Removing hydrocarbons from water

Produced water typically enters the water-treatment system from either a two or three phase separator, a free water knockout, a gun barrel, a heater treater, or other primary separation unit process. It probably includes small amounts of free or dissolved hydrocarbons and solids that must be removed before the water can be re-used, injected or discharged. The level of removal (particularly for hydrocarbons) and disposal options are typically specified by state, province, or national regulations. This article discusses techniques for the removal of free and dissolved hydrocarbons. See Removing solids from water for information on solids removal.

Separating free hydrocarbons from water

Produced water contains small concentrations (100 to 2000 mg/L) of dispersed hydrocarbons in the form of oil droplets. The oil-particle diameters will be very small (< 100 μm) because the water flows from this equipment through:

• Dump valves
• Control valves
• Chokes
• Pumps

Treatment equipment to remove dispersed oil from water relies on one or more of the following principles:

• Gravity separation (often with the addition of coalescing devices)
• Gas flotation
• Cyclonic separation
• Filtration
• Centrifuge separation

In applying these concepts, one must keep in mind the dispersion of large oil droplets to smaller ones and the coalescence of small droplets into larger ones, which takes place if energy is added to the system. The amount of energy added per unit time and the way in which it is added will determine whether dispersion or coalescence will take place.

Gravity separation

Stokes’ law, shown in Eq. 1, is valid for the buoyant rise velocity of an oil droplet in a water-continuous phase.

...............(1)

where

v = velocity of the droplet or particle rising or settling in a continuous phase, cm/s

Δρ = difference in density of the dispersed particle and the continuous phase, g/cm³

g_c = gravity acceleration constant, cm/s²

d_p = dispersed particle diameter, cm

μ_L = viscosity of the continuous phase (liquid), g/cm•s

Several immediate conclusions can be drawn from this equation.

1. The larger the diameter of an oil droplet, the greater its vertical velocity; that is, the bigger the droplet size, the less time it takes to rise to a collection surface and, thus, the easier it is to treat the water.
2. The greater the difference in density between the oil droplet and the water phase, the greater the vertical velocity; that is, the lighter the crude, the easier it is to treat the water.
3. The higher the temperature, the lower the viscosity of the water and, thus, the greater the vertical velocity; that is, it is easier to treat the water at high temperatures than at low temperatures.
4. Increasing the g-force imposed on the fluid (i.e., by centrifugal motion) will greatly increase the separation velocity.

The third conclusion requires further elaboration. Heat is the primary mechanism in oil-treating equipment to remove small water droplets from oil. The addition of heat significantly reduces oil viscosity, which prompts more rapid settling, and heat destabilizes water-in-oil emulsions. Heat is not commonly used in water treating because the percentage change in viscosity per degree of temperature change is much less in water than in oil. Water-in-oil emulsions tend to have a higher percentage of the dispersed phase than the oil-in-water emulsions; the dispersed phase tends to have larger-diameter droplets stabilized by heat-sensitive emulsifiers, and it takes twice as much heat input to raise a barrel of water as it takes to raise a barrel of oil to the same temperature.

Dispersion

Small oil droplets contained in the water-continuous phase are subject to the competing forces of dispersion and coalescence. An oil droplet will break apart when kinetic-energy input is sufficient to overcome the surface energy between the single droplet and the two smaller droplets formed from it. At the same time that this process occurs, the motion and collision of oil particles cause coalescence to take place. Therefore, it should be possible to define statistically a maximum droplet size for a given energy input per unit mass and the time at which the rate of coalescence equals the rate of dispersion.

Eq. 2 provides one relationship for the maximum particle size that can exist at equilibrium.

...............(2)

where

d_max = diameter of the droplet above which only 5% of the oil volume is contained

σ = surface tension

ρ_w = density

Δp = pressure drop

t_r = retention time

The greater the pressure drop (and, thus, the shear forces that the fluid experiences in a given time period), the smaller the oil droplets will be. Large pressure drops, which occur in small distances through chokes, control valves, and instruments, result in small oil droplets and water that is harder to treat. A pressure drop of 50 to 75 psi will result in a maximum particle size of 10 to 50 μm. To mitigate the negative effects of large presure drop across the valves, low shear valves can be employed.

Theoretically, the dispersion process is not instantaneous; however, from field experience, it appears to take place very rapidly. For conservative design purposes, it could be assumed that whenever large pressure drops occur, all droplets will disperse instantaneously.

Coalescence

Within water-treatment equipment, in which the energy input to the fluid is very small, the process of coalescence takes place; that is, small oil droplets collide and form bigger droplets. Because of the low energy input, these are not dispersed.

