You must log in to edit PetroWiki. Help with editing

Content of PetroWiki is intended for personal use only and to supplement, not replace, engineering judgment. SPE disclaims any and all liability for your use of such content. More information

Removing solids from water

Jump to navigation Jump to search

Produced or fresh water being treated may have suspended solids, such as formation sand, rust from piping and vessels, and scale particles, or dissolved solids (various chemical ions). For most uses or disposal methods, these solids may need to be removed. It may be necessary to remove these solids to prevent wear in high-velocity areas, prevent solids from filling up vessels and piping and interfering with instruments, and comply with discharge restrictions on oil-coated solids. This page discusses appropriate removal technologies and handling of the removed material.

Removing suspended solids

Suspended solids can be separated from the water stream by:

  • Gravity settling
  • Hydrocyclone desanders
  • Filters
  • Centrifuges

Gravity settling

Solid particles, because of their heavier density (compared to water) and net negative buoyant force, will settle to the bottom with a terminal velocity that can be derived from Stokes’ law, as shown in Eq. 1.



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

gc = gravity acceleration constant, cm/s2

dp = dispersed particle diameter, cm

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

This equation applies strictly to creeping flow regimes in which the Reynolds number is less than unity; this is mainly concerned with spheres of very small diameter surrounded by a liquid. For very small particles, the inertial forces are much less than the viscous forces because of the low particle mass, and the particle does not enter into a turbulent settling regime. This equation can be used to size any of the following to allow a particle of a certain diameter and specific gravity to settle under natural gravity conditions:

  • Tank
  • Vertical or horizontal pressure vessel
  • Rectangular sedimentation chamber
  • Device of any other configuration

Most sedimentation basins are rectangular flumes with length-to-width ratios of 4:1 or greater to limit crossflow. The width of the flow channel can be determined by setting the time required for a particle to settle from the top of the flume to the bottom equal to that required for the water to traverse from the inlet of the flume to the outlet, as shown in Fig. 1. This can be expressed as



b = width (breadth) of the flow channel (ft)

qw = water flow rate

Le = effective length

Note that the width and length of the settling chamber are independent of its depth. The API Manual on Disposal of Refinery Wastes[1] recommends a turbulence and short-circuiting factor of between 1.3 and 1.8, depending on the ratio of water velocity to solids-settling velocity. Using a factor of 1.8, Eq. 14 can be rewritten as


API also recommends that the water velocity be limited to 15 times the settling velocity or 3 ft/min, whichever is less. The settling velocity can be calculated from Eq. 1, and the water velocity can be calculated from



vw = velocity of water

hf = height of the flume

For practical considerations, b should be between 6 and 20 ft, and the ratio of hf to b should be between 0.3 and 0.5.

The flume can be concrete-lined or constructed as a soil pit; solids that settle in the bottom of the flume can be cleaned out with a bucket. A mechanical sludge scraper run on chain could be installed to concentrate the solids in one location for easy removal.

Desanding hydroclones

Desanding hydrocyclones, called desanders, offer the highest throughput-to-size ratio of any solids-removal equipment. Fig. 2 shows the basic operation of a desander. By definition, all hydrocyclones operate by pressure drop. The feed, a mixture of liquids and solids, enters the cyclone through the volute inlet at the operating feed pressure. The change in flow direction forces the mixture to spin in a radial vortex pattern. Because of the angular acceleration of the flow pattern, centrifugal forces are imparted on the solid particles, forcing them toward the internal wall of the cone. The solids continue to spin in a radial vortex pattern, down the length of the cone, and discharge through the apex, creating the underflow stream. Because of cone convergence, the liquid flow is reversed and sent upward through the vortex finder to create the overflow stream. The solids that exit through the apex collect into an accumulation chamber and are periodically purged, while the overflow discharges continually.

