ARNOLD, K. (1999). Design of Gas-Handling Systems and Facilities (2nd ed.) Episode 1 Part 7 ppsx

The temperature or number of trays can then be varied until the culated outlet liquid composition equals the assumed composition, cal-and the vapor pressure of the liquid is equal to or

with the liquid is known, the composition of the feed to this tray isalso known This is the composition of the liquid falling from Tray 1.

5 Calculate the temperature of Tray 1 From an enthalpy balance, thetemperature of the liquid falling from Tray 1, and thus the tempera-ture of the flash on Tray 1, can.be calculated The composition isknown, the enthalpy can be calculated Enthalpy must be main-tained, so the enthalpy of the liquid of known composition fallingfom Tray 1 must equal the sum of the enthalpies of the liquid andgas flashing from it at known temperature

6 This procedure can then be carried on up the tower to Tray N, whichestablishes the temperature of the inlet and the gas outlet composition

7 From the composition of the inlet and gas outlet the liquid outlet, position can be calculated and compared to that assumed in step 1

com-8 The temperature or number of trays can then be varied until the culated outlet liquid composition equals the assumed composition,

cal-and the vapor pressure of the liquid is equal to or less than that

assumed If the vapor pressure of the liquid is too high, the bottomstemperature must be increased

DISTILLATION TOWER WITH REFLUX

Figure 6-5 shows a stabilizer with reflux The well fluid is heated withthe bottoms product and injected into the tower, below the top, where thetemperature in the tower is equal to the temperature of the feed Thisminimizes the amount of flashing In the tower, the action is the same as

in a cold-feed stabilizer or any other distillation tower As the liquid falls

Figure 6-5 Stabilizer with reflux and feed/bottoms heat exchanger.

through the tower, it goes from tray to tray, and gets increasingly richer

in the heavy components and increasingly leaner in the light components.The stabilized liquid is cooled in the heat exchanger by the feed streambefore flowing to the stock tank

At the top of the tower any intermediate components going out withthe gas are condensed, separated, pumped back to the tower, and sprayeddown on the top tray This liquid is called "reflux," and the two-phaseseparator that separates it from the gas is called a "reflux tank" or "refluxdram," The reflux performs the same function as the cold feed in a cold-feed stabilizer Cold liquids strip out the intermediate components fromthe gas as the gas rises

The heat required at the reboiler depends upon the amount of coolingdone in the condenser The colder the condenser, the purer the productand the larger the percentage of the intermediate components that will berecovered in the separator and kept from going out with the gas The hot-ter the bottoms, the greater the percentage of light components will beboiled out of the bottoms liquid and the lower the vapor pressure of thebottoms liquid,

A condensate stabilizer with reflux will recover more intermediatecomponents from the gas than a cold-feed stabilizer However, it requiresmore equipment to purchase, install, and operate This additional costmust be justified by the net benefit of the incremental liquid recovery;less the cost of natural gas shrinkage and loss of heating value, over thatobtained from a cold-feed stabilizer

CONDENSATE STABILIZER DESIGN

It can be seen from the previous description that the design of both acold-feed stabilizer and a stabilizer with reflux is a rather complex andinvolved procedure Distillation computer simulations are available thatcan be used to optimize the design of any stabilizer if the properties ofthe feed stream and desired vapor pressure of the bottoms product areknown Cases should be run of both a cold-feed stabilizer and one withreflux before a selection is made Because of the large number of calcula-tions required, it is not advisable to use hand calculation techniques todesign a distillation process There is too much opportunity for computa-tional error

Normally, the crude or condensate sales contract will specify a mum Reid Vapor Pressure (RVP) This pressure is measured according to

maxi-a specific ASTM testing procedure A smaxi-ample is plmaxi-aced in maxi-an evmaxi-acumaxi-atedcontainer such that the ratio of the vapor volume to the liquid volume is 4

to 1 The sample is then immersed in a !00°F liquid bath The absolutepressure then measured is the RVP of the mixture

