Carry over Problems in CO2 Removal Units

The carry over of potassium carbonate solution from the top of the desorbers has been a significant
problem, especially when the ammonia plant is operated at above 100 % of design production rates.
The designer did not account for such higher plant loads. The result has been entrainment (carry
over of solution from desorber top). This has been studied to determine the causes and alternative
solutions. It was determined that the main reason for carry over of potassium solution was related to
poor liquid distribution on the packing, especially the upper distribution for the 4th bed. Therefore,
we recommended replacing the old design (pan distributors or riser tubes design RTD) to gallery
trays. Also the CO2, cleaning could be improved significantly by replacing the old design washing tray
from a thermal tray to a bubble cap tray. The bubble cap tray pressure drop and washing condensate
quantity show promising results. Replacing the demister along with careful leveling of the wash water
tray and the liquid distributor reduces carry over and reduces chemical and biological oxygen
demand (COD and BOD) in ammonia plants.

Ibrahim Abdel Wahab

MOPCO, Egypt
[email protected]

Problem Definition
Introduction
Carry over of Potassium solution contaminating
the CO2 leaving the desorbers for plant I and II

A
case study was performed in a Fertilizer
Company that operates three Uhde has been a noteworthy problem in the ammonia
designed ammonia plants: a plant I with plant’s Benfield Solution unit. Analysis of the
a capacity of 1100 MT/day Ammonia, a condensate leaving the downstream separators
plant II with a capacity of 1000 MT/day (after desorber) show potassium (K+) levels of
and the third pant III with a capacity of 1200 about 400 ppm for plant I and 180 ppm for plant
MT/day. A solvent carryover problem in the II compared to a design figure of < 12 ppm. The
CO2 desorber systems caused operating and chemical consumption in this unit is
environmental concerns in plant I and plant II. significantly high, hence increasing operating
Alternative it appears that plant III was a better, cost. Mechanical problems that resulted from
more recent design. The causes for the solvent this includes:
carryover were analyzed, and solutions were • Wash water tray deposits and demister
implemented during a shutdown in September blockage.
of 2000 with good results.
• Increasing fouling tendency in the
downstream CO2 compressor

2007 37 AMMONIA TECHNICAL MANUAL

Effect on the Environmental It's quite clear that consumption was reduced
conditions after the corrective overhauls in September of
2000.
Process condensate from the separators of the
Benfield units of plant I and II (12 m3/hr. at 45 k2 co3 consumpiton plant ||
o
C [423 ft3/hr at 113 oF]) is contaminated with 500
ammonia, methanol and traces of methyl amines

g / to n a m m o i n a
400
[Ref 2]. The majority of the contaminants are k 2 co3
300 consumtions
formed in the plant’s I front end, then
condensed and separated in I and II. Methanol is 200 design k2
co3
formed in the presence of copper based low 100
temperature shift catalyst and H2 and CO2. This 0 after shutdown
methanol reacts forming methylamines. The 1992 1994 1996 1998 2000 2002 2004
total concentration of methanol, ammonia, year
methylamines at the exit of the LTS
concentrations is about 50 ppm max. Small Drawing 2. K2CO3 consumption in plant II
quantities of contaminates are also found in
condensate from the CO2 removal unit The potassium carbonate consumption figures in
separators (after desorber). Typically this flow plant II are illustrated in above Drawing 2. Also
is 39 m3/hr at 60 oC (1,377 ft3/hr, 140 oF) which in this plant the K2CO3 consumption figure was
is highly contaminated with K+. If the carry over reduced significantly after the overhauls. COD
problem is solved, process condensate could be and BOD were also reduced as shown on
transferred to the condensate stripping units in Drawing 3.
plant II or plant III and thereafter, COD and
BOD figures would be controlled.
COD and BOD
Consumption and COD/BOD figures 500

400

K2 Co3 consumption in plant I

m g \ lit

300
700 COD
200
600
BOD
100 q; 100
q; 60
500
0
g \ to n a m m o i n a

400 00 00 00 00 00

300
act.consumption year
Design
200 Drawing 3. COD & BOD consumption figures
100

0
after shutdown Problem Diagnosis
1975 1980 1985 1990 1995 2000 2005

year The following sketches show the desorber

Drawing 1. K2CO3 consumption in plant I. internals. The tower may be divided into two
sections:
The carbonate consumption figures in plant I are • Sketch 1 for CO2 cleaning
illustrated in Drawing 1. (Washing tray. Take off tray, demister)

AMMONIA TECHNICAL MANUAL 38 2007

 • Sketch 2 for CO2 regeneration
(Packing. hold down grids, distributor,
and redistributors)

Sketch 2. CO2 regeneration

Comparing the desorbers in plant I, II and III it

is remarkable that the liquid distributor in the
Sketch 1. CO2 cleaning desorber in plant I is an old-fashion-design
(complicated pan distributors or riser tube
CO2 regeneration design)

