A gas-processing plant, is designed to recover ethane, propane, butane, and other natural gas liquids from the gas stream. A condensate stabilizer also recovers some portion of these liquids. The colder the temperature of the gas leaving the overhead condenser in a reflux stabilizer, or the colder the feed stream in a cold-feed stabilizer, and the higher the pressure in the tower, the greater the recovery of these components as liquids. Indeed, any stabilization process that leads to recovery of more molecules in the final liquid product is removing those molecules from the gas stream. In this sense, a stabilizer may be considered as a simple form of a gas-processing plant.
It is difficult to determine the point at which a condensate stabilizer becomes a gas plant. Typically, if the liquid product is sold as a condensate, the device would be considered a condensate stabilizer. If the product is sold as a mixed natural gas liquid stream (NGL) or is fractionated into its various components, the same process would be considered a gas plant. The least volatile NGL stream has an RVP between 10 and 14 and has sufficient light hydrocarbons such that 25% of the total volume is vaporized at 140°F.
It should be clear from the description of LTX units that the lower pressure separator in an LTX unit is a simple form of cold-feed condensate stabilizer. In the cold, upper portion of the separator some of the intermediate hydrocarbon components condense. In the hot, lower portion some of the lighter components flash.
An LTX unit is not a very efficient stabilizer because the absence of trays or packing keeps the two phases from approaching equilibrium at the various temperatures that exist in the vessel. In addition, it is difficult to control the process. Typically, for a 100-psi to 200-psi operating pressure, a 300°F to 400°F bottoms temperature is required to stabilize completely the condensate. The heating coil in an LTX separator is more likely to be in the range of 125°F to 175°F, and thus complete stabilization will not occur even if the flash were capable of reaching equilibrium.
There may be some additional recovery from an LTX unit than would be realized from a straight two-stage flash separation process, but this increment is normally small and may not justify the increased equipment cost and operating complexity associated with an LTX unit.
Next to sales contract specifications, corrosion protection ranks highest among the reasons for the removal of acid gases. The partial pressure of the acid gases may be used as a measure to determine whether treatment is required. The partial pressure of a gas is defined as the total pressure of the system times the mole % of the gaseous component. Where CO2 is present with free water, a partial pressure of 30 psia or greater would indicate that CO2 corrosion should be expected. If CO2 is not removed, inhibition and special metallurgy may be required. Below 15 psia, CO2 corrosion is not normally a problem, although inhibition may be required.
H2S may cause hydrogen embrittlement in certain metals. Figures 7-1 and 7-2 show the H2S concentration at which the National Association of Corrosion Engineers (NACE) recommends special metallurgy to guard against H2S corrosion.
In the sulfide stress cracking region, appropriate metallurgy is required in line piping, pressure vessels, etc. There is a listing of acceptable steels in the NACE standard. Steels with a hardness of less than 22 Rockwell C hardness should be used in areas where sulfide-stress cracking is a problem.
The concentration of H2S required for sulfide-stress cracking in a multiphase gas/liquid system (Figure 7-2) is somewhat higher than in pure gas streams (Figure 7-1). The liquid acts as an inhibitor.
In addition to heavy hydrocarbons and water vapor, natural gas often contains other contaminants that may have to be removed. Carbon dioxide (CO2), hydrogen sulfide (H2S), and other sulfur compounds such as mercaptans are compounds that may require complete or partial removal for acceptance by a gas purchaser. These compounds are known as “acid gases.” H2S combined with water forms a weak form of sulfuric acid, while CO2 and water forms carbonic acid, thus the term “acid gas.”
Natural gas with H2S or other sulfur compounds present is called “sour gas,” while gas with only CO2 is called “sweet.” Both H2S and CO2 are undesirable, as they cause corrosion and reduce the heating value and thus the sales value of the gas. In addition, H2S may be lethal in very small quantities. Table 7-1 shows physiological effects of H2S concentrations in air.
At 0.13 ppm by volume, H2S can be sensed by smell. At 4.6 ppm the smell is quite noticeable. As the concentration increases beyond 200 ppm, the sense of smell fatigues, and the gas can no longer be detected by odor. Thus, H2S cannot always be detected by smell. Even if H2S cannot be smelled, it is possible that there is sufficient H2S present to be life threatening. At 500 ppm, H2S can no longer be smelled, but breathing problems and then death can be expected within minutes. At concentrations above 700 to 1,000 ppm, death can be immediate and without warning. Generally, a concentration of 100 ppm H2S or more in a process stream is cause for concern and the taking of proper operating precautions.