Coalescence can also occur in the pipe downstream of pumps and control valves. However, in such instances, the process of dispersion will govern the maximum size of stable oil droplets that can exist. For normal pipe diameters and flow velocities, particles of 500 to 5000 μm are possible.

The process of coalescence in water-treatment systems appears to be more time-dependent than the process of dispersion. When two oil droplets collide, contact can be broken before coalescence is completed because of turbulent pressure fluctuations and the kinetic energy of the oscillating droplets.

The time required to “grow” a large droplet from a relatively small droplet in a “quiet” gravity-settling tank is approximated by Eq. 3.

...............(3)

where

d_d = droplet diameter

f_V = volume fraction of the dispersed phase

K_s = empirical settling constant

While it is very difficult to determine Ks for an actual installation, the following qualitative conclusions can be drawn:

• Doubling the residence time will cause an increase in droplet diameter of only 19%.
• The more dilute the dispersed phase, the greater the residence time needed to grow a given particle size (that is, coalescence occurs more rapidly in concentrated dispersions).

The addition of surfactant chemicals to the water stream can modify the surface tension of the oil droplets to aid in coalescence.

Gravity separation devices

Water-treating equipment that makes use of gravity separation includes:

• Skim tanks
• API separators
• Plate coalescers
• Skim piles

These devices are very simple and inexpensive; however, because of the large residence times necessary for separation, they are heavy and require large footprints. These devices are commonly used on both land-based and offshore fixed-structure facilities; however, they are motion-sensitive and find limited use on floating facilities.

It is necessary to know both the oil concentration in the influent water and the particle-size distribution to properly design a gravity separator to meet a certain effluent quality. This information can be determined accurately only by sampling the treated water stream. Laboratory testing can provide indicative data for scaleup and correlation, while curves, such as those shown in Fig. 1, can provide an initial estimate from which to work. These data will vary with the oil and water properties and process conditions.

• 

  Fig. 1—Form of empirical curves for oil concentration and droplet-size distribution (representative estimate).

Skim tanks and vessels

The simplest form of treatment equipment is a skim tank or pressure vessel. These are normally designed to provide large residence times during which coalescence and gravity separation can occur. They can be either pressure vessels or atmospheric tanks.

Skim tanks can be either vertical or horizontal in configuration. They may be set up for vertical downward flow of water with or without inlet spreaders or outlet collectors. They may also be designed as horizontal vessels in which the water enters on one side and flows over a weir on the far end.

In vertical vessels, the oil droplets must flow upward against the downward velocity of the water. For this reason, horizontal vessels are more efficient in gravity separation of the two liquid phases. In spite of this, vertical vessels and tanks are sometimes used for the following two reasons:

• Sand and other solids particles can be handled more easily in vertical vessels with either a water outlet or a sand drain off the bottom.
• Vertical vessels are less susceptible to high-level shutdowns caused by liquid surges. Internal waves resulting from surging in horizontal vessels can trigger a level float even though the volume of liquid between the normal operating level and the high-level shutdown is equal to or larger than that in a vertical vessel.

Tracer studies have shown that large skim tanks, even those with carefully designed spreaders and baffles, exhibit poor flow behavior and short circuiting. This is probably caused by such factors as:

• Sensity and temperature differences
• Seposition of solids
• Corrosion of spreaders
• Flow dynamics

In one case, a tank with a design mean residence time of 33 hours had a breakthrough of the tracer with a peak within minutes of tracer injection.

As discussed previously, providing residence time to allow for coalescence does not appear to be cost-efficient. However, a minimum residence time of 10 minutes to 1 hour should be provided to ensure that surges do not upset the system and to provide for some coalescence.

Horizontal pressure vessel sizing

The required diameter and length of a horizontal cylinder operating half full of water can be determined by the following equation:

...............(4)

where

d_i = vessel internal diameter (ID)

q_w = water flow rate

μ_w = water viscosity

d_d =oil-droplet diameter

L_e = effective length in which separation occurs (for design use of 75% seam-to-seam length)

Δγ_ow = difference in specific gravity between oil and water

While Eq. 4 will govern the design, it is also necessary to check for adequate retention time.

...............(5)

where

t_r = retention time.

Vertical cylindrical vessel

The required diameter of a vertical cylindrical pressure vessel or tank can be determined from

...............(6)

where

F = a factor to account for turbulence and short circuiting. For small-diameter vessels (48 in. or less), F = 1.0. For larger diameters, F depends on the type of inlet and outlet spreaders, collectors, and baffles that are provided. Large tanks (10 ft or more in diameter) should be considered to have an F > 2.0, depending on the inlet and outlet conditions.

