Professional Documents
Culture Documents
EXCHANGERS
DOUBLE PIPE
COUNTERFLOW:
Return head
part of this apparatus is the double pipe exchanger shown in Fig. 6.1.
The principal parts are two sets of concentric pipes, two connecting Tees,
and a return head and a return bend. The inner pipe is supported within
the outer pipe by packing glands, and the fluid enters the inner pipe
through a threaded connection located outside the exchanger section
proper. The Tees have nozzles or screwed connections attached to them
to permit the entry and exit of the annulus fluid which crosses from one
leg to the other through the return head. The two lengths of inner pipe
are connected by a return bend which is usually exposed and does not
provide effective heat-transfer surface. When arranged in two legs as in
Fig. 6.1, the unit is a hairpin.
102
COUNTERFLOW 103
-10)ETE)-10 (6.1)
of the heat-transfer path before mixing
occurs.
where L is the total length t12
mean deviations of approximately
Equation (6.1) gave maximum 2100 except for water. Beyond
the
100 to Re =
b 0,027 ( (6.2)
104
Slope-
2100
DG
Fra. 6.2. Tube-side heat-transfer ourve.
wetted perimeter
4r(D - D)D-D (6.3)
(6.3)
4TD1 D
In pressure-drop caleulations the friction not only results from the resist-
ance of the outer pipe but is also affected by the outer surface of the inner
pipe. The total wetted perimeter is r(D2 + D:), and for the pressure
pipe.
drop in annuli
4X flow area
D, 4r(DD)
Frictional wetted perimeter 4r(D + D)
=
D, -
Di (6.4)
This leads to the anomalous result that the Reynolds numbers:
same flow conditions, w, G, and 4, are differentfor
heat transfer and pressure drop since D, might
oipe
be above 2100 while D, is below 2100. Actually
both Reynolds numbers should be considered
only approximations, since the sharp distinction
between streamline and turbulent flow at the Quer
PIP
Reynolds number of 2100 is not completely valid
in annuli.
Film Coefñcientsfor Fluids in Annuli. When Fra. 6.3 Annaua, dism-
eters and location of coefi-
the equivalent diameter from Eq. (6.3) is substi- cients.
tuted for D, Fig. 24 in the Appendix may be con-
is the out-
sidered a plot of DG/4 vs. (hoD./E)(cu/k)-%(4/4«)=a14, h,
same manner as à; by
side or annulus coefficient and is obtained in the
D differs from D he is
multiplication of the ordinate. Even though
In double pipe
effective at the outside diameter of the inner pipe.
outside surface of the inner pipe
exchangers it is customary to use the
as the reference surface in Q=
UA AL, and since h; has been determined
based on the areacorrespond-
for A,and not A, it must be corrected. his ID.
diameter where the surface per foot of length is rX
ing to the inside and
the surface per foot of length X OD;
is
On the outside of the pipe
referred to the outside diameter,
again lettingh, be the value of hi
ID (6.5)
P R O C R S S H E A T T R A N S P E R
transfer
106 heat
process conditions may be determined 1rotm
overall
coefficient
of require
ren
the F o uirom
r i d the
The
Fouling Factors. Caleulated f
fulfll the known and and atare
Burface is not know
tion when the surface A
is
0/A
M.
AL. If the own U
Then U Q/A equation from the ttwo
process conditions.
independently
of the
Fourier
film
ohtained resistance,
= R +, t (6.6)
or
hioho (6.7)
hie t+ h are shown in Fio
coefficients and temperatures
The locations of the and Q and A
obtained from values of hio and ho are
When U has been the surtace 4 required for
caleulated from the process conditions, 18 known as desion
the
of A .
The calculation
process can be computed.
been in service irom some time, hom.
When heat-transfer apparatus has
and outside of the pipe, addi.
erer, dirt and scale deposit
on the inside ding
two more resistances than were
included in the caleulation of U by R
(6.6). The additional resistances reduce the
original value of 0, and the required amount
of heat is no longer transferred by the original
oner surface A; T2rises above and ta falls below the
pipe
desired outlet temperatures, although h; and h,
Dipe remain substantially constant. To overcome
Pipe
this eventuality, it 1s customary in design-
ing equipment to anticipate the deposition o
FiG. 64. Location of foul dirt and scale by introducing a resistance R
ing factors and heat-transfer
coefficients. acalled the dirt, scale, or fouling factor, or
r
sistance. Let Ra be the dirt factor for the
inner pipe fuid at its inside diameter and Rio the dirt factor for
the
annulus fuid at the outside diameter of the inner pipe. These may be
considered very thin for dirt but may be appreciably thick for scale, which
has a higher thermal conductivity than dirt.
