shell diameter is, however, lower in this case as compared to a triangular pitch layout (see tube
count tables 3A to 3F in the Appendix). From Figure [1], it can be seen that for a layout like
this,
(1)

Where

tube pitch normal to flow

tube pitch parallel to flow

In a rotated square layout (Figure 2), the number of tubes within a shell diameter
does not differ much from that in a square pitch layout, but since the tubes are staggered, the
layout provides larger shellside heat transfer coefficient. The shellside pressure drop shall
nevertheless be higher (demanding higher pumping cost ) and the tubes shall be less accessible
for external cleaning. For this type of layout, and shall be equal in magnitude, but not
9
 Advances in Chemical Engineering

equal to . From Figure (2), based on simple geometry,

(2)

A triangular pitch arrangement (Figure 3), as stated earlier, contributes the largest
shellside heat transfer coefficient. This layout also accommodates the largest number of tubes
within a given shell diameter and thus provides large heat transfer surface to the exchanger.
On the other hand, the shellside pressure drop shall be of larger magnitude and this leads to
increased operating cost. The accessibility of tubes for external cleaning when fouled shall
also be lower. From figure (3), it can be easily deduced that for a triangular pitch layout like
this, and are related to as

(3)

= 0.5 (4)

Figure 2: Rotated squre pitch layout

Figure 3: Triangular pitch layout

The tube sheet layout must be thus selected keeping all the above pros and cons in mind.

10
 Advances in Chemical Engineering

1.2. Tube sheet Construction

Fixed tubesheet construction (Figure 4) is the simplest and cheapest mode of construction
for these heat exchangers. Obviously, it is the first choice of the manufacturers. However,
this type of construction becomes unreliable when the temperature difference handled by the
exchanger is too large. At high working temperatures, the tubes tend to expand and this could
lead to fracture ( or cracking ) of tubes. In such cases, an exchanger with a floating head
at one end and a fixed tubesheet at the other end could be used (Figure 5). In such a pull
through floating head construction, the tubes are free to expand and this differential expansion
between the shell and the tube bundle shall not cause damage to the exchanger. The tube
bundle is removable for inspection, repair and replacement. However, this type of construction
is much more expensive than the conventional fixed tubesheet construction and is therefore
recommended only when large scale differential expansion of tubes is anticipated or when
frequent mechanical cleaning of tube surfaces due to fouling is imperative.

A still further alternative is to use the U – tube construction. In U – tube exchangers,

each tube is bent in the shape of the English letter U (U – shaped tube) and these U – tubes
are enclosed in the shell. The tubes are supported only at one end using a fixed tube sheet, the
U – ends of the tubes remain free or floating. However, the total number of tubes that can be
accommodated within a given shell diameter shall be less in this case since tubes cannot be
bent to form a sharp U (they tend to crack).

1.3. Baffles and Baffle Pitch

Baffles are installed in the shells of practically all shell and tube heat exchangers. These
are mostly circular plates with a number of holes punched (drilled) on them, through which
the tubes pass. But, each baffle does not occupy the entire cross – section of the shell. 25% cut
segmental baffles are most popular, which are circular discs with 25 per cent of surface being
chopped off. The height of the baffle thus becomes three – fourth (75 %) of the shell diameter.
The distance between the bottom tip of the baffle and the shell wall is called the baffle cut (
). For 25% cut segmental baffles, = .

Baffles are seldom welded to the shell wall. They are held in position by means of tie
rods and spacers. The spacing between two adjacent baffles is called the baffle spacing or
baffle pitch ( ) and is an important design parameter. If is the effective length of each tube,
then the number of baffles ( ) shall be

)–1 (5)

This is based on the assumption that a uniform baffle spacing or baffle pitch has been
used through the length of the exchanger. Often, a larger baffle spacing may have to be used at

11
 Advances in Chemical Engineering

the inlet and also at the outlet to accommodate inlet and outlet shellside nozzles. If is the
baffle spacing employed at the inlet and that at the outlet, then

) ]+1 (6)

No doubt, it is most preferable to use a uniform baffle spacing throughout, as far as

practicable. The shellside fluid flows over the tubes, between two baffles. This flow space
between two adjacent baffles is called the crossflow section. The shellside fluid thus flows
up or down each crossflow section and thereby moves from one end of the exchanger to the
other see (Figure 4). The smaller the baffle pitch ( ) used ( and thereby the larger the number
of baffles used ), the smaller will be the flow area between baffles and the larger the flow
velocity of shellside fluid. Consequently, the shellside Reynolds number shall be of
higher magnitude and this enhances the shellside heat transfer coefficient .

It is due to the presence of baffles the shellside fluid tends to execute more and more
crossflow ( between baffles ) and the heat transfer coefficient in crossflow is much higher than
that in countercurrent flow or co-current flow ( parallel flow ).

Baffles also act as support plates for tubes and help in minimizing tube vibrations. Tubes
tend to vibrate when shellside fluid flows over them. If these vibrations are of large amplitude,
then the tubes tend to undergo fracture or fatigue failure. By supporting tubes between baffles,
chances of such fatigue failure of tubes are minimized. To note that the maximum unsupported
length of each tube is equal to the baffle pitch ( ).

However, it is not all smiles with respect to the use of baffles. When the baffle pitch or
baffle spacing chosen is small ( or the number of baffles installed is large ), the flow velocity
of shellside fluid increases (as stated earlier) and this leads to increase in the shellside pressure
drop as well. It is to be noted that the shellside pressure drop is proportional to the square
of the shellside fluid velocity ( see equations discussed subsequently in this Chapter ) and
consequently, a small increase in flow velocity could cause a substantial increase in the
pressure drop penalty. A large increase in the shellside pressure drop means large pumping
power requirement and increased operating cost.

The baffle spacing ( ) and the number of baffles to be installed must be, therefore,
judiciously chosen. The number of baffles must be sufficiently large (the baffle spacing
sufficiently small) so as to maintain the shellside heat transfer coefficient sufficiently large
and also to ensure adequate support to tubes, but it should also be not too large such that
the shellside pressure drop penalty does not exceed the maximum permissible limit. TEMA
(Tubular Exchangers Manufacturers Association) specifies the following criterion for the
selection of baffle spacing / baffle pitch in a shell and tube heat exchanger:

12
 Advances in Chemical Engineering

(7)

In other words, the baffle spacing must be so chosen that it never falls below 20% ( one
– fifth) of the shell diameter ( and should also never exceed the shell diameter itself. Thus

(min) = (8)

(max) = (9)

During the design of the exchanger, it is common practice to choose the minimum baffle
spacing at the outset and subsequently increase it if the shellside pressure drop is found to
exceed the maximum permissible limit.

There are occasions where tubes are avoided in the baffle window. This is called the no –
tubes – in – baffle window construction. The baffle window is the space between two alternate
baffles and adjacent to the shell wall see (Figures 4 and 5). It is therefore obvious that the
maximum unsupported length of each tube in the baffle window is ( 2 ) and not one baffle
pitch as is the case with other tubes in the crossflow section. Consequently, the tubes in the
baffle window tend vibrate at larger amplitudes when the shellside fluid flows over them and
the chances of fatigue failure of these tubes become larger. However, by avoiding tubes in the
baffle window, the effective number of tubes in the exchanger gets reduced, thereby bringing
down the heat transfer surface available. The no – tubes – in – baffle window construction
must be therefore used only when the shellside mass velocity is too large and large scale tube
vibrations are anticipated in the baffle window.

1.4. Multipass Construction

Most of the industrial shell and tube heat exchangers employ multipass construction.
For example, a 1 – 2 exchanger ( in which the number of shellside passes = = 1 and the
number of tubeside passes = = 2 ) is what is sketched in Figure (4). This uses one pass
partition at one end of the exchanger. The tubeside fluid enters at this end, flows through all the
tubes above the pass partition ( there could be 50 to 500 or more tubes in this section ) and after
reaching the other end of the exchanger, flows back through the remaining tubes below the
pass partition and is discharged from the end – 1 itself. Since the fluid traverses the length of
the exchanger twice, the number of tubeside passes becomes equal to 2 ( = 2). The shellside
fluid, on the other hand, enters at one end of the exchanger, flows over tubes in each crossflow
section and is discharged from the other end, thereby constituting only one pass ( = 1).

In a similar way, a 2 – 4 heat exchanger ( = 2, = 4) is what is sketched in figure (5).

There are three tubeside pass partitions in the exchanger, two at one end (where the tubeside
fluid enters) and one at the other end. The tubeside fluid is thus made to traverse the length
13
 Advances in Chemical Engineering

of the exchanger four times (each time through one – fourth of the total number of tubes),
thereby executing four tubeside passes ( = 4). On the shellside, there is one longitudinal pass
partition ( along the axis of the shell ) which forces the shellside fluid to execute two shellside
passes ( = 2 ).

Multipass constructions provide higher heat transfer coefficients and thereby help in
attaining improved heat transfer effectiveness for the exchanger. However, such exchangers
are more expensive to fabricate, install and maintain. Both the tubeside and shellside pressure
drop penalties shall be higher. There shall be additional pressure drop due to flow reversal.

Figure 4: Schematic of a 1 – 2 shell and tube heat exchanger (with fixed tube sheets)

Figure 5: Schematic of a 2 – 4 shell and tube heat exchanger (with floating head construction)

A construction with larger number of passes must be therefore employed only at high
capacities, when the amount of fluid to be handled (the amount of fluid being heated or cooled)
is large. It is usual practice to start with an exchanger with one shellside pass ( = 1 ) and two
or more tubeside passes ( = 2, 4, 6 etc ) and if it is found unsuitable for the purpose, then go
for a 2 – 4 construction or an exchanger with two shellside passes ( = 2) and four or more
tubeside passes ( = 4, 8, 12 etc). During the estimation of the required heat transfer surface
(discussed later under the CAD package in Section – 2), we do get signals regarding the
suitability of the pass arrangement chosen. For example, during the computation of the heat
transfer surface using method, if the computed value of factor happens to be negative
or indeterminate (logarithm of a negative quantity appears in the expression), then it means
14
 Advances in Chemical Engineering

that the chosen pass arrangement is non-operable and an alternate pass arrangement is to be
selected. In the same way, under - NTU Method, if the computed value of is
found to be negative or indeterminate, then again it means that the pass arrangement considered
is unsuitable.

