P1: OTA/XYZ P2: ABC

c16 JWBS027-Lieberman May 18, 2010 13:20 Printer Name: Yet to Come

Chapter 16
VaporLiquid Separator
Entrainment Problems

I was driving west on I-10. I had a project at the Jupiter refinery in Baytown and
was running late. Just past the Lake Charles exit, I noticed an old Ford pickup with
a flat tire. A young-looking unshaven guy was sitting in the grass near the truck. I
pulled over.
Whats up? I asked, as he stood up and walked over to my car.
Mister. I got me a flat and no spare. He was dressed in old, worn-out coveralls
and dirty sneakers, and had unkempt bushy hair.
Do you want me to call a tow truck? You belong to Triple A? I asked.
Naw. I got me no money. Anyway, I is just bout outta gas. This ole trucks
finished. Alternators stopped working, too. But thanks for stopping, Mister.
So what are you going to do?
Dont rightly know. Nothin, I guess, he answered, as he sat back down in the
weedy grass.
Look. Get in and Ill take you to the next gas station. Lets get some breakfast.
Okay?
Thanks, he answered, as he sat next to me.
I guess youre not working.
No, sir. I done lost me my job. Was workin at that there refinery in Norco. But
they done gone and shut down and I got my ass laied off. Was workin as an assistant
operator on that there gas oil hydrotreater. I had me a nice little house down by the
river, too.

Process Engineering for a Small Planet: How to Reuse, Re-Purpose, and Retrofit Existing
Process Equipment, By Norman P. Lieberman
Copyright

C 2010 John Wiley & Sons, Inc.

171
P1: OTA/XYZ P2: ABC
c16 JWBS027-Lieberman May 18, 2010 13:20 Printer Name: Yet to Come

172 VAPORLIQUID SEPARATOR ENTRAINMENT PROBLEMS

Would you believe it, I said in surprise, Im on my way to a job to troubleshoot

a problem on a gas oil hydrotreater at the Jupiter refinery.
Whats their problem, mister? By the way, my names Leroy. Whats yours?
Norm. The problem Leroy is that the recycle gas, reactor effluent vertical separator
vessel is carrying over. Leroy, you know what entrainment means?
Say Mister! You aint that there Norm Lieberman guy, is you? That dude who
writes bout troubleshootin and stuff?
Thats me, I said, both amazed and flattered.
Sures I knows what that there entrainment is. Thats when that foamy liquid pukes
over, even when the levels okay. Those Jupiter boys! They should have had them a
horizontal, not a vertical separator. Thatll slow down them vapors. Its them vapors
moving too fast that entrains droplets. Guess them Jupiter cats aint none too smart.
Look, I got an idea. Come with me to the Jupiter plant now. I need someone
to help me field check pressures and temperatures. You know how to connect up
pressure gauges and change out dial thermometers? Ill pay you $60 for today.
Cash money? Leroy asked. Man. I is ready.

LEVEL PROBLEMS IN V-603

Leroy stood next to the separator in my spare boots and extra Nomex and studied the
liquid level in the gauge glass in V-603 (Figure 16-1).

Figure 16-1 Foam level is higher than level indicated.

P1: OTA/XYZ P2: ABC
c16 JWBS027-Lieberman May 18, 2010 13:20 Printer Name: Yet to Come

FOAM-INDUCED CARRYOVER 173

You all know somethin, Mr. Norm, that there level is too high.
I had made a mistake bringing Leroy into the Jupiter plant. Mr. Anderson, the
operations manager, did not approve of Leroy.
Lieberman, he said, Your company has been retained to provide a detailed
vessel sketch for a new and larger replacement vessel for V-603, the reactor effluent
high-pressure vaporliquid separator. We have budgeted $800,000 for the new 1800-
psig-rated vessel. We do not appreciate your associate, Mr. Leroy Johnson, lecturing
us that we should have used a horizontal separation drum, with a neutron backscatter,
level detection system to deal with foam.
Speaking about Leroy Johnson, I bill his time out at $1000 a day. Is that alright?
I asked.
Fine! Just see that he fills out his contractor time sheet correctly at the end of the
week. But kindly advise him to stop telling our hourly operators that our engineers
dont understand about level control. The level in the separator is not the problem.
The vessel is too small and needs replacement. I have budgeted $3,200,000 for the
design, purchase, and installation of this new vessel. Our engineering headquarters
division in Virginia has already approved of the replacement.
Okay, Ill tell Leroy. Sorry about it.
Of course, we at Jupiter are an equal opportunity employer and are very proactive
in promoting minority employment, Mr. Anderson added.
So are we, I agreed. Thats why I actively recruited Leroy myself for our staff
this year.