The resistances are shown
in Fig. 64. The value of U obtained in
Eq. (6.7) only from 1/h» and
1/h may be considered the clean overall coeficient
show that dirt has not been taken into designated by Ue to
account. The coeficient which
includesthe dirt resistance is called the design or dirty overall
The value of A
corresponding to Up rather than coeficient
on which
equipment is Uo provides the bass
two
ultimately built. The
overall coefficients Uo and Un is relationship between t
Ra+Rdo=Ra
1 (6.9)
+R (6.10)
that
Thus supp0se th for a double pipe
exchanger,
Thued to be 300 and i00 respectively, then his and h, have been
computed
s)F)/Btu
(ft)(°F)/Btu will deposit annually inside the
pipe and Rao 0.0015 =
1
Up 0.0158 = 63.3 Btu/(hr)(ft)(°F)
U 0.0158
The Fourier equation ior suriace on which dirt will be deposited becomes
Q = UpA At
(6.11)
If it is desired to obtain A, then hio and h. must first be calculated from
equations such as Eq. (6.1) and (6.2) which are independent of the extent
d the surface but dependent upon its form, such as the diameter and
fluid flow area. With these, Uc is obtained from Eq. (6.6) and U» is
obtained from Uc using Eq. (6.10). Sometimes, however, it is desirable
to study the rates at which dirt accumulates on a known surface
A.
alter the
Ue Will remain constant if the scale or dirt deposit does not
obvi-
mass velocity by constricting the fluid flow area. Un and At will
of thee
OuSly ehange as the dirt accumulates because the temperatures until
the surface is freshly placed
in service
Vary vary from the time
from observed temperatures
instead
becomes fouled. If At is caloulated
determine Ra
be used to
prOCess temperatures then Egq. (6.11) may
ACualy Ba should be referred to the outside diameter as Raio R(A/4i).
diameter, as a layer oldirt, 1ts resia
ylinder is very thin compared with its As shown by Eqs. (2.34)
and
e is nearly the same as that through a flat Wal
introduce a negligible error,
usually well
to correct to the outside will
beloi 1re
below 1 per cent. Por be appreeiable.
scale, however, the error may
thick
PROCESS H E A T TRANSFER
108
for a given fouling period. Then from Eq. (6.10)
RaUp Ue (6.12)
which can also be written
=Uo-U
RaUcU, (6.13)
When Ra (deposited)>Ra (allowed), as after a period of service, the
the
apparatus no longer delivers a quantity of heat equal to the proces
ess
requirements and must be cleaned.
Numerical values of the dirt or fouling factors for a variety of procese
services are provided in Appendix Table 12. The tabulated
fouling
factors are intended to protect the exchanger from delivering less than
the required process heat load for a period of
about a year to a year and a
half. Actually the purpose of the tabulated fouling factors should he
considered írom another point of view. n designing a process
plant
containing many heat exchangers but without alternate or spare pieces of
heat-transfer equipment, the process must be discontinued and the
equip-
ment cleaned as
soon as the first exchanger becomes fouled. ItItisis
impractical to shut down every time one exchanger or another is fouled,
and by using the tabulated fouling
factors, it can be arranged so that all
the exchangers in the process become
of service. At that time all can be
dirty at the same time regardless
dismantled and cleaned
shutdown. The tabulated values may differ from those during single
a
encountered by
experience in particular services. If too frequent cleaning is
greater value of Ra should be kept in mind for future design. necessary,
a
niping, fittings, control regulators, and the pressure drop in the exchanger
pi the end of the
itself. To this head mustorbe added the static preseure atvessel.
itself
Once s
line such as
the elevation pressure of the final receiving
allowance has been designated for an exchanger
definite pressure drop
of a pumping cireuit, it should always be utilized as completely
A & part
in the exchanger, sire it will otherwise be blown off
or
9s possible
expanded through a pressure reducer. Since in Eq. (3.44)
.159
The presure drop per hairpin will be 3.2/(2 >X 32.2)= 0.1 159 ftft of
ot
or 0.07 psi. Unlese the velocity is well above 3 fps, the
the entra
entrae
Luue and
exit lossee may be neglected. Values of V/2g'are plotted directlv.d
1.0 n
e maas velocity for a fluid with a specific gravity ol Fig.27
Appendix.
