Professional Documents
Culture Documents
SUMMARY
High amine losses, high H2S in the fuel gas, and poor unit reliability were the norm for the amine
regeneration units at the Aruba refinery. In 2003, the three amine units were revamped to increase
capacity, to improve unit operation and reliability, and to minimize the losses. The three high
pressure amine absorber systems at the hydrotreaters were also revamped. These changes
(revamps) resulted in significant reduction in amine losses; the amine purchases went from
1,791,060 (2002) Lbs/yr to 178,120 Lbs/yr (2005) after the revamp. No amine was purchased in
2006 and 2007, this resulted in significant amine solution savings of $406,000/Year.
Subsequent to the revamps and unit restarts, frequent rich amine cartridge filter replacements and
high Heat Stable Salts (HSS) were observed. An AMIPUR (HSS removal skid) was installed to
remove the HSS, and rich amine acid gas loadings were targeted to 0.35 mol H2S/mol MDEA
maximum. The AMIPUR operation and the new operating targets for the amine system resulted
in significant filter costs savings of 65% on the rich amine filters and 34% on the lean amine
filters. A savings of $119,000 per month amounting to annual savings of about $1.5 Million were
achieved by continuing to operate the units on lower rich amine acid loadings and lower HSS
contents.
The amine losses also reduced further due to less filters being changed out. Besides the
economical benefits, fewer filter change-outs resulted in safety and environmental benefit because
of the reduced potential for H2S exposure and release.
265
BACKGROUND
The Aruba refinery amine system uses a 45 wt% MDEA solution for removing H2S from various
gas streams and an LPG stream.
The Amine regeneration units, M1AR, M2AR and M3AR, have a designed capacity of 1450 BPH
(barrels per hour) or 1015 GPM each. These units regenerate the amine solution, removing
hydrogen sulfide gas (H2S) from a solution of water and amine (MDEA), and provide lean amine
solution for H2S absorbers. The high pressure H2S absorbers are at the hydrotreaters: D2AR (Gas
oil Hydrotreater unit) and D3AR (Gas oil Hydrotreater unit). The medium pressure absorbers are
at the D1AR (Diesel Hydrotreater unit), N1AR (Purge gas unit) and the LPG absorber. The low
pressure absorbers are the two DC1AR and DC2AR Coker units absorbers and the two FS1AR
and FS2AR fuel gas absorbers.
The following tables (Table 1.0, 2.0) show the design and operating conditions of the regenerators
and absorbers at the Aruba refinery:
Table 1.0: Aruba refinery amine units design data
Amine units
Circulation
Regen. Tower
(regenerators) Flow (BPH)
Pressure (Psig)
M1AR
1450
13
M2AR
1450
13
M3AR
1450
13
Amine
units
M1AR
M2AR
M3AR
Table 2.0: Aruba refinery absorbers data for max crude sulfur case.
Absorber
Lean
Rich amine
Rich amine,
Absorber
Amine
H2S loading,
mol H2S per
Pressure
Flow, BPH
lbs/hr
mol MDEA
(Psig)
D1AR
D2AR
FS1AR
FS2AR
N1AR
LPG
DC1AR
DC2AR
D3AR
611
719
1330
11,199
16,151
27,350
606
476
128
46
1,256
8,588
6,748
2,885
708
18,928
373
429
617
1,419
6,481
7,465
13,847
27,793
266
420
730
0.46
75
75
400
320
0.39
80
80
730
0.43
267
268
PROBLEM DIFINITION
The Aruba refinery observed high amine losses, high H2S in the fuel gas, and poor unit
reliability at the amine regeneration units. High corrosion rates resulted in leaking of the
lean/rich exchangers and also the lean amine coolers; for example, the leaking lean amine
cooler resulted in seawater contamination of the amine at M3AR unit. After exchanger
tube repairs, the unit equipment were washed with clean water and the unit was
replenished with fresh amine. Contamination and foaming caused great losses of the
solution and off spec treated gas.
AMINE UNIT REVAMP
In an effort to correct the many problems at the amine units, and at the same time
accommodate higher projected sulfur loads (at higher sulfur crude rates) while
minimizing cost, it was agreed to revamp the existing amine units. The units were
originally designed for MEA but were revamped for MDEA. The revamp directive
included higher amine circulation rates and higher MDEA concentration of 45 wt. %.