Since a portion of the liquid was vaporized to the vapor space, the uid will have lost some of its lighter components This effectivelychanges the composition of the liquid and yields a slightly lower vaporpressure than the true vapor pressure of the liquid at 100°F Figure 6-6can be used to estimate true vapor pressure at any temperature from aknown RVP

liq-The inherent error between true vapor pressure and RVP means that astabilizer designed to produce a bottoms liquid with a true vapor pressureequal to the specified RVP will be conservatively designed The vaporpressures of various hydrocarbon components at 100°F are given inTable 6-1

The bottoms temperature of the tower can be approximated if thedesired vapor pressure of the liquid is known The vapor pressure of amixture is given by:

where VP = vapor pressure of mixture, psia

VPn = vapor pressure of component n, psia

MFn = mole fraction of component n in liquid

To estimate the desired composition of the bottom liquid, the vaporpressures of the different components at 100°F can be assumed to be ameasure of the volatility of the component Thus, if a split of n~C4 isassumed, the mole fraction of each component in the liquid can be esti-mated from:

(equation continued on page 140)

Figure 6-6 Relationship between Reid vapor pressure and actual vapor pressure.

{From Gas Processors Suppliers Association, Engineering Data Book, 9th Edition.)

Table 6-1 Vapor Pressure and Relative Volatility of Various Components

Component

C,

c>

C 3 i-C 4

CO 2 N,

H 7 S

Vapor Pressure

At !00°F psia 5000 800 190 72.2 51,6 20.4 15.6 5.0 -~0.1

—

— 394

Relative Volatility 96,9 15.5 3.68 1.40 1.00 0.40 0.30 0.10 0.00 Infinite Infinite 7.64

(equation continued from page 138)

where Fn = total number of moles of component n in feed

Ln = total number of moles of component n in the bottomliquid

(fl-C4 split) = assumed moles of component n-C4 in bottom liquid

divided by moles of n-C4 in feed

RVn = relative volatility of component n from Table 6-1

To determine the compositon of the bottom liquid, assume a split of

n-C4 and compute MFn from Equations 6-2 and 6-3 The vapor pressurecan then be computed from Equation 6-1 If the vapor pressure is higherthan the desired RVP choose a lower number for the n-C4 split If the cal-culated vapor pressure is lower than the desired RVP, choose a highernumber for the n~C4 split Iterate until the calculated vapor pressureequals the desired RVP

The bottoms temperature can then be determined by calculating thebubble point of the liquid described by the previous iteration at the cho-sen operating pressure in the tower This is done by choosing a tempera-ture, determining equilibrium constants from Chapter 3, Volume I , andcomputing:

If C is greater than 1,0, the assumed temperature is too high If C islower than 1.0, the assumed temperature is too low By iteration a tem-perature can be determined where C = 1.0.

Typically, bottoms temperatures will range from 200~4QO°F depending

on operating pressure, bottoms composition, and vapor pressure ments Temperatures should be kept to a minimum to decrease the heatrequirements, limit salt build-up, and prevent corrosion problems,

require-When stabilizer operating pressures are kept below 200 psig thereboiler temperatures will normally be below 300°R A water glycol heat-ing medium can then be used to provide the heat Higher stabilizer oper-ating pressures require the use of steam- or hydrocarbon-based heatingmediums Operating at higher pressures, however, decreases the flashing

of the feed on entering the column, which decreases the amount of feedcooling required In general, a crude stabilizer should be designed tooperate between 100 and 200 psig

TRAYS AND PACKING

The number of actual equilibrium stages determines the number offlashes that will occur The more stages, the more complete the split, butthe taller and more costly the tower Most condensate stabilizers will nor-mally contain approximately five theoretical stages In a refluxed tower,the section above the feed is known as the rectification section, while thesection below the feed is known as the stripping section The rectificationsection normally contains about two equilibrium stages above the feed,and the stripping section normally contains three equilibrium stages.Theoretical stages within a tower are provided by actual stage devices(typically either trays or packings) The actual diameter and height of thetower can be derived using manufacturer's data for the particular device.The height of the tower is a function of the number of theoretical stagesand of the efficiency of the actual stages The diameter of the tower is afunction of the hydraulic capacity of the actual stages

frays

For most trays, liquid flows across an "active area" of the tray andthen into a "downcomer" to the next tray below, etc Inlet and/or outletweirs control the liquid distribution across the tray Vapor flows up thetower and passes through the tray active area, bubbling up through (andthus contacting) the liquid flowing across the tray The vapor distribution

is controlled by (1) perforations in the tray deck (sieve trays), (2) bubblecaps (bubble cap trays), or (3) valves (valve trays),

