This is the first article in a two-part series. It discusses methods for determining the size of the three-phase separator and quantifying the oil/water separation performance in the horizontal separator based on the estimation of the volume of the dispersed phase and the associated droplet size distribution. .
Historically, the size of three-phase separators, especially for oil/water separation, has been based on the specifications of residence time standards. Unfortunately, this is an inaccurate method, because the residence time alone cannot capture many parameters/variables that obviously affect the liquid/liquid phase separation performance, such as feed pipe conditions, inlet equipment type, Phase flow distribution, container length/diameter ratio, fluid characteristics, and dispersed phase droplet size. Therefore, the residence time criterion cannot and will not be reliably related to the actual quality of the separated liquid phase, such as the amount of water remaining in the oil (WIO) and the remaining water The amount of oil (OIW).
A more analytical method is to determine the size of the separator based on the specified separable droplet size, such as removing water droplets larger than 250 microns from the oil phase, or removing oil droplets of 150 microns and larger from the water phase . There are two main problems here:
What is the basis for selecting the size of the separable droplets?
How to perform the required calculations to achieve the target separable droplet size? A large number of variables and assumptions are involved.
Ideally, the separable droplet size should correspond to a given volume of dispersed phase remaining in the separated continuous phase. For example, removing water droplets larger than 250 microns in diameter from the oil phase will result in about 3% v/v water remaining in the oil, or removing oil droplets larger than 150 microns from the water phase will result in about 1,000 ppmv oil remaining in the water. Obviously, the assumption here is that the relationship between droplet size and dispersed phase entrainment volume is known or at least approximate. Are these relationships (one is oil and the other is water) fixed? In fact they are not. There are many variables involved.
In addition to the quantification of the droplet size distribution/dispersed phase (entrainment) volume estimation, the utilization of the target separable droplet size requires many assumptions to be made to perform the actual droplet separation calculations.
This article has two main points:
Development of an approximate method for estimating the relationship between entrainment and droplet size in all three stages
A method to perform actual droplet separation calculations based on container size, internal component selection, fluid characteristics, etc., to achieve the specified target outlet fluid quality
This discussion mainly involves the oil/water separation aspects of the three-phase separator design and operation. The author's previous three-part series (Part 1, Part 2, and Part 3) introduced similar gas/liquid separation methods.
The estimation of the dispersed phase droplet size and entrainment volume starts from the feed tube conditions.
There are two main aspects:
Estimate the three-phase flow pattern in the feed pipe of the separator. The main purpose of this step is to estimate the amount of water dispersed in the oil phase in the form of droplets and the amount of oil dispersed in the water phase in the form of droplets.
Estimate the distribution of the three phases (gas, oil and water) in the feed pipe, the cross-sectional area of the feed pipe occupied by each phase (assuming stratified flow), and the actual velocity of each phase. The main purpose of this step is to obtain the actual velocities of the oil and water phases, which are then used to calculate the maximum stable droplet size of each dispersed phase (water droplets in oil and oil droplets in water). The calculated maximum droplet size is then used to characterize the two droplet size distributions.
In addition, calculations will be performed to determine the influence of any fittings (such as elbows, tees, etc.) present in the feed tube on the droplet size distribution of the feed tube. For the purpose of this analysis, two different methods are used for these two steps. This is an approximation and can produce results that are considered reasonable. Future improvements of this approach will use a single "unified" approach to achieve the above two goals, similar to the goals proposed by Zhang and Sarica (2005).
Characterization of inlet flow stability (or lack) is also a key requirement for proper separator size, but it is outside the scope of this article.
Currently, there is very little published information that can be used to estimate the amount of dispersed phase (WIO, OIW) entrained in each continuous phase of a liquid/liquid stream. The method used in this work is based on the three-phase flow pattern and related flow pattern proposed by Keskin et al. (2007), as shown in Figure 1.
The three-phase flow mode is described by two identifiers, such as ST-DC. ST stands for stratification, and it describes the gas/liquid flow pattern. DC stands for dual continuous, and it describes the liquid/liquid flow pattern.
Keskin et al. provide a more detailed description of the flow pattern.