The height of the water column can be determined from retention-time requirements as follows:

...............(7)

where

Z_w = the height of the water column

API separators

An API separator is the name given to a horizontal, rectangular cross-section, atmospheric oil skimmer that follows the sizing equations and guidelines included in the API Manual on Disposal of Refinery Wastes. The equations for sizing an API separator and their derivations are discussed in Removing solids from water#Gravity settling. API separators find limited use in offshore facilities because of their large size.

Plate coalescers

Various configurations of plate coalescers have been devised and are commonly called:

• Parallel plate interceptors (PPI)
• Corrugated plate interceptors (CPI)
• Crossflow separators

These coalescers depend on gravity separation to allow the oil droplets to rise to a plate surface, where coalescence and capture occur. Flow is split among a number of parallel plates spaced a short distance apart. To facilitate capture of the oil particles, the plates are inclined horizontally.

As shown in Fig. 2, an oil droplet entering the space between the plates will rise in accordance with Stokes’ law; at the same time, it will have a forward velocity equal to the bulk water velocity. By solving for the vertical velocity that a particle entering at the base of the flow needs to reach the coalescing plate at the top of the flow, the resulting droplet diameter can be determined. A restriction is placed on the Reynolds number for the water to ensure that turbulence in this flow does not affect the oil layer on the coalescing plate.

• 

  Fig. 2—Oil-droplet flow pattern in plate coalesce.

General sizing equation

For a plate coalescer with flow either parallel or perpendicular to the direction of flow, the general sizing equation for the droplet-size removal is

...............(8)

where

d_d = design oil-droplet diameter

q_w = bulk water flow rate

L_p = perpendicular distance between plates

μ_w = water viscosity

θ = angle of the plate with the horizontal

Z_p = height of the plate section perpendicular to the axis of water flow

B_p = width of the plate section perpendicular to the axis of water flow

L = length of plate section parallel to the axis of water flow

Δγ_ow = difference in specific gravity between oil and water

Experiments have indicated that on the basis of the hydraulic radius as the characteristic dimension, the Reynolds number for the flow regime cannot exceed 400. Thus, the maximum flow rate is given by

...............(9)

PPI

The first form of plate coalescers was the PPI. This involved installing a series of plates parallel to the longitudinal axis of an API separator (a horizontal, rectangular cross-sectioned skimmer). The plates form a V when viewed perpendicular to the flow axis so that the oil sheet migrates up the underside of the coalescing plate and to the sides. Sediments migrate toward the middle and down to the bottom of the separator, where they are removed.

CPI

The most common PPI form used in production facilities is the CPI. This is a refinement of the PPI in that it takes up less platform space (length) for the same particle-size removal and has the added benefit of making sediment handling easier. Fig. 3 is a typical design with a corrugated plate.

• 

  Fig. 3—CPI separator schematic and flow pattern (courtesy of Natco).

In CPIs, the parallel plates are corrugated (like roofing material), with the axis of the corrugations inclined to an angle of 45°. The bulk water flow is forced downward; the oil sheet raises upward counter to the water flow and is concentrated in the top of each corrugation. When the oil reaches the end of the plate pack, it is collected in a vertical channel and brought to the oil/water interface. CPIs require frequent cleaning of the plate packs in which large amounts of sediment are handled.

Crossflow devices

CPI configurations are available for horizontal water flow perpendicular to the axis of the corrugations in the plates. This allows the plates to be put on a steeper angle to facilitate sediment removal and to enable the plate pack to be packaged more conveniently in a pressure vessel. The latter benefit may be required if gas blowby through an upstream dump valve could cause relief problems with an atmospheric tank (see the discussion on gas blowby in Safety systems).

Crossflow devices can be constructed in either horizontal or vertical pressure vessels. The horizontal vessels require less internal baffling because the ends of nearly every plate conduct the oil directly to the oil/water interface and the sediments to the sediment area below the water flow. The vertical units, although requiring collection channels on one end to enable the oil to rise to the oil/water interface and on the other end to allow the sand to settle to the bottom, can be designed for more efficient sand removal. Crossflow separators are used where sand is a considerable problem and are not removed in the process upstream of the CPI.

Practical limitations

Stokes’ law theory should apply to oil droplets as small in diameter as 1 to 10 μm. However, field experience indicates that 30 μm sets a reasonable lower limit on the droplet sizes that can be removed. At less than this size, small pressure fluctuations and platform vibrations tend to impede the rise of droplets to the coalescing surface; thus, the practical limit for sizing-plate coalescers is 30-μm removal.