The particle size that is separated depends on the pressure drop through the desander, and the pressure drop, in turn, is dependent on the flow rate. Thus, there is a minimum flow and pressure drop that must be provided for each desander to settle a certain particle size. For general comparison information, in a produced-water-treatment application, a 0.5-in. (10-mm) desander will have a separation size of 5 μm, while a 30-in. (750-mm) desander will have a separation size of 100 μm. The practical limit for sand separation from water by a hydrocyclone is 10 μm. An estimation of particle separation size by a desander can be calculated from the following equation:[2]



x98 = particle size at 98% efficiency

D = internal diameter of the cyclone

c = solids concentration

Qf = feed volumetric flow rate

ρs = solid density

ρl = liquid density

The previous equation is based on a fixed hydrocyclone geometry relationship, and the desander geometry is unique to each manufacturer. The design is typically scalable in that the following are proportional to the internal diameter (D):

  • Inlet area
  • Vortex-finder diameter
  • Length
  • Apex diameter

A desander acts as a fixed orifice in the flow stream, with the pressure drop proportional to the flow rate. Each manufacturer can provide a pressure-drop curve, so that the pressure drop is known for a given flow rate. As such, the pressure drop and flow rate are used interchangeably.

The desander pressure drop is the main operating parameter that practically can be changed on line to influence separation efficiency. The desander can operate in a wide pressure-drop range and, thus, inherently has a high turndown ratio. The minimum operating pressure drop for a desander is 5 psig. At less than this value, the fluid does not contain enough energy to form the proper vortex flow pattern. No theoretical maximum pressure drop exists, but 100 psi is recommended as a practical maximum to balance wear and recovery.

Fig. 3 shows the two types of desanders that are commonly used in water systems:

  • Vessel style
  • Liner style

The criteria in Table 1 are used to select the proper style.[3]

Because of the simpler design (i.e., without ceramic liners or tube sheets), the vessel desander has a lower capital cost when compared to the liner design. In most applications, though, the deciding factor is the required separation size, which is why most conventional desanders in oil and gas production are of the liner design.


To avoid plugging the injection formation, it may be necessary to separate small-diameter suspended particles by filtration. Filters cannot handle the volume of solids that can be handled by sedimentation and desanders, but they are the only practical method for separating very fine particles (< 10 μm). By properly choosing the filter element, filters can remove fine solids in the 0.5- to 50-μm range and are used as a form of secondary treatment. The three types of filters commonly used are cartridge, media, or diatomaceous-earth filters. Because filtration is more commonly used with injection of surface water, this technology is covered in further detail here.


Centrifuges are used on drilling rigs to separate low-gravity drill solids and to reclaim high percentages of heavy solids. They have not found wide use in producing operations because of the high maintenance associated with their use. Normally, if it is desirable to separate solid particles with a diameter less than that resulting from sedimentation or desanders, filters are used.

Solids handling

Once solids separation is identified as a need in a produced-water system, facilitating the solids handling becomes an important need in overall system design. Systems are based on modular designs that can be built to handle a very wide range of process needs for either land-based or offshore systems. These systems should be manufactured to require minimal operator intervention and, in case of hazardous disposal, minimal contact.

Solids handling can be broken down into five areas[3]:


Separation is defined as diverting the solids and liquids contained in a mixed slurry stream to different locations. The solids are removed from the produced-water stream by one of the following:

  • Gravity vessel (tank bottoms or vessel drain)
  • Desander
  • Sand-jet system
  • Filter dump


Collection is defined as gathering all separated solids into a central location and physically isolating them from the production process. By collecting the solids in one location, a simpler system can be designed to isolate the solids from the process. Collection can be as simple as either of the following:

  • Desander accumulator vessel
  • Dedicated sump tank


In many cases, the sand may require cleaning of adsorbed oil or chemicals before further handling. Sand-cleaning systems are handled in one of two ways:

  • Offered as modular add-on packages
  • Integrated into the separation system


The total volume of sand slurry to be transported and disposed of can be greatly reduced by a dewatering step, which involves removing the liquids from the collected (cleaned) solids slurry. A range of systems is available to provide dewatering from a sand-drainage bag to a filter press or screw classifier. The goal is to reduce the liquid to less than 10% by volume.