Trays operate within a hydraulic envelope At excessively high vaporrates, liquid is carried upward from one tray to the next (essentially back-mixing the liquid phase in the tower) For valve trays and sieve trays, acapacity limit can be reached at low vapor rates when liquid falls throughthe tray floor rather than being forced across the active area into thedowncomers Because the liquid does not flow across the trays, it missescontact with the vapor, and the separation efficiency drops dramatically.Trays are generally divided into four categories: (1) sieve trays, (2) valvetrays, (3) bubble cap trays, and (4) high capacity/high efficiency trays

Sieve Trays

Sieve trays are the least expensive tray option In sieve trays, vaporflowing up through the tower contacts the liquid by passing throughsmall perforations in the tray floor (Figure 6-7b) Sieve trays rely onvapor velocity to exclude liquid from falling through the perforations inthe tray floor If the vapor velocity is much lower than design, liquid willbegin to flow through the perforations rather than into the downcomer.This condition is known as weeping Where weeping is severe, the equi-librium efficiency will be very low For this reason, sieve trays have avery small turndown ratio

Valve Trays

Valve trays are essentially modified sieve trays Like sieve trays, holesare punched in the tray floor However, these holes are much larger thanthose in sieve trays Each of these holes is fitted with a device called a

"valve." Vapor flowing up through the tower contacts the liquid by

pass-ing through valves in the tray floor (Figure 6-1 c) Valves can be fixed or

moving Fixed valves are permanently open and operate as deflectorplates for the vapor coming up through the holes in the tray floor Formoving valves, vapor passing through the tray floor lifts the valves andcontacts the liquid Moving valves come in a variety of designs, depend-ing on the manufacturer and the application At low vapor rates, valveswill close, helping to keep liquid from falling through the holes in thedeck At sufficiently low vapor rates, a valve tray will begin to weep.That is, some liquid will leak through the valves rather than flowing to

the tray downcomers At very low vapor rates, it is possible that all theliquid will fall through the valves and no liquid will reach the downcom-ers This severe weeping is known as "dumping." At this point, the effi-ciency of the tray is nearly zero.

Bubble Cap Trays

In bubble cap trays, vapor flowing up through the tower contacts theliquid by passing through bubble caps (Figure 6-7a) Each bubble capassembly consists of a riser and a cap The vapor rising through the col-umn passes up through the riser in the tray floor and then is turned down-ward to bubble into the liquid surrounding the cap Because of theirdesign, bubble cap trays cannot weep However bubble cap trays are alsomore expensive and have a lower capacity/higher pressure drop thanvalve trays or sieve trays

Figure 6-7 Vapor flow through trays,

High Capacity/High Efficiency Trays

High capacity/high efficiency trays have valves or sieve holes or both.They typically achieve higher efficiencies and capacities by takingadvantage of the active area under the downcomer At this time, each ofthe major vendors has its own version of these trays, and the designs areproprietary

Bubble Cap Trays vs Valve Trays

At low vapor rates, valve trays will weep Bubble cap trays cannotweep (unless they are damaged) For this reason, it is generally assumedthat bubble cap trays have nearly an infinite turndown ratio This is true

in absorption processes (e.g., glycol dehydration), in which it is moreimportant to contact the vapor with liquid than the liquid with vapor,However, this is not true of distillation processes (e.g., stabilization), inwhich it is more important to contact the liquid with the vapor

