Chemical Engineering April 2022 - 53

pressure drops, especially
in subcooler E2, could only
be explained by the presence
of gas.
entry
The subcooled liquid
(by splashing
onto
the drum surface) would
generate a large surface
to absorb the gas. This
gas exited in the drum liquid.
Some ended in the
distillate product, the rest
was refluxed. In the tower,
the
the
gases absorbed
reflux stream
FIGURE 13. Fuel gas absorbed by the liquid in a reflux drum
causes pressure fluctuations in towers and condensers
area between the nitrogen and the
liquid is then the stagnant surface,
that will reach saturation and, as
long as it remains undisturbed, will
not absorb additional nitrogen.
A possible " band-aid " solution
would have been to raise the liquid
level, but this was undesirable because
the feed frequently fluctuated,
so the level was controlled at 50%.
Other temporary solutions would be
to keep the feed warmer or reduce
the drum pressure, which the plant
did until a permanent solution could
be implemented.
In another tower, a split-range
pressure controller on the reflux
drum would either bring fuel gas in or
vent gas to the flare (Figure 13). The
valve adding fuel gas was always
widely open and the valve to the
flare closed, which was desirable to
minimize flaring. The wide opening
means that a significant quantity of
fuel gas was absorbed by the drum
liquid. The pressure control kept the
drum pressure steady, but the tower
pressure would swing by as much
as 7 psi.
The tower overhead at 170ºF was
totally condensed in exchanger E1,
with an outlet temperature of 90ºF,
then subcooled by exchanger E2 to
70ºF. The pressure drop across the
condensers varied between 3 and
15 psi depending on the rates. As
the rates increased, the pressure
drop across each exchanger would
rise, reaching as high as 7 psi across
each exchanger. The high exchanger
in
were
desorbed, and gas blanketed
the condenser and
subcooler, causing the
high and fluctuating dP
that destabilized the tower
pressure control.
A solution would be to extend
the condensate pipe to near the
bottom of the reflux drum. This
would shield the condensate from
the fuel gas and mitigate the absorption
of fuel gas. The only absorption
would take place right at
the drum surface. As long as the
surface remained undisturbed, it
would quickly become saturated
and there will be little additional fuel
gas absorption.
Deficient or variable gas quantity
Tower design is usually based on a
range of expected quantities of noncondensable
gases entering the
tower. When the actual quantity deviates
from the design, the control system
may not keep up with it, causing
swings or other major operation issues.
Theunick's statement in Ref.
18 - " Effectively managing the
vapor inventory is the key to controlling
the pressure " - summarizes
this issue.
According to Equation (1), when
the product vapor pressure, VPliquid
is low, and at the same time there
is not enough non-condensable
gas to keep the pressure Pgas high,
the drum pressure, Pdrum, will fall,
at times below the desired setpoint
on the pressure controller. Lowering
the controller setpoint reduces
tower pressure, raising vapor velocities
in the tower, which in turn
may induce premature flooding
or
entrainment,
CHEMICAL ENGINEERING WWW.CHEMENGONLINE.COM
possibly
APRIL 2022
tower capacity. Attempting to keep
the controller at the higher (desired)
setpoint may induce cycles of accumulating
and discharging the little
non-condensable gas that is present.
The swings can be severe - in
one case described in Ref. 18, tower
pressure cycled by 20% every 6 min,
inducing temperature swings as high
as 90ºF, temperature inversions and
off-specification product. The problem
was generated by the absence
of a low-boiler, which acted as a
non-condensable gas. The problem
was solved by adding nitrogen
downstream of the condenser to
increase Pgas.
Another superb illustration of this
issue is in the vacuum tower described
by Van der Merwe in Ref.
17. Figure 14 is a simplified diagram
of the tower based on Ref. 17. The
tower had an internal large dimple
plate condenser (E-1) using a
treated condensate stream referred
to as warm tempered water (WTW)
entering at 130ºF, and a smaller
internal dimple-plate inerts cooler
(E-2) using colder WTW (95ºF), with
1/20th the duty of E-1. The main
pressure control on the tower was
by manipulating the WTW rate from
E-1. In case of excess pressure, the
inerts flow to the vacuum system
would be increased by the other
pressure controller. The operator
had the ability to manipulate the flow
of inerts to the vacuum via the HC.
As it turned out, the actual air ingress
into the system was 1/100th
of the design. For vacuum towers,
the design leakage rate is usually
limiting
FIGURE 14. Based on Ref. 17, a vacuum tower is shown
that experienced severe pressure-control problems
53
http://WWW.CHEMENGONLINE.COM
Chemical Engineering April 2022

Table of Contents for the Digital Edition of Chemical Engineering April 2022
