sierrata04_05

8/13/2019 SIerrata04_05

1/17

GPSA ENGINEERING DATABOOK ERRATA

(2004 SI Edition)

PAGE DESCRIPTION

18 Fig. 12 Change units for Energy conversion factor

715 Fig. 723 Delete commas

98 Text changes

910 Fig. 913 Corrections

1017 Fig. 104 Change equation

1018 Change example following Eq. 104

1320 Ex. 135 Change units on head

1322 Eq. 1328 Change equation

1326 Fig. 1333 Change units on head

1327 Fig. 1334 Change units on head


2019 Text changes

2031 Figs. 2063, 2064 Reference change


2042 Fig. 2084 Change axis label

215 Fig. 213 Change heats of reaction

8/13/2019 SIerrata04_05

2/17

wellheadThe assembly of fit t ings, valves, and controls located at thesurface a nd connected t o the flow lines, tubing, and casing ofthe w ell so as to control the flow from th e reservoir.

wet gas(1) A gas conta ining wa ter, or a ga s wh ich ha s not been dehy-dra ted. (2) A term synonymous w ith r ich ga s. Refer to defini-tion of "rich gas" .

Wobbe numberA number proportional to the heat input to a burner at con-sta nt pressure. In Brit ish practice, it is the gross heating va lueof a ga s divided by t he squa re root of its gra vity. Widely usedin Europe, together with a measured or calculated flamespeed, to determine int erchangeability of fuel ga ses.

Conversion Factors

In t hese ta bles, fa ctors for conversion, includin g conversionsto the Int erna tional S ystem of Un its (SI ), are based on ASTMStandard for Metric Practice, E380-91. The latest edition ofthis publication should be studied for more detail on the SIsystem, including definitions a nd sy mbols.

In calculat ing derived factors in th e ta bles that follow, exactconversions w ere used, when ava ilable, ra ther t ha n th e 7-digitround-offs listed in ASTM E380 conversion t a bles. Derived fa c-tors given below a re rounded to the sam e number of significan tdigits a s the source factors.

In a ny conversion of fundam enta l measurement un its, someconfusion may result due t o redefinition of units used in earlierta bles. For exam ple, in 1959 a sma ll refinement wa s ma de inthe definition of the yard, which changed its length from3600/3937 met er (or 1 inch = 25.4000508 mm) t o 0.9144 mexactly (or 1 inch = 25.4 mm exa ctly). The ta bles below a rebased on the new definition, but one should be aware thatwh ere U.S. land m easurements a re concerned, the old relation-ship a pplies. Refer to ASTM E 380-91, note 13, for more det a il.

Energy Units Conversion

Confusion may arise in the definition of units for heat orenergy. In the tables below, the Btu (IT) and calorie (IT) are

used. These are the hea t unit s recommended by the Int erna-tional C onference on th e Pr operties of Steam , as defined:

1 B tu (IT) = 1055.055 852 62 joule (exa ctly )1 Ca lorie (IT) = 4.186 800 joule (exact ly)

For information only, other definitions tha t ma y be usedelsewhere:

1 Bt u (Mean) = 1055.87 joule1 B tu (39F ) = 1059.67 joule1 B tu (60F ) = 1054.68 joule1 B tu (Ther mochem ica l) = 1054.350 joule1 calorie (Mea n) = 4.190 02 joule1 calorie (15C ) = 4.185 80 joule

1 calorie (20C ) = 4.181 90 joule1 calorie (Ther mochem ical ) = 4.184 000 joule

The fundam enta l relat ionship between the Bt u and th e calo-rie:

grampound relat ionship

FahrenheitCelsius scale relat ionship

or : B tu 453.592 29

1.8= calorie (IT, mea n, or other)

1 therm = 100,000 B tu = 105.5056 x 106J

= 105,505.6 kJ = 1.055 056 x 105kJ .

(B tu denotes Br itish Therma l Unit s)

(ref: Physical P roperties of Nat ura l Ga ses,G as U nie, 1988 p. 23)

Velocity(Lengt h/unit of time)

ft /sec ft/min Miles/hr (U .S . S t a t ut e) m/sec m/min km/hr

1 60 0.6818182 0.3048 18.288 1.09728

0.01666667 1 0.01136364 5.08 x 103

0.3048 0.018288

1.466667 88 1 0.44704 26.8224 1.609344

3.280840 196.8504 2.236936 1 60 3.6

0.05468066 3.280840 0.03728227 0.016667 1 0.06

0.9113444 54.68066 0.6213712 0.2777778 16.66667 1

Energy

Ft -lbf Kgf-met er B tu (IT) Kilo-ca lorie (IT) H p-hr Kilow a t t -hr joule (J )

1 0.1382550 1.285068 x 103

3.238316 x 104

5.050505 x 107

3.766161 x 107

1.355818

7.233014 1 9.294911 x 103

2.342278 x 103

3.653037 x 106

2.724070 x 106

9.806650

778.1692 107.5858 1 0.2519958 3.930148 x 104

2.930711 x 104

1055.056

3088.025 426.9348 3.968321 1 1.559609 x 103

1.163 x 103

4186.8

1980000 273744.8 2544.434 641.1865 1 0.7456999 2684520.

2655224 367097.8 3412.142 859.8452 1.341022 1 3600000.