Skim pile

Skim piles are gravity water-treating devices used offshore. As shown in Fig. 4, flow through the multiple series of baffle plates creates quiescent zones that reduce the distance a given oil droplet must rise to be separated from the main flow.

• 

  Fig. 4—Skim-pile flow pattern (courtesy of Natco).

Once in the quiescent zone, there is plenty of time for coalescence and gravity separation. The larger droplets then migrate up the underside of the baffle to an oil-collection system. Skim piles are used extensively to treat deck drainage of washdown or rainwater that has been contaminated with oil. They have the added benefit of providing some degree of sand cleaning. Sand traversing the length of a skim pile will abrade on the baffles and be water washed. This removes the free oil, which is then captured in the quiescent zone.

Skim pile sizing deck drainage

Field experience has indicated that acceptable effluent is obtained with 20 minutes of retention time in the baffled section of the pile. Using this and assuming that 25% of the volume is taken up by the coalescing zones,

...............(10)

where

d_i = pile internal diameter

L_bs = length of baffle section

q_w = produced-water rate if it is disposed in pile, B/D

A_d = deck area

q_r = rainfall rate

q_WD = washdown rate

Intermittent flow

During periods of no flow, oil droplets rise to the area of the quiescent zone and become trapped and protected from being swept back into the flow stream when flow is resumed. The net effect of the baffles is to reduce this rise distance. Each time that flow is stopped as the water traverses the baffled section, more oil particles are trapped in the quiescent zone.

This phenomenon can be used when a skim pile is used downstream of a skim tank or CPI for further treating. With a snap-acting water dump on the influent, intermittent flow is established in the pile.

If t = the time in seconds for the dump cycle,

...............(11)

where

N_c = the number of nonflow cycles that a particle sees as it traverses the baffle section

If t_c = the time the valves are closed, the removal efficiency on any cycle of a particular drop size is

...............(12)

The overall removal efficiency of that particle size can then be determined by

...............(13)

Gas flotation units

Flotation units do not rely on gravity forces for separating the oil droplets; in fact, the action of these units is independent of the oil-droplet size. In gas flotation units, large quantities of small-diameter gas bubbles are injected into the water stream. The bubbles attach to the oil droplets suspended in the stream, causing them to rise to the water surface and form a froth layer. Experimental results have shown that very-small-diameter oil droplets in dilute suspensions can be removed easily by flotation. High percentages of oil removal are achieved.

Two distinct types of flotation units have been used, distinguished by the method employed in distributing small gas bubbles throughout the water. These are dissolved- and dispersed-gas units.

Dissolved gas units

Dissolved gas designs take a portion of the treated water effluent and saturate the water with natural gas or air in a contactor. The higher the pressure, the more gas that can be dissolved in the water. However, most units are designed for a contact pressure of 20 to 40 psig. Normally, 20 to 50% of the treated water is recirculated for contact with the gas.

The gas-saturated water is then injected into the flotation chamber. The dissolved gas breaks out of solution in small-diameter bubbles when the flow enters the chamber, which is operated at near-atmospheric pressure.

Design parameters are recommended by the individual manufacturers but normally range from 0.2 to 0.5 scf/bbl of water to be treated and flow rates of treated plus recycled water of between 2 and 4 gal/min-ft². Retention times of 10 to 40 minutes and depths of between 6 and 9 ft are specified.

Dissolved-gas units have been used successfully in refinery operations in which air can be used as the gas and where large areas are available. In treating produced water for injection, it is desirable to use natural gas to exclude oxygen. This requires venting the gas or installing a vapor-recovery unit. Field experience with dissolved-natural-gas units has not been as successful as experience with dispersed-gas units.

Dispersed gas units

In dispersed-gas units, gas bubbles are dispersed in the total stream either by use of an inductor device or by a vortex set up by mechanical rotors. Fig. 5 shows a schematic cross section of such a unit.

• 

  Fig. 5—Dispersed-gas unit with inductor device.

Most dispersed-gas units contain three or four cells. Bulk water flow moves in series from one cell to the other by underflow baffles. Field tests have indicated that the high intensity of mixing in each cell creates the effect of plug flow of the bulk water from one cell to the next; that is, there is virtually no short circuiting or breakthrough of a part of the inlet flow to the outlet weir box.

