- Email: [email protected]
Condensate stabilization
CHAPTER
6
CONDENSATE STABILIZATION
6.1 INTRODUCTION Hydrocarbon condensate recovered from natural gas may be shipped without further processing but is stabilized often for blending into the crude oil stream and thereby sold as crude oil. In the case of raw condensate, there are no particular specifications for the product other than the process requirements. The process of increasing the amount of intermediates (C3 to C5) and heavy (C^) components in the condensate is called "condensate stabilization." This process is performed primarily in order to reduce the vapor pressure of the condensate liquids so that a vapor phase is not produced upon flashing the liquid to atmospheric storage tanks. In other word, the scope of this process is to separate the very light hydrocarbon gases, methane and ethane in particular, from the heavier hydrocarbon components (C^). Stabilized liquid, however, generally has a vapor pressure specification, as the product will be injected into a pipeline or transport pressure vessel, which has definite pressure limitations. Condensates may contain a relatively high percentage of intermediate components and can be separated easily from entrained water due to its lower viscosity and greater density difference with water. Thus, some sort of condensate stabilization should be considered for each gas well production facility. The purpose of this chapter is to describe the basic processes used to condensate stabilization and associated equipment design procedure.
247
248
Handbook of Natural Gas Transmission and Processing
6.2 STABILIZATION PROCESSES Stabilization of condensate streams can be accomplished through either flash vaporization or fractionation. 6.2.1 Flash Vaporization Stabilization by flash vaporization is a simple operation employing only two or three flash tanks. This process is similar to stage separation utilizing the equilibrium principles between vapor and condensate phases. Equilibrium vaporization occurs when the vapor and condensate phases are in equilibrium at the temperature and pressure of separation. Figure 6-1 shows a typical scheme of condensate stabilization through the flash vaporization process. As shown, condensate from the inlet separator after passing through the exchanger enters to the high-pressure flash tank, where the pressure is maintained at 600 psia. A pressure drop of 300 psia is obtained here, which assists flashing of large amounts of lighter ends, which join the sour vapor stream after recompression. The vapor can either be processed further and put into the sales gas or be recycled into the reservoir and used as gas lift to produce more crude oils. The bottom liquid from the high-pressure tank flows to the middle pressure flash tank operated at 300 psia. Additional methane and ethane are released in this tank. The bottom product is withdrawn again to the low-pressure tank
Flash gas to sweetening
To high pressure fuel gas system To low pressure fuel gas system
Condensate from Inlet separator
HP Flash tank
MP Flash tank
LP Condensatee Storage tank Flash tank stripper
Figure 6 - 1 . Schematic of condensate stabilization through ash vaporization process. H.P., high pressure; M.R, middle pressure; L.P., low pressure.
Condensate Stabilization
249
Operated at 65 psia. To ensure efficient separation, condensate is degassed in the stripper vessel at the lowest possible pressure prior to storage. This reduces excess flashing of condensate in the storage tank and reduces the inert gas blanket pressure required in it. Note that flash vaporization as a condensate stabilization method is old technology and is not used in a modem gas plant. However, variations of flash technology might also be found on oil production facilities stabilizing crude oil. 6.2.2 Stabilization by Fractionation
Stabilization by fractionation is a detailed process, very popular in the industry and precise enough to produce liquids of suitable vapor pressure. During the operation, light fractions such as methane-ethane-propane and most of the butanes are removed and recovered. The finished product from the bottom of the column is composed mainly of pentanes and heavier hydrocarbons, with small amounts of butane. The process actually makes a cut between the lightest liquid component (pentane) and the heaviest gas (butane). The bottom product is thus a liquid free of all gaseous components able to be stored safely at atmospheric pressure. Stabilization by fractionation is a modem operation and economically attractive next to flash vaporization. It is a single tower process, as only one specification product is required. The bottom product of the column is capable of meeting any kind of rigid specifications with the proper operating conditions. 6.2.2.1 Process
Description
Figure 6-2 shows a schematic condensate stabilization process. The liquid hydrocarbon (condensate) is brought into the system from the inlet separator and preheated in the stabilizer feed/bottoms exchanger before entering the stabilizer feed dmm.^ Liquid from the feed dmm is fed to the stabilization tower at approximately 50 to 200 psi depending on whether they are sour (sour stabilization is carried out at the low end of the range and sweet stabilization at the high end of the range). ^Sometimes the liquid is flashed down to a feed drum at pressure sHghtly above the tower pressure. This flashes off vapor so that the stabiUzation tower can often be a smaller diameter.
250
Handbook of Natural Gas Transmission and Processing To high pressure fuel gas system A
—^ Y i Condensate from inlet separator *
^ Exchanger
To low pressure ^uel gas system
gau
(Feed drum'
Reboiler & Cooler/ ^ ^ Storage tank Figure 6 - 2 , Schematic of a condensate stabilization system.