The Calculation of a Double Pipe Exchanger. All the equa!.
eveloped previously will be combined to outline the solution
ion of
of aad ations
d
pipe exchanger. The calculation consists simply of computing
TABLE 6.2. FLow AREAs AND EqtivaLENT DIaMETERS IN DoUBLE PiPe ExcHAKCPas
below.
Inner pipe
Equivalent diameter Da
a4xfow areaDDft
Di
(63
etted perimeter
PROCESS HEAT TRANSFER
112
w/as, lb/(hr) (ft) From D..
,D. ft,
2.42. From
(6) MasS velocity, G. centipoise X
=
T, or,
a at
l1b/(t)(hr)= the Reynolds number,
Obtain obtain
1D/(hr)(fE"), u
1b/(¬t) (hr)
Re
D
(¢u/R)T*u/4%)0.1 vs. Dg,
() From Fig. 24 in which jn-(h,D./k)
obtain j. compute (cu/k)%
obtained at T. or t,
From c, u, and k, all
(8) by (k/D.)(cu/k)¥ ($ = 1.0) or
( ) To obtain h, multiply ja
( x 1.0 - h, Btu/(hr)({t)(°F) (6.156)
Overall cocficients:
(11) Compute Ue =
hiho/(hso + he), Btu/(hr)(ft*)(°F). (6.7)
Compute U» from 1/U» 1/Uc + Ka.
=
(6.10)
(2) Compute
(13) A from Q = UnA At which may be translated into length.
If the length should not correspond to an integral number of hair.
pins, a change in the dirt factor will result. The recalculated dirt
factor should equal or exceed the required dirt factor by using the
next larger integral number of hairpins.
Calculation of AP. This requires a knowledge of the total length of
path satisfying the heat-transfer requirements.
Inner pipe:
(1) For Re, in (6) bove obtainffrom Eq. (3.46) or (3.476).
(2) AP,= 4fG L/2gp*D, ft. (3.45)
APp/144 = AP», psi.
Annulus:
(1) Obtain D,= L (D,- Di). (6.4)
Compute the frictional Reynolds number, Re, = DIG./4. For Re
obtain f from Eq. (3.46) or (3.47b).
(2) AP.= 4fG°L/2gp'D, ft. (6.14)
(3') Entrance and exit losses, one velooity head per hairpin:
AP ft/hairpin
(AP+AF)p/144= AP, psi.
There is an advantage if both
fluids are computed side by
use of the outline in this manner will side, and the
be demonstrated in Example 6.1.
COUNTERFLOW 18
Solution:
(1) H e a t b a l a n c e :
80+120 (Fig. 2
Benzene, lav -100°F
c0.425Btu/0b) (°F)
2
Q-9820 x 0.425(120 80)= 167,000 Btu/hr
160+1 130°F c 0.44 Btu/(lb)(°F) (Fig. 2
Toluene, Tav 2
167,000 = 6330 Ib/hr
W 0.44(160 100)
80 20
100 Lower temp 20 At At
(5.14)
- 28.8°F
28.8°F
A t -A1 4%0
LMTD=
2.3 log Ata/Al12.3 log streams will show that
neither is viscous
= 130°F av
ta K120 +80)=
H(120
Tav (160 +100) indicates
indieates that the fow
ares
a
Table 6.2
6.2 beasene
A check of larger stream,
pipe. Place the
Proceed n o w
to the inner of the
annulus.
