Page 244 - Applied Process Design For Chemical And Petrochemical Plants Volume III

66131_Ludwig_CH10G 5/30/2001 4:37 PM Page 207

Heat Transfer 207

Surface Area Required When a liquid is vaporized in horizontal tubes, the initial
overall coefficient is several times the value for forced con-
Q 5,000,000 Btu>hr vection single-phase heat transfer. As the amount of vapor
5,000,000 increases up to 100%, the coefficient falls off, down to a gas
Area a b 130 ft 2 convection coefficient. The work of McAdams 84, 85 is some of
38,600
the limited literature in this type of heat transfer.
Use 1130211.102 143 ft 2
Reboiler Piping
Tube length corrected for tubesheets 5.73 ft
11432 The mechanical design of thermosiphon reboiler piping
Number of tubes 96 tubes
10.2618215.732 must be carefully examined for (a) system pressures and (b)
elevation relationship between the liquid level in the distil-
A 16-in. O.D. shell is required.
lation column and the vertical or horizontal reboiler. Kern 199
provides an excellent presentation on this topic, including
Vapor Nozzle Diameter 200
the important hydraulics. Abbot also presents a computer
program for this topic.
(16)(0.667) 10.7 in. approximately
Use a 10-in. vapor nozzle
Film Boiling
Liquid Inlet Nozzle Diameter
Normally the designer does not try to establish film boil-
gpm 12.5>tube2196 tubes2 240 gpm ing conditions for the vaporizers or reboilers. However, for
systems set by other controlling processing conditions, these
4-in. is too small; use 6 in.
film conditions may be imposed. In such cases, they should
be recognized and handled accordingly. The principles of
Design Notes
design for the equipment are the same as other such equip-
When steam pressures in the chest are near atmospheric, ment, and only the actual value of the coefficient affected
condensate can rise in the shell and drastically reduce avail- needs special attention.
able surface—if the trap is too small to dump steam into the
condensate return system or if the condensate return pres-
sure is greater than the calculated chest pressure required. Vertical Tubes, Boiling Outside, Submerged 14
In these cases, the steam pressure will have to rise in the
chest to overcome this error, if steam pressure is available. If Conditions: Above critical temperature dif-
not, the reboiler will not deliver design flux. ference for nucleate boiling
3
1
Proper condensate removal is important. An inverted split Tube sizes: / 8 -in.– / 2 -in. O.D. (data of
cup inside the shell, with the upper capped end above the correlation)
3
nozzle and the lower open end / 4 -in. above the bottom Reynold’s number: 800–5,000
tubesheet, should be used to cover the outlet nozzle. This Range of error: 14% (greatest single value in
can be made by splitting a pipe that is one size larger than correlation data is 36%)
the condensate outlet down the centerline. In this case, a 2-
in. split is adequate. This cup must be fully seal welded (not 4w 0.6 2 1>3
h a ¿ 0.002 c d c d (10-202)
3
3
tack welded) to force condensate down to the / 4 -in. clear- D o
k a v 1 L v 2g
ance above the bottom tubesheet. A common error is to
allow 6 in. or more above the tubesheet for the centerline of where w maximum vapor flow per tube, lb/hr
D o tube O.D., ft
the condensate outlet. In this case, 6 in. of tube is 10% of the

viscosity of vapor, lb/(ft) (hr)
surface. If the cup is not used, add 10% more tubes to cor-
k a thermal conductivity of vapor, Btu/hr (ft)(°F)
rect for the dead liquid space near the bottom. This is in 3
v density of vapor, lb/ft
addition to the 10% safety factor. L density of liquid, lb/ft 3
In the region of low t (less than about 10), the heat flux g 4.17 10 , ft/hr 2
8
is about 10 times as great for water under forced convection h a average heat transfer coefficient over entire tube,
82
2
(agitation) as for natural circulation. This does not hold at Btu/hr (ft )(F)
the higher t values. The critical t is practically unaffected
by agitation or increased velocity over the value at natural The data for organic fluids and low temperature nitrogen
convection. In general, the heat flux is lower for a given t fit. However, methanol data gives coefficients many times
at lower pressures. Likewise, the peak Q/A is lower. higher.