Efficient design of a dispersed-gas unit requires:

• High gas-induction rate
• Small-diameter induced-gas bubble
• Relatively large mixing zone

Thus, the design of the rotor and internal baffles is critical to the unit efficiency. Field tests and theory both indicate that these units operate on a constant percent-removal basis. Within normal ranges, their oil-removal efficiency is independent of inlet concentration or oil-droplet diameter.

There are many different proprietary designs on the market. The most common designs have three to five separate cells in which gas is induced into the water stream. Each cell is designed for an approximately 1-minute retention time to allow the gas bubbles to break free of the liquid and form the froth at the surface. Field experiments show that the designs can remove 50% of the oil in each cell. From Eq. 13, a three-cell unit is expected to have an overall efficiency of 87%, and a four-cell unit should have an efficiency of 94%. In practice, the typical efficiency of an installed four-cell unit is only approximately 90%.

Because the unit recycles the gas, a natural-gas blanket can be maintained easily with little or no venting. The low required retention times make this an ideal choice for offshore facilities, where space and weight are at a premium. However, in motion-sensitive installations, sloshing within each cell degrades the performance of the flotation cell.

Deoiling hydrocyclones

Deoiling hydrocyclones, or “deoilers,” provide the highest throughput-to-size ratio of any water-treating technology and are insensitive to motion or orientation. For a given capacity of water to be treated, deoilers will provide the smallest footprint and size of any water-treating technology. Deoilers use fluid-pressure energy to create rotational fluid motion, as shown in Fig. 6. This rotational motion causes relative movement of materials suspended in the fluid, thus permitting separation of these materials, one from another or from the fluid. In the case of produced-water deoiling, this process can remove small oil droplets from the water stream.

• 

  Fig. 6—Flow pattern of a deoiler hydrocyclone.

To obtain maximum benefits from a hydrocyclone system, it must be understood that the system must be incorporated differently into the overall process. Because hydrocyclones are pressure driven, they ideally should be located as close as possible to the oily water outlet of the three-phase separator. This location results in the simplest and most cost-effective installation with minimum operating cost. At the highest-pressure location in the process, the hydrocyclone will have maximum capacity because there is maximum pressure available. Furthermore, installing the hydrocyclones in this manner will return the best separation performance because the oily water has not yet been exposed to droplet-shearing pressure drops across level control valves.

An added advantage of deoiling hydrocyclones is the simplicity of their control systems, which normally use the standard interface-level control system of the three-phase separator, as shown in Fig. 7. The hydrocyclones are simply installed on the water outlet of the three-phase separator with the interface-level control valve downstream of the hydrocyclones on their water outlet. This control valve normally will control the flow rate through the deoiler in a proportional control mode. On/off control can be used for very low flow rates but is generally not recommended. In on/off control, there will be a surge of untreated water when the valve is first opened, until the flow stabilizes and there is adequate pressure drop through the hydrocyclone.

• 

  Fig. 7—Deoiler (a) throughput control and (b) pressure-ratio control.

The overflow can be controlled by installing a valve on the overflow line that operates in parallel with the interface-level control valve or by using a simple pressure-control system that ensures a constant overflow rate, as shown in Fig. 7. This latter control system is referred to as the pressure-ratio control because it keeps a constant ratio between the two pressure drops in the deoiler: the inlet-to-overflow pressure drop and the inlet-to-underflow pressure drop.

If the existing process pressure is less than 25 psig, it is recommended to boost the pressure by pumping the water to the hydrocyclone system. It is imperative that strict guidelines are followed when designing a pumped hydrocyclone system. Pump selection and operation are critical because the wrong pump type, or even the correct pump type operated incorrectly, can introduce considerable droplet shear.

• Low-shear pumps provide the best performance in this application.
• Positive-displacement pumps can be low shear if they are equipped with low-shear check valves, as in the case of reciprocating pumps.
• Progressing cavity pumps can be excellent low-shear pumps, but they may require more maintenance than is acceptable.
• Centrifugal pumps are not typically low-shear ones, but some standard centrifugal pumps can be used very effectively for low-shear service. Because of their simplicity, reliability, and relatively low cost, centrifugal pumps are the recommended pumps to feed hydrocyclone systems.

When selecting a centrifugal pump for a deoiler system, almost any brand can be used as long as the following are observed.

• Pump is of closed impeller design.
• Pump operation is near the maximum on the efficiency curve (at least 70%).
• Maximum speed is 1,750 rpm.
• Maximum pressure boost per stage is 75 psi.

Several companies have reported about developments of low shear centrifugal pumps. Common similarity is the utilization of mentioned guidelines.

