Gas

Dehydration


1

Gas Dehydration
Contents:
Principles of Gas Dehydration ..........................................................................3
Water Content of Gases......................................................................................4
Hydrate Formation...........................................................................................15
Factors Promoting Hydrate Formation............................................................17
Procedres and !esorces for Determining Hydrate"Formation
#emperatres of $%eet and $or Gas $treams ..............................................&'
#emperatre Control (ethods and )*ipment +sed #o ,nhi-it
Hydrate Formation in a .atral Gas $tream....................................................&&
(ethanol ,n/ection !ate !e*ired #o ,nhi-it Hydrate
Formation in a .atral Gas $tream.................................................................&0
Glycol Dehydration .........................................................................................31
Descri-ing the Glycol Dehydration Process ...................................................41
Glycol Dehydration $ystem Components .....................................................43
Process2Design 3aria-les ................................................................................44
5ptimi6ing and #ro-leshooting Dehydrator 5perations................................51
$olid Desiccant Dehydration...........................................................................54
7dsorption Calclations...................................................................................57
Process Flo% and the fnction of the ma/or components of $olid
Desiccant Dehydrators.....................................................................................01
Fnction of (a/or Components of $olid Desiccant Dehydrators....................00
,nstrmentation................................................................................................7'
5ptimi6ing 7dsorption"#ype Dehydrators......................................................70

&

PRINCIPLES OF GAS DEHYDRATION
Background
Liquid water and/or water vapor are reoved !ro natura" #a$ to%
8 Pre9ent formation of hydrates in transmission lines.
8 (eet a %ater de% point re*irement of a sales gas contract.
8 Pre9ent corrosion.
#echni*es for dehydrating natral gas inclde:
8 7-sorption sing li*id desiccants.
8 7dsorption sing solid desiccants.
#hrogh a-sorption; the %ater in a gas stream is dissol9ed in a relati9ely pre li*id
sol9ent stream. #he re9erse process; in %hich the %ater in the sol9ent is transferred
into the gas phase; is <no%n as stripping. #he term regeneration is also sed to
descri-e stripping =or prification> -ecase the sol9ent is sally reco9ered for rese
in the a-sorption step.
7-sorption and stripping are fre*ently sed in gas processing and most gas
s%eetening operations; as %ell as in glycol dehydration.
#he second ma/or process -y %hich %ater 9apor is remo9ed from a gas stream is
called adsorption. 7dsorption is a physical phenomenon that occrs %hen molecles
of a gas are -roght into contact %ith a solid srface and some of them condense on
the srface.
Dehydration of a gas %ith a dry desiccant is an adsorption process in %hich %ater
molecles are preferentially held -y the desiccant and remo9ed from the stream.
&ater Content o! H'dro(ar)on
?ased on e@perimental data; Fig"& sho%s the sol-ility of %ater in s%eet
hydrocar-on li*ids. ,n sor hydrocar-on li*ids; %ater sol-ility can -e
s-stantially higher.

3

Fig 2 - Solubility of Water in Liquid Hydrocarbons

4

&ATER CONTENT OF GASES
#he satrated %ater content of a gas depends on pressre; temperatre; and
composition. #he effect of composition increases %ith pressre and is particlarly
important if the gas contains C5& and2or H&$. For lean; s%eet natral gases
containing o9er 7'A methane and small amonts of hea9y hydrocar-ons;
generali6ed pressre"temperatre correlations are sita-le for many applications.
Fig."3 is an e@ample of one sch correlation %hich has -een %idely sed for many
years in the design of Bs%eetC natral gas dehydrators. #he gas gra9ity correlation
shold ne9er -e sed to accont for the presence of H&$ and C5& and may not
al%ays -e ade*ate for certain hydrocar-on effects; especially for the prediction of
%ater content at pressres a-o9e 15'' psia. #he hydrate formation line is
appro@imate and shold not -e sed to predict hydrate formation conditions.
#he follo%ing e@amples are pro9ided to illstrate the se of Fig."3:
E*ap"e +,-. D Determine the satrated %ater content for a s%eet lean
hydrocar-on gas at 15'EF and 1;''' psia.
From Fig. "3;
W = &&' l-/((scf
For a &0 moleclar %eight gas;
Cg = '.41 =Fig. "3>
W = ('.41)(&&') = &10 l-/((scf
For a gas in e*ili-rim %ith a 3A -rine;
Cs = '.43 =Fig. "3>
W = ('.43)(&&') = &'5 l-/((scf

5


0

Fig- 3 : Water Content of Hydrocarbon Gas

7

Water Content of High CO2/H2S Gases
?oth H&$ and C5& contain more %ater at satration than methane or s%eet
natral gas mi@tres. #he relati9e amonts 9ary considera-ly %ith temperatre and
pressre. Fig. "4 and"5 display satrated %ater content of pre C5& and H&$;
respecti9ely; at 9arios temperatres and pressres. Fig. "0 sho%s the satrated %ater
contents of pre CH4; C5& and selected mi@tres 9s. pressre at 1''EF. Fig. "7
e@hi-its the satrated %ater content of pre CH4; C5& and a 45A C5& F 5A CH4
mi@tre 9s. pressre at 1''EF. $e9eral significant o-ser9ations can -e made from
these figres and other a9aila-le data.
1. $atrated %ater content of pre C5& and H&$ can -e significantly higher than
that of s%eet natral gas; particlarly at pressres a-o9e a-ot 7'' psia at
am-ient temperatres.
&. Corrections for H&$ and C5& shold -e applied %hen the gas mi@tre
contains more than 5A H&$ and2or C5& at pressres a-o9e 7'' psia. #hese
corrections -ecome increasingly significant at higher concentrations and
higher pressres.
3. #he addition of small amonts of CH4 or .& to C5& or H&$ can dramatically
redce the satrated %ater content compared to the pre acid gas.

1
FG-!
Water content of C"2


4
FG-#


1'
FG-$

7cid gas %ater content is a 9ery comple@ s-/ect. #he data and methods
presented here shold not -e sed for final design. Fig. "4; "5; "0 and "7 are all
-ased on e@perimental data. 7 crsory stdy of these figres re9eals the
comple@ities in9ol9ed. 7n accrate determination of %ater content re*ires a carefl
stdy of the e@isting literatre and a9aila-le e@perimental data. ,n most cases
additional e@perimental data is the -est %ay to 9erify predicted 9ales.
?elo% 4'A acid gas components; one method of estimating the %ater content ses
)* "1 and Fig. "3; "1; and "4.
W = yHC WHC + yC5& WC5& + yH&$WH&$ %q -&

11
FG-'

%here:
W G $atrated %ater content of gas stream; l- H&52(($CF
W@@ G )ffecti9e satrated %ater content of each
component; l- H&5 2(($CF
y@@ G (ole fraction of component in gas stream
.ote that Fig. "1 and "4 pro9ide 9ales for %hat is termed the Beffecti9eC
%ater content of C5& and H&$ in natral gas mi@tres for se only in )* "1. #hese
are not pre C5& and H&$ %ater contents.
7nother method for estimation of the satrated %ater content of acid gas
mi@tres p to 0''' psia ses Fig. "1' and "11. With gases containing C5&; the C5&
mst -e con9erted to an Be*i9alentC H&$ concentration. For prposes of this
method; it is assmed the C5& %ill contri-te 75A as mch %ater to the gas
mi@tre; on a molar -asis; as H&$.

1&
FG-( FG-)


13
FG-&*

E*ap"e +,-+ D Determine the satrated %ater content of an
1'A C1; &'A C5& mi@tre at 10' EF and &''' psia. #he e@perimentally
determined %ater content %as 17& l-2((scf.
+ethod "ne
WHC = 107 l-/((scf =Fig. "3>
WC5& = &4' l-/((scf =Fig. "4>
W = ('.1')(107) + ('.&')(&4')
= 11& l-/((scf

14
FG-&&

+ethod ,-o
First the composition mst -e con9erted for se %ith Fig. "1'.
yH&$ ( psedo ) = ('.75)(yC5&) = ('.75)('.&') = '.15
W = '.44 --l/((scf =Fig. "1'>
Density of %ater = 35' l-/--l
= ('.44)(35') = 17& l-/((scf
Fig. "1& compares the t%o methods presented for satrated %ater content
determination of high C5&2H&$ gas mi@tres %ith some of the a9aila-le
e@perimental data. #he last for data points sho%n in Fig. "1& indicate the dangers
in9ol9ed %ith e@trapolation to higher C5& or H&$ contents. ,n one case; the
estimated %ater content agrees %ithin 11A of the e@perimental 9ale. ,n another
case; the e@perimental 9ale is o9er 0 times the estimated %ater content.
Fig "1& Comparison of )@perimental 9s. Calclated Water Contents for 7cid Gases
(i@tre #; EF P; psig
Water Content l-2((scf
)@perimental )* "1 Fig."1' H"11
11A C5&214A C1 1'' &''' 4'.0 4& 34.&
11A C5&214A C1 10' 1''' &10 &77 &17
&'A C5&21'A C1 1'' &''' 4'.0 43 44.1
&'A C5&21'A C1 10' 1''' &1& &71 &17
1A H&$24&A C1 13' 15'' 111 1'5 11&
&7.5A H&$27&.5A C1 10' 1307 &47 &51 &73
17A H&$213A C1 10' 1''' &4& &71 &4'
C12C5&2H&$
3'A20'A21'A
1'' 11'' 11 7& .7
C12C5&2H&$
4A21'A211A
1'' 14'' 44& 7& .7
5.31A C1244.04A C5& 77 15'' 1'4.& 31 .7
5.31A C1244.04A C5& 1&& &''' 104.0 1'5 .7

15

H'drate Foration
7 hydrate is a physical com-ination of %ater and other small molecles to prodce a
solid that has an Iice"li<eI appearance; -t possesses a different strctre than ice.
#he conditions that promote hydrate formation are:
Primary Considerations
Gas mst -e at or -elo% its %ater de% point %ith IfreeI %ater present.
Jo% temperatre.
High pressre.
Secondary Considerations
High 9elocities.
Pressre plsations.
,ntrodction of a small crystal of the hydrate.
Physical site for crystal formation sch as a pipe el-o%; an orifice;
thermo%ell; or line scale.
7ll of these primary and secondary considerations shold -e minimi6ed %hen forced
to operate near a possi-le hydrate region.
Deterinin# t/e H'drate-Foration Teperature$ o! Sweet and
Sour Ga$ Strea$
C/ei(a" 0a1eup o! H'drate$
Hydrates form %hen free %ater com-ines %ith the follo%ing gases:
?tane =C4>
Car-on dio@ide =C5&>
)thane =C&>
Hydrogen slfide =H&$>
(ethane =C1>
.itrogen =.&>
Propane =C3>
Ji*id %ater has a mo-ile lattice strctre. #his lattice strctre has t%o 9acant
lattice positions. When gas molecles fill these 9acancies; the lattice is immo-ili6ed;
and the gas and %ater form a solid strctre.
Co./osition
Hydrocar-ons %ith fi9e or more car-on atoms =C5K> do not fit into these lattice
9acancies; so they do not form hydrates. Pre n"-tane does not form hydrates -y

10

itself; -t n"-tane does e@ist in hydrates %hen other smaller hydrate"forming
molecles are present. Hydrates are 4' %tA %ater %ith specific gra9ities that range
from '.40 to '.41. #herefore; hydrates float on %ater; -t sin< in hydrocar-on
li*ids.
Structure
Hydrates form one of t%o different crystalline strctres: $trctre , or $trctre ,,.
Figre"13 smmari6es the characteristics of hydrate crystalline strctres. Figre"14
sho%s $trctre , and $trctre ,, lattices.

$#!+C#+!) , $#!+C#+!) ,,
& small and 1 large 9oids 10 small and 0 large 9oids
Generally formed -y CH
4
; C
&
H
0
; H
&
$;
C5
&

Generally formed -y C
3
H
1
; i"C
4
H
1'
;
CH
&
Cl
&
; CHCl
3
5 324 %ater molecles per gas molecle
(7L
17 %ater molecles per gas molecle
(7L
F,G+!) 13: CH7!7C#)!,$#,C$ 5F HMD!7#) C!M$#7J $#!+C#+!)$
F,G+!) 14: HMD!7#) $#!+C#+!)$

17

Factors 0ro.oting Hydrate For.ation
0ri.ary
#he follo%ing are the primary conditions that promote hydrate formation:
Free %ater =Gas is at or -elo% its de% point.>
High pressre
Jo% temperatre
Secondary
$econdary conditions that promote hydrate formation inclde the follo%ing:
High 9elocities.
Physical sites %here crystals might form sch as pipe el-o%s; orifices; or line
scale.
Pressre plsations.
$mall crystals of hydrates that may act as seed crystals.
#r-lence in gas streams =promotes crystal gro%th -y agitating spercooled
soltions>.
"ther Considerations
Gas composition greatly affects hydrate"formation temperatres. High H&$
concentrations promote hydrate formation. #he presence of ethane and propane also
promotes the formation of hydrates. Gases %ith higher specific gra9ities form
hydrates at lo%er pressres.
Hydrates ta<e se9eral hors to form. ,n one test; hydrate crystals too< more than 1'
hors to form in gases sper"cooled to 14EF.
%ffects of Co./osition on Hydrate For.ation Conditions
C& ,hrough C3
#he smaller natral gas molecles; methane =C1> and ethane =C&>; form sta-le
strctre , hydrates. Ho%e9er; e9en small concentrations of propane =C3> or ethane
strongly promote the formation of hydrates in gas streams. ,n a test performed -y
Deaton and Frost; a 1A soltion of propane in methane at 5'EF lo%ered the hydrate
formation pressre of methane nearly 4'' psi. Deaton and Frost fond ethane
prodced similar reslts as propane.
H2S and C"2

11

H&$ and C5& form sta-le $trctre", hydrates. H&$ in a gas stream strongly
promotes the formation of hydrates. 7-o9e 15EF H&$ does not form hydrates. #his
critical temperatre a-o9e %hich hydrates do not form is higher for H&$ than for
other components typically fond in natral gas streams. #herefore; in gas streams
of e*al densities; the gas stream %ith H&$ forms hydrates more readily.
1sing Gra/hical ,echniques to 0redict Hydrate-For.ation Conditions
Generally; hydrate formation temperatres shold -e e9alated anytime a gas stream
containing %ater and hydrate"forming components is cooled -elo% 1'EF. ,t is not
necessary to e9alate hydrate formation temperatres -elo% 3&EF since pre %ater
free6es any%ay.
Figre 15 plots hydrate"formation temperatres of pre light gases. Hydrate
formation %ill occr in the region a-o9e and to the left of the cr9e for a gi9en
compond. #he discontinities in the lines correspond to changes in phase of the
nonhydrate phases.
F,G+!) 15: HMD!7#) F5!(7#,5. C5.D,#,5.$ 5F P+!) J,GH# G7$)$