0.7375621 0.1019716 9.478171 x 104

2.388459 x 104

3.725061 x 107

2.777778 x 107

1

FIG. 1-2

Conversion Factor Tables

1-8

8/13/2019 SIerrata04_05

3/17

where the two immiscible liquid phases separate within thevessel by the d ifferences in density of the liquids. Sufficientretention time must be provided in the separa tor to allow forthe gravity separation to take place. The second category isdefined as coalescing separa tion. This is w here sma ll part i-cles of one liquid pha se must be separa ted or removed from alarge quantity of another liquid phase. Different types of in-terna l construction of separa tors much be provided for eachty pe of liquid-liquid sepa ra tors. The follow ing principles of de-

sign for liquid-liquid separa tion apply equa lly for horizonta lor vertical separators. Horizontal vessels have some advan-tage over verticals for liquid-liquid separation, due to thelarger interface area a vailable in the horizonta l s tyle, and theshorter dista nce particles must tra vel to coalesce.

There are two factors that may prevent two liquid phasesfrom separa ting d ue to differences in specific gra vity:

If droplet particles are so small that they may be sus-pended by Br ownian movement. This is defined as a ra n-dom motion tha t is greater tha n directed movement dueto gra vity for part icles less than 0.1 m in diam eter.

The droplets ma y car ry electric charges due to dissolvedions. These charges can cause the droplets to repel eachother rat her tha n coalesce into larger part icles and settle

by gra vity.Effects due to B rownian movement a re usually small and

proper chemical treat ment w ill usually neutra lize an y electriccharges. Then settling becomes a function of gravity an d vis-cosity in a ccorda nce wit h St okes La w. The sett ling velocity ofspheres through a fluid is directly proportional t o the differ-ence in densities of the sphere and the fluid, and inverselyproportional to the viscosity of the fluid a nd t he squa re of thediameter of the sphere (droplet), as noted in Eq 7-3. The liq-uid-liquid separation capacity of separators may be deter-mined fr om Eq ua tions 7-13 a nd 7-14, w hich were deriv ed fromEq uat ion 7-39. Va lues of C * ar e found in Fig. 7-23.

Vertica l vessels:

Wcl

=

C

S hl S ll

(0.785

)D

v

2 Eq 7-13

Horizontal vessel:

Wcl = C

S h lS ll

LlH l Eq 7-14

Since the d roplet size of one liquid pha se dispersed in a n-other is usually unknown, it is simpler to size liquid-liquidseparation based on retention time of the liquid within thesepara tor vessel. For gravity separa tion of tw o liquid pha ses,a large retention or quiet settling section is required in thevessel. Good separa tion requires sufficient t ime t o obta in anequilibrium condition between the two liquid phases at thetemperat ure an d pressure of separa tion. The liquid capacityof a separator or the settling volume required can be deter-

mined fr om Eq 7-12 using the r etent ion time given in F ig. 7-22.

The following example shows how to size a liquid-liquidseparator.

Example 7-3 Determine the size of a vertical separa tor toha ndle 100 m3/da y of 0.76 rela tive dens ity condensa te a nd 10m 3/da y of produced wa ter. Assume the w a ter pa rt icle size is200 m. Other opera ting conditions are a s follows:

Operat ing temperature = 25 COperating pressure = 6900 kPa (ga)Wat er relat ive density = 1.01

Condensat e viscosity = 0.55 mP a s @ 25 CCondensat e relative density = 0.76

Fr om Eq 7-14

Wc l = C

S h lS ll

(0.785)(D v)2

From Fig. 7-17 for free liquids w ith w at er part icle diamet er200 microns, C* = 1880.

100m3/day = 18801.010.76

(0.55)(0.785)(D v)2(106)

(D v)2 = (1.43)(105)

D v = 390m m

Using the alternate method of design based on retentiotime a s shown in Eq 7-16 would give:

U =W(t )1440

From Fig. 7-18, use 3 minutes retention time.

U =(110)(3)

1440 = 0.23m3

A 390 mm d iam eter vessel w ill hold about 0.12 m 3per 10mm of height. The sma ll volume held in the bottom hea d cbe discounted in t his size vessel. The shell height requ ired fthe retent ion volume required w ould be:

Shell height =0.23

0.12 = 1.9 m =1900 m m

Another para meter tha t should be checked when separ at iam ine or glycol from liquid h ydrocarbons is the interfa ce arbetween t he tw o liquid la yers. This a rea sh ould be sized so tglycol or a mine flow a cross th e interfa ce does not exceed a

proximately 100 m3

per day per m2

.

The above example indicates tha t a r elatively small separtor w ould be required for liquid-liquid separa tion. It should remembered tha t the separa tor must also be designed for tvapor capacity to be ha ndled. In m ost cases of high vapor-liuid loadings th at ar e encountered in gas processing equipmedesign, the vapor capa city required will dictate a much largvessel than would be required for the liquid load only. Tproperly designed vessel has to be a ble to handle both t he vpor and liquid loads. Therefore, one or the other will contrthe size of t he vessel used.

EmulsionCharacteristic

DropletDiameter, m

Constant ,10

C*

Free Liquids 200 1880

Loose E mulsion 150 1060

ModerateEmulsion

100 470

Tight E mulsion 60 170

FIG. 7-23

Values of C* Used in Equations 7-13 and 7-14

7-15

8/13/2019 SIerrata04_05

4/17

8/13/2019 SIerrata04_05

5/17

ceeds 5C, consulting a specialist is recommended for a more

rigorous calculation procedure.