In addition some manufacturers address the importance of key ratios of the inlet area, outlet area and the proportions of the impeller hub. The right configuration of these elements can help to control fluid flow in such way that reduces the shearing effect. Additional parameters that have an influence of shearing are the recirculation zones inside the pump and the surface finish effect of the impeller.

Another reported case is using the multistage centrifugal pump. By spreading the energy input into the system throughout the several stages and specially designed impeller configuration, the pump allows to transfer fluids with minimal shearing. Multistage pump design allows increasing operational range for pressures up to 20 bar and for flows up to 250 m³/h.  

Pumps should be controlled by either a recycle loop or a variable-speed control, with the latter typically being more costly. Recycle control is the simplest way of preventing deadhead. The easiest method is for the pump to feed the hydrocyclones directly with the level-control valve from the source vessel located at the water discharge from the hydrocyclone.

Centrifuges

Disc-stack centrifuges for produced-water cleanup have been in use for the past 10 years. They are used primarily for difficult applications to remove very small droplets of oil and for cases in which the fine-oil droplets do not coalesce easily. As shown in Fig. 10, the disk-stack centrifuge consists of a:

• Frame
• Drive motor
• Transmission
• Separator bowl
• Inlet/outlet arrangements

• 

  Fig. 8—Disc-stack centrifuge for produced-water cleanup (courtesy of Alfa Laval Inc.).

The separation process takes place inside the rotating bowl at up to 6,000 g-forces. Produced water is introduced in the center of the bowl through the feed pipe and is accelerated to full rotational speed. The bowl is fitted with special inserts, which shorten the settling distance for oil droplets to 0.5 mm. The annular channels can be regarded as parallel separation vessels. The oil flows toward the center of the bowl against the upper side of the disc. Water and sediment flow outward against the underside of the disc. Disc-stack centrifuges can effectively remove oil droplets as small as 1 to 2 μm, and units are available at 15,000 B/D capacity each.

Walnut shell filters

Media-type filters are used for removal of fine solids from water. A specific type of media filter that uses walnut shells is used specifically for removing residual dispersed hydrocarbons from produced water. This type of filter is shown in Fig. 9.

• 

  Fig. 9—Walnut-shell filters applied in a steamflood operation.

Besides the media type, walnut-shell filters differ from sand filters in the method of backwashing. Recent improvements include the use of mechanical agitation or walnut-shell recirculation in the backwash cycles. The induced shearing or rubbing action removes most of the oil and solids from the walnut shells. Because of the effective oil- and solid-removal procedure, the use of surfactant chemicals generally has been eliminated. Another important improvement is that the walnut-shell filters have significantly reduced the waste volume to approximately 1% of the total throughput. This relatively small amount of backwash waste has made it possible for offshore application, where space for a waste-handling system is limited.

In low-temperature applications, heavy crude may be difficult to remove during the backwash cycle. Chemicals or heat are often required to clean the walnut shells before their reuse. In a steamflood field, the produced water is generally warm or hot, and a chemical or warm-water wash may not be required. For the previous reasons, walnut-shell filters have grown in acceptance with steamflood fields.

Removing dissolved hydrocarbons from water

When oil is produced from underground formations, water is also produced. In fact, the amount of water that comes to the surface will, with time, exceed the amount of oil. Because the produced water stream is considered a waste product, it is more economical for offshore operators to dispose of produced water near the producing platform in the ocean. It should be noted that dissolved or water-soluble hydrocarbons are present in all produced-water streams. The amount of water-soluble hydrocarbons varies depending on:

• Properties of the hydrocarbon
• Operating conditions (particularly temperature and pressure)
• Reservoir characteristics
• Other factors

It also should be noted that other kinds of dissolved organic matter are present in produced water. In general, water-soluble organic matter falls into one of the following classes:

• Aliphatic hydrocarbons
• Phenols
• Organic acids
• Aromatic compounds, such as benzene and toluene

Hydrocarbon discharges

Hydrocarbon discharges are, in most cases, regulated by a government agency such as the US Environmental Protection Agency (EPA). The goal of these regulatory bodies is to minimize the impact that produced water discharges have on the local environment. Therefore, the EPA and other regulatory bodies around the world set limits for the amount of total hydrocarbons, or “oil and grease,” contained in the produced water discharged from offshore platforms. These limits usually are expressed in terms of milligrams per liter (mg/L) of the contaminant in the disposal stream. The EPA currently sets the limit of oil and grease in water discharged into federally regulated areas of the Gulf of Mexico at 42 mg/L daily maximum and 29 mg/L monthly average, per National Discharge Pollution Elimination System (NPDES) General Permit #28000. Research into the environmental impact (toxicity) of dissolved hydrocarbons concludes that discharges into a body of water from responsibly operated facilities will have a negligible environmental impact. This conclusion is based on research showing that low levels of dissolved hydrocarbons are assimilated quickly into the ecosystem of the receiving body. Furthermore, responsibly operated facilities provide means for the contaminants in produced water to be diluted quickly to levels well within toxicity limits. As an example, discharged water from a North Sea platform is diluted between 50 and 400 times within 40 to 70 seconds after discharge.