In high capacity installations, it is also common practice to use exchangers in series or

in parallel.

1.5. Selection of Tube Side and Shell Side Fluids

Among the cold and hot fluids, the question of which one is to be placed on the tubeside
and which one on the shellside is mostly dictated by economic considerations. A few thumb rules
could be useful here. For example, the more corrosive or more fouling fluid is recommended
to be used on the tubeside, since cleaning and replacement of the large diameter shell shall
be more laborious and expensive. When the fluids are pumped at high pressure (mainly in the
case of gases), the high pressure fluid be used on the tubeside to avoid an expensive, thick-
walled, high pressure shell.

When there is large difference between the flow rates of the two fluids, the larger
stream be placed on the tubeside and the smaller stream on the shellside. This is because fully
developed turbulent flow can be achieved on the shellside at much lower Reynolds number
(at ), while Reynolds numbers exceeding 10000 are required on the tubeside
for maintaining fully developed turbulent flow. However, in such cases, special care should
be taken to ensure that the pressure drop penalty on the tubeside is well within the maximum
permissible limit prescribed.

The CAD package discussed may very well be re-executed considering both alternatives
and based on the results, the choice could be made.

2. CAD Preliminaries

Design of heat exchangers involves, broadly speaking, two steps:

(a) Estimation of the heat transfer surface requirement of the exchanger,

(b) Estimation of the pressure drop penalty in each fluid stream (in the cold fluid stream and
in the hot fluid stream).

For a well – designed heat exchanger, the heat transfer surface requirement must be
reasonably low. In other words, the exchanger must be able to perform the duty (must heat the
cold fluid to the specified temperature at the specified rate or cool the hot fluid to the desired
temperature at the desired rate) with a reasonably low heat transfer surface requirement.

15
 Advances in Chemical Engineering

By pressure drop penalty, we mean the pressure difference driving force required for
pumping the cold fluid / hot fluid at the required flow rate through the exchanger. The operating
cost (pumping cost of fluids) of the exchanger is thus decided by the pressure drop penalty and
for an economic operation of the exchanger, this penalty must be reasonably low.

On occasions, the above two conditions could contradict against each other and we
would have to make a compromise between the two. For example, to restrict the operating cost
(to maintain the pressure drop penalty in both fluid streams below the maximum permissible
limit), we may have to accommodate a larger heat transfer surface. Conversely, to retain the
heat transfer surface requirement of the exchanger at a reasonably low value, a larger pressure
drop penalty and thereby a larger operating cost may have to be tolerated.

Heat exchanger problems could be a sizing problem or a rating problem. In a sizing

problem, we design a heat exchanger for a specific duty, while in a rating problem, the heat
exchanger is available and we estimate whether the available heat exchanger is suitable for
performing the given duty. The design procedures are similar, though the sizing problem
demands an iterative (trial and error) procedure, while in a rating problem, the computations
are relatively straightforward.

2.1. Cad Package for Sizing Problem

Let us first consider a sizing problem. As stated above, here we design a shell and tube
heat exchanger for a specific purpose, such as for heating a cold fluid from temperature to
temperature at the rate of kg / hr using a hot fluid flowing at kg/hr or vice versa. The
step by step procedure is described below. This entire procedure has also been illustrated in all
details in the CAD flow sheet of this section (Figures 7A to 7p).

16
 Advances in Chemical Engineering

Figure 7A: Computer Aided Design of Shell and Tube Heat Exchangers (Sizing Problem)

17
 Advances in Chemical Engineering

Figure 7B: CAD of Shell and Tube Heat Exchangers (Sizing Problem)-continued

18
 Advances in Chemical Engineering

Figure 7C: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

19
 Advances in Chemical Engineering

Figure 7D: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

20
 Advances in Chemical Engineering

Figure 7E: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

21
 Advances in Chemical Engineering

Figure 7F: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

22
 Advances in Chemical Engineering

Figure 7G: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

23
 Advances in Chemical Engineering

Figure 7H: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

24
 Advances in Chemical Engineering

Figure 7I: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

25
 Advances in Chemical Engineering

Figure 7J: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

26
 Advances in Chemical Engineering

Figure 7K: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

27
 Advances in Chemical Engineering

Figure 7L: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

28
 Advances in Chemical Engineering

Figure 7M: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

29
 Advances in Chemical Engineering

Figure 7N: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

30
 Advances in Chemical Engineering

Figure 7O: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued

31
 Advances in Chemical Engineering

Figure 7P: CAD of shell and Tube Heat Exchangers (Sizing Problem)- countinued
Step 1: specification of father file parameters:

The father file is the memory file of the computer in which we enlist the initial problem
specifications such as the mass flow rate of shellside fluid and that of tubeside fluid ( ,
),the terminal temperatures of heat exchanger ( , , , ).

In most cases, one among these parameters could be unknown. For example, let the
problem specify ( mass flow rate of shellside fluid ), inlet and outlet temperatures of
shellside fluid ( , ) and also inlet and outlet temperatures of tubeside fluid ( , ).The
mass flow rate of the tubeside fluid ( ) is unknown. This is then evaluated from the overall
heat balance shown in step – 3.

32
 Advances in Chemical Engineering

It is also required to specify the maximum permissible pressure drop on shellside, (

and that on tubeside, ( and also the minimum overall dirt
factor prescribed, .

Step 2: Estimation of property values of process fluids

Since the physical and transport properties of the fluids (density, viscosity, thermal
conductivity) are functions of temperature, they are specified at the mean fluid temperatures (
or ). Here, is the mean temperature of tubeside fluid and is the mean temperature
of the shellside fluid.

If the fluid is a low viscous liquid such as water or aqueous solution, then its property values
may be specified at the arithmetic mean temperature ( ). Thus,

/2 (10)

/ 2 (11)

In the case of viscous liquids such as petroleum oils, the property values are better specified
at caloric mean temperature ( or ) rather than at the arithmetic mean temperature. The
caloric mean temperature is to be computed as given below:

(12)

(13)

The plus sign is to be used for cold fluid and minus sign for hot fluid.

= caloric fraction = ( )– (14)

Where,

[ /( - 1) ] (15)

}] (16)

( ) (17)

( ) temperature difference at the cold end of the heat transfer surface / heat exchanger.

( ) temperature difference at the hot end of the heat transfer surface / heat exchanger.

Let the tubeside fluid be the cold fluid and shellside fluid be the hot fluid. Then, from
figures (6a) and (6b), end -1 is the cold end ( where the cold fluid enters and the hot fluid
leaves ) and the end -2 is the hot end ( where, hot fluid enters and the cold fluid leaves ).
33
 Advances in Chemical Engineering

Accordingly,

(18)

(19)

This situation will be reversed if the tubeside fluid is the hot fluid. In such a case, end-1
shall be the hot end and end-2 will be the cold end. And,

= (20)

= (21)

Figure 6 (a): 1-2 Exchanger, indicating the two ends of heat transfer surface.

Figure 6 (b): 2-4 Exchanger, indicating the two ends of heat transfer surface.

34
 Advances in Chemical Engineering

The value of the parameter depends on the API gravity of the petroleum oil and the
temperature difference of the fluid. It may be computed from the correlation given below:
2
+ ( )3 (23)

The values of the correlation coefficients , , , depend upon the API gravity
of the oil and the temperature difference ( or ) and are listed in Tables (1A) to (1F).
Here,

= temperature difference of cold fluid

= temperature difference of hot fluid

For example, if the tubeside fluid is the cold fluid, then

= ( ) (24)

= (25)

For intermediate values of or the value of could be estimated by linear

interpolation. It is also important to keep in mind that if both the fluids are viscous fluids like
petroleum oils, then the value of is to be computed separately for each of them and the
larger value is to be used for the computation of caloric fraction, (from equation 14). For
example, let

(1) = the value of for tubeside fluid

(2) = the value of for shellside fluid

Then = larger of (1) and (2)

This has been illustrated in the CAD flow sheet (Figures 7B to 7C) under Caloric Mean
Temperature.