The appearance of various flow patterns is shown in the relevant three-phase flow pattern diagrams. Keskin et al. Provides five flow pattern diagrams with 20%, 40%, 50%, 60% and 80% liquid water content (based on volume). For illustration, the 50% water content diagram is shown in Figure 2.
Using the flow pattern diagram and description of the flow pattern, it is possible to estimate the concentration of the dispersed phase (WIO and OIW) for a given set of feed tube operating conditions. Adjust any fittings in the feed pipe (near the separator inlet) and the separator inlet device. This will be discussed later in this article.
For several liquid/liquid flow modes, it is necessary to define "minimum" or "residual" WIO and OIW levels for subsequent droplet separation calculations. For example, the oil and water phases in the stratified liquid/liquid stream are not "pure"—some relatively small amounts of water are dispersed in the oil in the form of droplets, and a small amount of oil is dispersed in the water in the form of droplets—despite these It is difficult to identify and quantify droplets in Keskin et al. Stratified liquid/liquid flow pattern. Similarly, for continuous oil flow and continuous water flow modes, it is necessary to define the minimum concentrations of OIW and WIO respectively for separation calculations.
The assumed “minimum” level of the dispersed phase fraction is shown in Table 1.
Please note that these numbers do not predefine the mass of a single liquid phase. They are only used to provide the lowest level of dispersed phase entrainment at the entrance of the liquid gravity separation section.
There is no major source of shear in the separator feed pipe within 100 feet upstream of the separator, such as a throttle valve or control valve
Reasonably high salinity produced water (1.03 <SG <1.1). Water with low salinity is more difficult.
Moderately high temperature (80–120°F). Lower temperatures are more difficult.
No special/expected lotion issues
"Easy" and "difficult" conditions can be defined relative to the above-mentioned "average" conditions. Procedure for estimating the entrained dispersion volume fraction:
Calculate the in-situ liquid water content under the conditions of the feed pipe.
According to 1 and 2 above, determine the three-phase flow pattern of the water supply pipe from the appropriate water-bearing flow pattern diagram. Use the moisture content map closest to the actual moisture content or interpolate between the moisture content maps.
Determine the WIO and OIW volume fractions based on the determined liquid/liquid flow pattern.
If necessary, adjust the volume of the dispersed phase entrained in the "basic" feed tube according to any other shear-causing elements (such as pipe fittings) that may be present near the inlet of the separator.
This section discusses the distribution of gas, oil, and water phases in the feed pipe, the cross-sectional area of the feed pipe occupied by each phase (assuming stratified flow), and the actual velocity of each phase.
The purpose of this step is to estimate the actual phase velocity in the feed tube, which is then used to calculate the dispersed phase droplet size. The author's previous article discussed the estimation of gas phase liquid entrainment and droplet size distribution.
Taitel et al. introduced the method used in a paper. (1995). A follow-up paper by Heydari and Sadeghi (2011) provides some additional details to help perform calculations (requires an iterative process, which can be easily performed via the solver add-in in Excel). These calculations assume stratified three-phase flow.
Although not all feed tube conditions will necessarily result in stratified three-phase flow, this situation is very common and, in fact, is usually more suitable for three-phase separation applications. In particular, high gas flow rates and corresponding high gas velocities will result in high liquid velocities, which in turn will result in smaller dispersed phase droplet sizes that are more difficult to separate.
The Hinze equation can be used to estimate the maximum dispersed phase droplet size:
For pipe flow, the turbulent energy dissipation term ε can be defined as:
For the purpose of these calculations, the values used for the phase friction coefficient, velocity, and hydraulic diameter are outlined by Taitel et al. There are Haidari and Sadji.
Substitute this expression into the equation. 1 gives:
ρc = density of continuous phase
fc = sector friction coefficient of the continuous phase
Vc = continuous phase velocity
D = effective hydraulic diameter of the continuous phase
Although the Hinze equation is considered best for dilute dispersions, such as <1-2% of the dispersed phase volume, it will be used here for dilute dispersions and concentrated dispersions. For dense dispersions, the maximum droplet size may be underestimated by Equation 3. 3. Therefore it is conservative. This form of Hinze equation also ignores the influence of the dispersed phase viscosity, which may be more important for OIW calculations involving heavy oil.