The condensate stabilizer reduces vapor pressure of the condensate by removing the lighter components. The stabilization is typically carried out in a reboiled absorber, with tray type internals. However, if a better separation is required, typically the column is changed from a top feed reboiled absorber to a refluxed distillation tower. As the liquid falls into the column, it becomes leaner in light ends and richer in heavy ends. At the bottom of the tower some of the liquid is circulated through a reboiler to add heat to the tower. As the gas goes up from tray to tray, more and more of the heavy ends get stripped out of the gas at each tray and the gas becomes richer in the light ends and leaner in the heavy ends. Overhead gas from the stabilizer, which would seldom meet market specifications for the natural gas market, is then sent to the low-pressure fuel gas system through a back-pressure control valve that maintains the tower pressure to set point. Liquids leaving the bottom of the tower have undergone a series of stage flashes at ever-increasing temperatures, driving off the light components, which exit the top of the tower. These liquids must be cooled to a sufficiently low temperature to keep vapors from flashing to atmosphere in the condensate storage tank.
Condensate Stabilization
251
Most gas processing plants are implementing advanced process control^ (APC) on their condensate production systems in order to maximize condensate yields, improve stability of the condensate stabilization process, and ensure that product quality limits are adhered to at all times. However, most often, the main reason to implement APC on this unit is product quality and control, where the gas suppliers sell the condensate product with a Reid vapor pressure (RVP) defined by the customer. In this instance, a distributed control system is used by the APC system to maintain RVP within quality limits and push this to a higher specification; therefore, generating greater revenues as condensate throughput is increased (Hotblack, 2004). From a control standpoint, some optimization schemes include: • Nonlinear level control concepts on the feed drum can smooth the flow to the stabilizer column. This allows the feed drum to be truly used as a capacitance. The feed rate to the stabilizer is not changed if the level in the feed drum is within a dead band and not changing too quickly. As the level approaches the dead band limits, flow is gradually changed. As the level moves outside the dead band, then the feed rate is moved more aggressively. • Predictive models can be employed to anticipate the effects of feed rate and compositional effects on the bottom composition. • An inferential property for bottom RVP can be determined based on trays with sensitive pressure-compensated temperatures (PCT) to improve the control of the product quality. Alternatively, the most sensitive tray temperature or PCT can be cascaded to the flow of heat input to the reboiler. • Separation efficiency is improved as tower pressure is lowered. This can reduce heat requirements for reboiling while making a better separation between light and heavy key component in the tower. The pressure can be lowered subject to maximum valve opening of the pressure control valve or column flooding as indicated by differential pressure
^APC technology uses a multivariable control technique based on a Hnear dynamic process model. Using this predictive model, the controller is able to calculate an optimum set of manipulated variable moves, which minimize the error between actual and desired process behavior subject to process constraints. The controller is able to take account of process interactions and overcome process disturbances to reduce the standard deviation of key controlled variables.
252
Handbook of Natural Gas Transmission and Processing
measurement across the tower. The tower is more susceptible to flooding at lower pressures. If an overhead compressor is employed to boost the tower overhead up to fuel system pressure, then the pressure can be lowered subject to the maximum speed of the compressor driver, rod load limitations for a reciprocating compressor, or surge considerations for a centrifugal compressor. 6.2.2.2 Design Considerations of Stabilization Column In most cases of lease operation, the stabilization column will operate as a nonrefluxed tower. This type of operation is simpler but less efficient than the refluxed tower operation. Because the nonrefluxed tower requires no external cooling source, it is particularly applicable to remote locations. A condensate stabilization column with reflux will recover more intermediate components from the gas than a cold-feed stabilizer. However, it requires more equipment to purchase, install, and operate. This additional cost must be justified by the net benefit of the incremental liquid recovery, less the cost of natural gas shrinkage and loss of heating value, over that obtained from a cold-feed stabilizer. When a condenser is used in a stabilization column, it will always be a partial condenser because of the quantities of methane and ethane that must be removed from the tower feed. The stabilization tower pressure depends on the amount of liquid to be stabilized and whether it is sweet or sour. For sweet stabilization, the pressure should be as high as possible to minimize overhead vapor recompression, as this gas is remixed with the separator vapor. This also tends to decrease the cost of reflux cooling, if it is used. However, relative volatility of the components also decreases with pressure and, as stated previously, driving H2S overhead requires a relatively low pressure. Figure 6-3 shows the maximum recommended feed temperature to a stabilizer as a function of operating pressure of the stabilizer. An exception to this may be the case where either sour or small quantities of liquid are being handled and where first cost therefore is very critical. In these cases, many times a 40- to 70- psia working pressure for a nonrefluxed tower may show an economic advantage (Campbell, 1992). In some cases the reboiler for the stabilizer will be an indirect salt bath heater or a steam-fired heat exchanger. Figure 6-4 shows suggested bottom (reboiler) temperatures for producing a specified Reid vapor pressure product. In fact, the temperature used on the bottom is limited by
Condensate 260
Stabilization
253
I [ I I I I I I I I I I I I I I [[
I [ I I I I T=x=n
220 I I I I I M I I I I I I i I M I I M M I il I I 180
i
140
CO CO
9>
100
60
20 0
10
20
30
40
50
60
Feed Temperature, T Figure 6-3. Maximum recommended feed temperature to a cold-feed stabilizer (Campbell 1992).