Gas sales contracts for natural gases will limit the concentration of acid compounds. In the United States, typically, gas sales contracts will permit up to 2 to 3% carbon dioxide and 1A grain per 100 scf (approximately 4 ppm) of hydrogen sulfide. The actual requirement for any sales contract may vary, depending upon negotiations between seller and purchaser.
Numerous processes have been developed for gas sweetening based on a variety of chemical and physical principles. These processes can be categorized by the principles used in the process to separate the acid gas and the natural gases as follows:
1. Solidbed absorption
The Sulfatreat Company: Iron Sponge, Sulfa Treat, Zinc Oxide.
Union Carbide Corporation: Molecular Shieve.
2. Chemical solvent
Monoethanol Amine (MEA), Diethanol Amine (DEA), Methyldiethanol Amine (MDEA), Diglycol Amine (DGA), Diisopropanol Amine (DIPA), Hot Potassium Carbonate, Proprietary Carbonate System.
3. Physical Solvent
Flour Daniel Corporation: Flour Flexsorb, Shell Sulfinol.
Norton Co., Chemical Products: Selexol
Lurg, Kohle & Mineraloltechnik Gmbh & Linde A.G: Rectisol.
4. Direct conversion of H2S to Sulfur
ARI Technologies: Claus, LOCAT.
Ralph M, Parsons Co: Stretford.
Exxon Chemical Co: Sulfa-check.
Institute Francais du Petrole: IFP
5. Sulfide scavenger
6. Distillation – Amine-aldehyde condensate.
7. Gas permeation
The list, although not complete, does represent many of the commonly available commercial processes. New proprietary processes are being developed. The design engineer is cautioned to consult with vendors and experts in acid gas treating before making a selection for any large plant.
A fixed bed of solid particles can be used to remove acid gases either through chemical reactions or ionic bonding. Typically, in solid bed absorption processes the gas stream must flow through a fixed bed of solid particles that remove the acid gases and hold them in the bed. When the bed is saturated with acid gases, the vessel must be removed from service and the bed regenerated or replaced. Since the bed must be removed from service to be regenerated, some spare capacity must be provided. There are three commonly used processes under this category: the iron oxide process, the zinc oxide process, and the molecular sieve process.
The iron sponge process uses the chemical reaction of ferric oxide with H2S to sweeten gas streams. This process is applied to gases with low H2S concentrations (300 ppm) operating at low to moderate pressures (50-500 psig). Carbon dioxide is not removed by this process.
The reaction of H2S and ferric oxide produces water and ferric sulfide as follows:
The reaction requires the presence of slightly alkaline water and a temperature below 110°F. If the gas does not contain sufficient water vapor, water may need to be injected into the inlet gas stream. Additionally, bed alkalinity should be checked daily. A pH level of 8-10 should be maintained through the injection of caustic soda with the water.
The ferric oxide is impregnated on wood chips, which produces a solid bed with a large ferric oxide surface area. Several grades of treated wood chips are available, based on iron oxide content. The most common grades are 6.5-, 9.0-, 15.0-, and 20-lb iron oxide/bushel. The chips are contained in a vessel, and sour gas flows through the bed and reacts with the ferric oxide. Figure 7-3 shows a typical vessel for the iron sponge process.
The ferric sulfide can be oxidized with air to produce sulfur and regenerate the ferric oxide. The reaction for ferric oxide regeneration is as follows:
The regeneration step must be performed with great care as the reaction with oxygen is exothermic (that is, gives off heat). Air must be introduced slowly so the heat of reaction can be dissipated. If air is introduced quickly the heat of reaction may ignite the bed.
Some of the elemental sulfur produced in the regeneration step remains in the bed. After several cycles this sulfur will cake over the ferric oxide, decreasing the reactivity of the bed. Typically, after 10 cycles
the bed must be removed from the vessel and replaced with a new bed.
In some designs the iron sponge may be operated with continuous regeneration by injecting a small amount of air into the sour gas feed. The air regenerates ferric sulfide while H2S is being removed by ferric oxide. This process is not as effective at regenerating the bed as the batch process. It requires a higher pressure air stream, and if not properly controlled may create an explosive mixture of air and gas.