Unless the liquid distribution of the potassium Any failure in any part of the distributor will
carbonate solution is good, the liquid flow cause liquid carry-over. The inspection group is
distribution through the entire packed bed will required to check the distributor integrity during
be poor causing vapor-liquid channeling and each planned shut down overhaul.
then carry-over of liquid to the tower top and
sometimes down stream of the desorber. CO2 cleaning
th
The initial liquid distributer in the 4 bed is very The CO2 cleaning is done in the desorber of
important because it is responsible for evenly plant I by using one single tray (old design). It is
distributing the liquid on the top bed. If the remarkable that the design quantity of washing
distributor is out-of-level (even 1 inch or 25 condensate is 2.1 m3/hr (9.25 gpm) which is not
mm) the carryover problem will be much worse. suitable for this CO2 gas flow. Because of the
minimum liquid to gas ratio required from
calculation [Ref.2],

2007 39 AMMONIA TECHNICAL MANUAL

Photo 1. Water tray old fashion Photo 2. Demister failure

It was found that 11 m3/hr (48 gpm) is a Comparison of three plants

minimum required to wash CO2 gas (97 t/h at
100 % load). If there is any out-of-level on this In the table below a comparison is given of the
tray, this will lead to bad washing and CO2 removal units of the three ammonia plants
inefficient tray performance (Photo 1). The I, II and II.
inspection group has checked this point to
guarantee the horizontal tray lay out.

Particulate Problems

If there are any solution particles flowing with

the gas, the demister will be plugged as a result
of particles sticking to the demister fibers. In
this case, the pressure drop of vapor will be
increased. But the pressure drop can not
increase a lot, because the demister will break
[Ref.2]

Demister failure (Photo 2) promotes high

localized velocities which create more
entrainment than we would have without any
demister, Therefore, the inspection of the
demister is mandatory.

Table 1. Comparison of the CO2 removal units

of 3 Egyptian ammonia plants

AMMONIA TECHNICAL MANUAL 40 2007

For comparison also a sketch is given below of The space was available for installation. It must
the top section of the CO2 desorber column of be in horizontal and level position.
ammonia plant III
Alternatives to improve CO2 gas cleaning

(A) Replacing the demister. (Must be checked

by inspection group)

(B) Replacing the old design washing tray by a

bubble-cap one just as in desorber plant III

(C) Replacing both demister and washing tray.

Alternative (C) is recommended to guarantee

excellent gas cleaning. Therefore, any slight
entrainment will be controlled.

The bubble cap tray depends on special

configuration of the caps which consists of
risers that act as to prevent liquid from flowing
Sketch 3. Top of desorber of plant III down through the caps. However, it is clearly
evident that the tray leveling is important

Additionally, the following points must be

checked (Annex 1), [Ref.1]
Steps to minimize carry over
problem 1. Minimum wash water flow to be
maintained.
Alternatives to improve distribution. 2. Condensate inlet and outlet nozzles sizes
Two options were considered to improve the and availability to erect the new piping of
distribution: the new size.
(A) Replacing all liquid distributors on the four 3. New P on the tray.
beds by the new-design distributor just as in 4. Tray configuration and some specific
plant III. dimensions.
(B) Replacing only the liquid distributor on the
th
4 bed (pan to gallery type) [Ref.6] Annexes
Alternative (A) is recommended to guarantee Check the configuration of the new trays
even liquid distribution on all beds. The new (bubble cap type )
distributor has the following advantages:
v High efficiency due to very good liquid Using TRAY RATING software KTD [Ref.1],
distribution (Sketch 3) the configuration of bubble cap trays was
checked as follows
v Longer life time and easy to maintain.
v Lower pressure drop than the old one. (1) The system factor (constant depending on
tray type) is an important parameter related

2007 41 AMMONIA TECHNICAL MANUAL

 to the process. For CO2 Regenerator one single washing tray. The distance
(Desorber), it must be taken as 0.85 between this tray and take off tray is 23 in
which is quite safe
(2) Cap density: It was found to be 145 cap/m2
(13.5 cap/ft2) of tray area this depends on (6) Decreasing shroud width can cause choking
the tower inside diameter. at the tower inlet resulting in massive
(3) Weir Height: This should be taken between entrainment
24 mm (1 inch) as a minimum in vacuum (7) Increasing the total distance from the inlet
towers and 75 mm (3 inches) as a nozzle to the tray reduces the entrapment of
maximum in pressurized towers. Our weir smaller droplets
height of 42 mm (1.65 inch) is OK
(8) Increasing nozzle velocity appears to have
(4) Down comer clearance: If the down comer no effect on entrainment
clearance is too great, the down comer can (9) Increasing the total length of shroud and
becomes unsealed and vapor flows up the
opening (see Sketch 5) to 145 inch (3683
down comer causing flooding. (Sketch 4)
mm) gave 1 micron particle entrainment of
Down comer clearance should be 0.5 inch 10 to 18 % of total carryover particulates,
(13mm) below weir height
with minimum 75 inch (1905 mm) at
This dimension should be carefully checked shroud length and 70 inch (1778 mm) at the
when the tower is opened for inspection. It opening area. (Ref. 6)
is quite easy for sloppy tray installation to
distort this critical factor
Executive steps taken at plant I
(Sept. 2000)

Photo 3. Replacing the blocked demister.