The condensing of a pure component occurs at a constanttemperat ure equal to the sa turat ion temperature of the in-

coming va por strea m. Frequently a vapor enters a condenser

superheated and must have the sensible heat removed from

th e vapor before condensa tion can occur. If the conden sing sur -

face temperature is greater tha n the incoming vapor sa tura-

tion temperatur e, the superheat in th e vapor is tra nsferred to

the cold surface by a sensible heat tra nsfer mecha nism (dry-

wa ll condition). If th e condensing sur face tempera ture is less

tha n the sa tura tion temperatu re of the incoming va por, a con-

densate film will be formed on the cold surface. The sensible

heat is removed from the vapor a t th e condensa te-vapor inter-face by vaporizing (flashing) condensate so that the heat ofvaporization is equal to the sensible heat removed from thevapor. Under this w et w all condition, the effective tempera-ture of the vapor is the sat ura tion temperatu re, an d the effec-tive heat transfer mechanism is condensation. Thedetermination of the point in the desuperheating zone of acondenser where drywall conditions cease and wet wallconditions begin is a tria l an d error procedure. A method fre-

quent ly employed to give a sa fe approximat ion of the requiredsurface is to use the condensing coefficient and the CMTDbased on the vapor saturation temperature to calculate thesurface required for both the desuperheating zone and t he con-densing zone.

The following Exa mple 9-2 will illustra te t he use of the heatrelease curve to calculat e the surfa ce required and the LMTDfor each zone in a condenser for a pure component a pplication.

Example 9-2 A propane r efrigerant condenser is requ iredto condense the vapor stream using the heat release curve asshown in Fig. 9-14. This stream enters the condenser super-heat ed a nd leaves t he condenser a s a subcooled liquid. Assumethat a single-tube pass, single-shell pass, counterflow ex-changer is used so tha t L MTD correction fa ctors do not apply.Note tha t t he propane is on th e shell side. The overa ll heattra nsfer coefficients for each zone are a s follows:

Desuperheat ing: (82 C to 42C)

U v = 396.6 W/(m2 C)

hv = 630.4 W/(m2 C )

C on densin g: (42 C to 42 C )

U c = 794.4 W/(m2 C )

S ubcooling: (42 C to 35 C )

U L = 649.7 W/(m2

C )

Solution Steps1. Ca lcula te the surface temperat ure (outs ide wall) on the

vapor side at the refrigerant stream inlet using the fol-lowing equation:2

Leaka ge use TEMA toleran ces

tube hole = 576 19.05 0. 4 = 13 789 mm2

shell crack = 7624.445 0.60 = 6 385 mm2

20 174 mm2

Window flow a rea:

window area = 0.237 (762)2

= 137 613 mm2

tubes in window = 208 0.785(19.05)2

= 59 255 mm2

78 358 mm2

lea k 20 174 mm2

98 532 mm2

Cross f low a rea:

Free area = 171.45 mm 473.08 = 81 109 mm2

Lea ka ge = 20 174 mm2

Net cross flow a rea = 101 283 mm2

BP = baff le pitch = 473.08 mmFD = free diameter = 171.45 mm

Note: window area cross flow a rea

FIG. 9-14

Propane Condensing Curve

FIG. 9-13

Heat Exchanger Detail Design Results

9-10

8/13/2019 SIerrata04_05

6/17

Sound power level cannot be measur ed directly but must becalculated from sound pressure levels (SPL) dB. In metricterms, this is known as L w.

Sound pressure level, known a s SP L, or Lp in metric termi-nology, is the audible noise given in decibels that relates tointensity at a point some dista nce from t he noise source. It ca nbe related to octave bands or as an overall weighted leveldB (A).

Weighted sound levels rela te t he decibel (loudness) to a fr e-quency. Ears can easily pick up high-frequency noises (bothintensity an d direction) but ar e relatively insensitive to low-frequ ency noise. For a stereo system h igh-frequency speakersmust be very carefully locat ed in a room for best results, butlow-frequency bass speakers ca n be placed an yw here, even outof sight.

There are three basic weighting systems: A, B and C. The"A" system, dB (A), most closely relates to our ear, the "B " sys-tem, dB (B ), has some specific uses a nd t he "C " system, dB (C),is considered unweighted.

The dB(A) is the most common weighting system. It ex-presses sound levels wit h a single number instead of ha vingto specify the noise of each octave ba nd.

Note tha t t he sound ran ge of ACH Es (at close range) is typi-cally betw een 80 and 105 dB (A).

ACHE Noise

Whether the concern is for overall plant noise or the noiseexposure of plant workers in t he vicinity of the fan s, a differentty pe of noise specificat ion must be used.

Overall, noise limitations from an ACHE are typically asound power level (P WL) specificat ion for each u nit. This lim -its the contribution of each unit (typically two fans) to theplant noise as a whole. This is usua lly needed w here noise at

the plan t bounda ry is considered. Contributions of each paof the plant must be car efully controlled if overall plant noiis limit ed. PWLs can be expressed a s w eighted level dB (A)sometimes even by limita tions on each octa ve band.

If worker protection is the main concern, a limitation sound pressure level at 1 m below the bundle will imposed as "SPL dB(A) at 1 m". The OSHA limitation is dB (A) for eight -hour exposu re, but 85 dB (A) dow n t o 80 dB (is not uncommon.

Predicting Fan Noise

Each fan manufacturer has proprietary equations for prdicting fan noise. AP I G uidelines use the general formula:

PWL = 56 + 30 log

MP M

304.8

+ 10 log HP Eq 10

This calculates PWL as dB(A)

Proprietary noise equations are based on actual tests various speeds and operating conditions considering the folowing effects:

Fan diam eter

Fan tip speed

B lade Type

Bla de pitch angle

Inlet conditions

Horsepower

Note: Logs a re common logs (base 10).