14

)mpirical hydrate formation graphs may -e sed to determine the hydrate formation
conditions of a gas stream. #he gra9ity graphic method and the N"9ale method are
the t%o graphical methods sed to predict hydrate formation conditions.
Compter programs are more commonly sed; -t the gra9ity graphic method is still
sefl for roghly appro@imating hydrate formation conditions.
Gra2ity Gra/hic +ethod
Figre 10 plots the hydrate"formation conditions of gases -ased on their specific
gra9ity relati9e to air =(W G &4>.
?ecase hydrocar-on mi@tres -elong to the same chemical family; the methodOs
se of specific gra9ity ma<es some ad/stment for the composition of the gas
stream.
F,G+!) 10: Pressre"#emperatre Cr9es for Predicting Hydrate Formation

&'

#he gra9ity graphic method roghly appro@imates hydrate formation conditions.
#his method shold not -e sed on gas streams %ith s-stantial concentrations of
non"hydrocar-ons =.&; C5&; and especially H&$>.
0rocedures and 3esources for Deter.ining the Hydrate-For.ation
,e./eratures of S-eet and Sour Gas Strea.s 4Gra2ity Gra/hic +ethod5
1. Calclate the %eight of component per mole of gas mi@tre -y mltiplying the
mole fraction of each component -y the moleclar %eight of each component.
!ecord the partial moleclar %eight of each component in the right colmn of the
ta-le pro9ided %ith the e@ercise.
&. Calclate the total moleclar %eight of the gas mi@tre. #o ma<e this calclation;
total the partial moleclar %eights in the right colmn of the ta-le pro9ided %ith
the e@ercise.
!ecord the total %eight at the -ottom of the right colmn of the ta-le pro9ided
%ith the e@ercise.
3. +se )*n. 0 to calclate the specific gra9ity =relati9e to air> of the gas mi@tre.
sp. gr.=gas> G =)*n. &>
%here: sp. gr.=gas> G $pecific gra9ity of the gas stream
(Wgas G (oleclar %eight of gas stream; l-2mole
(Wair G (oleclar %eight of air
G &4 l-2mole
4. +se Figre 10 to determine the hydrate"formation temperatre of the gas stream.
#he follo%ing sample pro-lem demonstrates the gra9ity graphic method
$ample Pro-lem: Determine the Hydrate"Formation #emperatres of $%eet and
$or Gas $treams =Gra9ity Graphic (ethod>
Calclate the appro@imate temperatre at %hich the gas stream entering a chill do%n
train at a Gas Plant forms hydrates. +se the gra9ity graphic method. #he left colmn
of Figre 17 lists the composition of the gas stream and the right colmn is pro9ided
to help organi6e the calclations.
Gi9en:
Pressre G 4&4 psig

&1
(Wgas
(Wair
P Q

#emperatre from dehydrator G 1'EF
C5(P5.).#
(5J)
F!7C#,5.
(5J)C+J7!
W),GH#
l-2(ole 5F
(,L#+!)
.& '.''00 &1.' '.115
C5& '.'''3 44.' '.'13&
H&$ '.' 34.3 '.''
C1 '.0317 10.' 1'.1
C& '.&111 3'.1 0.35
C3 '.1'11 44.1 4.1'
i"C4 '.''71 51.1 '.453
n"C4 '.'&4& 51.1 1.41
i"C5 '.''31 7&.& '.&&4
n"C5 '.''41 7&.& '.340
n"C0 '.''14 10.& '.1&1
C7K '.'''& 1''.& '.'&'
#5#7J G7$ $#!)7( 1.'' "" &4.'
F,G+!) 17: #7?J) F5! C7JC+J7#,.G #H) (5J)C+J7! W),GH# 5F
#H) G7$ $#!)7(
$oltion:
1. #he right colmn of Figre 17 sho%s the calclation of the %eight of each gas
component per mole of gas stream.
&. #he %eight of each gas component per mole of gas stream is totaled at the -ottom
of the right colmn of the ta-le pro9ided in Figre 17.
3. #he se of )*n. & to calclate the specific gra9ity =relati9e to air> of the gas
mi@tre reslts in the follo%ing:
sp. gr. G
G
G '.1&1 =)*n.&>
4. From Figre 10; the hydrate"formation temperatre of the gas stream at 4&4 psig
=434 psia> is determined to -e 57EF.
7ns%er:
Hydrates can form in this gas stream at appro@imately 57EF.
6llo-able Gas %7/ansions
Graphical methods also predict permissi-le gas e@pansions. #hese graphs se the
gra9ity graphic method to determine the allo%a-le e@pansion =decrease in pressre>

&&
(Wgas
(Wair
P Q
&4.' l- 2 mole
&4.' l- 2 mole
P Q

at 9arios initial temperatres. 7s %ith the gra9ity graphic method; these graphs are
sefl for initial estimates; -t shold not -e sed for design.
Teperature Contro" 0et/od$ and Equipent 2$ed To In/i)it H'drate
Foration in a Natura" Ga$ Strea
Heating a natral gas or depressri6ing it =ths cooling it> %hile it is nder hot
conditions can inhi-it hydrate formation. ,n a-o9e grond operations; the
temperatre drop cased -y depressri6ing =e@panding> a gas can reslt in the
temperatre of the gas stream dropping -elo% its hydrate"formation temperatre.
?ecase of the high temperatres ndergrond; a gas stream can -e e@panded
ndergrond %ithot the reslting temperatre dropping -elo% its hydrate"formation
temperatre. #herefore; e@panding a gas stream in a %ell -ore helps pre9ent hydrate"
formation in do%nstream processing.
#he t%o main pieces of e*ipment sed to control gas stream temperatre and
inhi-it hydrate formation are do%nhole reglators and indirect heaters. Do%nhole
reglators inhi-it hydrate formation -y e@panding gas streams %hile they are in the
%ell-ore. ,ndirect heaters inhi-it hydrate formation -oth at %ellheads =%ellhead
heaters> and along flo%lines =flo%line heaters>.
,ndirect heaters are often sed to inhi-it hydrate formation cased -y e@pansion or
to replace heat lost -y a flo%line to the srronding air and grond.
Do-nhole 3egulators
#he se of do%nhole reglators to inhi-it hydrate formation -y controlling gas
stream temperatres is generally feasi-le %hen the gas %ell has the follo%ing
conditions:
7 high reser9oir pressre that is not e@pected to decline rapidly
)@cess pressre
High capacity
#he temperatre and pressre of a gas stream as %ell as its composition determine
%hether hydrates %ill form %hen gas is e@panded into the flo%lines. Cooling occrs
as gas is e@panded across the cho<e. Do%nhole reglators lo%er the pressre of the
gas stream from %ell pressre to near"salesline pressre in the %ell-ore. 5perating
conditions reslting from the e@pansion of the gas are otside the hydrate"formation
range of the gas stream -ecase of the high temperatres in the %ell.
Do-nhole 3egulator Design
Do%nhole reglators contain a spring"loaded 9al9e and stem that otside 9endors set
from the srface -y sing a %ireline =%ire sed to lo%er tools into the %ell-ore> rn

&3

throgh the %ell-ore t-ing. #he pressre drop across the reglator remains constant
and does not depend; %ithin a -road range; on the flo% rate of the %ell.
#he design of do%nhole reglators re*ires sing comple@ calclations that mst
accont for the follo%ing:
Do%nhole pressres and temperatres
Well depth
Well-ore configration
ndirect Heaters
#%o types of indirect heaters are sed to inhi-it hydrate formation: %ellhead and
flo%line.
#he e@pansion of gas streams at or near %ellheads often reslts in the formation of
hydrates.
Wellhead heaters <eep the temperatres of these gas streams a-o9e their hydrate"
formation temperatres.
Flo%lines some parts of the %orld often lose enogh heat to the srronding air and
grond to lo%er the temperatre of the gas stream -elo% its hydrate"formation
temperatre. Flo%line heaters inhi-it hydrate formation -y replacing this lost heat
and <eeping the temperatre of the gas stream a-o9e its hydrate"formation
temperatre. Flo%line heaters also inhi-it hydrate formation -y heating gas streams
e@panded or cho<ed do%nstream from the %ellhead.
ndirect Heater Design
Different heater designs accomplish the same prpose: to heat the gas. Flo%line
heaters do not re*ire the cho<es and high"pressre safety 9al9es that %ellhead
heaters need.
,ndirect heaters are 9essels that contain a fire t-e and a coil immersed in a heat
transfer flid =sally %ater or a glycol and %ater mi@tre> %ithin a heater shell. #he
fire t-e is sally fired -y gas. #he coil contains the flid =the gas stream> to -e
heated and operates at fll gas pressre. #he heater shell operates at atmospheric
pressre. Figre 11 sho%s a typical indirect heater.
Flo-line Heaters " Flo%line heaters heat gas streams a-o9e their hydrate"forming
temperatres. ,n many cases; properly designed and placed %ellhead heaters pro9ide
sfficient heat to eliminate the need for flo%line heaters.
ndirect Heater Si8ing
#he determination of the si6e of a heater depends on the follo%ing conditions:

&4

7monts of gas; %ater; oil; or condensate e@pected in the heater
,nlet temperatre and pressre
5tlet temperatre and pressre =to a9oid hydrate"forming conditions>
#he si6e of heater coils to se depends on the 9olme of flid flo%ing throgh the
coil and the re*ired heat"transfer load.
When heater coils are si6ed; it is important to consider operating conditions in
addition to normal; steady"state operating conditions. #ransient startp of a sht"in
%ell may re*ire e@tra heating capacity. #he temperatre and pressre conditions of
a sht"in %ell and the e@tra li*ids accmlated %hile the %ell %as sht in may
increase the heating load. 5ften; heaters are necessary only %hile %ells are -eing
started p. ,nstalling preheat coils ahead of cho<es is generally practical for %ells
operated only intermittently.
F,G+!) 11: #MP,C7J ,.D,!)C# H)7#)!
6d2antages and Disad2antages of ,e./erature Control +ethods
Downhole Regulators
Do%nhole reglators ha9e the follo%ing ad9antages:
Jo% initial in9estment
Do not re*ire rotine ser9ice
Do%nhole reglators ha9e the follo%ing limitations or disad9antages:

&5

#hey may not inhi-it hydrate formation dring startp. ,t may -e necessary to
inhi-it hydrate formation -y in/ecting either methanol or glycol ntil the gas
flo% and temperatre sta-ili6e.
When %ell otpt falls -elo% normal prodction le9els; processors mst
remo9e and replace do%nhole reglators %ith another hydrate inhi-ition
method.
When %or< is performed inside a %ell-ore; the %ell may -e permanently
damaged.
Indirect Heaters
#he ad9antages of sing indirect heaters to inhi-it the formation of hydrates inclde
the follo%ing:
(inimal maintenance or attention re*ired
3ery lo% chemical re*irements
#he disad9antages of sing indirect heaters to inhi-it hydrates inclde the follo%ing:
Difficlty of spplying clean and relia-le fel to remote locations
Jarge operating =fel> costs if cheap fel is not a9aila-le
Potentially large capital costs
$ignificant plot space re*ired
$pecial safety e*ipment needed -ecase of fire ha6ard
Co./arison of ,e./erature Control +ethods
Figre 14 compares the se of do%nhole reglators and %ellhead heaters to inhi-it
hydrate formation. #he high capital costs of heaters generally limit their se to large
hydrate inhi-ition installations. Do%nhole reglators %or< -est in large reser9oirs
%ith high gas pressres that are not e@pected to decline rapidly.
D)$,G. F7C#5!$
D5W.H5J)
!)G+J7#5!$
W)JJH)7D
H)7#)!$
,n9estment 3ery lo% 3ery high
Fel .one 3ery high
5perating (aintenance Jo% Jo%
Chemicals .one 3ery lo%
Plot 7rea .one 3ery high
Ha6ards High High
F,G+!) 14: C5(P7!,$5. 5F #)(P)!7#+!) C5.#!5J ()#H5D$

&0

0et/ano" In3e(tion Rate Required To In/i)it H'drate
Foration in a Natura" Ga$ Strea
Chemical Injection
Crrently; methanol =(e5H> and monoethylene glycol =()G> are the t%o
chemicals most commonly in/ected into gas streams to inhi-it hydrate formation.
Consider the se of chemical in/ection to inhi-it hydrate formation for the follo%ing:
Gas pipelines in %hich hydrates form at locali6ed points
Gas streams operating a fe% degrees a-o9e their hydrate formation
temperatre
Gas"gathering systems in pressre"declining fields
$itations %here hydrate pro-lems are of short dration
Hydrate inhi-itors act similarly to antifree6e. 7dding a <no%n *antity of an
inhi-itor to a <no%n *antity of pre li*id redces the hydrate"formation
temperatre -y a calcla-le amont.
Methanol
(ethanol %or<s %ell as a hydrate inhi-itor -ecase of the follo%ing reasons:
,t can attac< or dissol9e hydrates already formed.
,t does not react chemically %ith any natral gas constitents.
,t is not corrosi9e.
,t is reasona-le in cost.
,t is sol-le in %ater at all concentrations.
(ethanol significantly depresses hydrate"formation temperatres.
Methanol Applications
?ecase methanolOs material cost is so lo% and its 9apor losses so high; methanol is
often not reco9ered. .ot re*iring a reco9ery system significantly redces capital
costs. #herefore; methanol in/ection is generally economical for temporary
installations; sitations %ith lo% gas 9olmes; or sitations %ith mild; infre*ent; or
seasonal hydrate pro-lems.
Methanol Injection System
Figre &' sho%s a simplified schematic of a typical methanol in/ection system. #his
system inhi-its hydrate formation at a cho<e or pressre"redcing 9al9e. 7 gas"

