Académique Documents
Professionnel Documents
Culture Documents
SPE 13728
- CDC 4 ax M,@a EZ@ 01 T~hn~~l conference and Exhibition held in Bahrain, March 11-14, 1985. The Material iS
Tiiis paper was p~&$8iii6d 61 i!k . . ..- ,------- . __
subject to correction by the author. Permission to copy is restricted to an abatract of not more than ~ WOrdS. ‘%%3 ~P=, kx m, ‘!d?eK!Srr%
Texas 7508S-3S3S. Telex 7309S9 SEE DAL.
1.0 INTRODUCTION
The Abu Dhabi Gas LiquefactionCompany (ADGAS) 4. Feed Separation(by primary cooling)
hae since 1977 operated the only gas liquefa- ~..~f~=~~~Qp-~~~Q~
*
ction plant in the Middle East producingsimul- 5. CryogenicCooling and Liquefaction.
taneouslyIJUl,LPG and I?entaaes.“&e Plact
processes the associatedgases from Abu Dhabi Utilitiesand offaitesfacilitiesof storage
offshore oil fields as well as non associated and loading are common for both process traina.
gas. This is the only LNG facilityin the
world operatingon associatedgas. Following the commissioningof the plant, one
major problem identifiedwas the capacity of
The gas compositionof the 11 differentfeed the Acid Gas Removal Plants. The bottleneck
streams supplyingthe LNG plant varies broadly preventedeither train from matching steadily
and for instance the sour gaa content (C02+H2S) the LNG design throughputof 150 T/hr. on heavy
of each stream can vary between 3.3% and 34%. feed gas. LNG Plant performancesduring 1981
An average gas stream compositionsupplied to are gathered in Table-II.
the Acid Gas Removal Unit is given in Table-1.
Over a period of 16 monthe a high proportion
The Plant consistsof two identicalprocess of the effort of ADGAS TechnicalDepartment
traina in parallel arranged to permit safe ---->....-&- -+*,A.,
was uevuwd +0 ~F.e-.--= of $~ae ~CIJS~S
~f the
overhaul of one train while the other remains bottleneckand the means by which it could be
in service. alleviated.
The ~.--
nroK)OSed
objectivesfor imp??OVeIUentS
were porosity (void fraction)of the bed
assignedas follows: support ani hold dob= grid must be
equal to that of packing.
1. Improve the stabilityof the Plant.
2. Match design for acid gas removal. the liquid should be evenly distributed
3. Produce an extra 2.5 T/hr of LNG in a thin layer over the packing.
(17,500T/year)
the best efficiencyof the packing will
2.0 ACID GAS REMOVAL PLANT be at around 60$ flooding. It will have
a reasonableflexibilityof between
2.1 The Process (See Fig. 1) 35-70% .
The acid gas removal plant is designed 2.2 Analysis of the Problems
..” -.-~- eb~Orption,/regeneration.
as a b...- .+.~e
In practice the maximum acid gaa removal These causes were identifiedas being:
achievedwas around 92-93% weight of the
design capacity (heavy feed gas). feed gas pressure fluctuation
tower hydraulics
Investigationcarried out led us to solutionfoaming
conclude that the poor performanceof the
carbonateregeneratorwas responsiblefor As a result of our investigationwe identi-
the plant limitationunder these conditi- fied the reasons for poor performanceas
ons and the principalcauses were identi- being:
fied as follows:
change of operatingconditionsof the
Firstly, the recommendationof the carbonateregenerator,consequentupon
process licenser,to maintain in the non-commissioningof the Sulphur
the carbonatesolutiona concentration Pisnt.
of DEA around 3% could not be met. It
was found that this difficultyin main- The deteriorationof the ceramic packing
taining DEA concentrationwas due to which occurredat the high temperature
thermal degradationcaused by high .-a .t~nnn
=..-“ nI~BI~n~ty
...- --- of the carbonate
Plant instabilityled to sour gas break- Followingidentificationof the causes for the
through and thereby off specificationpro- poor performanceof the plant, ADGAS Technical
duct and/or fouling of the main cryogenic Departmentmade recommendationsto overcome the
exchanger. It usually necessitatedthe problems. Also, we drew-up the detailedprocess
operationof both trains at reduced specificationsfor the proposedmodifications,
throughput substantiallybelow those includingpacking and internalsof the columns,
which could have been achieved under as outlined thereafter.
steady operating”
conditions.