the thermal breakdown characteristics of the condensate and, of course, product specs. After the pressure has been chosen and the operating temperatures have been estabUshed through use of Figures 6-3 and 6-4, the spHt in the tower must be predicted. There are several methods in which this can be done, but one of the most convenient manual methods involves utilization of pseudo-equilibrium constant (K) values for each component between the top and the bottom of the tower. Using this concept, the separation that can be achieved across a nonrefluxed stabilizer can be estimated by use of the pseudo K values and a simple flash calculation. The vapor from the flash calculation will be the composition of the overhead product, and the liquid from the flash calculation will be the composition of the bottom liquid. Today, distillation and absorption/stripping calculations are done most often with process simulation computer software. Note that for estimating the desired composition of the bottom liquid if a split of nC4 (normal butane) is assumed, the mole fraction of each
Handbook of Natural Gas Transmission and Processing
254 360
340 ^^^^^^^^^^fflR|\ 'Vl^ml Tn 11! 111
m
320
[[ .|r|.|i.n.T^
gf 300 CD
JMHM IlillOI-lliTllHilt^f
I 280
[JifiTMrliilB^ffi
g o
tUiii
^^^B
260
1 j.i,i 11
240
^|d^^^W
1^^^^^^ffi mtiiTi 1 n 11 i 1 mir
220
200
itXfi- fflr
140
Wftftr :|:t:i:j:[ln|'lwffltftt fflffi+ffltntrf
160
180
200
220
240
m^ 260
280
Operating Pressure, Psia Figure 6-4. Estimation of proper bottom temperature of a nonre uxed stabilizer (Campbell, 1992).
component in the liquid can be estimated from the following equations: Li =
F,(nC4 split) RV, Xi =
Li
(6-1) (6-2)
ULi i=l
where X/ is the mole fraction of component / in the Hquid, Ft is total number of moles of component / in the feed, L/ is total number of moles of component / in the bottom liquid, nC4 split is assumed moles of component nC4 in the bottom liquid divided by moles of nC4 in the feed, RV/ is relative volatility of component / from Table 6-1, and n is number of components in the bottom liquid.
Condensate Stabilization
255
Table 6-1 RVP and Relative Volatility of Various Components (Reid ef aL, 1977) Component Ci C2 C3 i-C4 n-C4 i-C5 n-Cs C6 Cv^ CO2 N2 H2S
RVP psia
Relative volatility
5000 800 190 72.2 51.6 20.4 15.6 5.0 «0.1 — — 394
96.9 15.5 3.68 1.40 1.00 0.40 0.30 0.10 0.00 Infinite Infinite 7.64
The vapor pressure is the primary property used to make this spHt. It is assumed that the mole fraction of each component times its vapor pressure represents the contribution of that component to the total mixture vapor pressure. The total mixture vapour pressure can then be computed from Equation (6-3):
Pv = Yl (Pvi >< ^/)
(6-3)
/=i
where Py is vapor pressure of mixture, psia, and Pyt is vapor pressure of component /, psia. If the vapor pressure of the mixture is higher than the desired RVP of the bottom Uquid, choose a lower number for the nC4 split. If the calculated vapor pressure is lower than the desired RVP, choose a higher number for the nC4 split. Iterate until the calculated vapor pressure equals the desired RVP of the bottom liquid. The bottom liquid temperature can also be determined by calculating the bubble point of the liquid described by the previous iteration at the chosen operating pressure in the tower. This is done by choosing a temperature, determining pseudo K values from Figure 6-5, and computing
256
Handbook of Natural Gas Transmission and Processing
10.0
RVV
t
\
\ Q)
I
1.0
\ \
partii \l de- <^3
y
\
pakial d}B-C4
^
0.1
CD
^ de-
0.01
V > \ \ \ \ \
0.001 -200-100
0
100 200
L
300
400
Normal Boiling Point, T Figure 6 - 5 . Pseudo /C values for cold feed stabilizers (Campbell, 1992).
parameter C by the following equation:
C = J2(LixKi)
(6-4)
/=i
If C is greater than 1.0, the assumed temperature is too high. If C is lower than 1.0, the assumed temperature is too low. By iteration a temperature can be determined where C = 1.0. Typically, bottoms temperatures will range from 200 to 400°F depending on operating pressure, bottoms composition, and vapor pressure requirements. Temperatures should be kept to a minimum to decrease the heat requirements, Umit salt buildup, and prevent corrosion problems. 6.3 CONDENSATE STORAGE
Condensate is stored between production and shipping operations in condensate storage tanks, which are usually offloatingroof type (external and internal). If the condensate does not meet the specifications, the offspecification condensate may be routed to an off-specification condensate
257
Condensate Stabilization
Storage fixed roof tank (vertical and horizontal) until it is recycled to the condensate stabilization unit by the relevant recycle pump if the latter is available at the plant. The primary quality criterion for the condensate is its RVP, which is affected by atmospheric pressure (plant elevation) and maximum abmient temperature. To store the condensate in floating roof storage tanks, it is very crucial to control the RVP at the desired level (especially in warm seasons). Emissions from condensate storage tanks are normally categorized as occurring from breathing losses (standing storage losses) or working losses. The term breathing loss refers to those emissions that result without any corresponding change in the liquid level within the tank. Most Ukely, these types of emissions result from hydrocarbon vapors that are released from the tank by expansion or contraction caused by changes in either temperature or pressure. Working loss represents those emissions that occur due to changes in the liquid level caused by either filling or emptying the tank itself (U.S. EPA's AP-42 manual). For floating roof tanks, breathing losses are a result of evaporative losses through rim seals, deck fittings, and deck seam losses. Withdrawal losses occur as the level drops, and thus the floating roof is lowered. Some liquid remains on the inner tank wall surface and evaporates when the tank is emptied. For an internal floating roof tank that has a column-supported fixed roof, some liquid also clings to the columns and evaporates. Evaporative loss occurs until the tank is filled and the exposed surfaces are again covered (SCAQMD supplemental manual, 2003). The working pressure required to prevent breathing and thereby save standing storage losses depends on the vapor pressure of the product, the temperature variations of the liquid surface and the vapor space, and the setting of the vacuum vent. When these factors are known, the storage pressure required to eliminate venting can be computed by the Equation (6-5) (GPSA, 1998): Ps = Pv,m^x + {PT -
PvMn)
7;mn+460
-P. atm
(6-5)
where Ps is required storage pressure, psia; Py,max is true vapor pressure of liquid at maximum surface temperature, psia; Py,min is true vapor pressure of Uquid at minimum surface temperature, psia; Tmax is maximum average temperature of vapor, °F; Tmin is minimum average temperature of vapour, °F; Patm is atmospheric pressure, psia; and Pj is absolute internal tank pressure at which vacuum vent opens, psia.