As part of the revamp, the capacity of each amine regeneration unit was increased to
1450 BPH (1015 GPM), from 1000 BPH (700 GPM) at M1AR and M2AR units, and
1250 BPH (875 GPM) at M3AR unit. The revamp included larger feed flash drum to
increase the residence time and prevent hydrocarbon carryover to the amine regenerator.
REGENERATION SYSTEM REVAMP
Revamp of each of the three amine units included replacement of existing rich amine
hydrocarbon flash drum with a larger drum to provide 25 minutes of residence time for
the rich amine entering the drum, new rich amine feed pumps, new rich amine cartridge
filters, a closed drain system for all amine drains in the amine unit, and new lean amine
coolers. The trays in the amine regeneration towers were replaced with high capacity
trays. Larger lean and rich amine piping were added, along with larger off gas piping
inside and outside unit battery limit. One additional low pressure amine pump was added
at M3AR, while one additional high pressure amine pump was added as a common spare
for M1AR/M2AR.
REVAMP OF ABSORBERS TOWERS.
In addition, the absorber towers at the three hydrotreaters (one Diesel HDS, and two Gas
Oil HDS) were replaced. The old absorber towers had a water wash section at the top,
then the amine section, then another water wash section, and a pseudo-three phase knock
out section at the tower bottom where the gas enters to be treated. As part of the absorber
revamp, a new high pressure gas fin fan cooler, new three phase separator drum upstream
of the new absorber, and a knock out drum downstream of the absorber were added.
Wash water is now added upstream of the new gas fin fan cooler; the sour water is
separated in the new three phase separator. The quantity of wash water has also been
increased to improve ammonium bisulfide salt removal.
LOWER FOAMING INCIDENCES EXPERIENCED.
The contamination and foaming of the amine was the result of hydrocarbon and
particulates. The larger rich amine flash drum with the higher amine residence time has
269
resulted in the ability to effectively separate heavy hydrocarbon from the rich amine.
This reduces the residual heavy hydrocarbon content in the rich amine. The net effect is
the significant reduction of amine foaming in the absorbers/regenerators.
The change-out of the activated carbon filter on a more regular (6 month) basis was also
done to clean the amine solution; prior carbon change-outs were once every 1-2 years.
The lean amine cartridge filters (50 micron size) were replaced with the smaller 10
micron cartridges to help in the removal of particulates.
AMINE SOLVENT LOSSES
As part of the revamp, a closed drain system was also installed to recover amine drains
and minimize the amine losses.
The industry average MDEA consumption is at 10 Lbs/MMSCF. 1 Aruba Refinery
amine consumption exceeded this industry average, but since the revamp, the MDEA
losses were reduced significantly as shown in Figure 2.0. The MDEA purchases for 2002
were 1,791, 060 Lbs; this was reduced to 178,120 Lbs in 2005 as shown in Table 3.0. In
2006 and 2007, no MDEA was purchased because the stored high HSS amine was
reintroduced to the amine unit and treated by the HSS removal unit (AMIPUR skid).
With the reduction in amine losses after the revamps, an increase in filter change-outs
and iron content in amine were soon observed.
Total MDEA Consumption
140
100
80
60
40
20
Date (month)
270
Fe
b05
ec
-0
4
D
ct
-0
4
O
Au
g04
Ju
n04
Fe
b04
Ap
r04
ec
-0
3
D
ct
-0
3
O
Au
g03
Ju
n03
Fe
b03
Ap
r03
ec
-0
2
D
ct
-0
2
MDEA Consumption
(Lbs/MMscf)
120
271
During the start up of the amine units, 100 micron filter cartridges were initially used in
the rich amine filters. Over a five week period after startup, the cartridge size was
reduced stepwise from 100 micron, 75 micron, 40 micron, 25 micron, and eventually to
10 micron which required filter cartridge change-out twice in each 8 hour operating shift.
The decision was made to return to 25 micron size cartridges, which required change-out
once every day.
At two of the amine units treating the hydrotreaters and the Coker units, the filter change
out frequency was high: on average once a day. The total rich amine cartridge filter
replacement for the three amine units was approximately 1700 cartridge filters per month.
