Fig.

1
Computer input form
ITEMA fixed-Iube-sheel design)
Item No.
:5AMPLlS. I Shell side
Shell design Mela1 temp. Elastic (oeff. of Woll thickness
pressure minus 70 F. mod. x 10-
6
expo x 10
6
Shell D.d. (corroded)
2.5"0 36Z.S 27.3 6.9 If?o ,I
{26.S for steel {7.0 For steel
01 500
0
F,I 01 500
0
F.J
Tube side
Tube design Metal temp. "Elastic (oeff. of Tube length,
press. minus 70F. modx. x 10-
6
expo x 10
6
Tube D.d. Tube wall in.
7-5"0337.5 tJ.7.G 6.B 0.75 .060 .;1gB
Tube sheet
Eloslie
Tube {Qunt Tube pitch mod. x 10-
6
Stress G Ml M2
tJ.937S 17500 17,58 11191'/'1- Illh7/6'1
rnslIle ciiameter Operating Bolting
of integral moment
moment
Expansion joint
Inside pressure port
diom. of
expo joint
.9-F]
1.0 wilhuu, IUIIII
0"
o with ioint
Computers help design tube sheets
DALE L. GULLEY
Fig. 2
Gulley Computer Associates
Heat flow through tube walls Tulsa
Fluid lemp. inside lube
IN THE LATEST issue of TEMA, the
calculation procedures for designing
lube sheets in a fi1\ed-tube-sheet ex
changer are greatly '. The
new calculatior prosedure takes into
Temp. drop - turbulent
account differential thermal expan
region inside lube
sion. It also takes into account the
Temp. drop Ihrough
viscous loyer inside lube interaction of the various components
Temp. drop Ihrough dirt loyer, inside
in a fixed-lube-sheet exchanger aDd
Temp. drop fhrough lube woll
Temp, drop Ihrough dirl loyer, oUlside their effect on the tube sheet.
.Temp. drop Ihrough viscous
The result is a more -accurate anal
loyer outside lube
Temp. drop Ihrough lurbulent
ysis of the tube sheet, but a more
region, 9uIside
lengthy calculation procedure. It also
Fluid lemp.
provides the basis to check to see if
oulside lube
an expansion joint is needed. ThisJlew
trial-and-error procedure is a good
computer application.
Instead of using the design' p.essure
t:: Slognonl film, Ro
for calculating the tube-sheet thick
ness, special effective design pres
OGJ
sures are calculated for both the shell
and tube sides. For, TEMA R pro-
THE OIL AND GAS JOURNAL-MAY 20:1974 70
cedures, methods can be found on p.
40 and 41.'
Input
Fig. I is a computer input form.
There are four main categories: shell
side,. tube side, ,tube sheet, and ex
pansion jOint. It is necessary to know
the metal temperatures of both sides.
n is tempting to simplify things and
use the average fluid temperature on
each side, but this would give you a
thicker tube sheet than' you really
need.
If the thermal resistance of the
tube-side fluid is relatively high, the
two metal temperatures will be much
closer together than the average fluid
temperatures indicate.
Normally, the heat exchanger is
designed for operating conditions, but
if the nature of the process would give
unusual start-up conditions, then these
should certainly be taken into ac
count.
For metal temperature on the shell
side, use the average shell fluid tem
pertaure. The calculation of the tube
side metal temperature requires the
knowledge of the heat-transfer coef
ficients.
For the clean condition an average
tube-side metal temperature is cal
culated (refer to Fig. 2):
1. If shell side is hottest
TMavg = Tavg +
LTMD,XRio
(Ia)
(Rio+Ro+Rw)
2. If shell side is coldest
TMavg = Tavg -
LTMD,XRio
(Ib)
(Rio+Ro+Rw)
WHERE:
LTMD, = Corrected log me a n s
temperature difference, 'F.
Rio = Inside heat-transfer resist
ance referred to outside surface.
Ro = Outside heat-transfer resist
ance.
Rw = Tube-wall resistance.
T,.g = Average tube-fluid temper
ature, OF.
TMllvg = Average tube-metal tem
perature, of.
The resistances used in these equa
tions are Simply reciprocals of the
heat-transfer coefficients. They have
been put in the form of resistances to
more easily understand the equations.
Computer output
r,
Fig. 3
TUSE SHEET THICKNESS FOR FIXED TUBE SHEET
INPUT SAi-1PLE
DESIGN PRESSURE=
TEt".PERATURE-70.
ELASTIC MODULUS X 10 -6
COEFFICIENT OF EXP. X 10 +6
O. D .
THICKNESS
t. D. OF EXPANSION
TUBESHEET
TUBECOUNT
PITCH
ELASTt C r-'ODULUS
AllOWAS'LE STRESS
G
OPE RAT ING IlO"AENT
90l Tl NG ,"ImlENT
JOINT
X 10 -6
SHELL TUBE
250. 250.
362. 337.
27.30 27.60
6.90 6.80
18.0000 0.7500
0.2190 0.0600
28.00
260.
0.9375
28.00
17500.
17.5800
119144 .
116716.
OUTPUT -------------------------------------------
J= o. K= 0.357 Fa= 2,399
PRESSURE OF DIFFERENTIAL THERMAL EXPANSION
o.
SHELL BOLTING PRESSURE
133.
PS'= -192.
PT'= 250.
EFFECTIVE SHELL DESIGN
-192.