258
Handbook of Natural Gas Transmission and Processing
For the condition where Pv,rmn is greater than P j , Equation (6-5) can be represented as Ps = Pv,max - P^tm
(6-6)
With the situation represented by Equation (6-5), gas is admitted to the vapor space through the vacuum vent, where a gas vent system should be employed for preventing safety problems. However, for the situation represented by Equation (6-6), gas may be purged and kept out of the tank (Campbell, 1992). Maximum liquid surface temperatures vary from 85 to 115°F. Sufficient accuracy will generally result from the assumption that it is 10°F higher than the maximum temperature of the body of the liquid in a tank at that location. Note that true vapor pressure (TVP) may be the most difficult term in the aforementioned equation to calculate. A nomograph has been devised that relates TVP to both the Reid vapor pressure and the storage temperature (Ts). Numerically, the relationship between TVP, RVP and storage temperature can be expressed as follows (TRW Environmental, Inc., 1981). TVP = (RVP) EXP { Co (ci - j ^ \
j
(6-7)
where Co is constant dependent on the value of RVP, Ci = l/(rs + 460), and T^ is temperature of the stored liquid , °F. In the aforementioned equation, the term Co is dependent on the given value of RVP as shown in Table 6-2 (TRW Environmental, Inc., 1981). Once the parameters Co, Ts, and RVP have been determined, one is technically able to calculate a value for TVP. It should be noted, however, that an error was discovered in the API nomograph-calculated values of TVP, such that the RVP was not equal to the TVP at 100°F as should be expected, given the general definition of RVP. Using linear regression methods, a correction factor (Cp) was developed that should be added to the calculated values of TVP in order to obtain correct TVP values as follows (TRW Environmental, Inc., 1981). Corrected TVP = Calculated TVP + Cp
(6-8)
Condensate Stabilization
259
Table 6-2 Co for Different RVP Values RVP 2 < RVP < 3 RVP=3 3
Co
RVP
Co
-6439.2 -6255.9 -6212.1 -6169.2 -6177.9 -6186.5 -6220.4 -6254.3 -6182.1
RVP = 7 715
-6109.8 -6238.9 -6367.9 -6477.5 -6587.0 -6910.5 -7334.0 -8178.0 -9123.2
The correction factor was found to be dependent on RVP according to the following equations: RVP < 3 : CF = 0.04(RVP) + 0.1 RVP > 3 : CF = EXP{2.3452061og(RVP) - 4.132622}
(6-9) (6-10)
It is important to mention that the methods used for calculating each tank emissions are described in detail in the U.S. EPA's AP-42 manual. In this manual, equations are developed to calculate emissions for fixed roof tanks. By assuming that open top tanks and tanks with open holes or roof openings do not have emissions greater than those for fixed roof tanks, the State California Air Resource Board was able to use the AP-42 equations to calculate emissions for all of the storage tanks in a particular oil production field. REFERENCES
Campbell, J.M., "Gas Conditioning and Processing," 3rd Ed. Campbell Petroleum Series, Norman, OK (1992). GPSA Engineering Data Book, 11th Edition, Vol. 1, Section 6, Gas Processors Suppliers Association, Tulsa, OK (1998). Hotblack, C , BG Tunisia's advanced process control improves condensate product stabiUty. World Oil 225, 9 (2004).
260
Handbook of Natural Gas Transmission and Processing
Reid, R., Prausnitz, J.M., and Sherwood, T., "The Properties of Gases and Liquids," 3rd Ed. McGraw-Hill, New York (1977). SCAQMD Supplemental Manual, "Supplemental Instructions for Liquid Organic Storage Tanks and References." Annual Emissions Reporting Program, South Coast Air Quality Management District, CA (June 2003). TRW Environmental, Inc., "Background Documentation for Storage of Organic Liquids." EPA Contract No. 68-02-3174, NC (May 1981). U.S. EPA's AP-42 Manual, "Compilation of Air Pollutant Emission Factors," Chapter 7, Vol. 1, 5th Ed. U.S. Environmental Protection Agency, Washington, DC.