These frequent replacements were indicative of high levels of iron (FeS) in the amine
solution due to corrosion. The economic effect of these change-outs included: high filter
cost and amine losses. The safety impact due to the potential H2S exposure (during filter
change outs) and the environmental impact due to disposal of spent filter cartridges were
of concern.
HEAT STABLE SALTS (HSS)
After the revamp, the Heat Stable Salts (HSS) buildup rate was 1.3 wt.% per month. This
increase in HSS buildup rate since the 2003 revamp can be attributed to better amine
operation and lower amine losses.
The primary HSS's are formate and thiocyanate anions, which are degradation byproducts
from the delayed coker units. The upper limit recommended for formate is 500 ppmw, but
numbers as high as 19,600 ppmw have been measured at the amine unit. These heat
stable salts are formed with MDEA in the presence of CO or HCN. The amine stream
treating the Coker gas showed a high amount of HSS between 3.0 and 8.5 wt.%.
Although DuPart, Bacon and Edwards mentioned in their 1993 paper that HSS should
not exceed 10% of total amine circulation, we believe that this number should be much
lower for the Aruba refinery based on our observations. 2 Rooney, Bacon, and DuPart
have stated in a later 1997 article: Total HSAS level should not exceed 0.5% of the total
solution.
Table 4.0: HSS in MDEA Solution
Acetate (ppmw)
Formate (ppmw)
Thiocyanate (ppmw)
HSS (total) (wt%)
Aruba
Refinery data
2070
19,690
7000
3.0 8.5
Upper limit 4
recommended
1000
500
10,000
0.5
The iron in solution had increased from 11 to 40 ppmw since the revamp, and was an
indication of increased corrosion in the amine system. The dark color of the amine and
frequent filter replacements also confirmed the corrosion mechanism, due to high HSS.
272
The data below (accumulated from the Aruba refinery) shows that the required amount of
filter cartridges triples from 200 to 600+ cartridges when the HSS is above 4 wt.%.
900
800
700
600
500
400
300
200
100
0
0
4
HSS (wt%)
10
The improved water wash project at the Coker units was commissioned to help remove
cyanides and other acid salt precursors from the Coker gas stream upstream of the amine
absorbers; however no significant improvement has been observed.
RICH AMINE ACID GAS LOADING
By closely analyzing the plant operating data, a correlation was observed between the
rich amine acid gas loading and the amount of iron in the amine. At a rich amine acid gas
loading of 0.30 mol H2S/mol MDEA, significantly less rich filters have to be replaced
compared to a rich amine gas loading of 0.42 as illustrated in Figure 5.0 below.
The amine units were designed to operate at a maximum rich amine gas loading of 0.50.
However after the assessment, a new target of 0.35 mol H2S/mol MDEA was established
to minimize corrosion. As has been published in various articles, high acid loading
increases the corrosion tendency of the amine. 3
Due to lower refinery crude rates (220-240 KBD) and resultant lower H2S production, the
amine circulation rate was able to be increased in order to lower the rich amine gas
loading to < 0.35 mol H2S/mol MDEA. The lean amine strength was also increased to
45 wt.% MDEA. Consequently less iron in amine solution was measured and filter
change out frequency dropped as shown in Figure 6.0.
273
1000
800
600
400
200
0
0.20
0.25
0.30
0.35
0.40
0.45
0.50
9
Total HSS, wt%
Iron
90
80
7
70
60
5
50
4
40
3
30
2
20
0
14-Jan-04
10
23-Apr-04
01-Aug-04
09-Nov-04
17-Feb-05
28-May-05
0
05-Sep-05
274
275
BIBLIOGRAPHY:
1) Stewart, E.J. and Lanning R.A. Reduce amine plant solvent losses,
Hydrocarbon Processing, 1994.
2) DuPart, M.S, Bacon, T.R, and Edwards, D.J Understanding Corrosion in
Alkanolaine gas treating plants, Part 2 Hydrocarbon processing, 1993.
3) Rooney, P.C., Bacon, T.R, and DuPart, M.S Effect of heat stable salts on MDEA
solution Corrosivity, Part 1, Hydrocarbon Processing, March 1996.
4) Rooney, P.C., Bacon, T.R., and DuPart, M.S., Effect of heat stable salts on
MDEA solution Corrosivity, Part 2, Hydrocarbon Processing, April 2007.
276