EFF=CTtvE SHELL DESIGN
-192.
TUBE BOLTING PRESSURE
136.
PRESSURE-BENDING
PRESSURE-SHEAR
EFFECTIVE TUBE PRESSURE
578.
TU9E SHEET THICKNESS-SHEAR
0.364
TU9E SHEET THICKNESS-SENDING
1.598
If the inside and outside fouling are
equal, it doesn't matter if you evalu
ate the exchanger in the clean or
dirty condition. Under certain condi,
tions the difference in metal temper
ature will be greater when the ex
changer is fouled. This will be when
the ratio of the shell-side fouling to
the tube-side fouling is greater than 1.
For the foul condition the average
tube-side metal temperature is cal
culated:
1. If shell side is hottest
TMllvg =
Tovg+U (Rio+Rif) (LMTD, (2a)
2. If shell side is coldest
avg =
Tovg-U (Rio+Rif) (LMTD,) (2b)
OGJ
WHERE:
Rif = Inside fouling resistance
U = Overall heat-transfer coeffi
cient
After the metal temperatures are
calculated, 70' is subtracted from
them for input data.
The modulus of elasticity can be
found on p. 192.' For the shell wall
thickness, the t hi nne s t condition
should be used. Therefore, from the
wall thickness when new subtract- the
corrosion allowance. If the shell is
constructed from pipe, also subtract
the mill tolerance of 1216 % .
The tube-wall t hie k n e s s and the
length are expressed in inches.
G is defined as the inside diameter
of the integral pressure part. For
TEMA type BEM, use the shell Ld.
THE OIL AND GAS JOURNAL-MAY 20. 1974 72
;' ,
-'"", '7"';1;'-'-
;,{.

For TEMA type CEN, use the channel
i.d.
MI and M2 are the operating and
bolting moments, respectively. For an
integral type tube sheet (CEN), this
will be zero. For other types this
comes fmm the corresponding flange
calculation.
Fig. I is a completed input sheet
for this example:
Type TEMA ,BEM
Shell design pressure, 250 psi
Shell IS-in. std. wt steel pipe with
Va in. C.A.
Channel design pressure, 250 psi
Channel IS in. std. wt steel pipe with
Va in. C.A.
Tubes, A,214 s tee I with 0.060 in.
wall, 260' %in. on 15/16 by 24 It long
Shell temp. in = 525' Temp. out
= 340' F.
Tube temp. in = 320' Temp. out
= 445' F.
Shell heat-transfer coefficient = 231
Fouling = 0.002
Tube heat-transfer coefficient = 132
Fouling = O.ODI
28 in. Ld. expansion jOint in shell
Calculation of metal temperatures:
Shell
For shell use average temp.: (525
7A
Fig. 4
+340)12 = 432.5'
For computer input: 432.5 - 70' F.
= 362.5'
Tube
1. Calculate MTD
GTD = 525 - 445 = 80
LTD = 340 - 320 = 20
Difference 60
MID = 43.2
2. Calculate resistances
Rio (l/tube h) = 1/132= 0.00758
Ro (I/shell h) = 1/231 = 0.00433
Rif = 0.001 X 0.75/0.63 = 0.00119
Rof = 0.00200
3. Since shellside is hottest, use
Equation 2a.
TM""g = T""g + U (Rio + Rif)
(LTMDC)
= (320 + 445)12 + 66.1
(0.00758 + 0.00119) 43.2
= 407.5' F.
For computer input: 407.5-70' F.
= 337.5' F.
Shell wall thickness.
Subtract out corrosion allowance
0.375 - 0.125 = .25
Allow for mill tolerance for piping
0.25 X 0.S75 = .219
Metal temperature for tube sheet.
To be conservative.
fluid temperature and tube fluid
temperature at coldest tube sheet.
(Large va l.u e s of E, give thicker
tube sheets).
Temp. = 330'
G-For TEMA type BEM this would
be the Shell Ld.
Calculate this using above shell
wall thickness
J.D. = D.D. - 2 X TK. = IS
2 (0.219)
= 17.58
J = 0 since there is an expansion
joint.
Output
Fig. 3 shows the computer output.
The top part reflects the values used
on input. This provides a record for
verification. The bottom part of the
output shows the important values
computed.
The last two variables shown are
the calculated values for thickness us
ing. the two different equations. The
deSigner uses the la.rgest value.
Interpretation
Fig. 4 is a flow chart for the com
puter logic. The tube-sheet thickness
is_ determined by trial and error. The
starting thickness is determined from
TEMA equation R: 7.122. The pressure
used is the highest of the two design
pressures.
Usually the thickness calculated ac
cording to the bending equation will
be controlling. The shear equation
controls at high design pressure.
The bolting moments imposed on
the tube sheet can substantially in
crease the calculated tube-sheet thick
ness. At high design pressures where
the attached flange gives high mo
ments, it may be better to use an in
tegral tube sheet such as a TEMA
type CEN. The integral tube sheet,
which has zero bolting moment, cal
culates to only a fraction of the thick
ness that a combined flange and tube
sheet unit would.
With a few modifications, this pro
gram can be made to analyze for
expansion-joint reqUirements. It uses
the same effective design pressures
and pressure of differential thermal
expansion to calculate the stresses.
Reference
1. Standards of TEMA, Fifth Edition,
1968 and 1970 addenda.
THE OTT. ANn GAS JOHRNAT._MAV?O _1Q7.:1