Sketch 4. Down comer clearance

(5) Tray spacing : The normal tray spacing is
600mm (23.6 inches) In our case, we have

AMMONIA TECHNICAL MANUAL 42 2007

 Photo 4. Mech. and chem. cleaning of wash Photo 6. Covering the open area in the inlet
tray. chamber with additional plates.
• Install and secure all supports of wash tray,
then change the packing to high quality type
and adjust leveling of the tray.
• Make modifications for solution internal
area by the following steps, illustrated by the
photos 5-8
• Tighten, level, check dimensions according
to the technical drawing
Check all internal beds for integrity.

Photo 7. Valve trays

Photo 5. Drilling additional (7) holes at the

liquid inlet distributor.
Photo 8. Inspection of liquid distributor and it’s
leveling (very important)

2007 43 AMMONIA TECHNICAL MANUAL

 packing in between. This packing however
became the weak point for leaking on carryover.
This design was changed to another consisting
of 8 parts, but with each part fabricated by
bending rather than with packing or bolts to
avoid leaks on carryover. See photo 11 and
sketch 5, 6 and 7 for more detailed info.

Photo 9. Change the way of tightening the hold-

down grid from stainless steel (ss) wire to ss
bolt for a good fixing

Executive steps taken at plant II

• Replace the blocked demister. Photo 11. Gallery type (old liquid distributor)
• Replace the valve trays… (photo 7)
• … with bubble cap type (photo 10) and
adjusting for the leveling

Sketch 5. Layout of the 8 segments of the

Photo 10. Bubble caps distributor plate.

The liquid distributor (Photo 11) was replaced

with a new design similar to the one in plant III.
The previous plant II gallery tray was in 8 parts.
Each part consisted of two L shape segments
that were tightened together by bolts with

AMMONIA TECHNICAL MANUAL 44 2007

Sketch 6. Design of the old distributor segment
Sketch 8. Shroud area

Executive steps taken at plant III

1. Check the demister

2. Check bubble cap tray integrity and also its
leveling
3. Check the liquid distributor
4. Check the dimensions of the shroud area by
(by our software).
5. Check all internal bed parameters (tighten,
leveling, dimension acc. to the technical
drawing)

Sketch 7. Design of the new distributor segment Reduction of Carryover

• After checking the dimension of shroud area Before After
by K.T.D. software, 25 cm (10 inches) from shutdown shutdown
shroud length (Sketch 8) was cut to correct
K+ as potassium 400 ppm 2.2 ppm
the opening area of the gallery tray weir.
• All important internal bed parameters should Table 2: Comparison of chemical analysis
be checked (tighten, leveling, dimensions).
Design K2CO3 consumption / month = 116
g/ton ammonia.
Note: the design figure is 12 ppm.

2007 45 AMMONIA TECHNICAL MANUAL

 C3: Average K2CO3 after modification
K2 Co3 consumption in plant I
700
= 65 g/ton ammonia
= 2 ton/month
600
= 2 * 12 * 3000
500
= 72,000 USD/yr
g \ to n a m m o i n a

400

300
act.consumption The average savings on K2CO3 consumption by
200 Design installing the modifications is 468,000 USD per
100
year (C1-C3).
after shutdown
0
1975 1980 1985 1990 1995 2000 2005 Conclusions / Recommendations
year

Drawing 1. K2CO3 consumption For increasing the tower efficiency these are
steps or parameters that should be taken into
It's quite clear from comparing the K+ analysis consideration
in the HP separators from before the shutdown
(Drawing 1) and the results after the corrections 1. Change the wash water tray (thorman-type)
to bubble cap type for the desorber
,shown in Table 2 that the consumption was
reduced after the over hauls in plant I. Also the 2. Change the demister for absorber
carbonate consumption figure for plant II was 3. Change the wash water tray for the absorber
under the design figure. to a bubble cap type
4. Check leveling for liquid distributor
Economics 5. Check hold-down grid support
6. Check leveling for packing-bed
The K2CO3 that is lost in the CO2 removal unit 7. Packing type
has to be replaced by new solution at a cost
8. Pressure drop
price of 3000 USD per ton. The total annual
costs savings by reducing the K2CO3 losses 9. Gas cleaning at desorber top
amount to almost half a million USD per year. 10. Plant should be working on 100 % front load
See below for the calculations.

The average K2 CO3 consumption before the References

modifications were 400 g/ton ammonia. With an
ammonia production of 34000 ton per month (1) Working guide to process equipment
this is a total loss of about 15 ton K2CO3 per (Chapter 2, 3, and 7) 1993, by Norman
month. P. Lieberman

C1: K2CO3 costs /year = 15 * 12 * 3000 USD (2) Chemical Engineering Design {Vol.6}
= 540,000 USD/yr (3) Steel Manual Key {1998}
(4) Training Manual {Uhde}
C2: Design K2CO3 consumption / month
= 116 g/ton ammonia (5) Chemical Analysis For the CO2
=3 ton/month Removal Unit
= 3 * 12 * 3000 (6) UOP Confidential Reference
= 108,000 USD/yr (Sept. 2000 by J.P. Brady)

AMMONIA TECHNICAL MANUAL 46 2007