For example:

4.267 meter fa n

237 RP M (3177 m/min tip speed)

Altitude above Sea Level Barometer Atmospheric Pressure

meters feet mm Hg abs. in Hg abs. kPa (abs) psia

0 0 760.0 29.92 101.325 14.696

153 500 746.3 29.38 99.49 14.43

305 1000 733.0 28.86 97.63 14.16458 1500 719.6 28.33 95.91 13.91

610 2000 706.6 27.82 94.18 13.66

763 2500 693.9 27.32 92.46 13.41

915 3000 681.2 26.82 90.80 13.17

1068 3500 668.8 26.33 89.15 12.93

1220 4000 656.3 25.84 87.49 12.69

1373 4500 644.4 25.37 85.91 12.46

1526 5000 632.5 24.90 84.32 12.23

1831 6000 609.3 23.99 81.22 11.78

2136 7000 586.7 23.10 78.19 11.34

2441 8000 564.6 22.23 75.22 10.91

2746 9000 543.3 21.39 72.39 10.503050 10 000 522.7 20.58 69.64 10.10

Reprinted with permission of J ohn M. Cam pbell and Company.

FIG. 10-20

Altitude and Atmospheric Pressures8

10-17

8/13/2019 SIerrata04_05

7/17

25.1 HP

Find sound power level

P WL= 56 + 30 log 10.4 + 10 log 25.1

= 100.5 dB (A)

When considering mult iple noise sources (fans) use th erelation:

P WLN= PWL + 10 log N Eq 10-5

The sound power level for 2 adja cent fa ns is t he P WL ofone fan plus 10 log 2 or P WL 2= P WL + 3.

A doubling of th e noise source ad ds 3 dB.

Noise attenua tes with dista nce by the equation:

SP L (at dista nce R) = P WL - 20 log (3.28R) Eq 10-6

Where R is in m eters from t he center of the source. Meas -ure R as a " line of sight" dista nce.

Consider the noise an observer at grade hears at 15.24 mfrom an operating ACH E w ith th e fan on a nd the line-of-sightdistan ce from gra de to the center of the fa n is a ctually 18.9 m.Remember what the ear hears is SPL, the noise energy isPWL.

Assum e P WL = 100.5 dB (A).

SP L = 100.5 - 20 log 62 = 64.7 dB (A)

This a lso assumes background noise is a t least 10 dB q uieter.

Note: If both fa ns w ere running, the SP L w ould have been67.7 dB (A).

When considering the noise at 1 m beneath the unit, thedrive system a nd motor noise become dominant at lower tipspeeds.

Factors tha t influence this noise are:

Motor noise

Belt or gear noise

Bearing noise

Reflected noise from supports

Ba ckground noise

Gea r noise is especially significant in a forced draft unit.

Noise Testing

Frequently, the ACHE must be tested for confirma tion tha tits noise does not exceed specifications imposed by the pur-chas er. There ar e two ba sic types of tests norm a lly performedbefore shipment:

a) Measure SP L dB (A) (Sound Pr essure Level) at " 1 m be-low the bundle or fan guard" depending on whetherthe unit is induced or forced draft .

b) Measure PWL (Sound Power Level) using "hemispheri-cal power level test." P WL is specified as either a d B (A)weighted va lue or by octave ban ds.

The "SP L a t 1 m" test is by far the most common and leastexpensive. Usually, only one or two measurements are re-quired. The answer is immediate and read directly from thenoise met er.

The hemispherical test is far more complicated and expen-sive. Several hours, many technicians and a large crane arerequired to perform t his test. Full deta ils of the test a re givenin API Recomm ended P ra ctice 631 M, J une, 1981.7

REFERENCES

1. A.P.I . Sta ndard 661, Air Cooled Heat Exchangers for GeneralRefinery S ervices.

2. Br iggs, D. E. , Young, E. H., Convection Heat Tran sfer and P res-

sure D rop of Air F lowing Across Tria ngula r P itch of Tubes,

Chemical Engineering Progress Symposium Series, Volume 59,

No. 41, 1963.

3. Cook, E. M., Air Cooled Heat Exchangers, Chemical Engineer-

ing, Ma y 25, 1964, p. 137; J uly 6, 1964, p. 131; and Augu st 3,

1964, p. 97.

4. Ga rdner, K. A., Efficiency of Extended Surfaces, Tran s ASME,

Volum e 67, 1945, pp. 621-631.

5. Robinson, K. K. , Briggs, D. E. , Pr essure Drop of Air Flowing

Across Triangular Pitch Banks of Finned Tubes, Chemical En-

gineering Pr ogress Sy mposium S eries, Volume 62, No. 64, 1966.

6. Rubin, Fra nk L. , Winterizing Air Cooled Heat Exchangers, Hy-

droca rbon P rocessing, Oct ober 1980, pp. 147-149.

7. API S tan dard 631 M, First Edit ion, "Measurement of Noise from

Air-Cooled Heat E xchanger s," J une, 1981.

8. J ohn M. Ca mpbell Co. "Ga s Condit ioning and P rocessing, Vol. 2, "

Eighth Edit ion.

10-18

8/13/2019 SIerrata04_05

8/17

Answer: t 2 = 130 C (a pproximat ely) from Fig. 13-32.

Fig. 13-34 gives the approximate brake power required forthe compression. It includes compressor efficiencies in thera nge of 60 to 70%.

Example 13-5 Given: Weight f low, w, = 30 000 kg/h

head = 150 kNm/kg

Find: Bra ke Power

Answer: G p = 1700 kW from Fig. 13-34.