I
.-9K
0..
.—
.
Performanceof the train after implementation Hereafter follows the main conclusionswhich
of recommendedmodificationscarried out during can be drawn from the above mentionedwork.
1983 Train Major Overhaulshow that our object-
ives have been achieved,as shown in Table-III. * Metallic packing mass transferefficiencyis
better than ceramic packing. An increase of
● Operationof the train is now highly stable, 40-60% in the mass transfercoefficientcan be
even at a throughputwell above the design achievedwhen using the proper size of metallic
capacity. packing instead of ceramic packing. This will
~ef~ec~ on the d~ensions of the absorbersand
* The design for the acid gas removal C02+H2S the regenerationcolumns which could be made
has been exceededby more than 10$. The C02 shorter and smaller in diameter in a new plant.
absorptionrate currentlyachieved repre-
sents around 40~ above the design. This * Life time of metallic packing,when selecting
remarkableperformancegauges the improve- the adequatematerial for the required service,
ment achieved in mass transferefficiency is 2 to 3 times longer than the life time of
hence c02 reactionwith carbonatesolution the ceramic packing. In addition,the circu-
is the controllingfactor in the absorption lated solutionson the metallic packing are
process. clearer and contain less suspendedmatter than
those of ceramic packing which reflect on the
* Steam consumptionof the sweeteningunits solution filtrationcost.
has been reduced considerablyper kgmole of
acid gas removed. Around 25% saving in * To overcomethe poorer liquid redistribution
steam consumptionhas been achieved. propertiesof the metallic packing, it is impe-
rative to utilize the proper distributorsand
* The debottleneckingof the units has refle- redistributorsabove all the packing beds.
cted on the LNG make both quantitywiseand
qualitywise. Dependingon the feed case * Hold down grids and bed support plates should
(heavy or light) and increaseup to 20% have a free area equivalentto the cross sect-
above the design LNG make haa been achieved. ional area of the column.
Qualitywiaethe aulphur content of the final
products haa dropped to 50% from the levels * Unit ancillariessuch as steam desuperheaters,
before modifications. condensatedi8posalsystems end sour gas dis-
posal systems should be carefullyconsidered
* Economically,the modification have increa- and checked to match the best performance
sed the performanceof the Plant. With some requirements.
extra maintenancecost and no increase in
capital or running costs, the plant produ- ACKNOWLEDGEMENTS
ction has increasedmaterially.
The Authora thank ADGAS General Management
for permissionto publish this paper.
[ .07
.JO/
. .
TABLS-I
GASSTBEAU
cOHPOSITIO?i
(swPLY
TOBVJIBTENIKC
VRIT)
TABLE - II
~& ,c=
wximumthroughput
0.21 0.14 (dailykSiS)
N2
c1 64.92 68.67
L~GProductionLosses 2.36
i ‘“’ 2.32
C4 3.60 2.62
LlfGDesign throughput 150
C5 3.48 0.69
mtium unitperformance
TOTAL 100 .0+3 100.00
~
TABLE - III
{2s Loadins
Kg mole/Kg mole of ~co3 0.24s o.ls2 0.230
DEA Absorber
s2S Loadins
Kg mole/Kgmole of DSA 0.03’7 0.074 0.101
C02 Loading
5g mole/Kg mole of DEA 0.290 0.191 0.225
Steam Consumption(Reboilers
)
FIEL GAS
v
r-
OSA ABSORISU
Guts
PROUSS W
Cw
64s —
Assonm
7 ,.L
Jx
i-
——
f
I
j
21 RsFlux
w
L1 WWRCO
MAKf
I
WAIUi
%f-
MGH PRfss.
PUHP
Sssi
Fsso
GAS -,
I LJ[ I ()) 1
IQ , MOM
(-o-j-+ol PRESS,
CARBWAlf
Pufv
WA PUMP
l--
w
-4
w
Oa