35
 Advances in Chemical Engineering

Table 1A
Table 1A: Values of Correlation Constants for Computation of KC – factor (Equation – 23) [Database -1 ]

or = 2780C
R2 value
Coefficients
API = 20 – 29 API = 29 – 39 API = 39 – 50 API = 50 – 70

6.3085 0.47 2.09 0.606

a.12 x 10-3
1.818 0.04 0.036
0.996

2.73 x 10-3 6.42 x 10-4 7.36 x 10-6 3.2 x 10-5

9.445 X 10-5 2 x 10-5 9.8 x 10-7 1.92 x10-6

Table 1B

or = 2220C R2 value
Coefficients
API = 13 –28.5 API =28.5 - 41.8 API = 41.8 - 55.5 API = 55.5-70

2.43 6.594 87.06 – 5.8768

– 0.025 5.743 x10-3 – 45.672 0.0337

0.991

– 3.858 x10-4 7.482 x10-6 0.879 1.486 x 10-4

– 1.552 x10-5 – 1.0289 x10-5 – 4.915 x 10-6 1.2856 x 10-5

Table 1C

or = 1670C
Coefficients R2 value
API= 13 - 26 API = 26 - 38.5 API = 38.5 - 53 API = 53 - 68.2

2.077 3.074856 4.0074 – 1.8865

– 0.0393 – 0.20098 – 0.0799 7.2938 x 10-3

0.991
2.674 x 10 -3
0.00377 – 9.643 x 10 -6
4.51 x 10 -6

– 1.18 x 10-4 – 0.00003 8.7598 x 10-8 2.647 x 10-6

36
 Advances in Chemical Engineering

Table 1D

or = 111°C R2 value
Coefficients
API = 10 –23 API = 23.7 –32 API = 33 – 42 API = 43 –55.5

0.917844 28.1986 – 9.4515 – 294.69

– 0.0435737 – 3.14375 0.620646 17.6774

0.991

0.00225156 0.116005 – 0.0142829 – 0.353604

0.00023358 – 0.0014481 0.000093020 – 0.00234255

Table 1E

or = 56°C R2 value
Coefficients
API = 10 – 20 API = 21 – 30 API = 31 – 48

– 1.48379 4.99609 13.1336

0.502394 – 0.5937 – 1.07437

0.996

– 0.039699 0.0217476 0.0264278

0.00086013 – 0.00030225 – 0.000223917

Table 1F

or = 28°C
Coefficients R2 value
API = 10 – 14 API = 15 – 20 API = 21 – 30 API = 31 – 35

– 16.2581 – 9.60895 – 14.3198 1.43606

4.54308 1.55559 1.77582 – 0.194326

0.990
– 0.40621 – 0.0829894 – 0.0761433 0.00416506

0.0116306 0.00133767 0.00103175 – 0.0000396

In many cases, correlations are available for the estimation of property values at any specified
temperature. As for example, let the tubeside fluid be water and its property values are being
specified at the arithmetic mean temperature , while the shellside fluid is a viscous petroleum
oil and its property values are being specified at the caloric mean temperature, . In other
words, = and = . Now, for the tubeside fluid,

[ ] (26)
37
 Advances in Chemical Engineering

[ ] (27)

[ ] (28)

[ ] (29)

Similarly, for the shellside fluid,

[ ] (30)

[ ] (31)

[ ] (32)

[ ] (33)

Where to , to , to and to are correlation constants.

Step 3: Overall heat balance:

As stated earlier, out of the six parameters such as the two flow rates ( , ) and the
four terminal temperatures ( , , , ), one of them could be unknown. This is evaluated
from the

following overall heat balance:

= = (34)

This step also computes the magnitude of all overall rate of heat transfer, .

Step 4: Initial choice of overall design heat transfer coefficient ( )

The recommended range of values of overall design heat transfer coefficient ( ) for
different process fluids are given by TEMA and these are listed in Table (2). The value of is
to be selected based on this table. The maximum value of is first selected (since this would
correspond to minimum heat transfer surface requirement for the exchanger) and the value of
is subsequently decreased if the computed value of overall dirt factor ( ) is found to be
below the minimum prescribed value, such as, .

For example, let the specified range of for the process fluids at hand be 425 – 850
. Then, the computations are started by assuming = 850 This value
of is decreased subsequently and computations repeated if (computed) is found to be
less than .

The selected value of may be specified as or . In the CAD flowsheet, it has been
38
 Advances in Chemical Engineering

specified as . It must be kept in mind that since ( )=( ), the final results shall
remain the same in spite of whether has been selected as or .

Step 5: Computation of heat transfer surface (

For the computation of the required heat transfer surface of the exchanger, there are
three alternate methods, such as

(a) Method

(b) - NTU Method

(c) Martin’s Method

All the above three methods are based on the heat balance equations written separately
for each pass of the exchanger and then clubbed together. Accordingly, each of the above
methods should predict the same value of the heat transfer surface ( ). The choice of
the method, therefore, lies on the convenience of the user. All of the above three methods are
illustrated in the CAD flowsheet.
Table 2: Recommended Values of Overall Design Heat Transfer Coefficient ( ) [ Database – 2 ]

Hot fluid Cold fluid

Water, methanol, ammonia,

Water 1420 – 2840
aqueous solutions
Water Brine 570 – 1135
Aqueous solutions Aqueous solutions 1420 – 2840
Light organics ( liquids with
viscosities less than 0.5 cp
like benzene, toluene, acetone, Light organics 227 – 425
ethanol, MEK, gasoline, naphtha
)
Light organics Water 425 – 850
Medium organics ( liquids with
viscosities 0.5 to 1.0 cp like Medium organics 113 – 340
kerosene, light gas oil )
Medium organics Water 284 – 710
Heavy organics ( liquids with
viscosities more than 1.0 cp like
Heavy organics 57 – 227
lube oils, fuel oils, reduced crude
oils, tars, asphalts )
Heavy organics Light organics 170 – 340
Heavy organics Water 30 – 425
Light organics Heavy organics 56 – 227

Steam Aqueous solutions ( 2 cp ) 1135 – 3975

39
 Advances in Chemical Engineering

Steam Aqueous solutions ( 2 cp ) 570 – 2840

Steam Light organics 570 – 1135
Steam Medium organics 284 – 570
Steam Heavy organics 34 – 340
Steam Water, methanol, ammonia 1135 – 3975
Steam Gases 30 – 284
Gases Water 12 – 284

Source: TEMA standards

(a) Method

This method utilises a correction factor FT , such that,

= (35)

where, logarithmic mean temperature difference

( / (36)

temperature difference at the hot end of the heat transfer surface / heat exchanger
(defined earlier)

temperature difference at cold end of the heat transfer surface / heat exchanger (defined
earlier)

The correction factor is to be computed as per the equations given below. For a 1-2
heat exchanger ( or for an exchanger with = 1 and = 2, 4, 6 etc ),

(37)

(38)

(39)

] (40)

(41)

(42)

In the above equations,

/ (43)

40
 Advances in Chemical Engineering

( / ) (44)

maximum temperature difference (45)

It is obvious that the maximum temperature difference in the case of any exchanger
shall be the difference between the inlet temperature of the hot fluid (highest temperature) and
the inlet temperature of the cold fluid (lowest temperature). It must be noted that the parameter
defined above is different from defined in equation (17) and used for the computation
of caloric mean temperature.

For a 2-4 heat exchanger (or for an exchanger with = 2 and = 4, 8, 12 etc),

(46)

(47)

(48)

(49)

[ [(1- / (1- )] (50)

[ )- ) ] (51)

[ )- ) ] (52)

(53)

) (54)

Once the value of has been computed, then the heat transfer surface required (
) can be estimated from equation (35).

As stated earlier, it is better to choose a 1 – 2 exchanger ( = 1, = 2) at the outset.

If the computed value of factor turns out to be negative or indeterminate, then it means
that such an exchanger is unsuitable and we have to proceed to design an exchanger of larger
number of passes ( such as a 2 – 4 exchanger ).

b) - NTU Method

Here, we define two parameters such as heat exchanger effectiveness ( ϵ) and number of
transfer units, . These are defined as given below:

=Q/[ ] (55)

41
 Advances in Chemical Engineering

/C ( )] (56)

/C ( )] (57)

where,

C( smaller of ( and ) (58)

The number of transfer units, , can be computed as described below.

For a 1-2 heat exchanger ( or for an exchanger with = 1 and = 2, 4, 6 etc ),

/ )–1–C ] (59)

/ )–1–C ] (60)

=(1/ ) (61)

where

C = / (62)

(63)

larger of ( and ( ) (64)

For a 2-4 heat exchanger (or for an exchanger with = 2 and = 4, 8, 12 etc),

(65)

(65)

/ )–1–C + (2/ ) + ] (66)

/ )–1–C + (2/ ) ] (67)

= (2/ ln ( ) (68)

Once the value of has been computed, then the heat transfer surface required
( ) can be estimated from equation (56) or (57). As stated under – method, in this
case also, if by considering a 1 – 2 exchanger, the computed value of is seen to
be negative or indeterminate, then it indicates that the selected exchanger is inadequate and we
have to go for a 2 – 4 exchanger.

C) Martin’s Method

The method proposed by Martin involves a trial and error procedure. A value of
42
 Advances in Chemical Engineering

is to be assumed at the outset and subsequently verified. The procedure is outlined below:

1. Assume a value of

For example, let

(69)

or,

(70)

2. Compute parameters as

( )/( ) (71)

= ( )/( (72)

= ( / ) (73)

Z (74)

3. Compute Ф ( ), Ф ( ) and Ф ( ) as

Ф( ) = / [ 1 – exp ( - Y ) ] (75)

Ф( ) = / [ 1 – exp ( - )] (76)

Ф( ) = / [ 1 – exp ( - ) ] (77)

4. Compute the parameter Ө as

(1 / Ө) = Ф ( ) + Ф ( ) – Ф ( ) + 0.5 [ ] (78)

5. Compute as

Ө) (79)

6. Compute exit temperature ( of tubeside fluid as

(80)

The plus sign is to be used if the tubeside fluid is cold fluid and the minus sign if the
tubeside fluid is hot fluid.

7. If the above – computed value of agrees with the value of specified in the problem
within 1 , then print Otherwise, increase ( for example,
43
 Advances in Chemical Engineering

and repeat the computations starting from Step 2.

Step 6: Computation of number of tubes required

Select exchanger specifications such as OD of tubes ( , tube wall thickness

, tube pitch and the tubesheet layout ( square pitch / rotated square pitch / triangular
pitch ). Select also the effective length of each tube ( ). Now, compute the number of tubes
required as,

[ (π )] (81)

[ (π )] (82)

The above calculated value of is to be rounded off to the nearest higher standard value
with reference to the standard tube count tables ( tables 3A to 3F which constitute database –
3). The value of )) and ) are to be recomputed based on the above chosen
value of . The internal dameter of shell is also retrieved from the tube count table
(database– 3). This has been clearly illustrated in the CAD flowsheet. Select also the baffle
spacing ( ). It is common practice to start the computations by choosing

= ( )=( /5) (83)

This would provide the largest magnitude of shellside heat transfer coefficient ( ).
However, the value would have to be increased subsequently if pressure drop considerations
demand so. This is discussed in one of the subsequent steps (Step – 14).