Haulage is a simple term used to define removing, hauling, and disposing of the solids. The design of the haulage system will be dependent upon location (land-based or offshore) and disposal requirements (i.e., disposal well, overboard, landfill, road surfacing, etc.). Offshore systems typically involve crane-to-boat-to-truck transport, while land-based systems may use a truck to a landfill.

Every solids-handling and -disposal system will be different because of:

  • Economics
  • Environmental and hazardous regulations
  • Location
  • Total solids to be handled

Removing dissolved solids from water

Various chemical compounds are dissolved in water as ions to form an aqueous solution. The term “dissolved solids” is used to describe these ions in water; some of the more common are:

  • Silica
  • Calcium
  • Magnesium

When water is thermally evaporated or treated with membranes, these ions become saturated and exceed their solubility in water. They will then precipitate or crystallize to form scale. Scale formation plugs piping and fouls the following equipment:

  • Water-handling system
  • Steam-generator tubes
  • Membranes

Scaling sometimes can be controlled with an inhibitor chemical; however, when this does not work, these ions should be removed from the system. The dissolved ions can be removed from water with:

  • Membranes
  • Ion exchange
  • Hot or warm softening


Membranes are predominantly used to remove species of salts and organics from water. Reverse osmosis (RO) can remove 95 to 99% of the metallic ions, such as sodium and potassium salts, as well as a relatively high percentage of organic material. Nanofiltration (NF) can remove most divalent ions, such as sulfate and nitrate, from water. An RO membrane can remove most of the dissolved solids or ions from the water, as shown in Fig. 4. Comparatively, NF membranes can remove divalent ions but allow monovalent ions to pass through. Ultrafiltration and microfiltration can remove only the submicron-size suspended particles and are not effective for removing soluble ions.

The performance of a membrane unit can be expressed by the following equation.



Qpf = permeate flow

Kf = fouling factor

KT = membrane temperature-correction factor

K = permeate flow coefficient at standard temperature

A = membrane area

Δpavg = average transmembrane pressure drop

Because of the presence of various impurities in water, the membrane gradually will become fouled, and the permeate flow will decrease, measured by the fouling factor, Kf. For a clean, new membrane, the fouling factor is 1.0. This value will decrease gradually, and the membrane element will need to be cleaned. Generally speaking, the membrane element requires cleaning when the fouling factor is decreased approximately 10 to 15%, or when it reaches a value of less than 0.85.

The water flow coefficient, K, is a function of the water/membrane chemistry interaction, especially pH. Normalized permeate rates are higher at a pH of 11 than at 7. Although in general, the water flow coefficient cannot be predetermined, the variation of KT with temperature is known for each membrane and can be used to normalize permeate-flow data at different temperatures. At any point in the RO system, the transmembrane pressure drop is the difference between the brine hydraulic pressure and the system osmotic pressure plus the permeate pressure. As the brine concentration increases from the feed to reject ends of the membrane system, the transient pressure decreases.

Various types of membrane configurations are available. The most common types are plate and frame, hollow-fiber, monolithic-tubular, tubular, and thin-film-composite spiral-wound, as shown in Figs. 5 through 11. The thin-film-composite spiral-wound membranes have been used successfully to treat brackish oilfield produced waters.[4][5][6]

It is desirable to produce a high fraction of good-quality permeate water to reduce the amount of concentrate water for disposal. For practical purposes, a 75% recovery of the permeate water can be achieved without seriously fouling the membranes. This level of removal efficiency is relatively high for most of the inorganic and organic materials, as shown in Table 2.[7] Trace metal removal is also relatively high, as shown in Table 3.[8]

Ion exchange

The ion-exchange process is used to remove specific ions from solution. Its primary application in produced-water treatment is the removal of calcium and magnesium ions, which make up the “hardness” in water. Ion exchange used for this purpose is called “water softening.” It also can be used to remove residual minerals and, in this instance, is called demineralization.