&7

dri9en pmp in/ects the methanol into the gas stream pstream of the cho<e or
pressre"redcing 9al9e. #he temperatre controller measres the temperatre in the
gas stream and ad/sts the po%er"gas control 9al9e. #he po%er"gas control 9al9e
controls the flo% of po%er gas; %hich controls the methanol in/ection rate.
F,G+!) &': ()#H7.5J ,.R)C#,5. $M$#)(
+ethod of n9ecting +ethanol " #he in/ection of methanol considera-ly pstream
of a hydrate"forming location allo%s the methanol to distri-te and 9apori6e
completely. ?ecase of methanolOs high 9olatility; no66le placement and design are
not as critical as they are for glycol in/ection. (ethanol in/ection no66les shold -e
located as follo%s:
+pstream of front"end e@changers
7t the inlets of tr-oe@panders
7t any refrigerated condensers in do%nstream fractionation

&1
(e5H
#emp. Controller
Po%er gas
Gas stream
(ethanol
pmp
,n/ection point
Cho<e

G"'(o" In3e(tion Rate Required To In/i)it H'drate
Foration in a Natura" Ga$ Strea
Ji<e methanol; glycol inhi-its hydrate formation %hen in/ected into gas streams.
Figre &1 compares the ad9antages and disad9antages of glycol and methanol
in/ection.
,.H,?,#5! 7D37.#7G)$ D,$7D37.#7G)$2 J,(,#7#,5.$
Glycol +sally lo%er operating cost
than methanol %hen -oth
systems reco9er in/ected
chemical
Jo% 9apor losses =lo% 9olatility>
High initial cost
Possi-ility of glycol contamination
Jimited se =only non"cryogenic
applications>
Cannot dissol9e hydrates already formed
(ethanol !elati9ely lo% initial cost
$imple system
Does not generally need to -e
reco9ered
Jo% 9iscosity
When in/ected; distri-tes %ell
into gas streams
Can dissol9e hydrates already
formed
High operating cost
Generally; se glycol in/ection if
methanol in/ection rate is o9er 3' gph
Jarge 9apor losses =high 9olatility>
F,G+!) &1: C5(P7!,$5. 5F CH)(,C7J ,.R)C#,5. ,.H,?,#5!$
Glycol does not e9aporate as easily as methanol. ,n some applications; glycol does
not dissol9e into li*id hydrocar-ons as easily as methanol. Glycol sol-ility in
hydrocar-on li*id increases %ith:
Glycol moleclar %eight
#emperatre increase
,ncrease in glycol concentration in %ater"glycol mi@tre
Glycol Concentration and Dilution:
,n addition to inhi-iting hydrate formation; yo also need to choose glycol
concentrations that do not free6e. Figre && sho%s the free6ing points of 9arios
a*eos glycol soltions:

&4

F,G+!) &&: F!))S,.G P5,.#$ 5F 7T+)5+$ GJMC5J $5J+#,5.$
.ote that soltions %ith glycol concentrations -et%een a-ot 0' %t A and 1' %t A
do not free6e. ?ecase of this; glycol soltions are generally <ept -et%een these
concentrations; e9en if lo%er concentrations are re*ired to depress the hydrate"
formation temperatre.
When glycol in/ection is performed -elo% &'EF; the glycol free6ing point mst -e
considered. Glycols crystalli6e; -t do not free6e solid; %hich inhi-its flo% and
proper separation. For this reason; it is common practice to <eep glycol
concentrations -et%een 0'"1' %t A.
#o a9oid the formation of emlsions; the %ater content of the in/ected inhi-itor =lean
glycol> soltion shold -e greater than &' %t A. #herefore; the in/ection rate of pre
glycol re*ired -y the system to inhi-it hydrate formation is first calclated and then
the in/ection rate of the lean glycol soltion is calclated.

3'

#o <eep the concentration of the glycol -et%een 0' %t A and 1' %t A; the e@tent to
%hich the free %ater diltes the in/ected glycol mst -e determined. Figre &3 lists
and smmari6es diltion restrictions
$,#+7#,5. 7JJ5W7?J) 5! !)C5(().D)D
D,J+#,5. 5F GJMC5J
+n<no%ns a-ot the system e@ist .ot o9er 5A to 1'A
$pot in/ection =in a heat e@changer; for
e@ample>
,f distri-tion of glycol is a pro-lem;
limit to a-ot 5A
Pipelines operating a-o9e &'EF +p to a-ot &'A
F,G+!) &3: 7JJ5W7?J) GJMC5J D,J+#,5.$
Selecting Glycol ,y/e
#he glycols normally sed for hydrate inhi-ition are the follo%ing:
()G
D)G
#)G
$election of the appropriate type of glycol depends on the composition of the gas
stream and on information pro9ided -y the glycol 9endor.
For instance; Do% Chemical recommends that its glycols -e sed at concentrations
of 7' %t A to 75 %t A to a9oid free6ing pro-lems. Do% Chemical also ma<es the
recommendations for selecting glycols listed in Figre &4.
$,#+7#,5.2C5.D,#,5. !)C5(().D7#,5.
.atral gas transmission in
%hich reco9ery is not important
+se ()G -ecase it depresses hydrate
formation temperatres the most.
,n/ected glycol contacts
hydrocar-on li*ids
+se ()G -ecase it has the lo%est sol-ility
of the glycols in high moleclar %eight
hydrocar-ons.
$e9ere 9apor losses +se D)G or #)G -ecase -oth glycols ha9e
lo%er 9apor pressres than the other glycols.
$e9ere 9apor losses and in/ected
glycol contacts hydrocar-on
li*ids
When -oth of these conditions are present;
D)G may -e the -est choice
F,G+!) &4: GJMC5J !)C5(().D7#,5.$

31

Glycol n9ection and 3eco2ery Syste. 1sing a ,hree-0hase Se/arator
Figre &5 sho%s a typical glycol in/ection and reco9ery system that ses a three"
phase separator. #he po%er"gas"dri9en pmp; the temperatre controller; and the
in/ection point sho%n in Figre 15 are similar to the methanol in/ection system
sho%n in Figre 5. 7 gas dri9en pmp in/ects the glycol into the gas stream
pstream from the cho<e or pressre redcing 9al9e. #he temperatre controller
measres the temperatre in the gas stream and ad/sts the po%er"gas control 9al9e.
#he po%er"gas control 9al9e controls the flo% of po%er gas; %hich controls the
in/ection rate.
F,G+!) &5: GJMC5J ,.R)C#,5. 7.D !)C53)!M $M$#)(

3&

F,G+!) &0: GJMC5J ,.R)C#,5. 7.D !)C53)!M $M$#)(

LTS
PC4-+,. D.
PC4-56+
E +,.
A/7
G"'(o"
Rea(/ G"'(o"
LC4 +,8 D+
LC4 569
E+,+
A/7
F"are
LC4 +,8 D.
LC4 5:.
S D 4 5 6 ,
S D 4 5 ; .
SD4 5;5
SD4 5:+
SD4 +,8
SD4 5:8
PC4 566
PC4 +,8D+
D-205
F"are
Lean
Glycol
From
D-204
Gas
from
D-201
LP Sa"e$
Ga$
F&'1
7
F&'1
?
D &1'
$<immer
D " &11
!e-oiler
Jean Glycol to
in/ection points
Jean Glycol to
in/ection points
#N &'1
Jean Glycol
$torage #an<
#N &'&
!ich Glycol
$torage #an<
Rich Glycol
?lan<et Gas
Closed Drain
P &'1 72?
F &'3
F &'4
Glycol Sump
7ir Pmp
P"&1'2&11
33

#he reco9ery side of the system sho%n in Figre &5 incldes a re-oiler and a three"
phase separator. #he glycol in/ection and reco9ery cycle is as follo%s:
#he in/ection no66le in/ects the lean glycol into the gas stream.
#he lean glycol a-sor-s the %ater and inhi-its hydrate formation in the cho<e
or pressre"redcing 9al9e.
#he three"phase separator separates the %ater and rich glycol from the
hydrocar-on gas and li*id.
#he separated components are piped to their respecti9e destinations.
#he re-oiler -oils off e@cess %ater from the rich glycol; and there-y prepares
it to -e in/ected again.
Glycol n9ection and 3eco2ery Syste. Co./onents
Separators
#he lo%"temperatre separator sho%n in Figre &5 separates the hydrocar-on gas
from the hydrocar-on condensate"rich glycol mi@tre. #he glycol"oil separator in
Figre &5 flashes the remaining hydrocar-on condensate"rich glycol mi@tre to a
lo% pressre and then separates ot the rich glycol.
$eparating the rich glycol from the hydrocar-on li*id is more difficlt than
separating hydrocar-on li*id from 9apor. Performing -oth separations in one 9essel
sacrifices some effecti9eness and efficiency. Generally; three"phase separators
re*ire longer residence times =&' to 4' mintes> and sffer higher glycol losses.
Reboiler
#he temperatre in the re-oiler depends on the type and concentration of the glycol
sed. !e-oilers in hydrate inhi-ition systems do not regenerate glycols to the same
high le9els of prity sed in dehydration systems.
Glycol simmer
7 t%o "phase separator %ith a 5 to 1' minte li*id retention time pre9ents e@cess
hydrocar-on 9apor from entering the stripping colmn. ,f hydrocar-ons enter the
stripping colmn; they can flash; increase glycol losses; and possi-ly -rea< the
ceramic pac<ing.
,f the specific gra9ity of the gas stream is high; then a three"phase separator
=containing 9apor; a*eos; and li*id hydrocar-on phases> %ith a &' to 3' minte
residence time is preferred.

34

F,G+!) &7: GJMC5J $N,(()!
Inhibitor Pump
7 drm on top of a typical po%er"gas"dri9en pmp contains the inhi-itor: methanol
or glycol. #he drm connects directly to the pmp =generally; a positi9e
displacement pmp>. (ethods for monitoring the inhi-itor in/ection rate inclde
inserting a cali-rated dipstic< throgh the top of the drm or pmping the inhi-itor
into a measred 9essel. Drms are replaced %hen empty.
Glycol !osses
Glycol in/ection systems that in9ol9e -oth hydrocar-on li*ids and gases generally
lose glycol to the follo%ing:
$ol-ility =normally a-ot '.3 to 3 gallons of glycol per 1''' -arrels of
hydrocar-on li*id prodced>
Jea<age
Carryo9er %ith hydrocar-on li*id and in the re-oiler
3apori6ation in the re-oiler and dring in/ection
"o##le Selection and Placement
.o66le selection and placement indirectly affect glycol in/ection calclations.
7lthogh calclated to inhi-it hydrate formation; in/ection rates may need to -e
ad/sted to maintain a flo% rate or pressre recommended for a particlar no66le
design or placement.
?ecase of glycolOs lo% 9apor pressre; no66le design is more critical for glycol
than it is for methanol. #o mi@ ade*ately %ith the natral gas; glycol re*ires a
fine; %ell"distri-ted mist. 7lso; to inhi-it hydrates flly; the no66le mst -e placed
to ensre fll co9erage.
,nstalling -ac<p no66les in parallel %ith the primary no66le allo%s no66le remo9al;
replacement; or inspection %ithot interrpting inhi-itor ser9ice.

35

;o88le Selection " .o66le design is especially important in the design of glycol
in/ection systems for cold separation facilities. #he criteria for selecting a no66le
inclde the follo%ing:
Capacity
$pray angle
$fficient pressre drop -et%een the no66le and the gas stream o9er the
e@pected range of operating conditions
.ormally; a pressre differential of 1'' psi to 15' psi sfficiently atomi6es glycol.
7lso; gas stream 9elocities a-o9e 1& ft2s help ensre atomi6ation.
;o88le 0lace.ent " .ormally; no66les are located /st pstream of the heat
e@changer or chiller %here hydrates form. #he spray from a properly located no66le
co9ers the entire t-e sheet of a heat e@changer.
,nade*ate atomi6ation cases the formation of glycol droplets that settle and flood
the -ottom of the heat e@changer. 7s a reslt; the glycol inhi-its hydrate formation in
the -ottom; -t not the top; of the heat e@changer. Flooding of the -ottom of the heat
e@changer also significantly decreases its effecti9eness.
,nade*ate co9erage can lea9e some t-es %ith a concentration of glycol that is too
lo%; %hich %ill reslt in the formation of hydrates. 7s sho%n in Figre 11; hydrates
plg the t-es; and there-y increase the differential pressre across the heat
e@changer.
F,G+!) &1: ,.C!)7$) ,. P!)$$+!) D!5P ?)C7+$) 5F HMD!7#) F5!(7#,5.

30

Figre &4 sho%s one no66le location -t three flo% rates. #oo lo% a no66le flo% rate
prodces the same reslt as a no66le located too close to the t-e sheet. #oo high a
no66le flo% rate prodces the same reslt as a no66le located too far from the t-e
sheet.
F,G+!) &4: .5SSJ) PJ7C)D 7# 5.) J5C7#,5.: #H!)) FJ5W !7#)$

37

&ATER RE0O4AL PROCESSES
Liquid/Solid esiccants
,n those sitations %here inhi-ition is not feasi-le or practical; dehydration mst -e
sed. ?oth li*id and solid desiccants may -e sed; -t economics fa9or li*id
desiccant dehydration %hen it %ill meet the re*ired dehydration specification.
Ji*id desiccant dehydration e*ipment is simple to operate and maintain. ,t can
easily -e atomated for nattended operationU for e@ample; glycol dehydration at a
remote prodction %ell. Ji*id desiccants can -e sed for sor gases; -t additional
precations in the design are needed de to the sol-ility of the acid gases in the
desiccant soltion.
$olid desiccants are normally sed for e@tremely lo% de% point specifications as
re*ired to reco9er li*id hydrocar-ons.
Gl!col eh!dration
Background
#he more common li*ids in se for dehydrating natral gas are diethylene glycol
=D)G>; triethylene glycol =#)G>; and tetraethylene glycol =#!)G>. ,n general;
glycols are sed for applications %here de% point depressions of the order of 0'EF
to 1&'EF are re*ired.
D)G %as the first glycol to -e sed commercially in natral gas dehydration and can
pro9ide reasona-le de% point control. With the e@ception of #)G; D)G is the -est
li*id a9aila-le.
Ho%e9er; %ith normal field e*ipment; D)G can -e concentrated to only 45A
prity; %hereas #)G concentrations can reach 41 to 41.5A %ithot special
e*ipment. 7lthogh -oth glycols perform sfficient dehydration in many sitations;
#)G is sed more commonly -ecase it re*ires lo%er circlation rates for a
compara-le de% point depression than D)G does and can reach lo%er de% points. ,t
is not ad9isa-le to se triethylene glycol for dehydration at lo% temperatres
=appro@imately 5'EF>; de to its high 9iscosity. #!)G is primarily sed %hen
dehydration conditions fall -et%een those encontered in normal #)G operations;
and those in %hich gas stripping or 9acm distillation -ecomes necessary.
7 process flo% diagram of a glycol dehydration nit is sho%n in Figre 3'. Good
practice dictates installing an inlet gas scr--er; e9en if the dehydrator is near a
prodction separator.
#he inlet gas scr--er %ill pre9ent accidental dmping of large *antities of %ater;
hydrocar-ons; and2or salt %ater into the glycol contractor. )9en small *antities of

31

these materials can reslt in e@cessi9e glycol losses de to foaming; redced
efficiency; and increased maintenance.