As vapor rates decrease, the tray activity also decreases There

eventual-ly comes a point at which some of the active devices (valves or bubblecaps) become inactive Liquid passing these inactive devices gets very lit-tle contact with vapor At very low vapor rates, the vapor activity will con-centrate only in certain sections of the tray (or, in the limit, one bubble cap

or one valve) At this point, it is possible that liquid may flow across theentire active area without ever contacting a significant amount of vapor.This will result in very low tray efficiencies for a distillation process Noth-ing can be done with a bubble cap tray to compensate for this

However, a valve tray can be designed with heavy valves and lightvalves At high vapor rates, all the valves will be open As the vapor ratedecreases, the valves will begin to close With light and heavy valves onthe tray, the heavy valves will close first, and some or all of the lightvalves will remain open If the light valves are properly distributed overthe active area, even though the tray activity is diminished at low vaporrates, what activity remains will be distributed across the tray All liquidflowing across the tray will contact some vapor, and mass transfer willcontinue Of course, even with weighted valves, if the vapor rate isreduced enough, the tray will weep and eventually become inoperable.However, with a properly designed valve tray this point may be reachedafter the loss in efficiency of a comparable bubble cap tray So, in distil-lation applications, valve trays can have a greater vapor turndown ratiothan bubble cap trays

Tray Efficiency and Tower Height

In condensate stabilizers, trays generally have 70% equilibrium stageefficiency That is, 1.4 actual trays are required to provide one theoreticalstage The spacing between trays is a function of the spray height and thedowncomer backup (the height of clear liquid established in the down-corner) The tray spacing will typically range from 20 to 30 in (with 24

in being the most common), depending on the specific design and theinternal vapor and liquid traffic The tray spacing may increase at higheroperating pressures (greater than 165 psia) because of the difficulty indisengaging vapor from liquid on both the active areas and in the down-comers,

Packing

Packing typically comes in two types: random and structured

Liquid distribution in a packed bed is a function of the internalvapor/liquid traffic, the type of packing employed, and the quality of theliquid distributors mounted above the packed bed Vapor distribution iscontrolled by the internal vapor/liquid traffic, by the type of packingemployed, and by the quality of the vapor distributors located below thepacked beds

Packing material can be plastic, metal, or ceramic Packing cies can be expressed as HETP (height equivalent to a theoretical plate),

efficien-Random Packing

A bed of random packing typically consists of a bed support (typically

a gas injection support plate) upon which pieces of packing material arerandomly arranged (they are usually poured or dumped onto this supportplate) Bed limiters, or hold-downs, are sometimes set above randombeds to prevent the pieces of packing from migrating or entrainingupward Random packing comes in a variety of shapes and sizes For agiven shape (design) of packing, small sizes have higher efficiencies andlower capacities than large sizes

Figure 6-8 shows a variety of random packing designs An earlydesign is known as a Raschig ring Raschig rings are short sections oftubing and are low-capacity, low-efficiency, high-pressure drop devices.Today's industry standard is the slotted metal (Pall) ring A packed bedmade of 1-in slotted metal rings will have a higher mass transfer effi-ciency and a higher capacity than will a bed of 1-in Raschig rings The

Figure 6-8 Various types of packing.

HETP for a 2-in, slotted metal ring in a condensate stabilizer is about 36

in This is slightly more than a typical tray design, which would require

34 in (1.4 trays x 24-in tray spacing) for one theoretical plate or stage

Structured Packing

A bed of structured packing consists of a bed support upon which ments of structured packing are placed Beds of structured packing typi-cally have lower pressure drops than beds of random packing of compa-rable mass transfer efficiency Structured packing elements are composed

ele-of grids (metal or plastic) or woven mesh (metal or plastic) or ele-of thin tical crimped sheets (metal, plastic, or ceramic) stacked parallel to eachother Figure 6-9 shows examples of the vertical crimped sheet style ofstructured packing

ver-Figure 6-9 Structured packing can offer better mass transfer than trays (Courtesy

of Koch Engineering Co., Inc.]

The grid types of structured packing have very high capacities and verylow efficiencies, and are typically used for heat transfer or for vapor scrub-bing The wire mesh and the crimped sheet types of structured packing typ-ically have lower capacities and higher efficiencies than the grid type