Water softening

The water-softening process is used in the oil industry for steamflood operations. To operate reliably at high temperatures and pressures, steam generators require a very low hardness content in the feed water. Scale precipitation can coat the heating tubes, causing localized overheating and tube failure. Fig. 12 shows a typical water-softener bank used in steamflood operations. For this case, the divalent ions, calcium and magnesium, are exchanged with the sodium ions from an ion-exchange resin. After the exchange, the ion-exchange resin is saturated with the divalent ions. It is regenerated with a higher-concentration brine solution, which is rich in sodium ions. This brine solution is generally a 10 to 20% salt (sodium chloride) solution.

Types of water softening resins

There are two types of water-softening resins commonly used in the oilfield for steam generation:

  • Strong acid resins
  • Weak acid resins

A typical strong acid resin is the sodium zeolite cation-exchange resin, consisting of a synthetic zeolite material that contains numerous cation-exchange sites. These sites primarily contain sodium (Na) ions. The zeolite resin is commonly expressed as “Z,” and its ion-exchange reaction with the hardness material, either calcium or magnesium, is shown in the following reactions:



The strong acid resin generally is used for removing the hardness materials from water with relatively low sodium contents or total dissolved solids (TDS). This type of resin attracts the calcium and magnesium ions and exchanges them with a sodium ion upon regeneration. However, as the sodium content in the feed water increases (on the basis of the previous formula), the reaction will go in reverse and start to regenerate during the softening process. This partial regeneration is called the “hardness leak.” The hardness leak is pronounced for the strong acid resins when the TDS of the feed water is greater than approximately 5,000 to 7,000 ppm. Hence, when the concentration of TDS exceeds this limit, strong acid resins are no longer effective.

The sodium ion used for regeneration is common salt (NaCl). During regeneration, the high concentration of salt in the brine exchanges the hardness material, such as calcium ion, with sodium, as shown in the following reaction.


The source for regeneration is generally sea salt, produced by the evaporation of seawater; however, rock salt from the land-mining process is sometimes used. The quality of salt is important, and its impurities should be studied before use to minimize operational problems.

The weak acid resins are generally used for removing hardness materials from water with higher sodium content or TDS. The weak acid resin can handle produced water with TDS up to 30,000 to 40,000 mg/L. Its regeneration program uses an acid, such as hydrochloric acid, followed by a base, such as sodium hydroxide. The capital cost of a process of weak acid resins increases in comparison with processes using strong acid resins because of the use of an acid and a base during regeneration, which requires using linings in the vessels and corrosion-resistant piping. Similarly, the operating cost of processes that use weak acid resins increases in comparison with processes using strong acid resins because of the higher cost of the regenerating materials (acids and bases for the weak acid resins vs. a salt solution for the strong acid resins).

Softening the economic breaking point between weak acid resins and strong acid resins is approximately 7,000 mg/L of TDS. When the TDS measures between 5,000 and 7,000 mg/L, there is slight hardness leakage, but the water with a small amount of hardness materials could be treated with either of the following:

  • A chelant, such as EDTA or NTA
  • A combination of chelant and polymeric scale-suppressant chemicals

However, when the TDS is greater than 7,000 mg/L, the hardness leakage would be too much, and a chelant program might become uneconomical.


Using ion-exchange resins to remove residual minerals from water is called demineralization. This process is generally used for polishing water after membranes, softening, or warm-lime treating. When the minerals are removed, this water can be used in boilers for steamflooding or in steam turbines for electrical generation. The demineralization process uses various types of ion-exchange resins to achieve different results, as summarized in Table 5.