FG13% 3*: 03"C%SS FL"W D6G36+ F"3 GL<C"L D%H<D36,"; 1;,
,n the glycol dehydration process; regenerated glycol is pmped to the top
tray of the contactor =a-sor-er>. #he glycol flo%s do%n throgh the contactor
contercrrent to the gas flo%. Water rich glycol is remo9ed from the -ottom of the
contactor; passes throgh the condenser coil; flashes off gas in a flash drm; and
flo%s throgh the glycol"glycol heat e@changer to the regenerator. ,n the regenerator;
a-sor-ed %ater is remo9ed from the glycol at atmospheric pressre -y heating. #he
regenerated glycol is cooled in the glycol heat e@changers and is recirclated to the
contactor -y the glycol pmp.
#)G %ill a-sor- a-ot 1 $CF of natral gas per gal at 1''' psig a-sor-er pressre.
#here %ill -e more a-sorption if aromatic hydrocar-ons are present. 7 three to fi9e
minte residence time in the flash drm is re*ired for degassing. )@cessi9e
hydrocar-ons in the glycol may case high glycol losses and foaming. #he o9erhead
9ent from the glycol regenerator may contain hydrocar-ons and shold -e piped to a
safe location.

34

#he separation of #)G and %ater in the regenerator is accomplished easily %ith only
internal refl@. #he separation of D)G and %ater is more difficlt de to D)GVs
higher 9apor pressre.


Fig-3& %=1L>31+ W6,%3 D%W 0";,S F"3 G6S%S ; C";,6C,
W,H ?63"1S C";C%;,36,";S "F ,%G
#o o-tain the high glycol concentrations re*ired for high de% point depressions;
stripping gas or 9acm distillation mst -e sed in the re-oiler portion of the
regeneration nit. #he amont of stripping gas re*ired to re"concentrate the glycol
to a high prity ranges from & to 1' ft3 per gallon of glycol circlated. ,f stripping
gas is sed; a reco9ery system may -e /stified.
#he de% point depression o-taina-le %ith triethylene glycol can -e estimated
from Figre 31 -ased on the contact temperatre and the concentration of the
reconcentrated glycol that is sed. Figre 31 sho%s the e*ili-rim %ater de% point
at different temperatres for gases in contact %ith 9arios concentrations of glycol.
#o se this plot; locate the contact temperatre; read p to the glycol concentration;
and then read across to find the e*ili-rim %ater de% point. ,n practice it is seldom
economical for actal gas de% points to approach e*ili-rim de% points closer
than &'EF.

4'

"SC#IBI$G %H" GL&COL "H&#'%IO$ (#OC"SS
,n general; the process of dehydrating natral gas streams %ith glycol is similar to
sing glycol in/ection to inhi-it hydrate formation. Ho%e9er; -ecase the glycol
mst not only a-sor-; -t also remo9e the %ater from the gas stream; dehydration
systems a-sor- the %ater in contactors =also called a-sor-ers> instead of -y
in/ection. Glycol dehydration systems also re*ire higher and more precisely
reglated temperatres in their re-oilers.

FG13% 32: GL<C"L D%H<D36,"; S<S,%+
Glycol Dehydration 0rocess
Figre 33 sho%s the simplified glycol dehydration system. Wet inlet gas enters the
-ottom of the contactor %hile lean glycol enters the top. 7s the %et gas stream flo%s
p%ard; it contacts the do%n%ard flo%ing lean glycol. Dring this contact; the

41

glycol a-sor-s %ater from the gas stream. Dry otlet gas lea9es the top of the
contactor and rich glycol e@its the -ottom.
#he rich glycol enters the top of the stripping colmn and contercrrently contacts
steam rising from the re-oiler. #he rich glycol then enters the re-oiler; %hich -oils
the %ater ot of the glycol. #he lean glycol lea9es the -ottom of the re-oiler and
enters the srge tan< for storage. #he pmp raises the glycol to system pressre;
preparing it for another dehydration cycle.
FG13% 33: GL<C"L D%H<D36,"; S<S,%+
(ost glycol dehydration systems are mch more comple@ than the one sho%n in
Figre 33. ,n addition to the contactor; re-oiler; still colmn; srge tan<; and glycol
pmp in Figre 33; most glycol dehydration systems se:
7 flash tan< separator
7 heat e@changer: lean gas2glycol and glycol2glycol
7n inlet scr--er
Filters
$tripping colmn

Glycol !e-oiler
Gly. ,n/. Pmp
Glycol 7ccmlator
Jean 2!ich Gly. )@ .
!ich Gly .
Flashed Gas
Gas flo% meter $ystem 3al9e
Glycol
Flash
Drm
Glycol
Cartridge
filter
Glycol Car-on
filter
Dehydration
to%er #"1'1
Sea Line
Produ(ed
Ga$
4&

#his additional e*ipment impro9es the efficiency and effecti9eness of the
simplified system . Figre 34 sho%s another glycol dehydration system.
FG13% 3!: 03"C%SS FL"W "F 6 GL<C"L D%H<D36,"; S<S,%+
Glycol Dehydration Syste. Co./onents
#his section descri-es the ma/or components in a glycol dehydration system and
ho% they fnction. #he follo%ing components are co9ered in this section:
,nlet scr--er
Contactor
Flash tan< separator
Filters
Glycol pmp
Heat e@changers
$till colmn
!e-oiler
nlet Scrubber

43

,nade*ate scr--ing cases a-ot half of all glycol dehydration system pro-lems.
,nlet scr--ers remo9e free %ater and many contaminants from the inlet gas stream.
,n addition to free %ater; these contaminants inclde:
5ils or hydrocar-ons
)ntrained -rine
Do%nhole additi9es
$olids; sch as sand and corrosion prodcts
Figre 35 smmari6es the pro-lems cased -y these contaminants if not remo9ed -y
the inlet scr--er.
C5.#7(,.7.# P!5?J)($
Free Water ,ncreases glycol recirclation; re-oiler heat dty; and fel
costs
,f the dehydration system -ecomes o9erloaded %ith %ater;
glycol can carry o9er from the contactor and2or still .
$ales gas specification may not -e achie9ed
5ils or
Hydrocar-ons
!edce the drying capacity of the glycol With %ater present;
can case foaming
+ndissol9ed oils can:
8 Plg a-sor-er trays
8 Fol heat transfer srfaces in the re-oiler
8 ,ncrease the 9iscosity of the glycol
8 Jight hydrocar-ons can flash in the stripping colmn
and case loss of glycol and2or damage to the pac<ing
)ntrained ?rine Dissol9es into glycol
Corrodes steel; especially stainless steel
Deposits on re-oiler fire t-es; casing hot spots and firet-e
-rnot
Do%nhole
7dditi9es
Case foaming; corrosion; and; if they deposit on fire t-es;
hot spots
$olids Promote foaming
)rode 9al9es and pmps
)9entally plg trays and pac<ing
FG13% 3# : 03">L%+S C61S%D >< C";,6+;6;,S

44

Contactor
Contactor to%ers contact the lean glycol %ith the %et gas stream. 7s Figre 5 sho%s;
contactor to%ers consist of three sections:
$cr--ing section in the -ottom
Drying =mass transfer> section in the middle
Glycol cooler and mist e@tractor in the top
.ote: #he scr--er section in the glycol contactorOs -ase and the glycol cooling coil
are optional items =restricted to small field nits> .
FG13% 3$ : C";,6C,"3 ,"W%3

45

Scrubbing Section " #he gas stream enters the -ottom section of the contactor and
then enters a second scr--er integrated into the contactor and a %ire mesh mist
e@tractor. #hese remo9e any contaminants and entrained li*ids not remo9ed -y the
inlet scr--er. #his second stage of scr--ing frther minimi6es the contamination
of the glycol and helps pre9ent the free %ater from o9erloading the system.
Drying Section - ,n the middle section of the contactor; the gas stream flo%s p%ard
and thoroghly contacts the do%n%ard flo%ing lean glycol throgh 9al9e trays;
---le caps; or pac<ing.
+ist %7tractor " 7 mesh or %o9en mist e@tractor redces carryo9er to less than 1 l-
li*id glycol2(($CF. For the mist e@tractor; 4 to 1 inches of stainless steel mesh
and 4 inches of Dacron are generally recommended.
Flash Dru. Se/arator
#he flash tan< separates hydrocar-on li*id and 9apor from the glycol. #)G a-sor-s
1 $CF2gal glycol at 1;''' psig and 1''EF; -t glycol a-sor-s hea9ier hydrocar-ons
mch more readily. Flash tan<s separate gas condensate and glycol -est -et%een
1''EF to 15'EF and -et%een 5' psig and 75 psig.
7 t%o"phase separator %ith a 5 to 1' minte li*id retention time pre9ents e@cess
hydrocar-on 9apor from entering the stripping colmn. ,f hydrocar-ons enter the
stripping colmn; they can flash; increase glycol losses; and possi-ly -rea< the
ceramic pac<ing. ,f the specific gra9ity of the gas stream is high; then a three"phase
separator =containing 9apor; a*eos; and li*id hydrocar-on phases> %ith a &' to
3' minte residence time is preferred.
Filters
Neeping the solids content in glycol -elo% '.'1 %t A helps pre9ent the %earing of
pmps; plgging of heat e@changers; foaming; foling of contactor trays and still
pac<ing; cell corrosion; and hot spots on re-oiler fire t-es. Placing a filter after the
glycol"glycol heat e@changer ta<es ad9antage of the redced 9iscosity of the heated
glycol.
#%o types of filters are typically sed for the glycol soltion. Cartridge filters
remo9e solids; that is; corrosion prodcts. 7cti9ated car-on filters remo9e
hydrocar-on imprities; for e@ample; %ell treating chemicals and compressor oils.
#he placement of a rich glycol filter after the lean2rich glycol e@change has the
ad9antage of a lo%er glycol 9iscosity; -t the lo%er 9iscosity is at the e@pense of a

40

potential increase to the foling rate of the rich side of the lean2rich glycol
e@changer.
Glycol 0u./
Glycol circlation pmps contain the only mo9ing parts in a glycol dehydration
system. #he three types of pmps sed in dehydration systems are:
)lectric"motor dri9en
High"pressre gas"operated
High"pressre li*id"operated
Jarger dehydration systems generally se electric"motor dri9en pmps. $maller
dehydration systems and those remotely located generally se high"pressre gas"
operated or high"pressre li*id"operated pmps.
,nstalling a second =spare> pmp capa-le of pro9iding fll glycol circlation ensres
continos dehydration if the primary glycol circlation pmp fails.
Heat %7changers
Glycol dehydration systems often se three heat e@changers:
Jean gas2glycol
Glycol2glycol
!efl@ coil located in still colmn
Lean Gas@Glycol Heat %7changers se lean gas to frther cool the glycol to 5EF to
15EF a-o9e the e@it temperatre of the gas stream -efore the glycol enters the top of
the contactor.
Generally; glycol dehydration systems se do-le"pipe or shell"and"t-e heat
e@changers for lean gas2glycol heat e@changers.
Glycol@Glycol Heat %7changers preheat the rich glycol lea9ing the -ottom of the
contactor -efore it enters the re-oiler and cools the lean glycol lea9ing the re-oiler
-efore it goes to the lean gas2glycol heat e@changer and the top of the contactor.
#he refl@ coil =sing cool; rich glycol> or the cooling fins on the still colmn
=%hiche9er is applica-le> maintains the temperatre at the top of the still colmn.
#he glycol2glycol heat e@changers essentially increase the energy efficiency of the
system. #he large difference -et%een the e@it temperatres of the lean and rich
glycol from these heat e@changers re*ires conter crrent flo% to pre9ent

47

temperatre cross. #herefore; larger glycol dehydration systems generally se t%o
do-le"pipe or plate"and"frame heat e@changers in series.
Still Colu.n
3eflu7 Coil - #he top of the still colmn contains a cooling coil that condenses
some of the steam rising from the re-oiler; pro9iding refl@ for the colmn. #his
cooling coil controls condensation and redces glycol losses.
,n addition to the re-oiler; the still colmn also reconcentrates glycol. $till colmns
contercrrently contact rich glycol %ith steam rising from the re-oiler. #his steam
strips %ater from the li*id glycol. $till colmns sally contain 4 to 1 ft of ceramic
pac<ing althogh some large nits se trays.
FG13% 3' : S,LL C"L1+; 06CA;G
3eboiler
!e-oilers pro9ide enogh heat to -oil the %ater ot of the glycol. Glycol
dehydration systems often se direct"fired heaters; -t not %hen the fire ha6ard they
present poses too mch of a threat. ,n sch locations; for e@ample; offshore
platforms; fire codes and safety re*ire the se of indirect"fired heaters.
#he -l< temperatre of #)G mst -e <ept -elo% 4''EF and the ma@imm fire"t-e
s<in temperatre -elo% 43'EF to help pre9ent thermal decomposition of the #)G.
#he +"shaped fire t-e in a #)G re-oiler mst -e si6ed for an a9erage heat fl@ of
0;''' to 1;''' ?t2hr"ft& to <eep the #)G and the re-oiler -elo% these
temperatres.
Surge ,anB 46ccu.ulator5
#he srge tan< =accmlator> holds glycol to compensate for flctations in the load
on the glycol dehydration systems. Dring normal operation; srge tan<s are only