Fig. 13-37 predicts the approximate number of compressorwheels required to produce the head. If the number of wheelsis not a wh ole number, use the next high est num ber.

Calculating Performance

When more a ccura te informa tion is requ ired for compressorhead, gas horsepower, and discharge temperature, the equa-tions in this section should be used. This method applies to agas mixtu re for which a P -H dia gra m chart is not ava ilable. To

calculate the properties of the gas, see Figs. 13-6 and 13-7. Allvalues for pressure and temperature in these calculation pro-cedures are the absolute values. Unless otherwise specified,volumes of flow in t his section a re a ctual volumes.

To ca lculat e the inlet volume:

Q =(w)(8.314)(T1)(Z1)

(M)(P 1)Eq 13-25

If we assume the compression to be isentropic (reversible

adiabatic, constant entropy), then:

H is =ZRT

M (k1)/k

P 2

P 1

(k 1)/k

1

Eq 13-26

Since these calculations will not be wheel-by-wheel, thehead will be calculated a cross the ent ire ma chine. For this, usethe average compressibility factor:

FIG. 13-25

Typical Centrifugal Compressor Cutaway

13-20

8/13/2019 SIerrata04_05

9/17

Za vg =Z1+Z 2

2

The heat capacity ra tio, k, is normally determined at the a v-erage suction and discharge temperature (see Figs. 13-7 and13-8).

Isentropic Calculation

To calcula te th e head:

H is =Za vgRT1

M (k 1)/k

P 2

P 1

(k 1)/k

1

Eq 13-27a

wh ich can a lso be written in th e form:

H is =8.314

M

Za v gT1

(k 1)/k

P 2

P 1

(k 1)/k

1

Eq 13-27b

The ga s horsepower ca n now be calculat ed from:

Ghp =(w)(H is)

(is)(3600)Eq 13-28

FIG. 13-28

Compressor Performance, Low Compression Ratio

FIG. 13-29

Compressor Performance, Higher Compression Ratio

FIG. 13-27

Compressor Head

13-22

8/13/2019 SIerrata04_05

10/17

FIG. 13-33

Head

Z = 1

13-26

8/13/2019 SIerrata04_05

11/17

FIG. 13-34

Power Determination

FIG. 13-35

Efficiency Conversion

13-27

8/13/2019 SIerrata04_05

12/17

The approximate theoretical discharge temperature can becalculated from:

Ti d e a l = T1

P 2

P 1

(k 1)/k

1

Eq 13-29

T2 = T1+Ti dea l Eq 13-30

The actual discharge temperature can be approximated:

Tactual = T1

P 2

P 1

(k 1)/k1

isEq 13-31

T2 = T1+Ta c t u a l Eq 13-32

Polytropic Calculation

Sometimes compressor ma nufa cturers use a polytropic pat hinstead of isentropic. Polytropic efficiency is defined by:

n

(n1) =

k

(k 1)p Eq 13-33

(See Fig. 13-35 for conver sion of isent ropic efficiency topolytropic efficiency.)

The equations for head and gas horsepower based uponpolytropic compression are:

H p =Za vg RT1

M (n 1)/n

P 2

P 1

(n1)/n

1

Eq 13-34a

which a lso can be writ ten in the form:

H p =8.314

M

Za v gT1

(n 1)/n

P 2

P 1

(n 1)/n

1

Eq 13-34b

Ghp =(w)(H p)

(p)(3600)Eq 13-35

Polytropic and isentropic head a re related by

H p = His pis

Eq 13-36

The approximate actual discharge temperature can be cal-culated from:

T2 = T1

P 2

P 1

n 1n

Eq 13-37

Mechanical Losses

After the gas horsepower has been determined by eithermethod, horsepower losses due to friction in bearings, seals,and speed increasing gears must be added.

Fig. 13-36 shows losses relat ed to th e sha ft speed and casing

size for conventiona l multist age unit s.

Bea rings a nd seal losses can also be roughly computed fromScheels equation:

Mechanical losses = (G p)0.4 Eq 13-38

To ca lculat e the bra ke horsepower :

Bhp = G p +mecha nicallosses Eq 13-39

Compressor Speed

The basic equa tion for estimat ing th e speed of a centr ifugalcompressor is:

N = (Nnominal)H t o t a l(No. ofw heels)(H ma x /w heel )Eq 13-40

w here th e number of wh eels is determ ined from Fig. 13-37.

Nominal speeds to develop 30 000 N m/kg of hea d/w heelcan be determined from Fig. 13-23. However, to calculate themaximum head per wheel, the following equation based onmolecular weight (or more accurately, density) can be used.

H m a x/stage = 15,000 1,500 (M)0.35 Eq 13-41

This equation will give a head of 30 000 N m/kg for a ga swh en M = 30 an d 33 000 N m/kg w hen M = 16.

Casing SizeMax Flow(inlet m

3/h)

Nominal S peed(rpm)

1 12 700 10 500

2 33 900 8 200

3 56 000 6 400

4 93 400 4 900

5 195 000 3 600

6 255 000 2 800

FIG. 13-36

Mechanical Losses

13-28

8/13/2019 SIerrata04_05

13/17

economic advantage versus methanol recovered by distilla-tion. At cryogenic conditions (below 40C) methanol usuallyis preferred because glycols viscosity makes effective separa-tion difficult.

Ethylene glycol (EG), diethylene glycol (DEG), andtriethy lene glycol (TEG ) glycols have been used for hy dra teinhibit ion. The most popular h a s been ethylen e glycol becauseof its lower cost, lower v iscosity, and low er solubility in liq uidhydrocarbons.