Treating produced water with a demineralization system as the final stage for polishing allows it to be used in cogeneration plants for electrical production. Additionally, the produced steam can be used for steamflooding. Because of the use of acid and base chemicals for regeneration, the cost of a demineralization system is relatively high. It can be used only for polishing good-quality water from other pretreatment processes and is generally uneconomical to apply to raw water treatment.

Hot and warm lime softening

Hot and warm lime softening processes are other technologies used to remove the hardness and silica ions from produced water for steam generation. The advantages of these systems are that they can process a large amount of water in a relatively small unit, and their applicability, unlike zeolite softeners, is not limited by the water’s TDS.

The design of one type of the hot-lime process is shown in Fig. 13. Normally, a residence time of 1 hour is specified. Produced water is heated up to or greater than 212°F by spraying at the top of the unit. The lime (calcium oxide) is premixed in water as slurry and fed immediately below the spray. It reacts quickly with the hardness materials in water to form precipitates; this mixture flows to the bottom part of the vessel through a downcomer pipe to contact with the remaining sludge. At the bottom, the flow is reversed, and the water rises slowly through a blanket of previously formed sludge. The intimate contact increases the efficiency of the softening process. The basic reaction is shown as follows.



The reaction products, calcium carbonate (CaCO3) and magnesium hydroxide [Mg(OH)2], are precipitated and removed as sludge. This reduces both the total hardness and TDS. When the carbonate is short in the feed water or the system contains noncarbonate hardness ions, soda ash should be used. Its reaction is shown as follows:



When soda ash is used, the final product still contains the soluble noncarbonate ions; hence, it does not reduce the TDS. The sludge or the precipitated solids from the lime process are periodically blown down from the unit. The treated water is processed with a filter before sending it to the steam generators. The backwash water from the filters is returned to the unit. When the unit is operated at a temperature of less than 212°F, it is classified as a warm-lime softener.

One type of warm-lime softener is shown in Fig. 14. This warm-lime unit has two compartments. In the reaction zone, or the first compartment, the lime and magnesium oxide slurry is fed together with the produced water. A mixer moves very slowly to keep the influent mixture in contact with the slurry in this compartment. This contact promotes precipitation of both hardness materials and silica. The treated water flows upward and overflows into the clarification zone, or the second compartment. The sludge separates from the water and sinks to the bottom of the clarification zone. The treated water flows upward through a set of clarification weirs to prevent any sludge flowing out of the system. The bottom sludge is recycled with an external pump back to the reaction zone. This unit can be used for removing both silica and hardness ions from produced water.

Silica removal

Silica fouls steam generators and reverse osmosis (RO) membranes when its concentration exceeds the solubility limit. Its solubility data are plotted in Fig. 15. The silica solubility also depends upon the pH of the water. The silica solubility increases significantly at higher pH values.[9] Silica scale is generally deposited on the inside of the radiation section tubes. This deposit acts as an insulator for the tube and reduces its heat transfer. More seriously, the tube can overheat from the flame and cause tube failures. Based on the previous analysis, it is necessary to remove silica and control its concentration in the feed water for the RO-membrane and steam-generation systems.

Silica can be removed with the warm- or hot-lime softeners described previously. In a steamflood operation, the produced water generally is at greater than 170 to 180°F; hence, the warm-lime-softening unit is adequate for handling it. However, the option of using the hot-lime process exists. The chemical reaction for silica removal is still unknown, and the key is to use an adequate amount of lime and magnesium oxide and recycle the slurry to promote an intimate contact with the incoming water for a better reaction and better use of the chemicals. Silica also can be removed at lower temperatures (60 to 80°F), generally referred to as the cold-lime process.