41

a-ot half fll of glycol. Dring shtdo%n; ho%e9er; they need to hold all of the
glycol that drains do%n from the trays; 9essels; and piping. Contact %ith air can
reslt in degradation of the glycol. 7 gas -lan<et sally protects the glycol from air
contamination.
Stri//ing Gas
7dding a stripping gas to the still colmn increases the glycol prity prodced -y
the glycol reconcentrator from a-ot 41.5A to 44.4A. Glycol dehydration systems
sing stripping gas generally in/ect dry; sales gas into the top of the re-oiler. #he
stripping gas flo%s p conter crrently to the glycol; frther stripping %ater from
the glycol. +sing sales gas increases the cost of the system -ecase of the loss of
re9ene.
(rocess/esign )aria*les
$e9eral process and design 9aria-les ha9e an important effect on the sccessfl
operation of a glycol dehydration system.
Gas ,e./erature
Plant performance is especially sensiti9e to the temperatre of the incoming gas. 7t
constant pressre; the %ater content of the inlet gas increases as this temperatre is
raised. Glycol 9apori6ation losses are also increased at the the higher temperatre.
Frthermore; pro-lems can reslt from too lo% a temperatre =-elo% 5'EF> -ecase
glycol -ecomes 9ery 9iscos.
Lean Glycol ,e./erature
#he temperatre of lean glycol entering the a-sor-er has a significant effect on the
gas de% point depression; and shold -e held to a minimm to achie9e the -est
operation. Ho%e9er; it shold -e <ept at least 1'EF a-o9e the inlet gas temperatre
to minimi6e hydrocar-on condensation in the a-sor-er and s-se*ent foaming.
Glycol 3eboiler ,e./erature
#he re-oiler temperatre controls the concentration of the %ater in the glycol. With a
constant pressre; the glycol concentration increases %ith higher re-oiler
temperatres. #he re-oiler temperatre shold ne9er -e allo%ed to remain at or
a-o9e the glycol degradation temperatres. When higher glycol concentrations are
re*ired; stripping gas can -e added to the re-oiler.
3egenerator ,o/ ,e./erature
#he temperatre in the top of the regenerator is also important. 7 high temperatre
can increase glycol losses de to e@cessi9e 9apori6ation. #he recommended

44

temperatre in the top of the colmn is a-ot &&5EF. ,f the temperatre in the top of
the colmn drops too lo%; too mch %ater can -e condensed and %ashed -ac< into
the regenerator to flood the colmn and fill the re-oiler %ith e@cessi9e li*ids.
3eboiler 0ressure
!edcing the pressre in the re-oiler at a constant temperatre reslts in higher
glycol prity. #his pressre redction lo%ers the %ater partial pressre in the 9apor;
increasing the dri9ing force nder %hich %ater lea9es the glycol soltion.
Glycol Concentration
#he %ater content of the dehydrated gas depends primarily on the lean glycol
concentration. #he dry gas lea9es the contactor approaching e*ili-rim %ith the
lean glycol. #he leaner the glycol flo%ing to the a-sor-er; the more efficient the
dehydration. Figre 31 sho%s the effect of glycol concentration on gas de% point.
Glycol Circulation 3ate
When the nm-er of a-sor-er trays and glycol concentration are fi@ed; the de% point
depression of a satrated gas is a fnction of the glycol circlation rate. Whereas the
glycol concentration mainly affects the de% point of dry gas; the glycol rate controls
the total amont of %ater that can -e remo9ed. 7 typical glycol circlation rate is
a-ot three gallons of glycol per pond of %ater remo9ed =se9en ma@imm>. #he
minimm circlation rate to assre good glycol"gas contacting is a-ot t%o gallons
of glycol for each pond of %ater remo9ed.
7 greater de% point depression is easier to achie9e -y increasing the glycol
concentration rather than -y increasing the glycol circlation rate =see Figre 4>. #o
se this plot; locate the glycol circlation rate; read p to the glycol concentration;
and then read across to find the de% point depression. 7n e@cessi9e circlation rate;
especially a-o9e the design capacity; o9erloads the re-oiler and pre9ents good
glycol regeneration. ,t also pre9ents ade*ate glycol"gas contacting in the a-sor-er;
increases pmp maintenance pro-lems; and can increase glycol losses.

5'
Figure 3(:
%ffect of ,%G circulation
rate and concentration on
de- /oint de/ression

Optii<in# and Trou)"e$/ootin# De/'drator Operation$
Gl!col +aintenance
5perating and corrosion pro-lems sally occr %hen the circlating glycol gets
dirty. #herefore; to achie9e a long; tro-le"free life from the glycol; it is necessary to
recogni6e these pro-lems and <no% ho% to pre9ent them. $ome of the ma/or areas
are discssed -elo%:
0et/ano"
(ethanol in the feed gas to a glycol dehydrator %ill -e a-sor-ed -y the glycol. #his
reslts in the follo%ing pro-lems:
(ethanol %ill add additional heat dty on the re-oiler and additional 9apor
load on the regenerator. High methanol in/ection rates and slg carryo9er can
case flooding.
7*eos methanol cases rst in car-on steel; so corrosion can occr in the
regenerator and re-oiler 9apor space.
(ost of the methanol a-sor-ed in the rich glycol soltion can -e remo9ed -y
flashing in the regenerator. 7cti9ated car-on filters are sed to adsor- methanol
from the lean glycol soltion to a9oid these pro-lems.
T/era" De(opo$ition
)@cessi9e heat; a reslt of one of the follo%ing conditions; %ill decompose glycol
and form corrosi9e prodcts:
High re-oiler temperatre a-o9e the glycol decomposition le9el.
Jocali6ed o9erheating; cased -y deposits of salt or tarry prodcts on the
re-oiler fired t-es or -y poor flame direction on the fired t-es.

/H Control
.e% glycol has a netral pH of appro@imately se9en. 7s it is sed; ho%e9er; the pH
al%ays decreases and the glycol -ecomes acidic and corrosi9e; nless pH
netrali6ers or -ffers are sed. #he e*ipment corrosion rate increases rapidly %ith
a decrease in the glycol pH. 7cids created -y glycol o@idation; thermal

51

decomposition prodcts; or acid gases pic<ed p from the gas stream are the most
tro-lesome of corrosi9e contaminants. 7 lo% pH accelerates the decomposition of
glycol. ,deally; the glycol pH shold -e held at a le9el of 7.' to 7.5. 7 9ale a-o9e
1.' to 1.5 tends to ma<e glycol foam and emlsify.
?ora@; ethanolamines =sally triethanolamine>; or other al<aline netrali6ers can -e
sed to control the pH. #hese netrali6ers shold -e added %ith great care "" slo%ly
and continosly "" for -est reslts. 7n o9erdose of netrali6er %ill sally
precipitate a sspension of -lac< sldge in the glycol. #he sldge cold settle and
restrict glycol circlation. Fre*ent filter element changes shold -e made %hile pH
netrali6ers are added.
Sa"t Containation
$alt deposits accelerate e*ipment corrosion; redce heat transfer in the re-oiler
t-es; and alter specific gra9ity readings %hen a hydrometer is sed to measre
glycol"%ater concentrations. #his tro-lesome contaminant cannot -e remo9ed %ith
normal regeneration.
#herefore; an efficient scr--er pstream of the glycol plant shold -e sed to
pre9ent salt carryo9er %ith the incoming gas. ,n areas %here large *antities of -rine
are prodced; some salt contamination %ill occr. #he remo9al of salt from the
glycol soltion is then necessary.
$alt contaminated glycol may -e reclaimed -y se9eral methods. $craped"srface
heat e@changers in con/nction %ith centrifges are sed in cases of e@treme
contamination. 5ther reclamation methods are 9acm distillation or ion e@change.
H'dro(ar)on$
Ji*id hydrocar-ons; a reslt of carryo9er %ith the incoming gas or condensation in
the a-sor-er; increase glycol foaming; degradation; and losses. #hey mst -e
remo9ed %ith a glycol"gas separator; hydrocar-on li*id s<immer; or acti9ated
car-on -eds.
S"ud#e
7n accmlation of solid particles and tarry hydrocar-ons 9ery often forms in the
glycol. #his sldge is sspended in the circlating glycolU o9er a period of time; the
accmlation -ecomes large enogh to settle ot. #his action reslts in the formation
of -lac<; stic<y; a-rasi9e gm that can case tro-le in pmps; 9al9es; and other
e*ipment; sally %hen the glycol pH is lo%. #he gmmy s-stance -ecomes hard
and -rittle %hen deposited on the a-sor-er trays; stripper pac<ing; and other places
in the circlating system. Good soltion filtration pre9ents a -ildp of sldge.


5&

Foain#
Foaming can increase glycol losses and redce plant capacity. )ntrained glycol %ill
-e carried o9er the top of the a-sor-er %ith the sales gas %hen sta-le foam -ilds p
on the trays. Foaming also cases poor contacting -et%een the gas and glycol;
decreasing the drying efficiency.
$ome foam promoters are:
Hydrocar-on li*ids.
Field corrosion inhi-itors.
$alt.
Finely di9ided sspended solids.
)@cessi9e tr-lence and high li*id"to"9apor contacting 9elocities sally case
the glycol to foam. #his condition can -e cased -y mechanical or chemical
pro-lems.
#he -est %ay to pre9ent foaming is proper care of the glycol. #his in9ol9es effecti9e
gas cleaning ahead of the glycol system and good filtration of the circlating
soltion. #he se of defoamers does not sol9e the -asic pro-lem; and ser9es only as
a temporary control ntil the conditions generating foam can -e identified and
remo9ed.
'nal!sis and Control of Gl!col
7nalysis of glycol is essential to good plant operation. (eaningfl analytical
information helps pinpoint high glycol losses; foaming; corrosion; and other
operating pro-lems.
7nalyses ena-le the operator to e9alate plant performance and ma<e operating
changes to o-tain ma@imm drying efficiency.
7 glycol sample shold first -e 9isally inspected to identify some of the
contaminants:
7 finely di9ided -lac< precipitate may indicate the presence of iron corrosion
prodcts.
7 -lac<; 9iscos soltion may contain hea9y hydrocar-ons.
#he characteristic odor of decomposed glycol =a s%eet aromatic odor> sally
indicates thermal degradation.
7 t%o"phase li*id sample sally indicates the glycol is hea9ily
contaminated %ith hydrocar-ons.


53

#he 9isal inspections shold ne@t -e spported -y chemical analysis. $amples of
the lean and rich glycol shold -e ta<en and rotine tests performed: salt analysis;
solids content; pH; iron content; foam test; and titration procedre =to determine the
amont of netrali6er necessary to raise the pH to a safe le9el>. #hese analyses
sally pro9ide sfficient information to determine the condition of the glycol.
Gl!col Loss (re,ention
Glycol losses can -e defined as li*id carryo9er from the contactor =normally '.1'
gal2 ($CF %ith a standard mist eliminator> pls 9apori6ation from the contactor and
regenerator; and spillage. Glycol losses; e@clsi9e of spillage; range from '.'5
gal2($CF for high pressre; lo% temperatre gases to as mch as '.3' gal2($CF
for lo% pressre; high temperatre gases.
,here are se2eral -ays to reduce glycol losses:
7 certain amont of glycol al%ays 9apori6es in the sales gas stream.
7de*ate cooling of the lean glycol -efore it enters the a-sor-er minimi6es
these losses.
.ormally; most of the glycol entrainment is remo9ed -y a mist eliminator in
the top of the a-sor-er. )@cessi9e gas 9elocities and glycol foaming in the
a-sor-er sharply increase the glycol carryo9er. 7 do%nstream gas scr--er
can pay for itself *ic<ly and sa9e mch money -y trapping the carryo9er
and reco9ering the e@cess glycol. #his gas scr--er also helps pre9ent
pro-lems do%nstream of the glycol plant.
3apori6ation losses in the stripper can -e held to a minimm %ith good
glycol condensation and control of the to%er top temperatre. Glycol
entrainment; or mechanical carryo9er; can -e redced %ith proper
maintenance of the stripper and re-oiler.
(echanical lea<s can -e redced -y <eeping the pmp; 9al9es; and other
fittings in good order. #he glycol from these lea<s shold -e collected and
reprocessed.
)@cessi9e entrainment losses may -e the reslt of foaming in the a-sor-er
and2or regenerator. Defoamers are sometimes sed.
Gl!col -iltration
Filters e@tend the life of the glycol pmps; and pre9ent an accmlation of solids in
the a-sor-er and regeneration e*ipment. $olids that settle on metal srfaces
fre*ently set p cell corrosion. Filters also remo9e the solids that contri-te to
foling; foaming; and plgging. $oc<"type filters are preferred; althogh fine screens
and cartridge filters are also sita-le. #he filters shold -e designed to remo9e all
solid particles o9er 5 microns in si6e.

54

#hey shold -e a-le to operate p to pressre drops of &' to &5 psi. For -est reslts;
filters shold -e placed in the rich glycol line; -t the lean glycol can also -e filtered
to help <eep the glycol clean. Fre*ent filter changes may -e needed dring plant
start"p; or %hen netrali6ers are added to control the glycol pH.

55

Solid esiccant eh!dration
7a(1#round
$ince solid desiccant nits cost more than glycol nits; their se is sally limited to
applications sch as 9ery sor gases; 9ery lo% %ater de% point re*irements;
simltaneos control of %ater and hydrocar-on de% points; and special cases sch
as o@ygen containing gases; etc. ,n cryogenic plants; solid desiccant dehydration
sally is preferred o9er methanol in/ection to pre9ent hydrate and ice formation.
$olid desiccants are also often sed for the drying and s%eetening of .GJ li*ids.
Desiccants in common commercial se fall into one of three categories:
7lmina " !egenera-le alminm o@ide -ase desiccant.
$ilica Gel " !egenera-le silicon o@ide adsor-ent.
(oleclar $ie9es " !egenera-le solid desiccants composed of crystalline
metal alminosilicates =6eolites>.
)ach desiccant category offers ad9antages in different ser9ices. #he -est choice is
not rotine.
7cti9ated almina has a strong affinity for %ater and high internal adsorption area
de to the presence of pores or 9ery fine capillaries. 7lmina condenses and holds
the %ater in the pores -y srface adsorption and capillary attraction. 7cti9ated
almina desiccant can -e sed for drying li*ids %hich do not contain nsatrates
sch as olefins or diolefins. ,t is less costly than moleclar sie9e desiccant -t its
capacity for a-sor-ing %ater also tends to -e lo%er; particlarly %hen attempting to
reach 9ery lo% %ater le9els; e.g. 5 %ppm in the prodct.
$ilica gel has a higher e*ili-rim adsorption capacity =see Figre 0> than almina
-ecase its a9aila-le srface is greater. De to silica gelVs higher price per pond;
almina is generally the economic choice. $ilica gel is not sed %here free %ater can
-e present; -ecase free %ater destroys silica gel. Free %ater o9er long"term
operation; either as droplets or slgs; %ill also damage moleclar sie9e and acti9ated
almina -y mechanical attrition and shold -e a9oided.
(oleclar sie9es ha9e the featre of niform pore si6e; %hich allo%s them to
e@clde molecles -ased on si6e. ?ecase different pore si6e moleclar sie9es are
prodced; selection of proper type of sie9e can alle9iate the pro-lem of ndesira-le
coadsorption.