Physical properties of methanol and methanol-water mix-tur es ar e given in F ig. 20-34 through Fig. 20-37. P hysica l prop-erties of the most common glycols an d glycol-wa ter m ixturesar e given in Fig. 20-38 through Fig. 20-49. Ta bula r inform a tionfor the pure gly cols a nd m etha nol is provided in Fig. 20-56.

To be effective, th e inhibitor m ust be present a t t he verypoint w here the wet ga s is cooled to its hydra te temperat ure.For exam ple, in refrigeration plan ts glycol inhibitors a re ty pi-cally spray ed on the t ube-sheet faces of the ga s excha ngers sotha t it can flow with t he gas thr ough the tubes. As wa ter con-denses, the inhibitor is present to mix with th e wat er and pre-vent hydrates. Injection must be in a manner to allow gooddistribution to every tube or plate pass in chillers and heatexchangers operating below the gas hydrate temperature.

The viscosities of ethylene glycol and its aqueous solutionsincrease significantly a s temperatur e decreases, and this mustbe allowed for in the ra ting of refrigeration-plant excha ngersan d chillers.

The inhibitor and condensed water mixture is separatedfrom the ga s s tream a long w ith a separa te l iquid hydrocarbonstrea m. At this point, the wat er dewpoint of the gas strea m isessentially equa l to the separa tion temperatur e. Gly col-wa tersolutions and liquid hydr ocar bons can emulsify when agit at edor when expanded from a high pressure to a lower pressure,e.g. , J T expan sion va lve. Ca reful separat or design will a llownearly complete recovery of t he diluted glycol for regenera tionan d reinjection. Fig. 20-57 shows a flow diagr am for a typicalEG injection system in a refrigerat ion plant.

The regenera tor in a glycol injection system should be opera tedto produce a regenerat ed glycol solution tha t w ill have a freezingpoint below the m inimum t empera ture encountered in t he sys-tem . This is t ypica lly 75-80 wt %. Fig. 20-58 shows t he freez ingpoint of var ious concent ra tions of glycol wa ter solutions.

The minimum inhibitor concentration in the free waterphase may be approxima ted by Ham merschmidts equation.25

d =K H X I

MWI(1 XI)Eq 20-5

XI=dMWI

KH+d MWIEq 20-6

Where K H (glycols) = 1297 to 2222 andK H (meth an ol) = 1297

The K H ra nge of 1297 to 2222 for glycols reflects t he un cer-ta inty in t he value of this para meter. At equilibrium, such asfor a la bora tory t est, 1297 is applicable as illustra ted on Fig.20-60. In some field operations, however, hydrate formationha s been prevented w ith glycol concentra tions correspondingwith K H va lues as h igh a s 2222. This is because h ydra te sup-pression with glycols depends on the systems physical andflow cha ra cteristics etc.) as w ell as t he properties of the ga s a nd t he glycol. There-fore, in the a bsence of reliable field-test d at a, a system sh ould

be designed for a K H of 1297. Once the syst em is operat ing, tglycol concentration can be reduced to tolerable levels.

Eq 20-5 an d E q 20-6 should not be u sed beyond 20-25 wtfor meth a nol an d 60-70 w t%for the gly cols. For meth a nol cocent ra ti ons up t o about 50%, the Nielsen-B ucklin equat ion26prvides better a ccura cy:

d = 72.0 ln(xH 2O) Eq 20

Note that xH 2O in Eq 20-7 is a m ole fraction, not a ma ss fra

tion. Expressing mole fraction in t erms of ma ss fraction, de

point depression is plotted against the weight percemet ha nol in F ig. 20-59.

Maddox et al .27presents a m ethod of estimat ing the requirinhibitor concentrat ion for both meth an ol and E G. The methis iterat ive but converges easily after a few itera tions.

Figs. 20-60 thr u 20-64 provide a comparison of var ious ihibitor correlations with experimenta l data .28,29,30Experimeta l data at very high inhibitor concentrat ions is limited.

On c e t h e r equ i r ed in h ib i t or c o n cen t r a t io n h a s b een c ac u l a t ed , t h e ma s s o f in h ib it o r r equ i r ed in t h e w a t e r p h ama y be ca lcu la ted f rom E q 20-8

m I =XRm H2O

XLX REq 20

The a mount of inhibitor to be injected n ot only must be sufficieto prevent freezing of the inhibitor water phase, but also must sufficient t o provide for the equilibrium va por pha se content of tinhibitor an d the solubility of the inhibitor in any liqu id hydrocarboThe vapor pressure of methanol is high enough that significaquantities will vaporize. Methanol vaporization losses may be esmated from Fig. 20-6531Fig. 20-65 is extrapolated above 4800 k(abs). Recent studies indicate Fig. 20-65 ma y underestimat e vapphase methanol losses at higher pressures. Glycol vaporizatilosses are generally very sma ll and are typically ignored in calcutions.

FIG. 20-30

Solution Sketch for Example 20-8

20-19

8/13/2019 SIerrata04_05

14/17

tor countercur rent to th e ga s flow. Wat er-rich glycol is removedfrom the bottom of the contactor, passes through the refluxcondenser coil, flashes off most of the soluble gas in the flashtank, and flows through the rich-lean heat exchanger to theregenerator. In the regenerator, absorbed water is distilledfrom the glycol at near at mospheric pressure by applicat ion ofheat. The regenerated lean glycol exits the surge drum, ispartly cooled in the lean-rich exchanger and is pumpedth rough t he glycol cooler before being recircula ted t o the con-

tactor.