6
CONDENSATE STABILIZATION
6.1 INTRODUCTION Hydrocarbon condensate recovered from natural gas may be shipped without further processing but is stabilized often for blending into the crude oil stream and thereby sold as crude oil. In the case of raw condensate, there are no particular specifications for the product other than the process requirements. The process of increasing the amount of intermediates (C3 to C5) and heavy (C^) components in the condensate is called "condensate stabilization." This process is performed primarily in order to reduce the vapor pressure of the condensate liquids so that a vapor phase is not produced upon flashing the liquid to atmospheric storage tanks. In other word, the scope of this process is to separate the very light hydrocarbon gases, methane and ethane in particular, from the heavier hydrocarbon components (C^). Stabilized liquid, however, generally has a vapor pressure specification, as the product will be injected into a pipeline or transport pressure vessel, which has definite pressure limitations. Condensates may contain a relatively high percentage of intermediate components and can be separated easily from entrained water due to its lower viscosity and greater density difference with water. Thus, some sort of condensate stabilization should be considered for each gas well production facility. The purpose of this chapter is to describe the basic processes used to condensate stabilization and associated equipment design procedure.
247
248
Handbook of Natural Gas Transmission and Processing
6.2 STABILIZATION PROCESSES Stabilization of condensate streams can be accomplished through either flash vaporization or fractionation. 6.2.1 Flash Vaporization Stabilization by flash vaporization is a simple operation employing only two or three flash tanks. This process is similar to stage separation utilizing the equilibrium principles between vapor and condensate phases. Equilibrium vaporization occurs when the vapor and condensate phases are in equilibrium at the temperature and pressure of separation. Figure 6-1 shows a typical scheme of condensate stabilization through the flash vaporization process. As shown, condensate from the inlet separator after passing through the exchanger enters to the high-pressure flash tank, where the pressure is maintained at 600 psia. A pressure drop of 300 psia is obtained here, which assists flashing of large amounts of lighter ends, which join the sour vapor stream after recompression. The vapor can either be processed further and put into the sales gas or be recycled into the reservoir and used as gas lift to produce more crude oils. The bottom liquid from the high-pressure tank flows to the middle pressure flash tank operated at 300 psia. Additional methane and ethane are released in this tank. The bottom product is withdrawn again to the low-pressure tank
Flash gas to sweetening
To high pressure fuel gas system To low pressure fuel gas system
Condensate from Inlet separator
HP Flash tank
MP Flash tank
LP Condensatee Storage tank Flash tank stripper
Figure 6 - 1 . Schematic of condensate stabilization through ash vaporization process. H.P., high pressure; M.R, middle pressure; L.P., low pressure.
Condensate Stabilization
249
Operated at 65 psia. To ensure efficient separation, condensate is degassed in the stripper vessel at the lowest possible pressure prior to storage. This reduces excess flashing of condensate in the storage tank and reduces the inert gas blanket pressure required in it. Note that flash vaporization as a condensate stabilization method is old technology and is not used in a modem gas plant. However, variations of flash technology might also be found on oil production facilities stabilizing crude oil. 6.2.2 Stabilization by Fractionation
Stabilization by fractionation is a detailed process, very popular in the industry and precise enough to produce liquids of suitable vapor pressure. During the operation, light fractions such as methane-ethane-propane and most of the butanes are removed and recovered. The finished product from the bottom of the column is composed mainly of pentanes and heavier hydrocarbons, with small amounts of butane. The process actually makes a cut between the lightest liquid component (pentane) and the heaviest gas (butane). The bottom product is thus a liquid free of all gaseous components able to be stored safely at atmospheric pressure. Stabilization by fractionation is a modem operation and economically attractive next to flash vaporization. It is a single tower process, as only one specification product is required. The bottom product of the column is capable of meeting any kind of rigid specifications with the proper operating conditions. 6.2.2.1 Process
Description
Figure 6-2 shows a schematic condensate stabilization process. The liquid hydrocarbon (condensate) is brought into the system from the inlet separator and preheated in the stabilizer feed/bottoms exchanger before entering the stabilizer feed dmm.^ Liquid from the feed dmm is fed to the stabilization tower at approximately 50 to 200 psi depending on whether they are sour (sour stabilization is carried out at the low end of the range and sweet stabilization at the high end of the range). ^Sometimes the liquid is flashed down to a feed drum at pressure sHghtly above the tower pressure. This flashes off vapor so that the stabiUzation tower can often be a smaller diameter.
250
Handbook of Natural Gas Transmission and Processing To high pressure fuel gas system A
—^ Y i Condensate from inlet separator *
^ Exchanger
To low pressure ^uel gas system
gau
(Feed drum'
Reboiler & Cooler/ ^ ^ Storage tank Figure 6 - 2 , Schematic of a condensate stabilization system.