50

(oleclar sie9es ha9e a higher design adsorption capacity than the other regenera-le
desiccants; -t this is often offset -y their considera-ly higher price per pond.
(oleclar sie9e dehydrators are commonly sed ahead of .GJ reco9ery plants
%here e@tremely dry gas is re*ired. Cryogenic .GJ plants designed to reco9er
ethane prodce 9ery cold temperatres and re*ire 9ery dry feed gas to pre9ent
formation of hydrates.
Dehydration to appro@imately 1 ppm% is possi-le %ith moleclar sie9es.
#%o types of moleclar sie9es; #ype 37 and #ype 47; are commonly sed for
drying hydrocar-on li*ids. #ype 47 sie9es are less costly than #ype 37 sie9es and
are sed for distillates %hich do not contain nsatrates. When nsatrates are
present in the feed; #ype 37 are sed to assre good regeneration.
$olid desiccants are sed in gas dehydrators containing t%o or more to%ers. Figre 7
is a simple t%o"to%er system. 5ne to%er is onstream adsor-ing %ater from the gas;
%hile the other to%er is -eing regenerated and cooled. Figre 1 sho%s a typical
moleclar sie9e dehydrator 9essel. Hot gas remo9es the adsor-ed %ater; after %hich
the to%er is cooled. #he to%ers are s%itched -efore the onstream to%er -ecomes
%ater satrated. Generally a -ed is designed to -e on line for 1 to &4 hors. When
the -ed is ta<en off"line; the %ater is remo9ed -y heating the -ed to 45'"0''EF. #he
regeneration gas sed to heat the -ed is sally a slipstream of dry process gas. #he
regeneration gas is retrned to the process after it has -een cooled and the free %ater
remo9ed. $ince heat is a ma/or operating cost; this is a ma/or design consideration.

57

FG13% 3) : S"LD D%SCC6;, D%H<D36,"3 ,W" ,"W%3 S<S,%+


51



FG13% !*: ,<0C6L +"L%C1L63 S%?% G6S D%H<D36,"; ?%SS%L
'dsor.tion Calculations
7dsorption calclations for a moleclar sie9e dehydrator are discssed -elo%. #he
allo%a-le sperficial 9apor 9elocity throgh the -ed is the first parameter that mst
-e estimated sing Figre 41. #o se this plot; locate the operating pressre; read p
to the type sie9e; then read across to find the allo%a-le sperficial 9elocity. 5nce the
allo%a-le sperficial 9elocity is estimated; the -ed diameter can -e calclated for a
design 9apor rate. #he design pressre drop throgh the -ed is calclated sing
)*ation 3 and shold -e a-ot fi9e psi. 7 design pressre drop higher than eight psi
is not recommended.

54

8 (oleclar sie9e pressre drop.
G ?W3 K CX 3
&
=)*n. 3>
%here:
DP G Pressre drop; psi.
J G Jength of pac<ed -ed; ft.
3 G 3apor 9elocity; ft2min.
W G 3apor 9iscosity; cP.
X G 3apor density; l-2ft3.
Constants:
Desiccant #ype ? C
121 in. -ead '.'50' '.''''114
121 in. e@trdate '.'7&& '.'''1&4
1210 in. -ead '.15& '.'''130
1210 in. e@trdate '.&31 '.'''&1'
FG13% !& 6LL"W6>L% ?%L"C,< F"3 +"L% S%?% D%H<D36,"3
#he ne@t step is to choose a cycle time and calclate the ponds of sie9e re*ired.
)ight to t%el9e hor cycles are common. Cycles greater than 1& hors may -e
/stified; especially if the gas is not %ater satrated. Jong cycles mean fe%er
regenerations and longer sie9e life; -t larger -eds and additional capital in9estment
are re*ired.

0'
DP
J



Dring the adsorption cycle; the -ed operates %ith three 6ones #he top 6one is called
the satration 6one. #he moleclar sie9e in this 6one is in e*ili-rim %ith the %et
inlet gas. #he middle or mass transfer 6one =(#S> is %here the %ater content of the
gas is redced from satration to Y 1 ppm. .ormally a system is designed so that
there is a moistre analy6er to indicate %hen the mass transfer 6one is li<ely to -rea<
throgh the end of the -ed. 7 gard -ed 6one =typically one to t%o feet deep> is
pro9ided after this point to pre9ent actal -rea<throgh -efore the system has a
chance to change to the regenerated -ed.
+nfortnately; -oth the %ater capacity and the rate at %hich the moleclar sie9es
adsor- %ater change as the moleclar sie9es age. #he o-/ect of the design is to
install enogh sie9e so that three to fi9e years into the life of the sie9e; the mass
transfer 6one %ill -e at the -ottom of the -ed at the end of the adsorption cycle.
,n the satration 6one; the moleclar sie9e is e@pected to hold appro@imately 13
ponds of %ater per 1'' ponds of sie9e. #his capacity needs to -e ad/sted %hen
the gas is not %ater satrated or %hen the temperatre is a-o9e 75EF. $ee Figres 4&
and 43 for the correction factors. #o determine the ponds of moleclar sie9e
re*ired in the satration 6one; calclate the amont of %ater to -e remo9ed dring
the cycle and di9ide -y the sie9e capacity =se )*ations 4 and 5>.

01
)*. 4
)*. 5

$ee fig 43
$ee fig 4&

)9en thogh the (#S %ill contain some %ater; the satration 6one is calclated
assming it %ill contain all the %ater to -e remo9ed. #he length of the mass transfer
6one can -e calclated sing )*ation 7 from Wor< 7id 4. #he total -ed height is
the smmation of the satration 6one; mass transfer 6one; and gard -ed 6one
heights. 7ppro@imately si@ feet free space a-o9e and -elo% the -ed is needed.
FG13% !2 +"L% S%?% C606C,< C"33%C,"; F"3
1;S6,136,%D ;L%, G6S
FG13% !3 +"L% S%?% C606C,< C"33%C,"; F"3
,%+0%36,13%

0&

Pro(e$$ F"ow and t/e !un(tion o! t/e a3or
(oponent$ o! So"id De$i((ant De/'drator$
+nli<e glycol dehydration systemOs continos processing; the adsorption process is
a -atch procedre %ith mltiple desiccant -eds sed in cyclic operation to dry the
gas on a continos -asis. 7dsor-er to%ers mst -e ta<en ot of drying ser9ice to -e
regenerated. ?ecase of this; solid desiccant dehydrators typically se t%o; three; or
for adsor-er to%ers. $ome applications se as many as t%el9e to%ers. #here are
economic and process ad9antages for three or more adsor-er to%er designs in some
sitations.
#he follo%ing three separate fnctions mst alternate in each adsor-er to%er:
7n adsor-ing or gas"drying cycle
7 heating or regenerating cycle
7 cooling cycle to prepare the regenerated -ed for another adsor-ing cycle
7t any gi9en time; one of the to%ers is on stream in the adsor-ing or drying cycle
%hile the other is in the process of -eing heated or cooled. $e9eral atomatically
operated s%itching 9al9es and a controller rote the inlet gas and regeneration gas to
the right to%er at the proper time. #ypically; a to%er is on the drying or adsor-ing
cycle for 4 to 1& hors; %ith 1 hr the sal time. #he to%er -eing regenerated is
heated for a-ot 5 to 0 hors and cooled dring the remaining & to 3 hors.
(rocess of #egeneration
7s the %et inlet gas flo%s do%n%ard throgh the to%er on the adsorption cycle;
each of the adsor-a-le components is adsor-ed at a different rate. #he %ater 9apor is
immediately adsor-ed in the top layers of the desiccant -ed. )ach of the light
hydrocar-on gases and hea9ier hydrocar-ons mo9ing do%n throgh the -ed are also
adsor-ed to a greater or lesser e@tent. With moleclar sie9es; minimal adsorption of
hydrocar-ons occrs de to the e@clsion from the smaller pores. Hea9ier
hydrocar-ons displace the lighter ones in the desiccant -ed as the adsor-ing cycle
proceeds. 7s the pper layers of desiccant -ecome satrated %ith %ater; %ater in the
%et gas stream -egins displacing the pre9iosly adsor-ed hydrocar-ons in the lo%er
layers.
For each component in the inlet gas stream; there %ill -e a section of -ed depth;
from top to -ottom; %here the desiccant is satrated %ith that component and %here
the desiccant -elo% is /st starting to adsor- it. #he depth of -ed from satration to
initial adsorption is the mass transfer 6one =(#S> descri-ed earlier.

03

7s the flo% of gas contines; the (#SVs mo9e do%n%ard throgh the -ed and %ater
displaces all of the pre9iosly adsor-ed gas ntil; finally; the entire -ed is satrated
%ith %ater 9apor. When the -ed is completely satrated %ith %ater 9apor; the otlet
gas is /st as %et as the inlet gas. ?efore the desiccant -ed has -ecome completely
satrated; the to%ers mst -e s%itched from the adsorption cycle to the regeneration
cycle =$ee Figre 44>.
5ne regeneration"gas"spply scheme consists of ta<ing a pressre drop across a
pressreredcing 9al9e that forces a portion =5A to 15A> of the entering %et"gas
stream throgh the regeneration system. ,n most plants; a flo% controller reglates
the 9olme of regeneration gas sed. #his gas is heated ntil it reaches 4''E to
0''EF; then it is piped to the to%er -eing regenerated. #he adsor-ed %ater -egins to
desor- at the start of the regeneration cycle if dry regeneration gas is sed. #he -l<
of the %ater is remo9ed from the moleclar sie9e at a temperatre of &4'EF to
&5'EF; -t the desorption contines; althogh at a continally diminishing rate; ntil
the end of the heating cycle.

7fter all the %ater has -een remo9ed; heating is maintained to dri9e off any hea9ier
hydrocar-ons and contaminants that do not 9apori6e at lo%er temperatres. #he
desiccant -ed is properly regenerated %hen the otlet gas =pea<"ot> temperatre has
reached 35'E to 55'EF.
7fter the heating cycle; the desiccant -ed is cooled -y flo%ing nheated
regeneration gas ntil the desiccant is sfficiently cooled.
7ll of the regeneration gas sed in the heating and cooling cycles is passed throgh
a heat e@changer %here it is cooled to condense the %ater remo9ed from the
regenerated desiccant -ed. #his %ater is separated in the regeneration gas separator;
and the gas is mi@ed %ith the incoming %et"gas stream. #his entire process is
continos and atomatic.
#he follo%ing sections co9er the process flo% of dehydrators sing t%o; three; and
for adsor-er to%ers.
(rocess -lo/ of Solid esiccant eh!drators
,-o-,o-er Dehydrators
$olid desiccant dehydrators sing t%o adsor-er to%ers offer only one flo%
arrangement. While one to%er dehydrates the process flid stream; the other to%er
regenerates. #he t%o to%ers alternate -et%een drying and regeneration. Figre 11

04

sho%s the process flo% of the process flid =gas or li*id> and the regeneration gas
in a t%o"to%er solid desiccant dehydrator.
FG13% !!: 03"C%SS FL"W "F ,W"-,"W%3 S"LD D%SCC6;, D%H<D36,"3
,n the dehydrator sho%n in Figre 44; the process flid passes throgh the inlet
separator. 7s in glycol dehydrators; the inlet separator remo9es free %ater and other
contaminants from the process flid. From the inlet separator; the process flid
flo%s to and do%n throgh %hiche9er adsor-er to%er is in the adsor-ing phase =,n
Figre 44; this is #o%er &.>. ,n the to%er; the solid desiccant adsor-s %ater from the
process flid. From the to%er; the dried process flid flo%s to the otlet filter and
ot of the dehydrator.
While #o%er & dries the process flid; the regeneration gas remo9es the %ater from
the solid desiccant in #o%er 1. #he regeneration gas flo%s throgh the regeneration
gas heater; p throgh #o%er 1; to the regeneration gas cooler; and to the
regeneration gas separator. 5nce the heated regeneration gas has dried the solid
desiccant; the regeneration gas -ypasses the regeneration gas heater and cools the
desiccant.

05

Gases used for regeneration include:
Dry prodct gas
Gas di9erted from the process gas stream =5A to 1'A of the main flo%>
Gases from otside the dehydrator; sch as demethani6er o9erhead
,n this e@ample; after the adsor-ent in #o%er & is satrated; 9al9es redirect the
process flid and the regeneration gas so that #o%er 1 adsor-s and #o%er &
regenerates.
,hree-,o-er Dehydrators
$olid desiccant dehydrators %ith three adsor-ing to%ers offer se9eral process flo%
9ariations.
#he process cycle of three to%ers may -e arranged so that t%o to%ers dehydrate in
parallel %hile one to%er regenerates. 5r; the process cycle may -e arranged so that
one to%er is adsor-ing; one to%er is regenerating; and one to%er is cooling.
0arallel 4S/lit-Flo- Design5 " Figre 45 sho%s a three"to%er dehydrator %ith t%o
to%ers adsor-ing in parallel. #he process flo% of the dehydrator sho%n in Figre 45
is similar to the flo% sho%n in Figre 44; e@cept that the flo% of the process flid is
split -et%een t%o to%ers.
7 three"-ed system =t%o -eds in parallel> may -e economical for a large installation
since there are three smaller -eds rather than t%o large -eds. ,f three -eds =9essels>
are less e@pensi9e; the 9essel sa9ings is then %eighed against the additional piping
and 9al9es re*ired for a three"-ed system. ,f economic considerations are
inconclsi9e; process ad9antages or past e@periences may dictate this three"to%er
system. For e@ample; this configration can -e operated continosly %ith only t%o
to%ers at redced throghpt; facilitating desiccant replacement on the rn.