Evaluation of a TEG system involves first establishing theminimum TEG concentra tion required to meet the outlet gaswa ter dew point specification. Fig. 20-68 shows the w at er dew-

point of a na tural ga s s tream in equilibrium with a TEG soltion at various temperat ures and TEG concentra tions. Fig.20-68 can be used to estima te t he required TEG concentra tifor a part icular applicat ion or the theoretical dewpoint depresion for a given TEG concentra tion and conta ctor tempera turActual outlet dew points depend on the TEG circulat ion raan d number of equilibrium sta ges, but ty pical a pproaches equilibrium are 611 C. Equilibrium dewpoints ar e relativein s en s i t i v e t o p r es s u r e a n d F ig . 20-68 ma y b e u s ed u p

10 300 kPa (a bs) w ith lit tle error.

Fig. 20-68 combines th e results published by P a rrish , et alcovering TEG purity up to 99.99 ma ss percent an d t hose rported by B ucklin an d Won in 198752covering TEG purity u

FIG. 20-61

Hydrate Inhibition with Methanol:

Hammerschmidt vs. Experimental Data28

FIG. 20-62


Nielsen & Bucklin vs. Experimental Data28

FIG. 20-63


Nielsen & Bucklin26

vs. Experimental Data29

FIG. 20-64


Maddox et al.27

vs. Experimental Data29

20-31

8/13/2019 SIerrata04_05

15/17

The solubilities of pure H 2S a n d C O2in pure TEG ar e givenin Fig. 20-76. These charts apply to pure TEG or to the leanTEG , which is essentia lly pure. For th e rich-TEG solution leav-ing the bottom of a contactor, it is approximately correct toapply them to the TEG portion.

Heavier paraffin hydrocarbons are essentially insoluble inTEG. Aromatic hydrocarbons, however, are very soluble inTEG , an d significan t a mounts of ar omatic hydrocarbons may

be absorbed in the TEG at conta ctor conditions. This ma y pre-sent an environmental or sa fety hazard when they are dis-charged from th e top of the regenera tor.

Vapor-liquid equilibrium constants (K-values) for benzene,toluene, ethylbenz ene, a nd o-xylene in TEG solutions ar e pre-sented in RR -131.43This data indicat es that a t typical contac-tor conditions a pproximat ely 10-30%of th e a romat ics in thegas st ream m ay be a bsorbed in the TEG solution. Ba sed on thedat a in RR-131 at 6900 kPa (abs), 38 C a nd a TEG circulat ionra te of 25 liters TEG /kg H 2O a bsorbed, the approxima te per-centa ge absorption of aroma tics from the ga s into the rich TEGare calcula ted a s :

B enzene 10%Toluene 14%

E t hylbenzene 19%O-xylene 28%

Aromat ic absorption increases with increasing pressure anddecreasing temperature. Aromatic absorption is directly re-lat ed to TEG circulat ion ra te. Higher circulation ra tes resultin increased a bsorption. Aromat ic absorption is essentia lly in-dependent of the number of contacts in the absorber so onemethod of minimizing ar omatic a bsorption is to use taller con-ta ctors and minimize TEG circulat ion ra tes.

Most of the ar omat ic components w ill be stripped from th eTEG solution in the r egenerator.

Flash tank sizing should be sufficient to degas the glycolsolution a nd skim ent ra ined liquid hy drocar bons, if necessary.

A minimum r etention tim e of 3-5 minutes is required for de-gassing. If liquid h ydrocarbons are t o be removed as w ell, re-tention times of 20-30 minutes m ay be required for a dequa teseparation. Flash tank pressures are typically less than 520kPa (abs).

Regenerator sizing requires establishing the reboiler dutyand, when high TEG concentrations are required, providingsufficient str ipping ga s.

A quick estima te of reboiler duty can be ma de using E q 20-10.

Q = (0.116)(L g) Eq 20-10

Eq 20-10 is approxima te a nd usua lly gives values which ar ehigher th an the a ctual dut y. A more rigorous determinat ion ofreboiler duty is shown in Example 20-12.

Example 20-12 Determ ine reboiler duty for condit ions in th eprevious example. Assume the rich TEG temperature enteringthe regenerat or is 150C and the reboiler temperature is 200 C.

G lycol Reboiler Duty: B asis 1 m3TEG

Sensible Heat:

Qs = mC pT

FIG. 20-69

Water Removal vs. TEG Circulation Rate at Various

TEG Concentrations (N = 1.0)

FIG. 20-70



FIG. 20-71



20-34

8/13/2019 SIerrata04_05

16/17

an d th e period of regeneration should be decreased by m ulti-plying it t imes the ratio V /Vmi n. A method more exact tha nFig. 20-88 for calculating the minimum superficial velocity isto use Eq 20-16, but to consider it in terms of (P /L)min a ndVmi ninstead of (P /L)ma xand Vma x.

General Comments

The regeneration cycle frequently includes depressur-ing/repressuring to ma tch th e regenerat ion ga s pressurean d/or to ma ximize the regenerat ion gas volume to meet thevelocity criterion. In th ese applicat ions, the ra te of depressur-ing or repress uring should not exceed 350 kP a per/minu te.Some a pplications, termed pressure swing ad sorption, regen-erat e the bed only w ith depressurization an d sweeping the bed

wit h ga s just above atm ospheric pressure.Moisture a na lyzers for very low w at er contents require care

to prevent da ma ge to the probes. When inserted int o the beds,sample probes and temperature probes must be installed toreach th e center of the gas pha se.

