Bài tập chế biến khí

The wet TEG leaving the bottom then flows to a reboiler type stripping column regenerator where the water absorbed is vaporized overhead, the lean TEG produced out the bottom being reci

15 DEHYDRATION The removal of water vapor from natural gas is a necessary part of any processing scheme Normal line flowing conditions, whether in gathering systems or transmission iines, reach the hydrate forming temperature When wet gas is transported, hydrate formation and consequent line plugging will result Even in areas where line temperatures will not reach the hydrate range, temperature variations lead to water condensation with consequent loss of line capacity and efficiency

Water may be removed from natural gas by absorption or adsorption A number of hygroscopic li- quids are capable of removing water but commercially one of the glycols or calcium chloride is used The latter is used only in special cases where small amounts of gas are involved Of the glycols, triethylene | glycol (TEG) is used most commonly Thus, all discussions herein will be limited to TEG

Adsorption, or solid desiccant dehydration, uses a variety of solids with a large surface area The primary materials are activated bauxite, activated alumina, gels and molecular sieves (marketed under various trade names)

Chapters 18 and 19 of Volume 2 provide a detailed summary of the systems and the dehydration materials used therein A portion of that information is repeated here for convenience

TRIETHYLENE GLYCOL DEHYDRATION The TEG dehydrator is a classical example of the absorption-stripping process discussed in previous chapters Lean TEG enters the top of the absorber and contacts the gas for water removal The wet TEG leaving the bottom then flows to a reboiler type stripping column (regenerator) where the water absorbed

is vaporized overhead, the lean TEG produced out the bottom being recirculated back to the top of the ab- sorber

Consequenily, one can use standard absorption-stripping calculation methods to design these two very critical portions of the unit The Kremser-Brown method is applicable for both absorption and gas stripping {if used} If no gas stripping is used, a standard binary distillation caiculation may be used for the regenerator

Whatever the approach, one must be able to predict accurately equilibrium K values for water in a TEG-water system References 15-3 and 15-7 represent the published data in this area

One can ignore any hydrocarbon present in the system (along with TEG and water) for equilibrium calculations The solubility of gas in these fluids is very small; likewise hydrocarbon liquid is very insol- uble in TEG and water The correctness of this assumption has been borne out in test units

Equilibrium data from various sources ts more uniform than hydrocarbon equilibrium data in general This is understandable when one recognizes that the water concentration in both the liquid and gaseous phases is small Furthermore, the concentration does not vary over a wide range

To measure accurately the water concentrations in both gas and liquid is difficult, partially because

of the low concentrations involved Thus, different investigators will show variations in the results presented

15-1

The early work of Wise (15-6) and Townsend (75-5) provided the basis for early designs (along with scat- tered data from glycol vendors) Scauzillo (75-4) presented the first complete set of K values which were based largely on available fugacity data but were influenced by then available data on actual operating plants The equilibrium data from Worley (75-7) were based on operating unit performance extrapolated to

an infinite number of trays and TEG circulation rates The work of Rosman, the last work cited, was based

on experimental data which was compared to that of previous investigators

The data of Scauzillo, Worley and Rosman (although differing in various degrees) have all been used successfully for the design of TEG dehydration systems For the usual type, whose function is to prevent liquid water in a line, the source of data will have little effect on the final design However, in special cases involving very high gas temperatures and/or large dew point depression requirements the data source used may have a discernible effect on some aspects of the design, particularly the circulation rate required

There are some general statements that describe the general performance of TEG dehydration systems

1 The outlet water dew point from the absorber is essentially independent of pressure (at least to

2500 psia) It is affected primarily by effective absorber contact temperature and concentration of lean TEG entering the top of the absorber

2 The water equilibrium contact (K = y/x) is essentially independent of lean TEG concentration (for water concentrations up to about 3 wt %)

3 The relationship between water dew point and K value depends on the water content correlation used

4 An infinite combination of theoretical trays and circulation rate may be used to achieve desired performance However, economic circulation rates normally fall within the range of 3-7 U.S gallon TEG per pound of water absorbed from the gas

The basic design concerns are the absorber and the regenerator Once these have been fixed the necessary pumps, heat exchangers, filters and associated equipment may be sized and chosen in a rela- tively routine manner

Absorber

Figure 15.1 is the correlation from Chapter 18, Volume 2, based primarily on the work of Worley (15-7) The left hand ordinate is the equilibrium water dew point of the dry exit gas that would be produced by an absorber containing infinite trays (or obtained in an equilibrium cell in the laboratory) The effective con- tactor temperature (on the abscissa) is normally taken as the temperature of the inlet wet gas Where the absorber pressure is 300 psia or greater, the temperature change throughout the absorber is less than 5°F, so that it is in effect a constant temperature device

The diagonal lines represent the effect of lean TEG concentration on equilibrium water dew point at

a given temperature The dashed line represents the concentration of lean TEG normally produced by a regenerator operating at 400°F and normal (sea level) atmospheric pressure At concentrations greater than 98.5 - 98.7 wt % TEG gas stripping and/or vacuum is required Table 15.1 below shows the effect of regenerator pressure on FEG concentration