00


FG13% !#: 03"C%SS FL"W "F ,H3%%-,"W%3 S"LD D%SCC6;, D%H<D36,"3

F-&
F-2
H"172?2C
C-!
6@>
6-&
D-&C D-&> D-&6
?-&
C

D
C

D
Dry Gas
To
Cooling
Train
Feed
Gas
07

FG13% !$ 03"C%SS FL"W "F ,H3%%-,"W%3 S"LD D%SCC6;, D%H<D36,"3
Function of +a9or Co./onents of Solid Desiccant Dehydrators
#his section descri-es the fnction; design; and operation of the ma/or components
of a solid desiccant dehydrator.
#he follo%ing are the essential components of any solid desiccant dehydration
system:
7n inlet gas stream separator2li*id coalescer.
#%o or more adsorption to%ers =contactors> filled %ith solid desiccant.
7 high"temperatre heater that pro9ides hot regeneration gas to reacti9ate the
desiccant in the to%ers.
7 regeneration gas cooler that condenses %ater from the hot regeneration gas.
7 regeneration gas separator =<noc<ot> that remo9es the condensed %ater
from the regeneration gas.
Piping manifolds; s%itching 9al9es; and controls that direct and control the
flo% of gases according to the process re*irements.
nlet Se/arator@Coalescer

F-&6 F-26
H"1D2)
C-!
C@D
D-&F D-&% D-&D
Dry Gas To
Cooling
Train
?&(
D-&
G
C

D
C

D
6-&
?-&
C

D
Feed
Gas
01

7s %ith glycol dehydrators; inlet separators protect the dehydrator from imprities
sch as free %ater; salt; compressor oils; hydrocar-on li*ids; paraffins; corrosion
inhi-itors; glycol; amines; rst; iron slfide; iron o@ide; fractionation sands; drilling
md; pipeline scale; and slfr. #hese imprities impact the desiccant -ed and case
-rea<ing and po%dering of the desiccant. (ethanol =sed for hydrate inhi-ition> can
also damage some adsor-ents.
.on9olatile li*ids coat the desiccant and -loc< its pores. $olid imprities plg the
-ed increasing the pressre drop and crshing the desiccant. 7ll of these effects
shorten the operating life of the desiccant. ,f the dehydration nit is do%nstream of
an amine nit; glycol nit; or compressors; a filter"separator or li*id coalescer =for
li*id ser9ice> may -e needed.
6dsorber ,o-er
#he adsor-er to%er holds the solid desiccant and contacts it %ith the process flid.
Figre 47 sho%s a typical adsor-er to%er.

04

FG13% !': 6DS"3>%3 ,"W%3
#hree pro-lems that fre*ently case poor operation are insfficient distri-tion;
inade*ate inslation; and improper -ed spports.
Distribution " Poor gas distri-tion at the inlet and otlet of the desiccant -eds can
case gas channeling and desiccant damage. #he inlet gas distri-tor shold -e

16 " Fill
hole
Supporting
screen
1/8 "
ceramic
ball
1/4 "
ceramic
ball
1/16 " mol. sieve
(h!"#
4 "
4 "
$ "
%.& .
7'

pro9ided %ith ade*ate -affling -efore the gas enters the desiccant -ed. .either the
gas to -e dehydrated nor the regeneration gas shold impinge directly on the -ed nor
shold there -e any sdden changes of direction near the srface of a desiccant -ed.
When the -ottom head of the 9essel is filled %ith spport -alls; a gas distri-tor may
-e needed -et%een the -alls and the lo%er portion of the desiccant -ed %hen
pflo%ing heating or cooling is sed. #his distri-tor shold -e sed particlarly on
9ery large diameter 9essels to pre9ent gas channeling and poor reacti9ation of the
desiccant. Channeling; high locali6ed 9elocities; and s%irling can case desiccant
dsting =-rea<age> and high pressre drop throgh the desiccant -ed. #he lodging of
the desiccant dst -et%een the standard"si6e particles increases the pressre drop.

,nert -alls are also re*ired on top of the adsor-ent -ed to pre9ent -ed mo9ement
reslting from high flid 9elocities. #he se of a 0I layer of these inert -alls is
recommended. 7 floating screen shold -e placed -et%een the inert -alls and the
adsor-ent to pre9ent the migration of inert -alls to the interior of the -ed. #his
migration is cased -y the cyclic natre of the adsorption process.
nsulation " ,nternal or e@ternal inslation for the adsor-er may -e sed. #he main
prpose of internal inslation is to redce the total regeneration"gas re*irements
and costs. ,nternal inslation redces heat transfer to the adsor-er 9essels. .ormally;
a casta-le refractory lining is sed for internal inslation.
#he refractory mst -e applied and properly cred to pre9ent liner crac<s. Pro9ision
mst also -e made for e@pansion and contraction of the internal inslation so that
there %ill -e no crac<ing or %eld failres. Jiner crac<s permit some of the %et gas
to -ypass the desiccant -ed. 5nly a small amont of %et; -ypassed gas can case
free6eps in cryogenic plants.
Jedges installed e9ery fe% feet along the 9essels %all can help eliminate -ypassing.
For heating cycles longer than t%o hors; pro9ision shold -e made for heat
transferred to the 9essel shell and heads.
>ed Su//orts " #he -ed spport can -e in the form of a mechanical grid; sch as
s-%ay type grating on ,"-eams ; #he grid mst -e of sfficient strength to spport
the adsor-ent -ed; inert -alls; and the -ed pressre drop. 7 set of %ire screens; si6ed
to retain the adsor-ent particles; shold -e fastened on top of the grid. #he screens
=sally stainless steel> are fastened in place -y tying them to the spport
grid%ithstainlesssteel %ire. #o pre9ent the loss of desiccant and spport -alls
throgh the gap -et%een the 9essel and the %ire screen; this gap is sealed sing
some sita-le rope"li<e material =typically as-estos> that can %ithstand the
regeneration temperatres and the process2regeneration gas flid.
$e9eral layers of inert spport -alls are re*ired -et%een the adsor-ent and the %ire
screens to pre9ent grinding -et%een them. #hese layers also pre9ent desiccant dst

71

or %hole particles from plgging the screen openings. .ormally; for 121"in.
desiccant; a &"3 inch layer of 12&"in. -alls is gently placed on the screen follo%ed -y
a smooth &"3 inch layer of 124"in. -alls. For 1210"in. desiccant =commonly sed for
li*id dehydration>; an additional layer of 121"in. -alls are placed on top of the 124"
in. -alls. #his complies %ith the &:1 ratio -et%een the layers as recommended -y the
desiccant 9endors.
3egeneration Gas Heater
#he regeneration gas heater heats the regeneration gas to a-ot 5''EF. $olid
desiccant dehydrators se many types of heaters inclding salt -ath; direct fired; hot
oil; and steam.
$mall nits =1 ((?t2hr> generally se indirect"fired; salt -ath heaters for safety
reasons.
Jarger nits tend to se direct"fired heaters. ,n addition; other sorces of heat are
sed inclding compressor"e@hast gases and %aste heat from tr-ines and other
heat sorces.


3egeneration Gas Cooler
!egeneration gas coolers redce the temperatre of the regeneration gas to condense
the adsor-ed %ater and; sometimes; hydrocar-ons. Cooling the regeneration gas also
prepares it for frther processing.
Coolers are heat e@changers that se air; %ater; or natral gas to cool the
regeneration gas.
#ypically; they se am-ient air to cool the regeneration gas to %ithin 15EF to &'EF of
the air temperatre.
3egeneration Gas Se/arator
!egeneration gas separators remo9e li*ids condensed -y the regeneration gas
cooler from the regeneration gas. ,f the li*id is primarily %ater; then a t%o"phase
separator; similar =e@cept smaller>; to the inlet separator is sed. ,f the li*id
contains s-stantial amonts of hydrocar-ons; then the dehydrator re*ires a three"
phase separator to remo9e the li*id from the gas stream and separate the li*id into
%ater and hydrocar-ons.
S-itching ?al2es
$%itching 9al9es direct the process flid and regeneration gas to the appropriate
component of the dehydrator. #%o"%ay 9al9es lea< less than three"%ay 9al9es.
$%itching 9al9es are in a harsh operating ser9ice as they mst operate %ith all

7&

com-inations of cold gas and hot gas on either side of the 9al9e. #his temperatre
cycling can case 9al9es to stic< and2or lea<.
3al9e se*encing and opening times are also important to pre9ent a sdden pflo%
of gas %hich cold flidi6e the -ed and damage the desiccant. $%itching 9al9es
re*ire fre*ent ser9icing to eliminate lea<age.

nstru.entation
#he monitoring of solid desiccant dehydrators re*ires a 9ariety of instrmentation
to measre or control the follo%ing process 9aria-les:
Flo% rate; temperatre; and pressre of the process flid
Water content of the prodct flid
Flo% rate and pressre of the regeneration gas
,nlet and otlet temperatres of the regeneration gas
#emperatre of the regeneration gas lea9ing the regeneration gas cooler
7dsor-ent differential -ed pressre
Cycle time controllers
(easring the flo% rate; temperatre; and pressre of the inlet gas monitors the load
on the dehydrator. (onitoring the temperatre and %ater content of the otlet gas
measres the performance of the dehydrator. (easring the conditions of the
regeneration gas and the regeneration cooler monitors the efficiency of the
regeneration cycle.
7 moistre sample pro-e shold -e located in the adsor-ers in cryogenic plants
se9eral feet from the otlet end of the -ed and e@tend to the center. #his pro-e; sed
in con/nction %ith the otlet"gas moistre pro-e; offers 9ala-le fle@i-ility in
stdying and sol9ing dehydrator pro-lems. ,f gas is channeling throgh the desiccant
-ed a high de% point is seen in the otlet gas. Ho%e9er; nless the gas is channeling
throgh the -ed at the e@act point %here the sample pro-e is located; the desiccant
srronding the pro-e %ill ensre that the pro-e %ill contine to BseeC only dry gas.
#he pro-e also permits capacity tests for optimi6ing drying cycle times. #hese tests
can -e condcted %ith reasona-le safety -ecase mo9ement of the %ater front can
-e detected prior to -rea<throgh.
#emperatre recording de9ices plot the temperatre of the regeneration gas entering
and e@iting the adsor-er to%er against the elapsed time of the cycles.
,t is important to monitor -ed differential pressre. 7n increase in differential
pressre can indicate desiccant pro-lems sch as e@cessi9e plgging or the
formation of fines.

73

"ffects of 0e! (rocess )aria*les on the o.eration of a Solid
esiccant eh!drator
#his section co9ers the follo%ing <ey process 9aria-les that affect the performance
and operation of solid desiccant dehydrators:
Direction of process flid and regeneration gas =do%nflo% 9s. pflo%>
Tality of inlet process flid =le9el of contamination; inclding free %ater>
#emperatre of process flid and regeneration gas
Pressre of process gas and regeneration gas
7dsorption; heating and cooling cycle times
3elocity of process flid and regeneration gas
!egeneration gas sorce =%et or dry>
$pecial li*id ser9ice considerations
Direction of Gas Flo-
#he follo%ing section discsses the effects of the direction of flo% of the process
flid on dehydrator performance dring the drying; regeneration; and cooling cycles.
#he flo% direction inflences efflent prity; regeneration gas re*irements; and
desiccant life.
Drying Cycle
.ormally; gas flo%s do%n throgh adsor-er to%ers dring the drying cycle.
,ncreasing the rate of p%ard flo% of gas throgh a to%er flidi6es the adsor-ent
-ed. 7ny sch nsettling or mo9ement can erode and2or crac< the desiccant %hich
decreases the drying performance and life of the desiccant; and generates desiccant
fines. 7lso; do%n%ard flo% allo%s higher 9elocities; %hich allo% smaller and less
e@pensi9e to%ers.
Con9ersely; hydrocar-on li*ids normally flo% p throgh adsor-er to%ers.
Hydrocar-on li*id streams fre*ently contain some gaseos components. +p%ard
li*id flo% allo%s any gas ---les to pass throgh the desiccant -ed. Do%n%ard
li*id flo% allo%s these gaseos components to accmlate at the top of the to%er.
#his accmlation redces the amont of desiccant e@posed to the hydrocar-on
li*id decreasing the effecti9e capacity of the to%er.
$ometimes do%n%ard li*id flo% is also sed to a9oid early -rea<throgh of %ater.
#his is particlarly tre %hen the %ater content of the li*id is high. Choice of
pflo% 9erss do%nflo% is also dictated -y the flo% magnitde of the streams to -e
dried.


74

3egeneration 4Heating5 Cycle
.ormally; regeneration gas flo%s in the direction opposite of the process flid in the
drying cycle. For drying hydrocar-on gas; this direction is p throgh the adsor-er
to%er. ,f the regeneration gas flo%s in the same direction as the process flid; then
regeneration gas mst displace the %ater and contaminants concentrated at the top of
the -ed do%n throgh the entire -ed. $ame direction flo% ris<s the contamination of
the rest of the adsor-ent -ed and can re*ire longer regeneration times. #ypical
regeneration flo% direction for li*id dehydrators is do%n%ard.
With regeneration gas flo%ing in the direction opposite of the process flid in the
drying cycle; the hot regeneration gas in the lo%er part =pper part for hydrocar-on
li*ids> of the -ed strips contaminants from the desiccant. ,f the hot regeneration gas
sfficiently increases the partial pressre of the contaminants; the contaminants %ill
desor- off the desiccant. #his flo% direction also prodces e@tremely dry adsor-ent
at the -ottom of the adsor-er to%er.
Dring the dehydrating cycle; this dry adsor-ent remo9es the last amonts of %ater
from the process flid and prodces efflent %ith 9ery lo% %ater contents. ,f the
contaminants do not desor- off the desiccant; they %ill -ild p and potentially co<e
or polymeri6e.
Cooling Cycle
,f a system ses %et gas for cooling; then the cooling gas flo%s in the same direction
as the process flid. #his partially preloads the desiccant -ed %ith %ater. #he
additional %ater load deposited dring cooling mst -e inclded %hen the amont of
desiccant re*ired is calclated.
,f a system ses dry prodct gas for cooling; then the cooling gas flo%s in the
direction opposite of the process flid. #his direction of flo% reslts in more
complete desiccant regeneration.