Solid desiccant t owers ar e insulated externa lly or possiblyinternally. Internal refractory requires careful installationan d curing, usually before the desiccan t is insta lled. It sa vesenergy but the greatest benefit is it can dramatically reducethe req uired heat ing a nd cooling times. This is often an impor-ta nt benefit for systems wh ere regenerat ion times are limited.The primary disa dvan ta ge is the potentia l for wet ga s bypass-ing the desiccan t th rough cracks and defects in the insulationduring the a dsorption cycle.

Example 20-13: 2.85 106

S m3

/day of natura l gas w ith a mo-lecular weight of 18 is to be processed for ethane recovery ina t urbo-expan der plant. It is wa ter sat ura ted at 4140 kPa (abs)and 38C and must be dried to 101C dew point. Determinethe water content of the gas , and the amount of water thatmust be removed; and do a prelimina ry design of a molecular-sieve dehydra tion system consisting of two towers with down-flow d ehydra tion in one tower a nd up-flow r egeneration in theother. Use 4A molecular sieve of 3.2 mm bead s (i.e., 4x8 mesh).The regenerat ion ga s is part of the plant s residue gas, which

FIG. 20-84

Mole Sieve Capacity Correction for Unsaturated Inlet Gas

FIG. 20-85

Mole Sieve Capacity Correction for Temperature

FIG. 20-86

Inlet and Outlet Temperatures During Typical Solid

Desiccant Bed Regeneration Period

FIG. 20-87

Allowable Velocity for Mole Sieve Dehydrator

20-42

8/13/2019 SIerrata04_05

17/17

Chemistry The overall equilibrium reactions applica-ble for H 2S a n d C O2 and primary a nd secondary amines areshown below w ith a primary amine 9. A qua litat ive estima tionof the velocity of the reaction is given.

For hydrogen sulfide removal

RN H 2 + H 2S RNH 3+ + H S

Fas t Eq. 21-1

RNH 2 + H S

RNH 3+

+ S

Fas t Eq. 21-2

The overall rea ctions betw een H 2S a nd a mines are s impsince H 2S rea cts direct ly and ra pidly w ith a ll amines to forth e bisulfide by E q. 21-1 and th e sulfide by E q. 21-2.

For carbon dioxide removal

2 RNH 2 + C O2R NH 3+ + R NH C OO

Fas t Eq. 21

RNH 2+ C O2+ H 2O R NH 3+ + H C O3

Slow Eq. 21

RNH 2 + H C O3

R NH 3+

+ C O3

Slow

Eq. 21

FIG. 21-3

Approximate Guidelines for Amine Processes1

MEA DEA(9) DGA Sulfinol MDEA(9)

Acid gas pickup, m3/100L @ 38 C,

normal range(2)

2.33.2 2.856.4 3.55.40 3.012.75 2.26.4

Acid gas pickup, mol/mol a min e,normal range

(3)0.330.40 0.200.80 0.250.38 NA 0.200.80

Lean solut ion residual acid gas,mol/mol a mine, norma l ra nge

(4)0.12 0.01 0.06 NA 0.0050.01

Rich solution acid gas loading,mol/mol am ine, normal ra nge

(3)0.450.52 0.210.81 0.350.44 NA 0.200.81

Solution concentr at ion, wt%, normalra nge 1525 3040 5060

3 components.Va ries. 4050

Approximat e reboiler hea t duty, kJ /Llean solution

(5)280335 235280 300360 100210 220250

Steam heated reboiler tube bundle,approx. average heat flux,

Q/A = M J /(h m2)(6)

100115 7585 100115 100115 7585

Direct fired r eboiler fire tube,average hea t f lux,

Q/A = MJ /(h m2)(6)

90115 7585 90115 90115 7585

Reclaimer, steam bund le or fire tube,average hea t f lux,

Q/A = MJ /(h

m

2

)

(6)

7090 NA

(7)

7090 NA NA

(7)

Reboiler temperature, normaloperat ing ra nge, C

(8)107127 110127 121132 110138 110132

Heat s of react ion;(10)

approximate:kJ /kg H 2S 1420 1290 1570 N/A 1230

kJ /kg C O2 1920 1700 2000 N/A 1425

NA not a pplicable or not a vaila ble

NOTES:

1. These data alone should not be used for specific design purposes. Ma ny design factors must be considered for actua l plant design.

2. Dependent upon acid gas partia l pressures and solution concentra tions.

3. Dependent upon acid gas part ial pressures and corrosiveness of solution. Might be only 60%or less of value shown for corrosive systems.

4. Var ies with stripper overhea d reflux ra tio. Low residual acid gas contents require more stripper trays a nd/or higher reflux ra tios yielding largerreboiler dut ies.

5. Varies w ith st ripper overhead reflux ratios, rich solution feed temperature t o stripper a nd reboiler temperature.

6. Maximum point hea t flux can rea ch 230 to 285 MJ /(h m 2) at h ighest flame tempera ture at t he inlet of a direct fired fire tube. The most sat isfactorydesign of firetube heating elements employs a zone by zone calculation based on thermal efficiency desired and limiting the maximum tube walltempera ture a s required by the solution to prevent therm al degra dat ion. The avera ge heat flux, Q/A, is a result of these calculations.

7. Reclaimers are not used in DEA and MDEA systems.

8. Reboiler tempera tures a re dependent on solution conc. flar e/vent line back pressure a nd/or residual C O2content r equired. It is good pra ctice to operat ethe reboiler at as low a temperature as possible.

9. According t o Total.

10. The heats of reaction vary w ith a cid gas loading a nd solution concentrat ion. The values shown are a verage.10.

21 5

sierrata04_05

Documents