If the baffle spacing at the inlet ( ) and that at the outlet ( ) are to be chosen
different from , then the values of and are also to be specified. As stated earlier,
larger baffle spacing is often required at the shell inlet as well as at the shell outlet in order to
accommodate the shell inlet nozzle and shell outlet nozzle. No doubt, it is always desirable to
employ a uniform baffle spacing such that

= = (84)

Once the tubesheet layout has been chosen, it is also necessary to specify the tube pitch
parallel to flow ( ) and that normal to flow ( ), based on equations (1) to (4).

Step 7: Computation of Tubeside heat transfer coefficient

The tubeside heat transfer coefficient depends on the tubeside Reynolds number (
and the Prandtl Number of tubeside fluid ( and these are defined below:

[ / (85)

44
 Advances in Chemical Engineering

where

mass velocity of tubeside fluid

( / ) (86)

tubeside flow area

( /4)( / ) (87)

number of tubeside passes

( / ) (88)

In most industrial shell and tube heat exchangers, the tubeside fluid is usually made
to execute fully developed turbulent flow ( 10,000 ) so as to maintain the tubeside heat
transfer coefficient at a high magnitude. In such cases, when is greater than 10000, the
value of tubeside heat transfer coefficient could be computed from the Dittus – Boelter equation
( modified by Sieder and Tate ) and this is reproduced below. This correlation is valid for a
Prandtl number range of 0.7 ≤ Pr ≤ 16700 :

0.027 ( (
0.8
)0.33 (89)

where

tubeside Nusselt number

(90)

viscosity correction factor

(91)

viscosity of tubeside fluid at the inside wall temperature ( ) of tubes

The above equation is applicable for the flow of all Newtonian fluids except water.
If the process fluid is water, then the value of tubeside heat transfer coefficient should be
estimated from the dimensional correlation reported by Perry [1] and subsequently modified
by Narayanan and Bhattacharya [2,3]. This correlation is based on the graphical data reported
by Eagle and Ferguson [4]:

( 1057.0 ) ( 1.352 + 0.02 )[ / ]( ) (92)

where 0.1864 0.22455 (93)

It must be kept in mind that the above correlation is dimensional in nature and all
45
 Advances in Chemical Engineering

parameters involved must be expressed in their corresponding SI units. This equation is also
valid for only fully developed turbulent flow of water through straight, cylindrical tubes (
10,000 ). However, in most commercial heat exchangers, the velocity of cooling water
through tubes is maintained at more than 1.8 m/s ( to minimize precipitation fouling ) and
consequently, the flow regime shall be in the fully developed turbulent zone.

Precipitation fouling is caused by the dissolved salts present in water such as sulfates,
silicates and hydroxides of calcium and magnesium which are called inverse solubility salts,
since the solubility of these salts decreases with increase in temperature. At high temperatures
therefore, these salts precipitate out and deposit on the heat transfer surfaces causing fouling or
scaling. The deposited scale being a poor conductor of heat offers additional resistance to heat
transfer and thus brings down the performance of the exchanger. At high fluid velocities, the
deposited dirt could get re-entrained into the flowing fluid stream and this helps in impeding
precipitation fouling. Also, the exit temperature of cooling water should not necessarily be
permitted to increase beyond 50C, since scaling occurs predominantly at high temperatures.

It is not yet fully understood why the Dittus – Boelter equation (equation – 89) is not
valid for water, though it is applicable to all other Newtonian fluids. A possible reason is that
the properties of water ( density, thermal conductivity ) exhibit unusual ( often, anomalous )
temperature dependence [1,2,3].

At the outset, the value of viscosity correction factor ( may be taken equal to unity
and the value of be computed from any of the correlations given above. This value of (
i.e, the value at is denoted as in the CAD flowsheet. The incorporation of
and the estimation of corrected value of is discussed in one of the subsequent steps. It may
also be noted that this correction factor ( ) is not a detrimental parameter. For water and many
aqueous solutions, this factor may be taken more or less equal to unity.

Step 8: Computation of Shellside Heat Transfer Coefficient,

As stated earlier, the flow of shellside fluid is, in fact, tortuous. It flows over the tube
bundle in the section between the baffles, thereby executing crossflow. But, as it flows from
one crossflow section to another, it executes countercurrent or co-current flow (depending on
the flow direction of tubeside fluid). The shellside fluid, thus, executes partly crossflow, partly
countercurrent flow and partly co-current or parallel flow. All of the experimental correlations
reported in literature are those which consider true crossflow or in other words, that consider
ideal crossflow section. Accordingly, the shellside heat transfer coefficient predicted by these
correlations is (ideal). Correction factors are to be, therefore, incorporated to take care
of supplementary effects and thereby to estimate the actual value of shellside heat transfer
coefficient ( ).

46
 Advances in Chemical Engineering

For the estimation of (ideal), one of the reliable correlations is that proposed by
Colburn [5]. This correlation is given below:

( )0.6 ( )1/3 (94)

shellside Nusselt number

(ideal) / (95)

shellside Reynolds number

.
( / ) (96)

mass velocity of the shellside fluid based on flow area,

( / ) (97)

minimum free flow area between baffles at the shell axis

( - ) / ( ) (98)

0.33, for staggered tubes ( for tubes that are in triangular pitch or rotated square pitch
arrangement )

0.26, for tubes in line (for tubes that are in square pitch arrangement)

Prandtl number of shellside fluid

/ (99)

= viscosity correction factor for shellside fluid

(99a)

= viscosity of shellside fluid at outer wall temperature ( ) of tubes

The above correlation is valid for 2000 32,000. An alternate correlation for the
estimation of has been proposed by Donohue [6]. This correlation uses a modified
shellside Reynolds number that is based on the geometric average of the mass velocity of
shellside fluid in the crossflow section ( ) and that in the baffle window ( ). Thus

= shellside Reynolds number

= ( ) / (100)

where
47
 Advances in Chemical Engineering

= (101)
= mass velocity of shellside fluid in baffle window

=( / ) (102)

= free area for flow of shellside fluid in the baffle window ( discussed subsequently in Step
– 14 under pressure drop computations )

Donohue’s correlation often predicts much lower value of (ideal) as compared to that
predicted by Colburn’s correlation. Though the approach used by Donohue is more renovated,
dubiousness does exist over the accuracy of employing a geometric average of and .
These two mass velocities are not always of comparable magnitude.

Alternate correlations have been proposed by McAdams [7] and also by Kern [8]. Kern
has defined an equivalent diameter for the shell and has used the same in the correlation.
However, the flow area used for defining the equivalent diameter is the free area (free space)
between tubes. Since shellside fluid does flow over the tubes (over the tube bundle), the
approach of Kern cannot be treated as fully accurate.

After comparing the different experimental correlations available, it is recommended that for
the usual case of shellside Reynolds number ( ) exceeding 3000, Colburn’s correlation
( equation – 94 ) be used for computing ( ideal ). No doubt, it is to be multiplied by the
appropriately defined correction factors (discussed subsequently) to obtain the actual magnitude
of shellside heat transfer coefficient, .

Step 9: Estimation of correction factors and actual shellside heat transfer coefficient (
)

As stated above, the value of shellside heat transfer coefficient computed from Colburn’s
correlation is that for ideal crossflow section, . In an industrial heat exchanger
however, supplementary effects come into play such as baffle configuration effect, baffle
leakage effect, bundle bypassing effect and that due to unequal baffle spacing. Correction
factors are to be incorporated to account for each of these effects. Thus,

= ( ) (103)

where

= correction factor that accounts for baffle configuration effect,

= correction factor that accounts for shell to baffle leakage and. tube tobaffle leakage,

= correction factor that accounts for bundle by passing effect and = correction factor
48
 Advances in Chemical Engineering

that accounts for unequal baffle spacing.

Elaborate graphical data have been reported by Bell [9] for the computation of
these correction factors. Bell’s graphical data have been fitted into analytical correlations by
Narayanan and Bhattacharya [2]. These are discussed below :

The correction factor is to take care of the fact that a portion of shellside fluid that
flows through the baffle window executes more or less countercurrent flow or co-current flow,
rather than crossflow. Since heat transfer coefficient is highest in crossflow, this tends to bring
down the overall magnitude of the shellside heat transfer coefficient. If tubes are avoided in the
baffle window ( no – tubes – in – baffle window construction ), then = 1.0. The correlation
developed by Narayanan and Bhattacharya [2] for the estimation of this correction factor is as
follows:

J c = co + c1 ( Fc ) + c2 ( Fc ) 2 + c3 ( Fc )3 (104)

where

= fraction of total tubes in crossflow

= 1+(2/ ) –(2 / ) (105)

= (106)

is the baffle cut and as stated earlier, for 25% cut segmental baffles that are popularly used,
= ( / 4 ). In the above equation (105), must be expressed in radians.

is called the outer tube limit and it depends on the type of exchanger construction
and the shell ID. It is to be kept in mind that in a shell and tube heat exchanger, tubes are laid
in the shell within and not within the entire cross – section of the shell. The values of
specified by TEMA are listed in Table (4) which constitutes Database – 4. It can be seen from
table (4) that for pipe shells ( lower diameter shells ), is around 11 mm less than the shell
diameter when a fixed tubesheet construction is used, while it is 29 mm less than the shell
diameter for a floating head construction. Similarly, in the case large diameter plate shells,
is 13 mm less than the shell diameter in fixed tube sheet exchangers, whereas in floating head
exchangers, it is 37 mm less than the shell diameter.