The condensate stabilizer reduces vapor pressure of the condensate by removing the lighter components. The stabilization is typically carried out in a reboiled absorber, with tray type internals. However, if a better separation is required, typically the column is changed from a top feed reboiled absorber to a refluxed distillation tower. As the liquid falls into the column, it becomes leaner in light ends and richer in heavy ends. At the bottom of the tower some of the liquid is circulated through a reboiler to add heat to the tower. As the gas goes up from tray to tray, more and more of the heavy ends get stripped out of the gas at each tray and the gas becomes richer in the light ends and leaner in the heavy ends. Overhead gas from the stabilizer, which would seldom meet market specifications for the natural gas market, is then sent to the low-pressure fuel gas system through a back-pressure control valve that maintains the tower pressure to set point. Liquids leaving the bottom of the tower have undergone a series of stage flashes at ever-increasing temperatures, driving off the light components, which exit the top of the tower. These liquids must be cooled to a sufficiently low temperature to keep vapors from flashing to atmosphere in the condensate storage tank.
Condensate Stabilization
251
Most gas processing plants are implementing advanced process control^ (APC) on their condensate production systems in order to maximize condensate yields, improve stability of the condensate stabilization process, and ensure that product quality limits are adhered to at all times. However, most often, the main reason to implement APC on this unit is product quality and control, where the gas suppliers sell the condensate product with a Reid vapor pressure (RVP) defined by the customer. In this instance, a distributed control system is used by the APC system to maintain RVP within quality limits and push this to a higher specification; therefore, generating greater revenues as condensate throughput is increased (Hotblack, 2004). From a control standpoint, some optimization schemes include: • Nonlinear level control concepts on the feed drum can smooth the flow to the stabilizer column. This allows the feed drum to be truly used as a capacitance. The feed rate to the stabilizer is not changed if the level in the feed drum is within a dead band and not changing too quickly. As the level approaches the dead band limits, flow is gradually changed. As the level moves outside the dead band, then the feed rate is moved more aggressively. • Predictive models can be employed to anticipate the effects of feed rate and compositional effects on the bottom composition. • An inferential property for bottom RVP can be determined based on trays with sensitive pressure-compensated temperatures (PCT) to improve the control of the product quality. Alternatively, the most sensitive tray temperature or PCT can be cascaded to the flow of heat input to the reboiler. • Separation efficiency is improved as tower pressure is lowered. This can reduce heat requirements for reboiling while making a better separation between light and heavy key component in the tower. The pressure can be lowered subject to maximum valve opening of the pressure control valve or column flooding as indicated by differential pressure
^APC technology uses a multivariable control technique based on a Hnear dynamic process model. Using this predictive model, the controller is able to calculate an optimum set of manipulated variable moves, which minimize the error between actual and desired process behavior subject to process constraints. The controller is able to take account of process interactions and overcome process disturbances to reduce the standard deviation of key controlled variables.
252
Handbook of Natural Gas Transmission and Processing
measurement across the tower. The tower is more susceptible to flooding at lower pressures. If an overhead compressor is employed to boost the tower overhead up to fuel system pressure, then the pressure can be lowered subject to the maximum speed of the compressor driver, rod load limitations for a reciprocating compressor, or surge considerations for a centrifugal compressor. 6.2.2.2 Design Considerations of Stabilization Column In most cases of lease operation, the stabilization column will operate as a nonrefluxed tower. This type of operation is simpler but less efficient than the refluxed tower operation. Because the nonrefluxed tower requires no external cooling source, it is particularly applicable to remote locations. A condensate stabilization column with reflux will recover more intermediate components from the gas than a cold-feed stabilizer. However, it requires more equipment to purchase, install, and operate. This additional cost must be justified by the net benefit of the incremental liquid recovery, less the cost of natural gas shrinkage and loss of heating value, over that obtained from a cold-feed stabilizer. When a condenser is used in a stabilization column, it will always be a partial condenser because of the quantities of methane and ethane that must be removed from the tower feed. The stabilization tower pressure depends on the amount of liquid to be stabilized and whether it is sweet or sour. For sweet stabilization, the pressure should be as high as possible to minimize overhead vapor recompression, as this gas is remixed with the separator vapor. This also tends to decrease the cost of reflux cooling, if it is used. However, relative volatility of the components also decreases with pressure and, as stated previously, driving H2S overhead requires a relatively low pressure. Figure 6-3 shows the maximum recommended feed temperature to a stabilizer as a function of operating pressure of the stabilizer. An exception to this may be the case where either sour or small quantities of liquid are being handled and where first cost therefore is very critical. In these cases, many times a 40- to 70- psia working pressure for a nonrefluxed tower may show an economic advantage (Campbell, 1992). In some cases the reboiler for the stabilizer will be an indirect salt bath heater or a steam-fired heat exchanger. Figure 6-4 shows suggested bottom (reboiler) temperatures for producing a specified Reid vapor pressure product. In fact, the temperature used on the bottom is limited by
Condensate 260
Stabilization
253
I [ I I I I I I I I I I I I I I [[
I [ I I I I T=x=n
220 I I I I I M I I I I I I i I M I I M M I il I I 180
i
140
CO CO
9>
100
60
20 0
10
20
30
40
50
60
Feed Temperature, T Figure 6-3. Maximum recommended feed temperature to a cold-feed stabilizer (Campbell 1992).