than that
greater
inner pipe is
of the denaene
inner pipe,
in the inner pipe. Cold uid:
toluene D-1.38/12-
0.115ft
annulus, (4)
Hot uid: Flow area,a p r D ° / 4
0.0104ft
(4) Flow sres, 0.115/4=
- 2.067/12-0.1725ft X
D1.66/12-0.138 ft
a(D-D)/4
0.138')/4-
0.00826 ft
0.1725
(9) h j
Eq. (6.156)] Eq. (6.15a)]
0.091
0.085 X 1.725 X 1.0
= 167 0.0762 236 X .1T X 1.78 X 1.0
= 333 Btu/(hr) (ft) (°F)
323 Btu/(hr)(ft)(°F)
(10) Correct h; to the surface at the OD
ID
hie hi X OD Eq. (6.5)1
333 1.38 276
U 149 + 0.002
UD-115 Btu/(hr)({t")(°F)
Summary
Uc 149
UD 115
(14) The surface supplied will actually be 120 X 0.435- 52.2 ft2
The dirt factor
wil aceordingly be grenter than required. The actual
design coeficient is
UD 167,000
52.2 288 1 Btu/(hr)(ft°)(°F)
149 111
R TU 149X 111 0.0023 (hr)(ft*) (°F)/Btu (6.13)
Pressure Drop
(1') D, for pressure drop differs from D.|(1) For Re, = 89,500 in (6) above
for heat transfer. 0.264
D, = (D; - D,)
Eq. (6.4) f =0.0035 + Tel. Eq.(3.476)
= (0.1725 - 0.138) = 0.0345 ft 0.264
-0.0035 t39,5004 0.0057
Re = DG. 8 0.88, P = 62.5 x 0.88 = 55.0
= 8.3 ft
ITable 6]
55.0
4fGL X
P, -8.3 144 3.2 psi
(2) AF.
2p D. Allowable AP, = 10.0 psi
X 120
4X 0.0071 x 767,0002
2x4.18 X10 X54.3 X 0.0345
= 23.5 ft
G 767,000 3.92 fps
) V
3600p 3600 X54.3
3.922
0.7 ft
Fi-327
=
-3
X2x322
AP,= (23.5+0.7)54.3 9.2 psi
144
Allowable AP.- 10.0 psi
From
respectively, and K, 0.17.
=
Obtained
drop of 10.0 psl. Suppose,
Suppose,j how.
against an allowable pressurewere 15 or 20 psi and exceeded 41
drop exceeded the
ever, that the calculated pressure the heat load be transferred with
available head. How then might so th
is the use ot a by-pa8s
ass so
available pressure head? One possibility
of the fluid flows through the exchange
exchane
the
three-quarters or two-thirds
only This does not provide an i
and the remainder through the by-pass. unfavorable changes in
the reduced flow causes several the
solution,
design. since
velocity G.
(1)and
Thethereduced
film coefficient
flow through
h. Since
the exchanger
both of the
reduces
coefficients
the are
decreases
reduction in Ga alone
nearly alike, 323 vs. 276, any sizable the annulus, it h.
by nearly G. (2) If less liquid circulates through
to be cooled over a longer range than from 160 to 100°F so that, uDon
t2
2
Fiu. 6.5. Double pipe exchangers in series. F1a. 6.6. Seriesparallel arrangenent.
mixing with the by-pass fluid, the process outlet temperature of 100°F
results. As an example, the portion circulating through the annulus
might have to be cooled over the range from 160 to 85°F depending upon
the perentage by-passed.The outlet temperature of 85°F is eloser to
the inner pipe inlet of 80°F than originally, and the new cold-terminal
difference Ali of only 5°F greatly decreases the LMTD. The two effects,
decreased Uc and LMTD, increase the required number of hairpins
greatly even though the heat load is constant. Reversing the location
of the streams by placing the benzene in the annulus does not provide a
solution in this case, since the benzene stream is larger than the
stream. The possibility of
toluene
reversing the location of the streans shoua
always be examined first whenever the allowable
pressure drop canno
be met.
A solution is still possible, however, even
when all the above n8
failed. When two double pipe exchangers are
arrangement is shown in Fig. 6.5. Suppose thatconnected
in series,isxtoo
the stream which
large to be accommodated in several exohangers in series is divided inhalf
and each half traverses but one exchanger through the in
inner pipe
COUNTERFLOW 117
Fig. 6.6. Dividing a
stream in half while
keeping the flow area constant
oroduces about one-cighth of the series pressure drop, since G and L will
be half and the product of G'L in Eq. (6.14) will be one-eighth. While
the film coefficient will also be reduced, the unfavorable
temperature
difference of by-passing can be circumvented. Where there is a
substal
tial unbalance between the weight flow of the two streams because one
operates Over a long range and the other over a very short range, the large
stream may be divided in three, four, or more parallel streams. In larger
services each parallel stream may also flow through several exchangers
in series in each parallel bank. The term "parallel streams" should not
be confused with "parallel flow." The former refers to the division of
the fow of one fuid, while the latter refers to the direction of fow
between two fluids.
