Ask the Experts
Distillation
Q. How can I design an optimized distilation column?
“The first step is to establish the optimum design procedure
fora standard cistiljation column, Tailor that desiga to opti-
‘mize column operation for flexibility, low capital eos, energy
efficiency or handling difficult feeds. In this article, we
assume thatthe feed isa mixture of hydrocarbon components
‘with near-ideal behavior
Fora standard design se the tower pressure such thatit
willed tothe design ofan economical condensing exchanger:
The cooling water or ai approach temperature usally sets the
condensing temperature. I the presure fs Set 00 low, lange
ann costly condenser wil be require. However, the higher the
pressure the lover the relative volaity between the ightand
heavy keys. In other word, separation of lighter componenis
requires higher column pressures in order to operate at reson
able condenser temperatures. Separation of heaviercomponents
can be accomplishes with lawer cokumn pressures. UOP com
monly uses a 30°F cold-end approach on aircoolrs Typically,
aircondenser outlets wll be designed at 130-145°F. Water con-
denser usually cool to100°F.
“The minimum number of sages and reflux required can be
calculated using the Fenske and Underwood equations, respec
tively, while the ililand corelation may be used to determine
the number stages vs. refux rates. Set the reflux rate to 10%
above the minimum reflux valu to find the intial (or minimum)
‘numberof sages. Run several rigorous column simulations to
establish the appropriate number of stages. Start with the mini-
‘mum numberof stages forthe frst rn and a 5% forthe see-
ondnin.
Sage efficiency may be determined by experimentation, cor-
relations, consulting literature o drawing onthe knowledge of
othe skilled engineers. Different sections of the column may’
‘operate at diferent tray efficiencies. Many towers designed at
UOP have stage efficiencies of 75-80%. Mos of the vacuum,
system designs have eficiencies of 15-20%. The drying section
‘of abenzene tower has an efficiency of 15%.
“To determine the rato of the light key tothe heavy key, iden-
tify the stage inthe tower where this ratio occurs in the Hiquid
‘phase. This shouldbe the optimum feed point.
exible columns
Ifthe feed composition and heat content are variable, such as
‘with naphtha or kerosene systems, stages may be added to the
column to achieve the desired separation. Additional trays
should be added to the tower section that needs more flexibility.
Kerosene towers often have extra trays (2.0t 3 stages) in each
column section, Columns that are designed to handle feed varia-
tions are not optimized for any one feed.
‘The feed-variability challenge can also be addressed with
additional reflux. Kerosene systems have few addtional trays
44 woncspmagaenecg Famuay200¢ CEP
rere
eae en a
ayes Fou B47 992.8 mae
'UoPant caret spas
Sis iactoneatn atosein
aha oft Crp, satn.
roe inch VOR ol Een Gey
foc ap 133 Ema i yeast
tenon ees on
Sea rowsi ‘Simei setcrtetcenhcb toni
th Unk of Messe Art en isa menber ofA
snl small additional capacity (159%) inthe eflu,
Ifthe product specifications are variable, the tower must be
‘capable of achieving the most difficult specification, The cost to
bail this capability into coftmn is high, and should be just
fied, High-purity products can require a large number of trays
(e4g. 205 fora polypropylene (PP) splitter) and large reftax
ratios (eg, inthe PP spliuer, RID= 12:1).
Lowest capital cost vs. minimum energy design
‘There isa tradcolf between the amount ofreflax and the
number of stages needed to achieve a desired separation. A
higher tex means that fewer stages are required. However,
‘more reboiler cay is required, meaning tht the column diame-
ters lager, and greater Surface area is required in the condenser
and reboles. There is an economic traect between energy
requltemeras dae to reflux changes and the capital cost Ussaly
the condenser and the reboiler are the most costly components in
a conventional fractionator, followed by the tower shell and
‘rays, High-efficiency trays (or packing) and close tay spacing
allow the tower diameter and tangent length to be minimized.
‘On the other hand, a tower with alower energy requirement
js designed at or near minimum reflux, Rigorous column simu-
Iations can be used to determine the minimum reflux require-
‘ment. The column should then be designed with a small margin
(typically 10%) for operational variability. Reducing the operat-
{ng pressure will minimize te reflux requirement.
Advanced options
In some cases, a better design can reslt by using the
advanced options, described below:
Dividing wall — UOP has commercialized several dividing
wal columns for the separation of kerosene and naphtha prod
vets (CEP, May 2002, p. 64). Energy savings of up to 30% and
capital savings of upto 25% have been obtained.
_Heat-pumped — Most of UOP’sheat-pumped towers ae
propylene/propane spliter, which achieve difficult propylene
purity and recovery specifications. The current design uses a
{two-stage heat-pump compressor with high-flux tubing, promot-
ing better heat transfer, and a reduction in heal-exchange area.
Pressure-taged — Pressue-staged fractionation columns
improve the energy efficiency of petrochemical complexes, One
‘pplication isto use one high pressure column in a complex to
‘provide the reboiler heat tothe ret ofthe columns in the com
plex. Another application i to split large vacuum fractionation
{ower into two small towers, one operating at 20 psi and the
‘other at 3 psia ll of the condenser duy in the high pressure
column is recovered in the reboier ofthe second tower. The
‘200s AICHE Spring National Mectin’s (New Orleans, LA;
‘Apr. 25-29) Seventh Topical Conference on refinery processing
will contain a session on energy efficiency, one of which will
‘etal the advanced distillation methaxls.