49
 Advances in Chemical Engineering

Table 4: Values of Outer Tube Limit recommended by TEMA Standards [Database – 4]

Shell diameter , mm

( ), mm Internal floating head

Fixed tubesheet
with split backing ring

Plate shells 637 mm – 13 mm – 37 mm

Pipe shells 610 mm – 11 mm – 29 mm

The values of correlation constants , , , are listed in Table (5) which constitutes
Database – 5.
Table 5: Values of Correlation Constants for Computation of Correction Factor ( equation –104 )

Correlation
constant 0.6 0.6 0.8 0.8 0.9 0.8 1.0

0.531428 0.6406 - 2.1616 557.71946

0.7737 0.588 7.37824 - 1793.534

0.0 0.0 - 4.11426 1925.5329

0.0 0.0 0.0 - 688.7156

is the correction factor to account for the leakage of shellside fluid through the shell-
to-baffle clearances and the tube-to-baffle clearances. Its value varies from 0.7 to 0.8. A portion
of the shellside fluid flows through the shell to baffle clearances and also through the tube to
baffle clearances. These are called the leakage streams. Due to these leakage streams, the
fraction of shellside fluid executing crossflow gets reduced and this penalizes the shellside
heat transfer coefficient. As specified by TEMA, the tube to baffle clearance ( ) ranges from
0.4 to 0.8 and the shell to baffle clearance ( ) varies from 2.54 for small diameter
pipe shells to as high as 10.8 for large diameter plate shells. The values of and
as specified by TEMA are listed in Table (6) which constitutes Database – 6. The correlation
developed by Narayanan and Bhattacharya [2] for the estimation of this correction factor is
given below:

J l = ao + a1 ( S r ) + a2 ( S r ) 2 + a3 ( S r )3 (107)

where

= ( + )/ (108)

= /( + ) (109)

50
 Advances in Chemical Engineering

= tube to baffle leakage area

=[π ( 1+ ] 2 (110)

= shell to baffle leakage area

= (πDs δ sb / 2)[1 − (θ / 2π)] (111)

= tube to baffle clearance

= shell to baffle clearance

= baffle cut angle ( in radians )

= 2 (112)

As stated above, for 25 % cut segmental baffles, = ( / 4 ) and therefore, =

1200 or (2π/3) radians.

= crossflow area at or near center line

= [ - +{ - )( - )/ }] (113)

If tubes are arranged on the tubesheet on a triangular pitch layout, then = 1.0 and if
they are laid on a square or rotated square layout, then = .

The values of correlation constants , , , are listed in Table (7) which constitutes
Database – 7.
Table 6: [Database 6] Recommended Values of Tube to Baffle Clearance ( ) and Shell to Baffle Clearance ( )

= 0.8 mm, if maximum unsupported tube length ( usually 2 ) 910 mm, = 0.4 mm,
if ( 2 ) 910 mm.

Shell diameter, mm
mm
203.2 – 336.5 2.54
355.6 – 439.1 3.175
Pipe shells
457.2 – 590.8 3.81
609.6 – 990.6 7.62
Plate shells 1016.0 – 1371.6 8.89
Above 1397.0 10.80

51
 Advances in Chemical Engineering

Table 7: [Database– 7] Values of Correlation Constants for Computation of Correction Factor ( equation – 107 )

0.0 0.997 – 2.54167 15.239 – 36.276

0.25 1.0 – 3.0845 17.2089 – 38.6776
0.50 0.9957 – 3.804 22.045 – 50.586
0.2
0.75 0.9952 – 4.0808 21.764 – 47.946
1.0 0.9916 – 5.0 29.0 – 66.532
0.0 0.8975 – 0.4375 0.0 0.0
0.25 0.87 – 0.55 0.0 0.0
0.50 0.8525 – 0.6625 0.0 0.0
0.2 0.7
0.75 0.825 – 0.775 0.0 0.0
1.0 0.7925 – 0.8375 0.0 0.0

The correction factor has been incorporated to take care of the bundle bypassing effect.
That portion of the shellside fluid which flows through the clearance between the outermost
tube and the shell wall has a tendency to flow adjacent to the shell wall and thereby bypass the
tube bundle ( it does not flow over the tube bundle ). In the case of fixed tube sheet exchangers,
the clearance between the outermost tube and the shell wall is usually maintained small and
hence, this effect is not predominant and the value of shall be quite high ( around 0.9 ).
However, in floating head exchangers, values as low as 0.7 have been reported. One of the
means of minimizing the bundle bypassing effect is to install sealing strips, which are typically
longitudinal strips of metal installed between the outside of the tube bundle and the shell and
fastened to the baffles. These strips force back the bypass stream into the main crossflow
stream and thereby reduce the bypassing effect and improve the heat transfer coefficient. It
must be, however, kept in mind that sealing strips are cumbersome to install and maintain. The
correlation developed by Narayanan and Bhattacharya [2] for the estimation of this correction
factor is given below:

= [ ( )] (114)

where

= ( - ) / (115)

The value of correlation constant depends on the values of and the ratio (
) and can be retrieved from Table (8) which constitutes Database – 8. Here,

= modified shellside Reynolds number based on

= ( / ) (116)

where
52
 Advances in Chemical Engineering

= shellside mass velocity based on

=( ) (117)

= number of sealing strips installed per cross flow section and = number of tube rows
crossed during flow through one crossflow section.

=( -2 )/ (118)

= tube pitch parallel to flow

It is thus clear from Table (8) that when the ratio ( ) is equal to 0.5 or more,
shall be equal to 1.0 and the bypassing effect shall be absent.

The correction factor is to take care of the effect of unequal baffle spacing on the
shellside heat transfer coefficient. It has been explained earlier that due to the presence of
nozzles, a larger baffle spacing is often required to be used at the inlet and at the outlet of the
exchanger ( , ). If = = ( which is most preferable), then

1.0 (119)

The value of thus obviously depends on the / and ratios and the
number of baffles used and can be estimated as follows:

) (120)

where

/ (121)

/ (122)

number of baffles

) ]+1 (123)
Table 8: Values of Correlation Constant for Computation of Correction Factor (Equation – 114)[ Database – 8 ]

100 100

0.0 1.2344 1.3433

0.05 0.6704 0.72975
0.10 0.5095 0.5811
0.167 0.37895 0.4324
0.30 0.1777 0.2055
0.5 and above 0.0 0.0

53
 Advances in Chemical Engineering

Once all the four correction factors have been evaluated, then the value of can be
computed from equation (103). To start with, viscosity correction factor may be assumed
equal to 1.0. Accordingly, the above – computed value of be designated as .

Step 10: Viscosity correction

As stated earlier, the viscosity correction factor ( , ) is not a controlling parameter.

For many systems, its value is very close to 1.0. To compute this correction factor, we need
to estimate the tube surface temperature ( , ). However, it must be kept in mind that
tube surface temperature varies from one end to the other end of the exchanger and it is also a
parameter that is difficult to record experimentally. But since, as stated earlier, or is not a
highly influencing parameter, we need to determine only an order of magnitude of and
. In the design computations therefore, approximate estimation of inner surface temperature of
tubes ( ) and outer surface temperature of tubes ( ) is performed from the following
approximate heat balance equations:

(124)

(125)

(126)

where

, = overall clean heat transfer coefficient ( value of U when fouling coefficients

or dirt factors are excluded ) based on and based on respectively

It is important to note that the above equation is approximate (since it does not accurately
define the temperature difference driving force) and should be used for the approximate
estimation of and only. Now, from the above equations,

(127)

(128)

The plus sign is to be used for the cold fluid and the minus sign for the hot fluid.

The computation of viscosity correction factor (or ) based on the above equations
involves a trial and error procedure, which is summarized below:

1. Assume the values of and . For example, if the tubeside fluid is cold fluid then,

= + 1.0 (129)

Similarly, if the shellside fluid is hot fluid, then

54
 Advances in Chemical Engineering

= 1.0 (130)

2. Put = (131)

and = . (132)

3. Compute (viscosity of tubeside fluid at ) and (viscosity of shellside fluid at

) from the available property value correlations and then, estimate and as,

(133)

(134)

4. Compute the corrected values of tubeside heat transfer coefficient ( ) and that of shellside
heat transfer coefficient ( ) as

(135)

(136)

5. Compute the clean overall heat transfer coefficient ( or ) as

= ( ) ( )( ) (137)

= ( ) ( )( ) (138)

6. Re-compute the tube surface temperatures ( and ) from equations (127 and 128),
using the above computed value of overall clean heat transfer coefficient, or .

7. Compute the deviations as

(139)

(140)

8. If either or has been found to exceed , then repeat the computations

starting from Step – 2. Otherwise, print the values of and .

Usually, the scheme shall converge within two to three iterations.

Step 11: Computation of overall dirt factor

Compute the overall dirt factor ( ) as

(141)

55
 Advances in Chemical Engineering

(142)

If the above-computed value of falls below , then proceed to Step – 12 for

re-computation of the heat transfer surface. Otherwise, proceed to Step – 13 for pressure drop
computations.. The recommended values of minimum dirt factor specified by TEMA are listed
in Table (8) which constitutes database – 9.

Step 12: Re-computation of heat transfer surface

Since the computed value of overall dirt factor has been found to be less than the
minimum required value of , the value of overall heat transfer coefficient ( or
) is to be decreased and the computations repeated as outlined below :

(i) Put = – 1.0 (143)

or

= – 1.0 (144)

(ii) Re-compute the heat transfer surface as (145)

or

[ (146)

(iii) Repeat the computations starting from Step – 6.