=uality of nlet Gas 40rocess Fluid5
#he relati9e satration of the inlet gas is the most important 9aria-le in determining
the %eight of desiccant to se. #he se of %et gas to regenerate a desiccant -ed does
not remo9e all of the %ater from the -ed. #herefore; the se of %et gas re*ires the
determination of the residal capacity of the desiccant -ed. #he residal capacity of
the desiccant -ed is the difference -et%een the dynamic capacity of the desiccant
-ed and the %eight percent of %ater left in the desiccant -ed after the %et gas has
regenerated it. #his 9aria-le is the dri9ing force that affects the transfer of %ater to
the adsor-ent. ,f satrated gas =1''A relati9e hmidity> is -eing dried; a mch
greater sefl capacity can -e e@pected for most desiccants than %hen partiallly

75

satrated gases are -eing dried. #he e@ception is the moleclar sie9e. ,ts e*ili-rim
cr9e is almost flat from &'A to 1''A relati9e hmidity.
Conta.inants
#he most important 9aria-le affecting the decline rate of desiccant capacity is the
chemical composition of the gas or li*id to -e dried. Compressor oils; corrosion
inhi-itors; glycols; amines; and other high"-oiling contaminants may -e present in
the feed gas. .ormal regeneration temperatres do not 9apori6e the hea9y materials.
#he residal contaminants slo%ly -ild p on the desiccantVs srface; redcing the
area a9aila-le for adsorption. (any corrosion inhi-itors chemically attac< certain
desiccants and permanently destroy their seflness. $ilica gels shatter %hen free
%ater or hydrocar-on li*ids are carried into the -eds.
(ethanol in the inlet gas is a ma/or contri-tor to the co<ing of moleclar sie9es
%here regeneration is carried ot at temperatres a-o9e 55'EF. Polymeri6ation of
methanol dring regeneration prodces intermediates %hich case co<ing of the
-eds. 7n )@@on affiliate that had -een sing s-stantial amonts of methanol to
inhi-it hydrate prodction in a t%o"phase system e@perienced rapid declines in sie9e
life -ecase of co<ing. Con9ersion to ethylene glycol in/ection for hydrate control
has greatly increased sie9e life and added at least 1'A to sie9e capacity.

,e./erature
#he higher the temperatre of a process flid; the greater its satrated %ater content.
$olid desiccants ha9e significantly higher adsor-ing capacity at lo%er temperatres.
!egeneration gas may -e com-ined %ith the incoming %et gas ahead of the
dehydrator. ,f the temperatre of these gases differs -y more than 15 to &'EF then
%ater and hydrocar-ons may condense ot of the %armer gas stream. High
regeneration"gas temperatres assre good desorption of %ater and contaminants.
!egeneration gas condenser temperatres shold -e as lo% as possi-le to condense
and remo9e %ater and hea9y hydrocar-ons from the gas.

,f a dehydrator ses %et gas to cool its desiccant; o9ercooling the desiccant cases
the solid desiccant to adsor- %ater from the cooling gas and presatrate the
adsor-ent -ed. #he cooling cycle shold -e ended %hen the desiccant -ed reaches
a-ot 1&5EF. ,f dry gas is sed for cooling the desiccant may -e cooled to %ithin 1'
to &'EF of the feed gas temperatre.
0ressure
#he adsorpti9e capacity of a desiccant -ed nit decreases as the pressre is lo%ered
%hen adsor-ents are operating -elo% satration capacity. 7lso; the %ater content of
a gas increases as the pressre is lo%ered. ,f the dehydrators are operated %ell -elo%
design pressre; the drying cycle time is shortened to conter the redced dri9ing

70

force and to remo9e the increased mass of %ater in order to maintain the desired
efflent de% point. #he same mass flo% rate of incoming gas at a redced pressre
increases gas 9elocity and increases the -ed pressre drop. )@cessi9e pressre drop
cases dsting =adsor-ent -rea<age> and damage to the desiccant. 7t pressres
a-o9e 1;3'' psia to 1;4'' psia; the coadsorption effects of hydrocar-ons sometimes
-ecome significant.
Cycle ,i.e
#he drying cycle time is the rnlength of the drier -efore it re*ires regeneration.
#he design drying cycle sets the drier %ater loading and; hence; drier si6e. $ince
desiccant capacity decreases %ith age =nm-er of regenerations> initial cycle times
are considera-ly longer than design cycle times. Design cycle times are sed to
esta-lish %hen desiccant replacement is necessary. #he design cycle time is
appro@imately e*al to the regeneration time at the design flo% rate. #herefore;
%hen less time is re*ired to satrate a desiccant than to regenerate it; either the
dessiccant mst -e replaced; or the flo% rate decreased.
#ypically; the adsorption cycle is operated on a fi@ed time. Fi@ed time cycles are
common as are dehydrator installations that s%itch -eds on %ater -rea<throgh
measrements. 7s the desiccant ages; the cycle time mst -e shortened to pre9ent
%ater -rea<throgh. (ore regenerations redce desiccant life; ths; a compromise
-et%een cycle time and cycle fre*ency mst -e achie9ed. #he cycle length shold
pro9ide an economic -alance -et%een cost of the drier and regeneration facilities;
and regeneration fre*ency; %hich affects desiccant life and opera-ility.

Jisted -elo% are the steps re*ired in the regeneration cycle and the appro@imate
time re*ired to perform these steps. $ome of these steps are ni*e to a particlar
ser9ice. #his information %ill aid the designer in esta-lishing a drying cycle; or if
the drying cycle; is already set; it %ill indicate the a9aila-le time for heating and
cooling.
7ppro@imate #imeZ
1. Depressre: 1 hor
&. Drain =for li*id drying only>: 1 hor
3. Cold Prge =not al%ays performed; discssed -elo%>: 1 hor
4. Heat to !egeneration #emperatre: 3aria-le
5. Heat $oa< =not al%ays performed; discssed -elo%>: 1"& hors
0. Cool to #emperatre 7pproaching Drying #emperatre: 3aria-le
7. Fill =for li*id drying only>2Pressre: 1 hor
Z Where atomated regeneration facilities are pro9ided for small si6e e*ipment
considera-ly shorter times may -e practical.

77

Gas ?elocities
nlet 40rocess5 Gas
Decreasing the gas 9elocity dring the drying cycle sally achie9es -oth lo%er
efflent moistre contents and longer drying"cycle times. Figre 11 sho%s the
general effect of gas rate on the e@tent of dehydration. (inimm flo% rates tili6e
the desiccant flly. Ho%e9er; lo% linear 9elocities re*ire to%ers %ith large cross"
sectional areas. ,n selecting the linear flo% rate; a compromise mst -e made
-et%een the to%er diameter and the ma@imm tili6ation of the desiccant. 7 high
linear flo% rate cases agitation of the granles; dsting and loss of capacity to
adsor-. ,n addition; flidi6ation can occr if the gas 9elocity =or li*id in the case of
li*id drying> p%ards throgh the -ed e@ceeds the flidi6ation 9elocity.
3egeneration Gas
Prodcing 9ery lo% efflent %ater contents =less than '.1 ppm> re*ires sfficiently
high regeneration gas 9elocities. Jo% gas 9elocities prodce channeling %hich
reslts in poor regeneration. Fre*ently; achie9ing 9ery lo% efflent %ater contents
re*ires regeneration gas 9elocities of at least 1' ft2min.
3egeneration Gas Source
#he sorce of gas for heating and cooling desiccant -eds depends on plant
re*irements and; possi-ly; on the a9aila-ility of a sita-le gas stream. +sing dry
regeneration gas prodces efflent %ith lo% %ater contents. +sing %et feed gas
reslts in moderate efflent %ater contents. Graphs plotting isoteres =lines of
constant %ater loading> can -e sed to predict the regeneration gas conditions
re*ired to achie9e a gi9en efflent %ater content.
#he effecti9eness of reacti9ation can also play a ma/or role in retarding the decline
of a desiccant adsorpti9e capacity and in prolonging its sefl life. .ot remo9ing all
of the %ater from the desiccant dring each regeneration sharply decreases its
seflness. For e@ample; if the dynamic adsorpti9e capacity of a thoroghly
reacti9ated desiccant is 1'A. 7 3A residal %ater remaining on the desiccant
-ecase of insfficient regeneration; %old case its capacity to drop from 1'A to
7A.
7lthogh gases rich in hea9ier hydrocar-ons may -e dried satisfactorily %ith
moleclar sie9es; the se of this same rich gas in a 5''E to 0''EF regeneration
ser9ice aggra9ates co<ing pro-lems. Jean dry gas is al%ays prefera-le for
regeneration.

71

O.timi1ing 'dsor.tion2%!.e eh!drators
De$i((ant Per!oran(e
5perating data shold -e monitored to try to pre9ent permanent damage to the
desiccant. Performance tests are fre*ently schedled on a rotine -asis; ranging
from monthly dring early operations; to si@ months or longer. #he si6e of the nit
and the *antity of the desiccant also affect the fre*ency of performance tests.
Desiccants decline in adsorpti9e capacity at different rates nder 9arying operating
conditions. (ar<edly different capacity"decline rates may -e e@perienced for the
same desiccant nder similar conditions of gas flo%; temperatre; pressre; %ater
remo9al re*irements; cycle times; and regeneration temperatres. Desiccant aging
is a fnction of many factors; inclding the nm-er of cycles e@perienced and
e@posre to any harmfl contaminants present in the inlet stream. (any of these
contaminants are not completely remo9ed dring normal reacti9ation. Contaminants
may -e the case of 4'A of nsatisfactory solid desiccant operations. #herefore; the
single most important 9aria-le affecting the decline rate of desiccant capacity is the
chemical composition of the gas or li*id to -e dried. Feed stream composition
shold al%ays inclde the contaminants.

#he capacity of a ne% desiccant %ill decline slo%ly dring the first fe% months in
ser9ice -ecase of cyclic heating; cooling; and %etting. Desiccant capacity sally
sta-ili6es at a-ot 55 to 7'A of the initial capacity. #o get ma@imm se ot of the
desiccant; a moistre analy6er can -e sed to optimi6e the drying cycle time. #hat
time can -e shortened as the desiccant ages. ?oth inlet and otlet moistre"analy6er
pro-es shold -e sed. (oistre analy6ers for 9ery lo% %ater contents re*ire care
to pre9ent damage to the pro-es. $ample pro-es and temperatre pro-es mst -e
installed to reach the center of the gas phase.
Proper conditioning of the inlet gas is important. Compressor oils; corrosion
inhi-itors; glycols; amines; and other high"-oiling contaminants present in the feed
gas case a frther decline in desiccant capacity; -ecase normal reacti9ation
temperatres %ill not 9apori6e the hea9y materials. #he residal contaminants
slo%ly -ild p on the desiccantVs srface; redcing the area a9aila-le for
adsorption. (any corrosion inhi-itors chemically attac< certain desiccants;
permanently destroying their seflness. 7 layer of less e@pensi9e desiccant can -e
installed on the top of the -ed to catch these contaminants.
7lthogh gases rich in hea9ier hydrocar-ons may -e dried satisfactorily %ith
moleclar sie9es; the se of this same rich gas in a 55' to 0''EF regeneration

74

ser9ice aggra9ates co<ing pro-lems. Jean dry gas is al%ays prefera-le for
regeneration; if it is a9aila-le.
(ethanol in the inlet gas is a ma/or contri-tor to the co<ing of moleclar sie9es
%here regeneration is carried ot at temperatres a-o9e 55'EF. Polymeri6ation of
methanol dring regeneration may prodce dimethyl ether and other intermediates
that %ill case co<ing of the -eds.
(onitoring -ed differential pressre is important. 7n increase in differential
pressre can indicate desiccant pro-lems sch as e@cessi9e co<ing or the formation
of fines. #he differential pressre along %ith the -ed rn length shold also -e
recorded %hen doing a performance test on a desiccant -ed.
#he sefl life of most desiccants ranges from one to for years in normal ser9ice. 7
longer life is possi-le if the feed gas is <ept clean. #he effecti9eness of reacti9ation
can also play a ma/or role in slo%ing the decline of a desiccantVs adsorpti9e capacity
and in prolonging its sefl life. 5-9iosly; if all the %ater is not remo9ed from the
desiccant dring each regeneration; its seflness %ill sharply decrease.


Performance data are sed for monitoring desiccant life and planning for desiccant
change ot. #he steps in9ol9ed are as follo%s:
Plot plant capacity and cycle time 9erss nm-er of cycles.
)@trapolate to determine %hen shortest cycle possi-le %ith e@isting
regeneration e*ipment %ill -e reached.
7t that point; or dring the nearest reglar plant trnarond preceding that
point; an adsor-ent change mst -e planned.
!egeneration gas not only spplies heat -t also acts as a carrier to remo9e %ater
9apor from the desiccant -ed. ,nsfficient reacti9ation can occr if the regeneration
gas temperatre or 9elocity is too lo%. #he desiccant manfactrer %ill generally
recommend the optimm regeneration temperatre and 9elocity for the prodct.
3elocity shold -e high enogh to remo9e the %ater and other contaminants *ic<ly.
#his measre %ill minimi6e the amont of residal %ater and protect the desiccant.
#o ma@imi6e desiccant capacity and to ensre the minimm efflent moistre
content; a higher reacti9ation temperatre or a drier reacti9ation gas; or -oth; may -e
needed. Higher reacti9ation temperatres may also -e sed to remo9e 9olatile
contaminants -efore they can form co<e on the desiccant. #he final efflent hot gas
temperatre shold -e held one or t%o hors to achie9e effecti9e desiccant
reacti9ation.

1'

Equipent Ite$
,n addition to the a-o9e process 9aria-les; engineers can optimi6e solid desiccant
dehydration e*ipment -y considering the follo%ing:
7n accrate estimation of -ed si6es in order to realistically e9alate
competiti9e -ids from desiccant 9endors.
5ptimal design of adsor-er internals =inlet gas distri-tor; internal inslation
and -ed spports>; s%itching 9al9es; and control systems.
Proper design of regeneration gas systems.
$ince mole sie9e can prodce dst; filters are fre*ently installed
do%nstream to protect s-se*ent e*ipment.


11