FOAM-INDUCED CARRYOVER

You see, Mr. Norm, that there level you all see in the gauge glass is not the real
level in V-603. The real levels higher. Probably up almost to the feed nozzle, said
Leroy. Look closer, Mr. Norm, at the level glass. See that liquid kind of dribbling
back through the top tap and running down the glass. Thats foam. The foam drains
through the top tap. Foam breaks down when it touches lots of surface area, like the
gauge glass. Thats the liquid part of the foam you see draining through the glass. The
top tap of the glass is only 1 foot below the feed nozzle. If the foam gets itself above
the feed nozzle, the vapor in the feed blows that foam into the hydrogen recycle gas.
I done seen it all before in Norco. Man, I sure do miss my little house on River Road
down by the levee.
Why all the foam? I asked. There was something odd about Leroy. Something
that I hadnt noticed before.
Well, Mr. Norm. At high pressures and low temperatures, hydrogen and heavier
liquid hydrocarbons like gas oil create a very stable foam. Wouldnt happen with jet
fuel or diesel. Wouldnt happen at lower hydrogen pressures. Wouldnt happen at
high temperatures. Has something to do with surface tensions.
What else, Leroy? I asked with growing suspicion.
Mr. Norm, the other problem is particulates. Like iron sulfides. You all see that
them Jupiter cats are bypassing the hydrogen feed caustic scrubber. That leaves them
chlorides in the hydrogen, which makes hydrochloric acid in the reactor. That causes
P1: OTA/XYZ P2: ABC
c16 JWBS027-Lieberman May 18, 2010 13:20 Printer Name: Yet to Come

174 VAPORLIQUID SEPARATOR ENTRAINMENT PROBLEMS

corrosion, which makes iron sulfide particulates. Particulates cause foam. Like dirty
water foams-up worse than clean water. You should tell them Jupiter boys to put that
there caustic scrubber back online.
But listen, Leroy. How can you be so sure the problem is a high foam level? You
dont have x-ray vision, do you?
Ill prove it, Leroy answered. He called over to the outside operator, Kevin, tell
Petey to cut that level in V-603 by 10% on the panel.
I watched the level in the glass drop by 8 in. But more to the point, the liquid flow
dribbling through the top tap stopped completely.
You understand now, Mr. Norm?
No, Leroy, whats happening here?
Ill explain, Leroy said. When the foam in the vessel rises above the top tap of
the gauge glass, the level in the glass does not represent the level in the vessel. The
level in the glass represents the density of the foam in the vessel, in proportion to the
density of the liquid in the glass. As foam density is just a small fraction of liquid
density, the foam level in the vessel must always be higher than the external liquid
level indicated in either the gauge glass or the level-trol.
So the level set point is too high, I concluded.
Correct, Mr. Norm, said Leroy.
And if we get the level too low? I asked. Whats going to happen then?
Then the flow of fuel gas (Figure 16-1) will jump up as the foam layer slips out
through the bottoms LRC. Mr. Norman Lieberman, perhaps I should summarize:

Step 1. Set the LRC to hold a level below that which causes liquid to dribble through
the top of the glass, but above the level that causes an increase in fuel gas flow.
Step 2. Put the NaOH scrubber back into service to keep the HCl out of the hydrogen
feed gas to the reactor.
Step 3. Inject a silicon defoamer into the separator feed (to the first vessel in Figure
16-1) at a concentration of 5 ppm to suppress carryover, but only on an emergency,
backup basis.
Step 4. Install a Nuetron backscatter level radiation detector [sold by K-Ray] to
monitor the true foam density and foam levels.

[I noticed a growing radiance behind Mr. Leroy Johnson. His bushy hair glowed
softly in the golden light. A smile spread across his youthful brown face.]

Step 5. Cancel the V-603 replacement vessel project. Mr. Norm. Its just a waste.