The procedure is to be continued until the computed value of overall dirt factor ( )
exceeds .
Table – 9: Minimum Recommended Values of Dirt Factor, [Database – 9]

Process fluid Dirt factor,

Fuel oil 0.0009
Machine oil, transformer oil 0.00018
Quenching oil 0.0007
Vegetable oils 0.0006
Organic liquids 0.0002
Refrigerating liquids 0.0002
Brine (cooling) 0.0002
Organic vapours 0.0001
Steam (non-oil bearing) 0.0
Alcohol vapours 0.0
Steam, exhaust (oil bearing) 0.0002
Air 0.0004
Coke oven gas, manufactured gas 0.002

56
 Advances in Chemical Engineering

Source: TEMA Standards

Note : In the case of water, with temperature of water ≤ 52 C and water velocity ≥ 1.2 m/s, the recommended value of
is 0.0001 for sea water, distilled water and treated boiler feed water, while it is 0.0002
for brackish water, clean river water and treated make-up water used in cooling towers. At the same
water velocity and temperature, specified for hard water ( over 15 grains / gal) is 0.0006
and that for muddy or silty river water is 0.0004 .

Step 13: Computation of tubeside pressure drop

Heat exchanger calculations are incomplete, unless the pressure drop in either stream is
evaluated and ascertained that neither of them ( pressure drop in the tubeside fluid or that in
the shellside fluid ) exceeds the maximum permissible limit. To note that the operating cost of
the exchanger is decided by the magnitude of pressure drop in the two streams. The tubeside
pressure drop includes frictional pressure drop ( due to skin friction between the tube wall and
the fluid layer ) which is predicted by the modified form of Fanning’s equation ( corrected
for non-isothermal flow) and the additional pressure drop due to flow reversal ( by virtue of
multipass construction ). Thus

(147)

where

additional pressure drop due to flow reversal four velocity heads per pass
(observed experimentally)

= (148)

= number of tubeside passes tube side friction factor ( for non-isothermal flow )

(149)

where and are empirical constants. The values of these constants are listed in table
(10) which constitutes database – 10.
Table – 10: Friction factor in Non – isothermal flow Values of Correlation Constants and (Equation – 149)
[Database – 10]

Smooth tubes Commercial pipes

18.0 1.0 18.0 1.0

1000
1000
0.12 0.272 0.105 0.243
to 105
105 to
0.087 0.2413 0.0423 0.164
106

57
 Advances in Chemical Engineering

The tubes used in shell and heat exchangers are relatively smooth. Accordingly, in the present
case, the values of correlation constants ( and are to be retrieved from column – 2 on
smooth tubes. Double pipe heat exchangers employ industrial pipes which have a given degree
of roughness on their inner surface. In the case of those exchangers therefore, the values of
and are to be read from column – 3 on commercial pipes. It is also important to keep in mind
that the conventional friction factor versus Reynolds number plots ( Moody’s plots ) are not
applicable here since those plots are for isothermal flow. The above correlation (149) is based
on the graphical data reported by Sieder and Tate and reproduced by Kern [8].

If the above computed value of tubeside pressure drop happens to exceed the maximum
permissible value, then computations are to be repeated after selecting a larger tube diameter,
starting from Step – 6.

Step 14: Computation of shellside pressure drop

The shellside pressure drop is more difficult to estimate accurately. This is because,
as discussed earlier, the flow of shellside fluid through the exchanger is too much tortuous,
it executes both crossflow and countercurrent flow as well as parallel flow. For flow over a
submerged object, the form drag comes into play, which is of higher magnitude than skin
friction. Since the shellside fluid flows over the tube bundle, the frictional resistance includes
form drag and it is more cumbersome to quantify.

For a reasonably reliable estimate of shellside pressure drop therefore, we first estimate
the pressure drop for flow through ideal crossflow section, ( and that for flow
through ideal baffle window section, . The actual value of shellside ptressure
drop is then computed by incorporating the correction factors, and which are similar
to the correction factors, , used for the estimation of shellside heat transfer coefficient.
Thus

( + (150)

where

combined correction factor

(151)

pressure drop in an ideal crossflow section

(152)

= shellside friction factor pressure drop in an ideal baffle

window section
58
 Advances in Chemical Engineering

(153)

= mass velocity of shellside fluid in baffle window

=( / ) (154)

= free area for flow of shellside fluid in the baffle window

= ( /4)– ( /4) (155)

= fraction of the shell cross-sectional area occupied by the baffle.

window

=( 1 / )[( /2)– ] (156)

= baffle cut angle ( in radians )

=2 (157)

= number of tubes in baffle window

= ( )( 1 ) (158)

= fraction of total tubes in crossflow ( defined earlier in equation – 105)

number of effective crossflow rows in each baffle window

( 0.8 ) (159)

correction factor to account for bundle by passing effect, baffle leakages and

unequal baffle spacing respectively on shellside pressure drop.

Bell and coworkers [9] have reported extensive graphical data for the estimation of
these correction factors as well and Narayanan and Bhattacharya [2] have converted them into
analytical correlations through rigorous regression analysis. The correlations developed by
them are reproduced below:

Rl = α o + α1 ( S r ) + α 2 ( S r ) 2 + α3 ( S r )3 (160)

The values of correlation constants , , , are listed in Table (11) which constitutes
Database – 11.

59
 Advances in Chemical Engineering

Table 11: [Database – 11] Values of Correlation Constants for Computation of Correction Factor (equation – 160)

0.0 0.995 – 4.94 26.952 – 58.77

0.2
0.25 0.9947 – 6.651 40.5936 – 95.67

0.50 0.9985 – 7.3934 37.7854 – 75.146

0.75 0.993 – 9.3936 56.934 – 132.37

1.0 0.995 – 11.256 71.358 – 170.295

0.0 0.7267 – 0.5737 0.0 0.0

0.2
0.25 0.66 – 0.71 0.0 0.0
0.7
0.50 0.5933 – 0.8476 0.0 0.0

0.75 0.5133 – 0.9506 0.0 0.0

1.0 0.4667 – 1.1476 0.0 0.0

The parameters and have been defined earlier ( see equations – 108, 109 ). The
correction factor which takes care of the effect of tube to baffle and shell to baffle leakages
on shellside pressure drop is thus analogous to factor defined earlier under computation of
shellside heat transfer coefficient. In a similar way, the correction factor is similar to
and it takes care of the effect of bundle bypassing effect on shellside pressure drop. It may be
computed from

= [ ( )] (161)

The value of correlation constant depends on the modified shellside Reynolds number,
and the ( ) ratio and can be retrieved from Table (12). This table constitutes
Database – 12. The dimensionless parameter has been defined earlier in equation (115).
As evident from table (12), when the number of sealing strips installed is large such that the
ratio ( ) is equal to or more than 0.5, = 1.0.
Table 12: Values of Correlation Constant for Computation of Correction Factor ( equation – 161)[Database – 12]

100 100

0.0 3.7041 4.3524

0.05 2.0245 2.4183
0.10 1.5270 1.8522
0.167 1.1684 1.30898
0.30 0.5944 0.72975
0.5 and above 0.0 0.0

60
 Advances in Chemical Engineering

The correction factor, has been incorporated to account for the effect of unequal baffle
spacing on shellside pressure drop. Evidently, its magnitude shall depend on the /
ratio and the ratio, as shown below:

= 0.5 (162)

The shellside friction factor is a non-linear function of shellside Reynolds number,

. It is also a function of the tubesheet layout chosen and the tube pitch ( , ). A reasonably
satisfactory estimate of can be obtained from the correlation proposed by Grimson [10]. It
is given below:

For staggered tubes,

(163)

where

(164)

For tubes in line,

(165)

where

(166)

(167)

Grimson’s correlation is valid within the Reynolds number range of 2000

40,000.

If the above computed value of shellside pressure drop happens to exceed the maximum
permissible value, then a larger value of baffle spacing ( ) is to be chosen and computations
repeated starting from Step – 8.

Step 15: Print Results

The entire procedure described above has been illustrated in all details in the CAD
flowsheet given in Figures – 7A to 7P.

It is needless to comment that the CAD package presented could very well be re-
executed with different pass arrangements and with different choices of son file parameters
and the most satisfactory design could be located from the results, keeping the heat transfer

61
 Advances in Chemical Engineering

surface requirement, fabrication cost and the pressure drop penalties ( both on tubeside as well
as on the shellside) in mind. This, in fact, forms the inherent flexibility of all types of CAD (
software ) packages.

We shall illustrate a numerical example here to demonstrate the applicability of the

above-described CAD package. The package is executed with the following Father File
parameters:

Shellside fluid : Petroleum Oil ( hot fluid ), Tubeside fluid : Water (cold fluid )

= mass flow rate of shellside fluid = 36300 𝑘𝑔 / hr

Inlet temperature of shellside fluid = = 1600C

Outlet temperature of shellside fluid = = 450C

Inlet temperature of tubeside fluid = = 20℃

Outlet temperature of tubeside fluid = = 42℃

(min) = 0.0005 ( m2. 𝐾 )/ 𝑊

= 60 𝑘𝑃𝑎

The results obtained are,

Type of exchanger recommended: 1 – 2 heat exchanger

Mass flow rate of water = = 104198.4 k𝑔 / hr

471.5 W/ (m2.K) [ finalized by trial, from the prescribed range of 284 – 710
W/ (m2.K) ]

Heat transfer surface required = = 111.08 m2

Heat exchanger specifications : = 19 mm, = 17 mm, = 5.0 m, tubesheet

layout = triangular pitch ( = 25.4 mm, = 22 mm, = 12.7 mm ).