Hydrocarbon liquids in the gas tend to coat the iron sponge media, inhibiting the reactions. The use of an adequately designed gas scrubber or filter separator upstream of the iron sponge unit will minimize the amount of liquids that condense on the bed. Sometimes the process can be arranged so that the scrubber operates at a lower temperature or higher pressure than the iron sponge unit, so that there is no possibility of hydrocarbon liquids condensing in the iron sponge unit.
Due to the difficulty of controlling the regeneration step, the eventual coating of the bed with elemental sulfur, the low cost of iron sponge material, and the possibility of hydrocarbon liquids coating the bed, iron sponge units are normally operated in the batch mode. The spent bed is removed from the unit and trucked to a disposal site. It is replaced with a new bed and the unit put back in service. The spent bed will react with the oxygen in air as shown in Equations 7-2 and 7-3 unless it is kept moist. In areas where iron sponge units are installed, service companies exist that can replace iron sponge beds and properly dispose of the waste material.
SulfaTreat process is similar to the iron sponge process. It uses a patented proprietary mixture of ferric oxide and triferric oxide to react with H2O to sweeten gas streams. In SulfaTreat process the iron oxides are supported on the surface of an inert, inorganic substrate forming a granular material, while in the iron sponge process the ferric oxide is impregnated on wood chips. The SulfaTreat starting material and the spent product are safe and stable. The spent product can be recycled or disposed in a landfill.
Two vessels arranged in series, a lead/lag arrangement, will allow the SulfaTreat material to be used more efficiently with no interruption in unit service and greater process reliability. The first vessel, the “lead” unit, acts as the “working” unit to remove all the H2S at the beginning of a treatment period with its outlet H2S increasing over time. The exit gas from the first vessel can go to the second vessel, the “lag” unit, for further polishing or bypass the second vessel as though the first vessel is operating in a single vessel arrangement. The second vessel is to be placed in operation as the lag unit to polish the H2S remaining in the gas when the lead unit outlet H2S starts to approach the specification.
Once the lead unit inlet and outlet concentrations are equal, the SulfaTreat material is considered spent or exhausted. Then, the gas flow is directed to the second vessel, which becomes the lead unit. The spent material is removed from the first vessel and the fresh SulfaTreat material reloaded to be placed into operation as the lag unit without gas flow interruption. Removal of the spent SulfaTreat material and the reload of the fresh material could be conveniently scheduled using the change-out “window” available without exceeding maximum outlet H2S concentrations when operating in this lead/lag mode.
In condensate stabilizers, trays generally have 70% equilibrium stage efficiency. That is, 1.4 actual trays are required to provide one theoretical stage. The spacing between trays is a function of the spray height and the downcomer backup (the height of clear liquid established in the downcorner). The tray spacing will typically range from 20 to 30 in. (with 24 in. being the most common), depending on the specific design and the internal vapor and liquid traffic. The tray spacing may increase at higher operating pressures (greater than 165 psia) because of the difficulty in disengaging vapor from liquid on both the active areas and in the downcomers.
Contactor column usually contain either tray random packing or structure packing. For large volume of gas, the contactor is usually a tray column consisting of 4 to 12 trays. The greater number of trays, more moisture can be removed.
Bubble cap column
Each tray has opening with bubble cap bolted over them. The up flowing gas is forced through these caps and bubbles evenly through the down flowing glycol. The gas gives up water and become dryer as passes upward through each subsiding tray. The glycol becomes more saturated with water as it flows downward over each tray. Weirs which dam like devices, maintain the level of glycol above the slot in the bubble cap. The downcomer carry the glycol to the tray below.
In smaller capacity unit gas contactor is having diameter 18 inches or less and random packing may be used instead of trays. The packing can be metal, plastic or ceramic. This structure provides large surface area for glycol solution to spread out and make better contact with the gas. Random packing is poured to the contactor on to a support grid. 4 feed of packing is usually standard and sufficient to achieve the dew point depression which is up to 45 – 65 F. If higher dew point depressions are needed, additional packing may be required.
Pack column utilize the same process as tray column where liquid glycol flows down over the packing and the gas flow up through the packing. Pack columns are less expensive however they tend to channel and have poor flow distribution. Channel is when liquid glycol flow in streams throughout the random packing. Channeling limits the surface area where the glycol and gas commit contact. To ensure that gas and glycol mixing will be continues throughout the packing and the glycol would not channel, a well designed glycol distribution header is installed above the packing.