Table 15.1

Abs Pressure of Still Column TEG — % by weight (T=400°F)

roduc

4009F,

70

a"

„

r

„+

⁄ | | LT Eee

“YO 40 x > a> Tế HS AS

70 <> Ie"

|

+1 Effective Contactor Temperature, OF

a Given Lean TEG Concentration and Effective Contactor Temperature (Ref 18.1,

15-3

Courtesy BS&B)

The equilibrium dew point temperature shown can be converted to a moi fraction of water in the vapor by a water content correlation of the type shown in Chapter 14 If the type of correlation illustrated

by Figure 14.7 is used, this dew point is entered on the abscissa and one reads water content in IbD/MM scf This can be converted to y°, the mol fraction of water in the vapor (at saturation) over liquid water containing no TEG

yÐ= (1b/MMscf)/ (47,484)

The glycol-water-natural gas system is a non-ideal one in the liquid phase This requires the use of activity coefficients for prediction of equilibrium vapor-liquid relationships The equilibrium relationship for water distribution between the gas phase and the glycol phase can be expressed as:

K = (y,) 0y”)

Where:

K = equilibrium vaporization coefficient

y° = mole fraction of water in the vapor phase at saturation

Y = activity coefficient for water-triethylene glycol system

Values of the activity coefficient for use in Equation 15.1 can be obtained from Figure 15.2

Table 15.2 Physical Properties of Glycols

Boiling Point at

Boiling Point at

Vapor Pressure at

Density at 25°C, g/cm? 1.110 1.113 1.119

Pounds Per Gallon

Viscosity in

Surface Tension at 25°C

Refractive Index at

Specific Heat at

(15.1)

.75

.70

.65

* an o

* la

Coefficient

„ 20

92 93 94 95 96 97 98 99

Weight Percent TEG In TEG-Water Solution

15-5

100

Once the equilibrium vaporization constants for the glycol-water system are evaluated, the absorber can be calculated by regular absorption procedures

Example Problem 15.1

A natural gas at 100°F and 1000 psia is saturated with water vapor The gas is to be dehy- drated to a dew point of 40°F using triethylene glycol What must be the glycol concentration and circulation rate if the contactor contains the equivalent of 1.0 theoretical trays?

Solution

Equilibrium constants can be determined from equation 15.1, if the activity coefficients are known The activity coefficient is composition dependent At the top of the contactor (for the lean glycol concentration of 97.5%) the activity coefficient from Figure 15.2 is 0.542 The value at the bottom of the contactor will depend upon the rich glycol concentration which is unknown, but depends on the water K-value To start, assume that the rich glycol will contain 95% by weight TEG, giving an activity coefficient of 0.657

From Figure 15.3, the water content of the saturated gas at 100°F and 1000 psia is 61 pounds/MMSCF The equilibrium constants are:

Saturated Water Content:

Y = 3638 + 61/718 ~ 1.28 x 10 ~ mol fr

Top (Lean Glycol):

K= 0.542 x 1.28 x 10 = 6.94 x 1074

Bottom (Rich Glycol):

K = 0.657 x 1.28 x 1072 = 8.41 x 107

Since the two equilibrium constants are different, the average of the two values will be used

to represent the average absorption factor in the contractor

The Kremser Brown equation is:

Y tl ~ Yy 7 antl -A

Because of the low water concentration in the gas phases, mol fractions can be substituted in the left side of Equation 10.12

The composition of the gas in equilibrium with the lean glycol is:

150.2 * 18 (also obtainable from Figure 15.4)

The composition of the 40°F dew point gas is:

W = 9#H,0/MMSCF (Figure 15.3)

7Ÿ ” 7638 + 9/18

Lb

F and

“60 “40 “20 o ¿0 40 60 80 10 64120 4O «#160 IBO 200 ¿40 «280

CORRECTION FOR GAS GRAVITY

40000 1 it: 3 : k — - 40000

>] GAS GRA’

l2 '

MOLECULAR WEIGHT

6000

9

2000

ee

<

=

=

Ỗ

oe

ba

Q

x

1000 bt

800

9 |

sco

400

200 Warning: Dashed lines are

meta-stable equilibrium - Actua! equilibrium is lower —

water content Angie is a”

function of composition

ph

100

80

40

HYDRAT

_is a function of gas

composition

I9g58

Woter contents of natural gases with Temperature, °F corrections for salinity and gravity After MeKetta and Wehe, Hydrocarbon Processing,

Rev 1976 *

Equation 10.12 becomes:

1.28 x 103 - 1.9 x10” - Ae a

This is beyond the range of Figure 10.4 but may be solved by trial and error to find

A= 16 = KỸ =

-4

L

n+]

=?