And then a gust of wind blew through the Jupiter refinery. A cloud of dust rose
and obscured V-603. And when the air cleared, Leroy was gone. As I drove home the
next day, I looked for his truck on the west-bound side of I-10. The truck, too, was
gone. Most unusual for the Parish Highway Department to be so prompt in removing
an abandoned vehicle [1].
P1: OTA/XYZ P2: ABC
c16 JWBS027-Lieberman May 18, 2010 13:20 Printer Name: Yet to Come

ENHANCING DEENTRAINMENT RATES 175

In conclusion, horizontal vaporliquid separators, not vertical separators, should

be used in this service. The horizontal vessel will have a larger liquid surface area, for
the same liquid residence time, than that of the vertical separator. The rate of foam
dissipation is proportional to the vaporliquid interface surface area in a vessel.

ENHANCING DEENTRAINMENT RATES

Entrainment velocity, V a , in ft/sec, is defined as

1/2
DL
Va = K (16-1)
DV

where
DL = liquid density (note: units cancel out)
DV = vapor density (note: units cancel out)
Va = vertical component of velocity, ft/sec

K is selected as follows:

r Low entrainment relying only on gravity settling: K = 0.10

r Moderate entrainment, relying only on gravity settling: K = 0.2
r High entrainment, relying only on gravity settling: K = 0.3
r Excessive entrainment, relying only on gravity settling: K = 0.4
r Moderate entrainment using a demister pad: K = 0.25 to 0.30
r Moderate entrainment using a demister pad and a vapor distribution device (such
as a vapor horn): K = 0.30 to 0.35

A demister is a woven metal mesh screen, typically 4 to 8 in. thick. Smaller

droplets of liquid impinge on the metal mesh and coalesce into larger droplets, which
drip off the pad. If the demister is used in a fouling service, or if the demister
corrodes, the pad will partially plug. This will lead to higher localized velocities
and thus the demister pad will make entrainment worse, not better. Thus, in many
refinery applications, demister pads should not be used. If you do use a demister pad,
select a metallurgy for the mesh that is not subject to corrosion. Monel, Hasteloy,
titanium, and 317 stainless steel are some possible choices (see Chapter 13 for
details).
For a demister to work, the vertical velocity using equation (16-1) must be in the
range calculated using a K factor between 0.20 and 0.3. Demisters are ineffective at
much lower velocities.
P1: OTA/XYZ P2: ABC
c16 JWBS027-Lieberman May 18, 2010 13:20 Printer Name: Yet to Come

176 VAPORLIQUID SEPARATOR ENTRAINMENT PROBLEMS

VAPOR DISTRIBUTION AS AN AID TO DEENTRAINMENT

The purpose of this chapter is to provide you with design advice so as to avoid the
purchase of new vessels used as vaporliquid separators. One such method is to
dissipate the momentum at the inlet nozzle of the feed by converting a single radiant
entry to a dual tangential entry, using the vertical baffle configuration described in
Chapter 8. Check the section of Chapter 8 on retrofitting of the Texaco Marine vacuum
tower for a sketch of this baffle. If the inlet nozzle velocities are less than 70 ft/sec.
I would not be concerned with any inlet vapor distribution. If inlet nozzle velocities
are greater than 150 ft/sec. I would be quite concerned with the even distribution of
the vapor feed.
An additional method to improve the vapor distribution is to place a chimney tray
distributor above the inlet nozzle, as detailed in Figure 16-2. The chimneys may be
either round or rectangular. The area of the chimneys should be calculated using a

Figure 16-2 Chimney tray used as a vapor distributor to promote deentrainment.

P1: OTA/XYZ P2: ABC
c16 JWBS027-Lieberman May 18, 2010 13:20 Printer Name: Yet to Come

REFERENCE 177

velocity so as to produce a pressure drop of a few inches of liquid:

DV 2

P = (0.8) V (16-2)
DL

where
DV , DL = as defined in equation (16-1)
V = vertical velocity of vapor flowing through chimneys, ft/sec

P = pressure drop of the vapor, inches of liquid at DL

Dont forget the overflow pipe. Otherwise, the deentrained liquid will overflow
the chimneys and be reentrained, thus defeating the purpose of the chimney tray
distributor. The overflow pipe should have a separate seal pot, as shown in Figure
16-2; or, extend the overflow pipe down to the liquid at the bottom of the vessel. If the
overflow pipe is not sealed in some way, vapor will blow up the pipe. Then liquid will
be unable to drain down the overflow pipe, countercurrently against the rising vapor
flow. This would then defeat the whole purpose of the overflow pipe, and flooding
could result.

REFERENCE

1. Adapted from Leo Tolstoys short story, What Men Live By.