The True Temperature Difference for Series-parallel Arrangements.
The LMTD calculated from T'1, T2, ti, and t2 for the series arrangement
will not be the same for a series-parallel arrangement. Halí of the pipe
fuid enters the upper exchanger II in Fig. 6.6 where the annulus fiuid is
hot, and half enters the lower exchanger I in which the annulus fluid has
already been partially cooled. While exchangers in series do not transfer
equal quantities of heat, the series-parallel relationship is even more
the total
adverse, the lower exchanger accounting for relatively less of
heat transfer. If the true temperature difference is called At, it will not
both of
be identical with the LMTD for the process conditions although
the exchangers operate in counterflow.
I and II. The
Consider the two exchangers in Fig. 6.6 designated by
of the parallel streams a
intermediate temperature is 1, and the outlets
designated by 4 and Their mixed temperature is a
For exchanger I, containing half the surface,
Q =
WC(T -
T) = X LMTD (6.16)
and
LMTD, 4T= (6.17)
Let
(T = 2WC
wc
UA
2WC n (6.10)
Similarly for exchanger II
Qn=WC(T1 - T)= X LMTDr
(71- )
2
- ( 7 - ti)
(6.20)
LMTD n (T-4/(T = ) (6.21)
Let
R-T 2WC wC
UA
(6.22)
Since c and C were assumed constant,
WC
R = R" = R =
2WC (6.23)
Let
S- M = R'S
M T-
Similarly let
UA In(R- 1D(71- ) + V T - 7 - 1
WC R V(T-h7%-h
(6.29)
At is the single value for the entire series-parallel arrangement; thus
UA At = WC(T1 - T) (6.30)
WC
AfUAU T1-7) (6.31)
It is conveníent in this derivation to employ a definition for the true
temperature difference in terms of the maximum temperature span
T1-h
(6.33)
Equating (6.31) and (6.32),
WC(T-T)
UA(TT-
ROCESS HEAT TRANSFER
120
M= (Ti- T,)/(71
-
- P"
) [G) +
If developed in a generalized manner it can be shown that, for one aer.
(6.34)
where
-2.3
=|()E)"+ + (6.35a)
R=1T
pi' 1/n
where
23TloE(R (P) +P" (6.35 )
h a (6.36)
(6.37)
Similarly for two resistances in series employing the viscosity corrections
for deviation from the isothermal the clean overall coefficient is again
Uchiehioo
+ho (6.38)
Bxample 6.3. Double Pipe Lube Oil-Crude Oil Exchanger. 6,900 Ib/hr ofa
26°API lube oil must be cooled from 450 to 350°F by 72,500 1b/hr of 34 API mid-conti-
nent crude oil. The crude oil will be heated from 300 to 310°F.
A fouling factor of 0.003 should be provided for each stream, and the allowable
pressure drop on each stream will be 10 psi.
A number of 20-ft hairpins of 3-by 2-in. IPS pipe are available. How many must
be
be used, and how shall they be arranged? The viscosity of the crude oil may
3.0
obtained from Fig. 14. For the lube oil, viscosities are 1.4 centipoises at 500°F,
to introduce an errorif
at 400°F, and 7.7 at 300°F. These viscosities are great enough
/#).14 1 is assumed.
=
Solution:
(1) Heat Balance:
6900X 0.62(450 350) 427,000 Btu/hr
Lubeoil, Q
-
Crude oil, Q =
72,500 X 0.585(310 300)=427,000 Btu/hr
2) A: Cold Fluid Diff.
Hot Fluid
310 140
450 Higher temp
or annulus,
to put the 72,500 lb/hr of orude into a single pipebe employed in
1t willbe impossible Assume, as a trial, that it will
is too small.