A = membrane area, ft2
c = solids concentration, fraction by volume
dd = droplet diameter, μm
dmax = droplet diameter, μm
dp = dispersed particle diameter
D = internal diameter of cyclone, in.
g = g-force acceleration factor
gc = gravity acceleration constant
K = permeate flow coefficient at standard temperature, gal/D-psi-ft2
Kf = fouling factor
Ks = empirical settling constant
KT = membrane temperature-correction factor
Le = effective length in which separation occurs, ft
qw = water flow rate, BWPD
Qf = feed volumetric flow rate, m3/s
Qpf = permeate flow, gal/D
vw = velocity of water, ft/s
x98 = particle size at 98% efficiency, m
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
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


  1. Oil-Water Separator Process Design. 1975. Manual on Disposal of Refinery Wastes, Volume on Liquid Wastes, Ch. 5. Dallas, Texas: API.
  2. Svarovsky, L. 1984. Hydrocyclones, 96-97. London: Technomic Publishing Co. Inc.
  3. 3.0 3.1 Rawlins, C.H., Staten, S.E., and Wang, I.I. 2000. Design and Installation of a Sand Separation and Handling System for a Gulf of Mexico Oil Production Facility. Presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas, 1-4 October 2000. SPE-63041-MS.
  4. Tao, F.T., Curtice, S., Hobbs, R.D. et al. 1993. Conversion of Oilfield Produced Water Into an Irrigation/Drinking Quality Water. Presented at the SPE/EPA Exploration and Production Environmental Conference, San Antonio, Texas, 7-10 March 1993. SPE-26003-MS.
  5. Tao, F.T. et al. 1993. Reverse Osmosis Process Successfully Converts Oil Field Brine into Freshwater. Oil & Gas J. (20 September): 88.
  6. VandeVenter, L.W., Ford, B.R., and Vera, M.W. 1989. Innovative Process Provides Cogeneration Power Plant with the Ability to Use Oil Field Water. Paper presented at the 1989 Annual International Water Conference, Pittsburgh, Pennsylvania, 23–25 October.
  7. Performance data of Filmtec FT-30 membrane, Dow Chemicals.
  8. Performance data of B-9 Hollow Fiber membrane, DuPont.
  9. Iler, R.K. 1979. The Chemistry of Silica, 42. New York City: John Wiley and Sons.

Noteworthy papers in OnePetro

Rawlins, C. H., Wang, I. I. 2001. Design and Installation of a Sand-Separation and -Handling System for a Gulf of Mexico Oil Production Facility. SPE Production & Facilities. SPE-72999-PA.

Muraleedaaran, S., Li, X., Li, L., Lee, R. L. 2009. Is Reverse Osmosis Effective for Produced Water Purification: Viability and Economic Analysis. Presented at the Society of Petroleum Engineers Western Regional Meeting, 24-26 March, San Jose, California, USA. SPE-115952-MS.

Webb, C. H., Nagghappan, L., Smart, G., Hoblitzell, J., Franks, R. 2009. Desalination of Oilfield-Produced Water at the San Ardo Water Reclamation Facility, CA. Presented at the Society of Petroleum Engineers Western Regional Meeting, 24-26 March, San Jose, California, USA. SPE-121520-MS.

Heins, W. F. 2010. Is a Paradigm Shift in Produced Water Treatment Technology Occurring at SAGD Facilities? Journal of Canadian Petroleum Technology. SPE-132804-PA.

Rawlins, H. 2013. Sand Management Methodologies for Sustained Facilities Operations. Presented at the SPE North Africa Technical Conference and Exhibition, 15-17 April, Cairo, Egypt. SPE-164645-MS.

Rawlins, C. H. 2013. Design of a Cyclonic Solids Jetting Device and Slurry Transport System for Production Facilities. Presented at the SPE Annual Technical Conference and Exhibition, 30 September-2 October, New Orleans, Louisiana, USA. SPE-166118-MS.

Online multimedia

Walsh, John M. 2013. Hydrocyclones for Water Treating—The Science and Technology.

Walsh, John M. 2012. 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 hydrocarbons from water

Facilities for steam generation

Water treating facilities

Water treating chemicals


Page champions

Hank Rawlins, PhD, PE