Determining removal

The measurement of oil and grease must be considered to understand the criteria for determining whether removal of dissolved hydrocarbon material is necessary. Using the EPA as an example, the current EPA-approved measurement techniques are EPA Method 413.1 or 1664. Both methods are gravimetric.

• Method 413.1 uses freon as a solvent
• Method 1664 uses n –hexane

These methods involve pH reduction to less than 2 (to precipitate the dissolved hydrocarbons), extraction of the dispersed and dissolved components from the produced water with the solvent, extract separation, and solvent vaporization, leaving a residue considered to be the “oil and grease” component.

However, there are two reasons that these methods do not accurately measure the total amount of hydrocarbons in the produced water.

1. The solvent removes not only the dispersed and dissolved hydrocarbons but also the nonhydrocarbon organic matter, such as phenols, organic acids, alcohols, and ketones
2. During the evaporation step of the measurement technique, low-molecular-weight volatile hydrocarbons flash off. These include alkanes and aromatic hydrocarbons (BTEX - benzene, toluene, ethylbenzene, and xylenes).

Hence, “oil and grease,” as defined by the EPA measurement techniques, excludes some hydrocarbon components from vaporization and includes some nonhydrocarbon components from extraction by a nonselective solvent.

It is apparent that for an offshore oilfield operator to meet the federally imposed discharge limits, most of the dispersed oil will have to be removed before the discharge point. The discussion above on separating free hydrocarbons from water contains a review of the types of equipment and processes commonly used to remove the dispersed-oil constituents. Furthermore, if the amount of dissolved organic matter is sufficiently high, it also may be necessary to remove the dissolved organic material to comply with the total oil and grease discharge limit. At present, most of the operating facilities worldwide are meeting compliance standards by removing only the dispersed hydrocarbons. However, there are operators that must also remove the dissolved organic material to be compliant. In these instances, some of the more common methods used for offshore applications are pH adjustment with stock mineral acids or patented mineral/acid blends and the use of adsorption/absorption materials (activated carbon). Each method has inherent advantages and disadvantages, as shown in Table 1.

• 

  Table 1

Nomenclature

A	=	membrane area, ft2
Ad	=	deck area, ft2
Bp	=	width of the plate section perpendicular to the axis of water flow, ft
c	=	solids concentration, fraction by volume
dd	=	droplet diameter, μm
di	=	vessel internal diamater, in.
di	=	pile internal diameter, in.
dmax	=	droplet diameter, μm
dp	=	dispersed particle diameter
D	=	ID of cyclone, in.
Erc	=	Eq. 12
Ero	=	Eq. 13
fV	=	volume fraction of the dispersed phase
F	=	factor accounting for turbulence and short-circuiting
g	=	g-force acceleration factor
gc	=	gravity acceleration constant
hf	=	height of flume, ft
K	=	permeate flow coefficient at standard temperature, gal/D-psi-ft2
Kf	=	fouling factor
Ks	=	empirical settling constant
KT	=	membrane temperature-correction factor
L	=	length of the plate section parallel to the axis of water flow, ft
Lbs	=	length of baffle section, ft
Le	=	effective length in which separation occurs, ft
Lp	=	perpendicular distance between plates, in.
L1	=	inlet-distribution section
L2	=	outlet-gathering section
Nc	=	number of nonflow cycles that a particle sees as it traverses the baffle section
qr	=	rainfall rate, in./hr
qw	=	water flow rate, BWPD
qWD	=	washdown rate, B/D
Qf	=	feed volumetric flow rate, m3/s
Qpf	=	permeate flow, gal/D
t	=	time, seconds
tc	=	time valves are closed
tr	=	retention time, minutes
v	=	velocity, ft/s
vs	=	settling velocity
vw	=	velocity of water, ft/s
x98	=	particle size at 98% efficiency, m
Zp	=	height of the plate section perpendicular to the axis of water flow, ft
Zw	=	height of water column, ft
Δp	=	pressure drop, psi
Δpavg	=	average transmembrane pressure drop, psi
Δγ	=	difference in specific gravity relative to water
Δγow	=	difference in specific gravity between oil and water
Δρ	=	difference in density of the dispersed particle and the continuous phase
σ	=	surface tension, dynes/cm
ρl	=	liquid density, kg/m3
ρs	=	solid density, kg/m3
ρw	=	density, g/cm3
μL	=	viscosity of the continuous phase (liquid)
μw	=	water viscosity, cp
θ	=	angle of the plate with the horizontal