Construction : Fixed tubesheet

Total number of tubes = = 416

Shell ID = = 590.8 mm

Baffle spacing = = = = 196.93 mm ( finalized by trial )

Baffle cut = = 147.7 mm ( 25 % cut segmental baffles )

62
 Advances in Chemical Engineering

Number of baffles = = 44

Number sealing strips per crossflow section = =2

Tubeside heat transfer coefficient = = 3228.46 W / ( m2. K )

Shellside heat transfer coefficient = = 700.92 W / ( m2. K )

(computed) = 0.000532 ( m2. K )/ W

Tubeside pressure drop = = 6.0 kPa

Shellside pressure drop = = 9.0 kPa

2.2. CAD Package for Rating Problem

As explained earlier, in a rating problem, a heat exchanger of known specifications is

available and we have to determine whether this exchanger is suitable for a specific purpose
( for performing the specified duty ). The overall design procedure is very similar to that
involved in the sizing problem, except that we do not have to resort to any trial and error
(iterative) computations here. The step by step procedure is summarized below:

Step 1: Specification of father file parameters

As discussed in Step – 1 of the sizing problem, in the father file, five among the six
parameters such as the mass flow rate of shellside fluid ( ) and that of the tubeside fluid (
), the four terminal temperatures ( , , , ) are specified. The sixth unknown parameter
is then estimated from the heat balance shown in Step – 3. For example, let the unknown
parameter be the mass flow rate of the tube side fluid ( ). This is then evaluated from the
overall heat balance as shown in step – 3.

Being a rating problem, the heat exchanger specifications are available and these are also
to be listed in the father file, such as number of tubeside passes ( ), number of shelleside
passes ( ), Inner and outer outer diameter of tubes ( , Tubesheet layout ( Triangular
/ Square / Rotated Square Pitch ), Tube pitch ( , , ), Effective length of each tube (
), Number of tubes ( ), Shell Diameter ( ), Baffle spacing ( ), Baffle cut
( ), Number of sealing strips installed per crossflow section ( ), Also to be specified are
the maximum permissible pressure drop on the shellside, ( and that on the
tubeside, ( and the minimum overall dirt factor prescribed, .

Step 2: Estimation of property values of process fluids

Estimate the property values of the tubeside fluid ( , , ) and those of the
shellside fluid ( , , , ) at the mean temperature and respectively, as
63
 Advances in Chemical Engineering

discussed in Step – 2 of sizing problem.

Step 3: Overall heat balance

Determine the unknown parameter ( here, ) from the overall heat balance equation, as
shown in Step – 3 of sizing problem. Compute also the overall rate of heat transfer ( ).

Step 4: Computation of tubeside heat transfer coefficient

Assuming the viscosity correction factor ( ) to be equal to unity, compute the tubeside
heat transfer coefficient from available correlations such as from equation (89) or (92), as
described in Step – 7 of sizing problem and denote it as .

Step 5: Computation of shellside heat transfer coefficient,

Assuming the shellside viscosity correction factor ( ) to be equal to unity, compute

the shellside heat transfer coefficient as discussed in Steps – 8 and 9 of sizing problem and
denote it as . The value of is to be computed first from Colburn’s correlation
( equation – 94 ) and thereafter the correction factors ( ) incorporated to obtain the
value of .

Step 6: Viscosity correction

Perform the viscosity correction as described in Step – 10 of sizing problem and estimate
the actual value of tubeside heat transfer coefficient ( and that of shellside heat transfer
coefficient ( ).

Step 7: Computation of overall heat transfer coefficient ( or )

Compute the overall heat transfer coefficient ( or ) as

( = ( ) ( )( ) (168)

(169)

= ( ) ( )( ) (170)

(171)

Step 8: Computation of Heat Transfer Surface

Compute the heat transfer surface required ( using any of the three methods such
as the FT Method, - NTU Method or Martin’s Method as described in Step – 5 of the sizing
problem.

64
 Advances in Chemical Engineering

Step 9: Computation of required tube length

Compute the effective tube length required, as given below

(172)

(173)

I f the above-computed value of exceeds the value of specified in the father

file ( Step – 1 ), then print “ the exchanger is not suitable for the purpose with respect to
heat transfer surface requirement”. Otherwise, proceed to Step – 10 for the computation of
tubeside pressure drop.

Step 10: Computation of tubeside pressure drop

Compute the tubeside pressure drop , as discussed in Step – 13 of the sizing

problem. If this value of exceeds the maximum permissible value, (max)
, specified in the father file ( Step – 1 ), then print, “ the exchanger is not suitable for
performing the given duty”. Otherwise, proceed to Step – 11 for the computation of shellside
pressure drop.

Step 11: Computation of shellside pressure drop

Compute the shellside pressure drop as described in Step – 14 of the sizing

problem. If this value of is found to exceed the maximum permissible value,
) , specified in the father file, then print, “ the exchanger is not suitable for
performing the given duty”. Otherwise, proceed to Step – 12.

Step 12: Print: The given exchanger is suitable for performing the specified duty.

The readers are encouraged to prepare the detailed CAD flowsheet for the rating problem
by themselves, as an interesting exercise.

3. Improved Design of Shell and Tube Heat Exchangers

Since shell and tube heat exchangers are quite popular in all process industries and
power plants, attempts have been made by many authors to propose improved design of these
exchangers. However, needless to comment, in many cases, though the heat transfer coefficient
( and thereby the heat transfer efficiency of the exchanger ) gets enhanced, there is simultaneous
increase in the pressure drop penalty ( and thereby in the operating cost ) and consequently, the
net benefit of employing the proposed design becomes marginal. In alternate cases, the modified
design demands complex and expensive construction or expensive accessories. Examples are
flow interception using corona discharge ( expensive accessories ), admitting process fluids
through multiple jets ( too high operating cost ), insertion of twisted tapes inside tubes ( too
65
 Advances in Chemical Engineering

cumbersome when a tube bundle composed of 500 – 1000 tubes are used and fouling fluids
are handled, net benefit marginal), installation of fins on tube surfaces ( high manufacturing
cost, simultaneous increase in pressure drop penalty tends to compensate higher heat transfer
coefficient attained unless used for gases ) etc.

A novel approach in this connection is the use of variable area construction for shell
and tube heat exchangers [11, 12]. A Variable Area Exchanger ( VAE ) employs a bundle of
diverging – converging tubes ( periodically constricted tubes ) instead of straight, cylindrical
tubes, as shown schematically in figures (8) and (8A). Each tube is composed of a number of
segments, each segment being made up of two frustums of cones joined base to base. .The tube
diameter or cross – sectional area thus varies continuously along the length of the tube. If is
the maximum diameter of each segment, the minimum diameter and the segment length,
then from simple geometry, the angle of divergence / convergence ( θ ) is predicted by

(174)

The optimum value of θ reported is 50 [11,13] or = (1/12). The geometry of

each tube thus deviates from straight, cylindrical geometry by only 50. If is the total number
of segments per tube, then the total effective length ( ) of each tube shall be

) (175)

The specific advantages of using such a construction are,

1.They provide substantially large heat transfer coefficient (350 to 400% higher than, or 3.5 to
4.00 times, that in a conventional heat exchanger of same heat transfer surface per unit length)
within a large range of flow rates (both in laminar flow and in turbulent flow), both under
constant wall temperature conditions as well as constant wall heat flux conditions.

2. The simultaneous increase in pressure drop penalty has been, however, observed to be
relatively negligible (only by 15 to 20 % or 1.15 to 1.2 times).

3. The performance efficiency of these exchangers is thus significantly high, but they do not
demand any large scale increase in the operating cost. This has been found to be true while
handling Newtonian fluids (water, aqueous solutions, petroleum oils) as well as while handling
Non – Newtonian fluids such as suspensions and polymer solutions [14].

4. Since the shellside heat transfer coefficient in a variable area exchanger is substantially
large, the shell of the exchanger need not have to be baffled. No doubt, a minimum number
of baffles may still be installed, keeping the baffle spacing at the maximum permissible
value , to act as support plates for tubes.. In the case of tubesheet layout,
it is recommended that the tube hole diameter be kept equal to (or slightly more than)
66
 Advances in Chemical Engineering

Figure 8: Schematic of Variable Area Shell and Tube Heat Exchanger ( showing 1 – 1 construction )

Figure 8A: Schematic of Diverging – Converging Geometry

(the maximum diameter of diverging - converging tube) so as to retain the flexibility of the
construction.

5. An additional interesting feature of these exchangers is that they exhibit lower tendency
to precipitation fouling. The tortuous wall geometry of the d-c (diverging – converging) tube
induces a degree of turbulence into the flow field and this tends to dislodge the deposited dirt
from the tube surface and gets it re-entrained into the flowing fluid. If fouling does occur, then
cleaning of the tube surface could be accomplished using high pressure liquid jets or by using
chemical solvents (chemical cleaning). Mechanical cleaning, no doubt, shall be relatively more
troublesome in the present case.

6. Supplementary effects such as bundle bypassing and baffle leakages that tend to diminish
the magnitude of shellside heat transfer coefficient shall not be significant in these exchangers.
For example, the bundle bypassing effect would not be significant in the proposed design due
to the fact that this bypass stream also tends to execute a tortuous flow owing to the diverging
– converging nature of the tube wall geometry. The baffle leakage effects (leakage of shellside
fluid through shell to baffle and tube to baffle clearances) will also not be predominant since

67
 Advances in Chemical Engineering

the shell is to be fitted with minimum number of baffles

The performance characteristics of variable area exchangers have been studied both
mathematically as well as experimentally [11 – 14]. Rigorous mathematical models (software
packages) have been developed which have been duly verified by comparing with extensive
experimental data compiled both on laboratory scale and pilot plant scale.

This construction has been successful not only for the improved design of shell and tube heat
exchangers, but also for the design of evaporators / condensers, solar flat plate collectors, solar
parabolic trough concentrators ( in which the absorber tube is made of variable area design)
and also in the case of mass transfer equipment such as gas – liquid absorbers, membrane
separation units and column reactors [12].

One of the major reasons for the attractive augmentation characteristics exhibited by these
exchangers stems from the fact that the tortuous wall geometry of the d-c (diverging –
converging) tube induces additional turbulence into the fluid stream and this increases the
intimacy of contacting between the fluid elements. This is substantiated by the fact that the
velocity profile in a d-c tube even at low Reynolds numbers ( Re ≤ 1500 ) has been observed to
be flat within the central core of the tube and the velocity is seen to fall sharply to zero at the
wall. Such flat velocity profile is obtained in straight cylindrical tubes only in fully developed
turbulent flow ( at Re ≥ 10000). It is also to be kept in mind that the onset turbulence in a d-c
tube occurs at a much lower Re.