1.23 x 10 x 150.2 x 2038 - 10.7 U.S gallon TEG/#,0

9 34(61-9) The concentration of the rich glycol wil! be:

(10.9 x 9.34 x 0.975) (10.7 x 9.34) +1

Figure 15.4:

0.30 We.% Water vs Mole Fraction Water

In TEG-Water Solutions,

025

hH

g

3

=

g

9

„

u

u

os

kt

fea

a

9

a

„10

.ũ

Weight Percent Water

The calculated circulation rate is greater than one can economically justify The number of 97.5 wt

% read from Figure 15.1 should be regarded as the minimum concentration required In actual practice one would produce a 98.5 - 98.7 wt % TEG If this concentration is used in the above calculation the cir- culation rate is reduced to 5 U.S gallons per pound of water removed from the gas

In this example one theoretical tray was used Using a tray efficiency of 25%, this represents the per- formance of 4 actua/ trays Figures 15.5 - 15.7 represent the general performance of glycol absorbers as a function of actual number of trays, lean TEG concentration and circulation rate (only at 100°F) These curves are not recommended for design purposes but they show clearly the relative effect of each variable The dew point depression is the inlet gas temperature minus the the outlet equilibrium dew point temperature

100 Ƒ 2 100°F contac! temperature Hồ 2 100°F, contact

u temperature 99.5%

Dã 9 Pr 5 woo OF

c "

ö ¬

a 50 L

z 60 È 98.5% a 70 |

a

M 5 Worley 1972 M 5 Worley 1972

40 L 1 4 1 L J L ah, L +L L L Ì L

“TEG rate, gal/Ib water in gas TEG rate, gal/lb water in gos

temperature

7 No,

¬

8

$ 100

a

=

£ 90°

a

e 80

70

M.S Worley 1972

60: Lộc BS HT l il

0 ’ 2' 3 4 5 6 7

TEG rate, gal/lb woter in gas

Figure 15.7 Dew-Point Depression vs

Glycol Rate

16-9

Figure 15.5 applies to the previous example Notice that for our 60°F dew point depression (100°F -40°F.) that a 98.5% TEG requires about 5 U.S gal./lb water removed while a 99 wt.% TEG would require only about 3 U.S gal./Ib water Which one could you choose? The one with the greatest economic benefit

The final choice of the combination of absorber contact and circulation rate should be the end result

of calculations which explore several combinations Figures 15.5 - 15.7 can serve to expedite this pro- cess Most commercial TEG units contain no more than 8 actual trays, with 4-6 being most common The Kremser-Brown expression (Equation 12.12) can be used to predict the performance of glycol ab- sorbers The equation applies particularly well to dehydration because there is very little water removed from the gas phase and the molar vapor rate through the column is very nearly constant The mass fates

of glycol circulation are large in comparison with the water removed However, the molar rates are closer because of the large molecular weight for the glycols and the small molecular weight for water Nonethe- less, Equation 10.12 will give a good representation of the behavior of glycol contactors

Frankly, plate-to-plate balance of the TEG absorber is seldom necessary For convenience, one can program the Kremser-Brown relationships and the K value correlations to produce a computer program that is very realistic Of course, one sub-routine of the entire package must be prediction of gas water contents, using an appropriate correlation of the type discussed in Chapter 14

For low pressure TEG units, particularly below 100 psia, a plate-to-plate balance may be desirable The water content is relatively high compared to the mass of gas flowing Consequently, the V/L ratio changes more drastically throughout the tower, as does the temperature However, this is the infrequent case

Regeneration (Still) Column

The required outlet water dew point or water content fixes the minimum TEG concentration needed for the lean solution entering the top of the absorber This concentration must be produced out the bot- tom of the regenerator reboiler

As always, the TEG leaving the reboiler will be at its bubble point Since the optimum reboiler temperature for TEG is about 400°F., the bubble point pressure is of primary concern

This reboiler operates at or below atmospheric pressure So, one can use ideal gas/solution relation- ships for prediction of performance If we use Raoult’s Law and assume only TEG and water are present, the bubble point is calculated in the usual manner In this case the K for each compound is equal to Pv/P, the vapor pressure divided by the total reboiler pressure (in absolute units) Thus

z(P_/P)x = 1.0 Furthermore, the X, (for TEG) plus the Xy (for water) equals 1.0 So, the bubble point calculation is quite simple

Because the vapor pressure for TEG is infinitesimal compared to that of water at 400°F., the vapor from the reboiler will be almost all water vapor if no stripping gas is added The relationship is

„- ()Œ)

P

Vv

For a given temperature Py is fixed The value of x, the mol % lean TEG leaving is affected by y (the mol fraction water vapor in the reboiler) and P (the reboiler pressure) A decrease in either y or P de- creases x Remember, we are talking about water concentration A decrease in x increases TEG concen- tration in the lean TEG solution

A decrease in the value of y and/or P thus increases lean TEG concentration: Operating at a vacuum increases this concentration If stripping gas is added, the mol fraction water in the vapor decreases be- cause part of the vapor is partially composed of that gas