Noteworthy papers in OnePetro

Use this section to list papers in OnePetro that a reader who wants to learn more should definitely read

Other noteworthy papers

Bassett, G. M. 1971. WEMCO Depurator TM System. Presented at the Society of Petroleum Engineers Rocky Mountain Regional Meeting, 2-4 June, Billings, Montana, USA. SPE-3349-MS. http://dx.doi.org/10.2118/3349-MS

Casaday, A. L. 1993. Advances in Flotation Unit Design for Produced Water Treatment. Presented at the Society of Petroleum Engineers Production Operations Symposium, 21-23 March, Oklahoma City, Oklahoma, USA. SPE-25472-MS. http://dx.doi.org/10.2118/25472-MS

Stroder, S. M., Wolfenberger, E. E. 1994. Hydrocyclone Separation: A Preferred Means of Water Separation and Handling in Oilfield Production. Presented at the Society of Petroleum Engineers Permian Basin Oil and Gas Recovery Conference, 16-18 March, Midland, Texas, USA. SPE-27671-MS. http://dx.doi.org/10.2118/27671-MS

Delgado, A., Lee, H. M. 1998. The Chronology of Water-Oil Handling Equipment. Presented at the Society of Petroleum Engineers International Petroleum Conference and Exhibition of Mexico, 3-5 March, Villahermosa, Mexico. SPE-39878-MS. http://dx.doi.org/10.2118/39878-MS

Van der Zande, M. J., Janssen, P. H., van den Broek, W. M. G. T. 2001. Size of Oil Droplets Under High-Water-Cut Conditions. Presented at the Society of Petroleum Engineers Production and Operations Symposium, 24-27 March, Oklahoma City, Oklahoma, USA. SPE-67250-MS. http://dx.doi.org/10.2118/67250-MS

Lee, C.-M., Frankiewicz, T. 2004. Developing Vertical Column Induced Gas Flotation for Floating Platforms Using Computational Fluid Dynamics. Presented at the Society of Petroleum Engineers Annual Technical Conference and Exhibition, 26-29 September, Houston, Texas, USA. SPE-90201-MS. http://dx.doi.org/10.2118/90201-MS

Rivet, C., Movafaghian, S., Chen, J. 2005. Introduction of a Dual-Cell Depurator for FPSO Application-Putting Theory to Practice. Presented at the Society of Petroleum Engineers Latin American and Caribbean Petroleum Engineering Conference, 20-23 June, Rio de Janeiro, Brazil. SPE-94655-MS. http://dx.doi.org/10.2118/94655-MS

Rawlins, C.H. 2010. Characterization Of Deep Bed Filter Media For Oil Removal From Produced Water, Presented at the 20th Annual Produced Water Society Seminar, January 20-22.

Husveg, R., Husveg, T., van Teeffelen, N., et al. 2016. Performance of a Coalescing Multistage Centrifugal Produced Water Pump with Respect to Water Characteristics and Point of Operation. NEL Produced Water Workshop, 7-8 June 2016, Abeerden, UK.

Husveg, T., 2007. Operational Contol of Deoiling Hydrocyclones and Cyclones for Petroleum Flow Control. PhD dissertation. University of Stavanger, Stavanger, Norway (October 2007)

Online multimedia

Walsh, John M. 2013. Hydrocyclones for Water Treating—The Science and Technology. https://webevents.spe.org/products/hydrocyclones-for-water-treatingthe-science-and-technology

Walsh, John M. 2012. Selection and Troubleshooting of Flotation Equipment for Produced Water Treating. https://webevents.spe.org/products/selection-and-troubleshooting-of-flotation-equipment-for-produced-water-treating

External links

Use this section to provide links to relevant material on websites other than PetroWiki and OnePetro

See also

Removing solids from water

Water treating facilities

Produced Oilfield Water

Water treating chemicals

Surface water treatment for injection

Materials for water treating equipment

PEH:Water-Treating_Facilities_in_Oil_and_Gas_Operations

Page champions

Hank Rawlins, PhD, PE