the thermal breakdown characteristics of the condensate and, of course, product specs. After the pressure has been chosen and the operating temperatures have been estabUshed through use of Figures 6-3 and 6-4, the spHt in the tower must be predicted. There are several methods in which this can be done, but one of the most convenient manual methods involves utilization of pseudo-equilibrium constant (K) values for each component between the top and the bottom of the tower. Using this concept, the separation that can be achieved across a nonrefluxed stabilizer can be estimated by use of the pseudo K values and a simple flash calculation. The vapor from the flash calculation will be the composition of the overhead product, and the liquid from the flash calculation will be the composition of the bottom liquid. Today, distillation and absorption/stripping calculations are done most often with process simulation computer software. Note that for estimating the desired composition of the bottom liquid if a split of nC4 (normal butane) is assumed, the mole fraction of each
Handbook of Natural Gas Transmission and Processing
254 360
340 ^^^^^^^^^^fflR|\ 'Vl^ml Tn 11! 111
m
320
[[ .|r|.|i.n.T^
gf 300 CD
JMHM IlillOI-lliTllHilt^f
I 280
[JifiTMrliilB^ffi
g o
tUiii
^^^B
260
1 j.i,i 11
240
^|d^^^W
1^^^^^^ffi mtiiTi 1 n 11 i 1 mir
220
200
itXfi- fflr
140
Wftftr :|:t:i:j:[ln|'lwffltftt fflffi+ffltntrf
160
180
200
220
240
m^ 260
280
Operating Pressure, Psia Figure 6-4. Estimation of proper bottom temperature of a nonre uxed stabilizer (Campbell, 1992).
component in the liquid can be estimated from the following equations: Li =
F,(nC4 split) RV, Xi =
Li
(6-1) (6-2)
ULi i=l
where X/ is the mole fraction of component / in the Hquid, Ft is total number of moles of component / in the feed, L/ is total number of moles of component / in the bottom liquid, nC4 split is assumed moles of component nC4 in the bottom liquid divided by moles of nC4 in the feed, RV/ is relative volatility of component / from Table 6-1, and n is number of components in the bottom liquid.
Condensate Stabilization
255
Table 6-1 RVP and Relative Volatility of Various Components (Reid ef aL, 1977) Component Ci C2 C3 i-C4 n-C4 i-C5 n-Cs C6 Cv^ CO2 N2 H2S
RVP psia
Relative volatility
5000 800 190 72.2 51.6 20.4 15.6 5.0 «0.1 — — 394
96.9 15.5 3.68 1.40 1.00 0.40 0.30 0.10 0.00 Infinite Infinite 7.64
The vapor pressure is the primary property used to make this spHt. It is assumed that the mole fraction of each component times its vapor pressure represents the contribution of that component to the total mixture vapor pressure. The total mixture vapour pressure can then be computed from Equation (6-3):
Pv = Yl (Pvi >< ^/)
(6-3)
/=i
where Py is vapor pressure of mixture, psia, and Pyt is vapor pressure of component /, psia. If the vapor pressure of the mixture is higher than the desired RVP of the bottom Uquid, choose a lower number for the nC4 split. If the calculated vapor pressure is lower than the desired RVP, choose a higher number for the nC4 split. Iterate until the calculated vapor pressure equals the desired RVP of the bottom liquid. The bottom liquid temperature can also be determined by calculating the bubble point of the liquid described by the previous iteration at the chosen operating pressure in the tower. This is done by choosing a temperature, determining pseudo K values from Figure 6-5, and computing
256
Handbook of Natural Gas Transmission and Processing
10.0
RVV
t
\
\ Q)
I
1.0
\ \
partii \l de- <^3
y
\
pakial d}B-C4
^
0.1
CD
^ de-
0.01
V > \ \ \ \ \
0.001 -200-100
0
100 200
L
300
400
Normal Boiling Point, T Figure 6 - 5 . Pseudo /C values for cold feed stabilizers (Campbell, 1992).
parameter C by the following equation:
C = J2(LixKi)
(6-4)
/=i
If C is greater than 1.0, the assumed temperature is too high. If C is lower than 1.0, the assumed temperature is too low. By iteration a temperature can be determined where C = 1.0. Typically, bottoms temperatures will range from 200 to 400°F depending on operating pressure, bottoms composition, and vapor pressure requirements. Temperatures should be kept to a minimum to decrease the heat requirements, Umit salt buildup, and prevent corrosion problems. 6.3 CONDENSATE STORAGE
Condensate is stored between production and shipping operations in condensate storage tanks, which are usually offloatingroof type (external and internal). If the condensate does not meet the specifications, the offspecification condensate may be routed to an off-specification condensate
257
Condensate Stabilization
Storage fixed roof tank (vertical and horizontal) until it is recycled to the condensate stabilization unit by the relevant recycle pump if the latter is available at the plant. The primary quality criterion for the condensate is its RVP, which is affected by atmospheric pressure (plant elevation) and maximum abmient temperature. To store the condensate in floating roof storage tanks, it is very crucial to control the RVP at the desired level (especially in warm seasons). Emissions from condensate storage tanks are normally categorized as occurring from breathing losses (standing storage losses) or working losses. The term breathing loss refers to those emissions that result without any corresponding change in the liquid level within the tank. Most Ukely, these types of emissions result from hydrocarbon vapors that are released from the tank by expansion or contraction caused by changes in either temperature or pressure. Working loss represents those emissions that occur due to changes in the liquid level caused by either filling or emptying the tank itself (U.S. EPA's AP-42 manual). For floating roof tanks, breathing losses are a result of evaporative losses through rim seals, deck fittings, and deck seam losses. Withdrawal losses occur as the level drops, and thus the floating roof is lowered. Some liquid remains on the inner tank wall surface and evaporates when the tank is emptied. For an internal floating roof tank that has a column-supported fixed roof, some liquid also clings to the columns and evaporates. Evaporative loss occurs until the tank is filled and the exposed surfaces are again covered (SCAQMD supplemental manual, 2003). The working pressure required to prevent breathing and thereby save standing storage losses depends on the vapor pressure of the product, the temperature variations of the liquid surface and the vapor space, and the setting of the vacuum vent. When these factors are known, the storage pressure required to eliminate venting can be computed by the Equation (6-5) (GPSA, 1998): Ps = Pv,m^x + {PT -
PvMn)
7;mn+460
-P. atm
(6-5)
where Ps is required storage pressure, psia; Py,max is true vapor pressure of liquid at maximum surface temperature, psia; Py,min is true vapor pressure of Uquid at minimum surface temperature, psia; Tmax is maximum average temperature of vapor, °F; Tmin is minimum average temperature of vapour, °F; Patm is atmospheric pressure, psia; and Pj is absolute internal tank pressure at which vacuum vent opens, psia.