8unce the flow area of each (6.35a)
two parallel streams. At-87.5F
P R O C E S S H B A T T R A N S P E R
122
(3) Caloric temperstures: F.0.396
(Fig. 17)
50 0.357 A -0.43
40 389.5°F
(5.28)
T 350 x 0.395(450
360)
304°F
(5.20)
t 300 x 0.395(310- 300)
Equiv diam, D, -
(D} -
D})/D
Eq. (6.3)
0.199)/0.199 = 0.13 ft
(0.256 -
Uc 34.0
UD 28.2
(13) Surface:
A 427,000
UpA 28.2 X87.6173 f
External surface/lin ft, a'" 0.622 ft
=
(Table 11)
Required length = 9173 278 lin ft
This is equivalent to more than six 20-ft hairpins or 240 lin feet. Since two
streams are employed, use eight hairpins or 320 lin. feet. The hairpins should parallel
have
the annuli connected in series and the tubes in two parallel banks of four
The corrected UD will be UD = Q/4 At = 427,000/320 x 0.622 x 87.5 = 24.5.
exchangers.
The corrected dirtfactor will be Ra = 1/Up - 1/Ue = 1/24.5 - 1/34.0 0.0114.
Pressure Drop
(1) D- (D, -D,) Eq. (6.4)1|(1) For Re, - 133,500 in (6) sbove
=(0.256-0.198)=0.058 ft f
0.264
0.0035+ s.50020a-0.005376
Re-DG/
-0.058 X 335,000/7.25 2680
s
[Eq.(3.476)
= 0.76, p = 62.5 x 0.76 47.5
0.264
f 0.0035 *28800p. 0.0132 Pig. 61
(Eq. (3.476)]|Halves of the tube fuid will flow throughh
-0.775, p= 62.5 X 0.775- 48.4 only four exchangers.
Pig. 61
g 4G
) AP,- ()A 2gpD
29p D 4 X 0.005375 X 1,560,000 X 160
4X 0.0132 X336,000 320 2X4.18 X10 X47.5 X0.172
2 X418 X 10X 48.4 X 0,058 25.7 ft
16.7 t au 25.7 X 7.08.5 psi
(8V G 386.000
() V- 3600 3500 X 48.41,9fps Allowable AP, 10.0 pai
0.45fT the fow had not been divided, thepreas
A8 -8( sure drop would be nearly eight times
X 85.8 pai great, or about 60 pai.
AP.L67+045)
14
Allowable AP. 10.0 pai
PROCESS HBAT T'RANSFER
124
PROBLEMS
Up =
20, (6) Uo
factor when (a) Uc =
30 and 60 and
= a.
clean streams?
Specify between two moderately data were available
oversized because no
6.2. A double pipo exchanger was w a s originally
the
designed to
rate at which dirt accumulated. The exchanger cool
acetic acid from 250 to 150°F by heating 19,00000 Ib /hr
3,000 l6/hr of 100 per cent
coefficient Up 85 was emploved
=
factor?
C0UNTKRFIDW 125
.11. 13,0M) }»/hr of 29°A Pl ga* oii
cooled from 40 ts A9°¥ by rnting 57°AP (s kxninpie 6.2 n text ior viscositics) 1s
in aN many 3- by 2 in. 1Pa dunhle
gaiine under pressure from 220 to 230°F
pip 2) ft
o 10 pmi are prernittai alorngwith a minimum hsirins s8 are required. Pressure drops
hairpins are required? (b) How *hall they he dirt factr of 0.04. (a) How many
factor? nrrangel? (c) What is the final dirt
6.12. 100,0XK) }b/hr of nitrobenzene in ty he eoled
fromn 325 to 275°F by benzene
heated from 100) to 300°V. Twenty-foot hairpinsR of 4- 3-in.
be employed, and by IP8 double pipe will
preneure droa of 10 pai are
0.004 in required. (a) low iany permiasible. A minimum dirt factor of
h»irping
nrranged ? () Whnt in the final dirt fauctor?
nre
required? () How shall they be
M
Number of parallel streams
OD Outaide diameter, ft or in.
dimensionles
P Temperature group (71-h)/(7'1- th), dimensionless
p Temperature group (T1-ta)/(T1- h),
AP Presure drop, pai
Hest fow, Btu/hr
7)/l%-), dimensionless
Temperature group (T%- Ti)/n(t-h), dimensionless
Temperature group (T dimensionless
Temperature group n(P, #/t%4),
HEAT
TRANSF ER
PROCESS
t1)/(T1
- -
t),
Temperature group (l -