Another gas dehydration method uses solid bed desiccant instead of liquid desiccant. The common example of this is the molecular sieves which are made of pellets that are electronically poured to water. When placed in line with the gas stream the polarity of the pellets attract the water out of the gas into molecule size pour on the surface of the pellets. The water is held there until the pellets are saturated. The pellets them self are then dehydrated by a small volume of heated gas so they can be used again.
Methanol injection gas dehydration method involves the injection of methanol in the gas stream to absorb water. The methanol and water mixture and then disposed off in the environmentally safe manner. Methanol injection is rarely used on dehydration because it is toxic, expensive and the disposal can be complicated.
Hydrocyclone separators, sometimes called enhanced gravity separators, use centrifugal force to remove oil droplets from oily water. As shown in Figure 7-16, static hydrocyclone separator consist of the following four sections: a cylindrical swirl chamber, a concentric reducing section, a fine tapered section, and a cylindrical tail section. Oily water enters the cylindrical swirl chamber through a tangential inlet, creating a high-velocity vortex with a reverse-flowing central core. The fluid accelerates as it flows through the concentric reducing section and the fine tapered section. The fluid then continues at a constant rate through the cylindrical tail section. Larger oil droplets are separated out from the fluid in the fine tapered section, while smaller droplets are removed in the tail section. Centripetal forces cause the lighter-density droplets to move toward the lowpressure central core, where axial reverse flow occurs. The oil is removed through a small-diameter reject port located in the head of the hydrocyclone separator. Clean water is removed through the downstream outlet.
Static hydrocyclone separator require a minimum pressure of 100 psi to produce the required velocities. Manufacturers make designs that operate at lower pressures, but these models have not always been as efficient as those that operate at higher inlet pressures. If a minimum separator pressure of 100 psi is not available, a low-shear pump should be used (e.g., a progressive cavity pump) or sufficient pipe should be used between the pump and the hydrocyclone separator to allow pipe coalescence of the oil droplets. As is the case with flotation units, hydrocyclone separator do not appear to work well with oil droplets less than 10 to 20 microns in diameter.
Performance is chiefly influenced by reject ratio and pressure drop ratio (PDR). The reject ratio refers to the ratio of the reject fluid rate to the total inlet fluid rate. Typically, the optimum ratio is between 1 and 3%. This ratio is also proportional to the PDR. Operation below the optimum reject ratio will result in low oil removal efficiencies. Operation above the optimum reject ratio does not impair oil removal efficiency, but it increases the amount of liquid that must be recirculated through the facility. The PDR refers to the ratio of the pressure difference between the inlet and reject outlets and the difference from the inlet to the water outlet. A PDR of between 1.4 and 2.0 is usually desired. Performance is also affected by inlet oil droplet size, concentration of inlet oil, differential specific gravity, and inlet temperature. Temperatures greater than 80°F result in better operation.
Although the performance of hydrocyclone separators varies from facility to facility (as with flotation units), an assumption of 90% oil removal is a reasonable number for design. Often the unit will perform better than this, but for design it would be unwise to assume this will happen. Performance cannot be predicted more accurately from laboratory or field testing because it is dependent on the actual shearing and coalescing that occurs under field flow conditions and on impurities in the water, such as residual treating and corrosion chemicals and sand, scale and corrosion products, which vary with time.
Hydrocyclones are excellent coalescing devices, and they actually function best as a primary treating device followed by a downstream skim vessel that can separate the 500 to 1,000 micron droplets that leave with the water effluent. A simplified P&ID for a hydrocyclones eparator is shown in Figure 7-17.
Advantages of static hydrocyclone separators include: (1) they have no moving parts (thus, minimum maintenance and operator attention is required), (2)their compact design reduces weight and space requirements when compared to those of a flotation unit, (3) they are insensitive to motion (thus, they are suitable for floating facilities), (4) their modular design allows easy addition of capacity, and (5) they offer lower operating costs when compared to flotation units, if inlet pressure is available.
Disadvantages include the need to install a pump if oil is available only at low pressure and the tendency of the reject port to plug with sand or scale. Sand in the produced water will cause erosion of the cones and increase operating costs.