258
Handbook of Natural Gas Transmission and Processing
For the condition where Pv,rmn is greater than P j , Equation (6-5) can be represented as Ps = Pv,max - P^tm
(6-6)
With the situation represented by Equation (6-5), gas is admitted to the vapor space through the vacuum vent, where a gas vent system should be employed for preventing safety problems. However, for the situation represented by Equation (6-6), gas may be purged and kept out of the tank (Campbell, 1992). Maximum liquid surface temperatures vary from 85 to 115°F. Sufficient accuracy will generally result from the assumption that it is 10°F higher than the maximum temperature of the body of the liquid in a tank at that location. Note that true vapor pressure (TVP) may be the most difficult term in the aforementioned equation to calculate. A nomograph has been devised that relates TVP to both the Reid vapor pressure and the storage temperature (Ts). Numerically, the relationship between TVP, RVP and storage temperature can be expressed as follows (TRW Environmental, Inc., 1981). TVP = (RVP) EXP { Co (ci - j ^ \
j
(6-7)
where Co is constant dependent on the value of RVP, Ci = l/(rs + 460), and T^ is temperature of the stored liquid , °F. In the aforementioned equation, the term Co is dependent on the given value of RVP as shown in Table 6-2 (TRW Environmental, Inc., 1981). Once the parameters Co, Ts, and RVP have been determined, one is technically able to calculate a value for TVP. It should be noted, however, that an error was discovered in the API nomograph-calculated values of TVP, such that the RVP was not equal to the TVP at 100°F as should be expected, given the general definition of RVP. Using linear regression methods, a correction factor (Cp) was developed that should be added to the calculated values of TVP in order to obtain correct TVP values as follows (TRW Environmental, Inc., 1981). Corrected TVP = Calculated TVP + Cp
(6-8)
Condensate Stabilization
259
Table 6-2 Co for Different RVP Values RVP 2 < RVP < 3 RVP=3 3
Co
RVP
Co
-6439.2 -6255.9 -6212.1 -6169.2 -6177.9 -6186.5 -6220.4 -6254.3 -6182.1
RVP = 7 7
-6109.8 -6238.9 -6367.9 -6477.5 -6587.0 -6910.5 -7334.0 -8178.0 -9123.2
The correction factor was found to be dependent on RVP according to the following equations: RVP < 3 : CF = 0.04(RVP) + 0.1 RVP > 3 : CF = EXP{2.3452061og(RVP) - 4.132622}
(6-9) (6-10)
It is important to mention that the methods used for calculating each tank emissions are described in detail in the U.S. EPA's AP-42 manual. In this manual, equations are developed to calculate emissions for fixed roof tanks. By assuming that open top tanks and tanks with open holes or roof openings do not have emissions greater than those for fixed roof tanks, the State California Air Resource Board was able to use the AP-42 equations to calculate emissions for all of the storage tanks in a particular oil production field. REFERENCES
Campbell, J.M., "Gas Conditioning and Processing," 3rd Ed. Campbell Petroleum Series, Norman, OK (1992). GPSA Engineering Data Book, 11th Edition, Vol. 1, Section 6, Gas Processors Suppliers Association, Tulsa, OK (1998). Hotblack, C , BG Tunisia's advanced process control improves condensate product stabiUty. World Oil 225, 9 (2004).
260
Handbook of Natural Gas Transmission and Processing
Reid, R., Prausnitz, J.M., and Sherwood, T., "The Properties of Gases and Liquids," 3rd Ed. McGraw-Hill, New York (1977). SCAQMD Supplemental Manual, "Supplemental Instructions for Liquid Organic Storage Tanks and References." Annual Emissions Reporting Program, South Coast Air Quality Management District, CA (June 2003). TRW Environmental, Inc., "Background Documentation for Storage of Organic Liquids." EPA Contract No. 68-02-3174, NC (May 1981). U.S. EPA's AP-42 Manual, "Compilation of Air Pollutant Emission Factors," Chapter 7, Vol. 1, 5th Ed. U.S. Environmental Protection Agency, Washington, DC.