Due to the improved radial mixing of fluid elements, the formation of any stagnant liquid film
or thermal layer at the wall of d-c tube is either absent or even if formed, its thickness is quite
low. This is evidenced by the nature of velocity and temperature profiles in these systems
which exhibit a boundary layer character. The velocity of the fluid falls sharply to zero at the
tube wall and the fluid temperature rises sharply in the close vicinity of the heated wall. Such
destruction of stagnant layer at the wall reduces the resistance to momentum and heat transport
and the transfer coefficient gets enhanced.

Due to the diverging – converging wall geometry of the tube, the flow direction of the fluid
varies along the length of the tube ( in the converging section, the fluid flows towards the tube
axis, while in the diverging section, it flows towards the wall ) and the average velocity of the
fluid also varies from section to section. This could be causing a type of pressure recovery, like
that in a venturi tube. This also helps in providing heat / mass transfer enhancement without
the expense of much additional pressure drop.

The thermal penetration distance from the heated wall into the fluid bulk is much larger in the
tubes of this geometry as is evident from the enhancement provided. This is in contrast to the
assumption usually involved with straight cylindrical tubes (while developing heat transfer
68
 Advances in Chemical Engineering

correlations) that the heat penetrates chiefly within a thin annular layer at the wall within
which the velocity distribution may even be assumed linear.

The fabrication cost of these exchangers shall be, no doubt, higher. However, this increased
initial investment could necessarily be recovered within 1 – 2 years since the exchanger operates
with enhanced performance efficiency with relatively little increase in the operating cost.

This design has been successfully adapted to quite a few industries. More large scale industrial
utilization of this design must be anticipated keeping in mind the attractive benefits / features
of this construction.

4. Nomenclature

= free area for flow of shellside fluid in the baffle window, m2

minimum free flow area between baffles at the shell axis, m2

tubeside flow area, m2

= inner heat transfer surface ( inside surface area of tubes ), m2

= crossflow area at or near center line, m2

outer heat transfer surface ( outer surface area of tubes), m2

= shell to baffle leakage area, m2

= tube to baffle leakage area, m2

= baffle spacing ( baffle pitch ), m

= baffle spacing ( baffle pitch ) at shell inlet, m

= baffle spacing ( baffle pitch ) at shell outlet, m

C( smaller of ( and ) , W/K

C( larger of ( and ) , W/K

C = / , dimensionless

specific heat of tubeside fluid, J/(kg.K)

specific heat of shellside fluid, J/(kg.K)

= inside diameter ( ID ) of tubes, m

69
 Advances in Chemical Engineering

= outer diameter ( OD ) of tubes, m

= outer tube limit, m

= inside diameter ( ID ) of shell, m

tubeside friction factor ( for non-isothermal flow ), dimensionless

= shellside friction factor ( for non-isothermal flow ), dimensionless

= bundle bypass coefficient (equation – 115), dimensionless

= fraction of total tubes in crossflow

= caloric fraction, dimensionless

= correction factor to LMTD for multipass construction, dimensionless

= mass velocity of shellside fluid in baffle window, kg/(m2.s)

= geometric average mass velocity of shellside fluid (equation – 101), kg/(m2.s)

= mass velocity of shellside fluid based on flow area , kg/(m2.s)

= mass velocity of shellside fluid based on flow area , kg/(m2.s)

= mass velocity of tubeside fluid, kg/(m2.s)

= tubeside heat transfer coefficient, W/(m2.K)

= value when viscosity correction factor ( , W/(m2.K)

= shellside heat transfer coefficient, W/(m2.K)

= value when viscosity correction factor ( , W/(m2.K)

(ideal) = shellside heat transfer coefficient for an ideal crossflow section, W/(m2.K)

= correction factor that accounts for baffle configuration effect, dimensionless

= correction factor to account for shell to baffle and tube to baffle leakages,
dimensionless

= correction factor that accounts for bundle bypassing effect, dimensioless

= correction factor that accounts for unequal baffle spacing, dimensionless

70
 Advances in Chemical Engineering

thermal conductivity of tubeside fluid, W/(m.K)

thermal conductivity of shellside fluid, W/(m.K)

= parameter defined in equation (23), dimensionless

= parameter defined in equation (44), dimensionless

= parameter defined in equation (17), dimensionless

= parameter defined in equation (43), dimensionless

= effective length of each tube, m

= mass flow rate of tubeside fluid, kg/s

= mass flow rate of shellside fluid, kg/s

= number of shellside passes

= number of tubeside passes

= number of baffles used

= number of tube rows crossed during flow through one crossflow section

number of effective crossflow rows in each baffle window

= number of sealing strips installed per crossflow section

= total number of tubes

= number of transfer units (maximum)

tubeside Nusselt number, dimensionless

shellside Nusselt number, dimensionless

tube pitch normal to flow, m

tube pitch parallel to flow, m

= tube pitch ( overall ), m

= Prandtl number of tubeside fluid, dimensionless

Prandtl number of shellside fluid, dimensionless

71
 Advances in Chemical Engineering

= overall rate of heat transfer, W

correction factor for bundle bypassing effect on shellside pressure drop, dimensionless

correction factor for baffle leakages on shellside pressure drop, dimensionless

correction factor for unequal baffle spacing on shellside pressure drop, dimensionless

= overall dirt factor, (m2.K)/W

= minimum required value of overall dirt factor, (m2.K)/W

= shellside Reynolds number, dimensionless

= shellside Reynolds number defined by Donohue (equation – 100), dimensionless

= modified shellside Reynolds number ( equation – 116), dimensionless

= tubeside Reynolds number, dimensionless

= dimensionless parameter defined in equation ( 108 )

= dimensionless parameter defined in equation ( 109 )

= inlet temperature of tubeside fluid, K

= outlet temperature of tubeside fluid, K

= inlet temperature of shellside fluid, K

= outlet temperature of shellside fluid, K

arithmetic average temperature of tubeside fluid, K

= arithmetic average temperature of shellside fluid, K

= caloric mean temperature of tubeside fluid, K

= caloric mean temperature of shellside fluid, K

= mean temperature of tubeside fluid, K

= mean temperature of shellside fluid, K

= inner surface temperature of tubes, K

= outer surface temperature of tubes, K

72
 Advances in Chemical Engineering

= clean overall heat transfer coefficient based on , W/(m2.K)

= clean overall heat transfer coefficient based on , W/(m2.K)

= overall design heat transfer coefficient, W/(m2.K)

= overall heat transfer coefficient based on , W/(m2.K)

= overall heat transfer coefficient based on , W/(m2.K)

Greek Letters

= tube to baffle clearance, m

= shell to baffle clearance, m

= tubeside pressure drop, N/m2

= additional pressure drop due to flow reversal, N/m2

= shellside pressure drop, N/m2

= pressure drop in ideal crossflow section, N/m2

= pressure drop in ideal baffle window, N/m2

= temperature difference of cold fluid, K

= temperature difference of hot fluid, K

maximum temperature difference, K

logarithmic mean temperature difference, K

( ) temperature difference at the cold end of the heat transfer surface / heat exchanger

( ) temperature difference at the hot end of the heat transfer surface / heat exchanger

= heat exchanger effectiveness, dimensionless

= baffle cut angle ( in radians )

= viscosity of tubeside fluid at temperature , kg/(m.s)

= viscosity of shellside fluid at temperature , kg/(m.s)

viscosity of tubeside fluid at the inside wall temperature ( ) of tubes, kg/(m.s)

73
 Advances in Chemical Engineering

= viscosity of shellside fluid at outer wall temperature ( ) of tubes, kg/(m.s)

density of tubeside fluid, kg / m3

density of shellside fluid, kg / m3

viscosity correction factor for tubeside fluid, dimensionless

= viscosity correction factor for shellside fluid, dimensionless

5. References
1. Perry, RH, Chemical Engineers’ Handbook, Sixth edition, McGraw Hill, New York. 1984.

2. Bhattacharya BC, Narayanan, CM, Computer Aided Design of Chemical Process Equipment, New Central Book
Agency, Calcutta. 1992.

3. Narayanan CM, Bhattacharya BC, Unit Operations and Unit Processes, Volume 1, CBS Publishers, New Delhi.
2006.

4. Eagle A. Ferguson RM, Proc Roy. Soc, A. 1930; 127: 540.

5. Colburn, AP, Trans Am. Inst. Chem. Eng. 1993; 29: 174.

6. Donohue, D.A., Ind. Eng. Chem.1949; 41: 2499.

7. McAdams, W.H., Heat Transmission, Third Edition, McGraw Hill, New York. 1954.

8. Kern DQ, Process Heat Transfer, McGraw Hill, New York. 1950.

9. Bell KJ, Delaware Method for Shell Design, in Heat Exchangers: Thermal Hydraulic Fundamentals and Design,
Kakac et. al (eds.), Hemisphere Publishing Co, New York. 1981.

10. Grimson ED, Trans. Am. Soc. Mech. Eng. 59, 583, 1937; 60, 381, 1938.

11. Narayanan, CM, J Chemical Engineering Japan. 1998; 31: 903 – 909.

12. Narayanan, C.M., Annual Technical Volume (interdisciplinary coordination committee), The Institution of Engineers
(I). 2017; 2: 128 – 139.

13. Narayanan CM, Bhattacharya BC, Industrial and Engineering Chemistry Research. 1998; 27: 149 – 155.

14. Narayanan CM. Heat and Mass Transfer ( Wärme und Stoffübertragung ). 2014; 50: 161 – 168.

74