For the purpose of this work, equation. 3 It will be applied to the oil phase and the water phase to obtain the largest droplet size in oil and the largest droplet size in water. It should be noted that the dmax value calculated from the equation. 3 Very sensitive to σ, σ is the value of the oil/water interfacial tension used. Since the dmax value will affect the droplet size distribution of the dispersed phase, thereby affecting the relationship between the dispersion volume of WIO and OIW and the droplet size, the selected interfacial tension value has a considerable influence on the calculation of the separator size.
Future improvements to the model will use correlation to estimate the oil/water interfacial tension based on fluid characteristics and operating conditions (such as temperature). The existing oil/water interfacial tension correlations in the literature are somewhat inconsistent and require further review. Obviously, if actual oil/water interfacial tension data is available, this is more preferable, because this parameter depends on the chemistry of the oil and water involved, which makes generalization difficult.
In the actual separator installation, it is very possible to install one or more pipe fittings (such as elbows, tee) in the feed pipe closer to the separator. These accessories generate additional shear forces, which may increase the entrainment of the dispersed phase, transfer the droplets to a smaller size, and also affect the flow distribution of the inlet device according to the type and direction of the accessory.
The joint is located within 15 times the diameter of the feed pipe from the separator. For fittings farther than this, it is assumed that the droplets will coalesce and move the droplet size back to the value calculated for the horizontal straight feed tube. The linear adjustment factor is used to describe the actual installation position between 0-15 diameters upstream of the separator inlet.
Assuming that the number of parts has no effect, only the part that produces the highest shear-the "worst case part" is considered. It is the fitting that produces the maximum value of ε, the turbulent energy dissipation rate. Note that if it is located in the feed pipe close to the separator, this method also shows the extreme negative impact of the throttle/throttle valve on the separation performance.
A modified version of the equation. 3 Used for worst-case fitting:
Le = equivalent length of the accessory
Lact = actual flow path length of the joint
For example, the equivalent length of a typical long radius 90o elbow is 16 times the pipe diameter, and the actual length is about 2.4 times the pipe diameter. For a 10-inch inner diameter (ID) of the feed pipe:
Comparison of equations Figure 3 and Figure 4 show that the droplet size "displacement factor" caused by fitting can be defined as
The turbulence/shear generated by the fitting is likely to cause part of the continuous phase (oil and water) to be entrained into another liquid phase in the form of droplets, thereby increasing the volume of the entrained dispersed phase relative to the "straight pipe" value. The impact is difficult to estimate and is not included in the results of this article; this is an area for future investigations.
At this point, it has been estimated that the amount of dispersed WIO and OIW has the maximum droplet size of each dispersed phase of the feed tube, including the influence of any accessories that are present. The next step is to generate the droplet size distribution of each phase. As described in part 1 of the author's previous gas/liquid separation series, the upper log normal (ULLN) distribution has been used for this purpose.
For feed tube applications, except for the maximum droplet size, the parameters used to describe the ULLN distribution are similar to those given by Karabelas (1978) and Simmons et al. (2001):
These parameters are used for WIO and OIW distribution. It should be noted that the separation calculation is very sensitive to these values.
Various computational fluid dynamics (CFD) studies (for example, Heijckers 2012) have shown that the elbow in the horizontal plane of the feed pipe within about 15D of the separator inlet will cause uneven flow distribution into and through the separator inlet device. This way, Depending on the direction of the elbow, a larger part of the flow (especially liquid) will move to the left or right. This in turn leads to increased uneven flow distribution in the separator, increased local velocity, and generally impaired separation performance.
Similar to the influence of the accessory on the droplet size and the entrainment fraction of the dispersed phase, the influence on the uneven flow distribution is adjusted according to the distance of the accessory from the inlet of the separator.
Various types of separator inlet devices and their effects on general separation performance, especially on gas/liquid separation, are discussed in detail in Part 1 of the gas/liquid separation series, and will not be repeated here.
Regarding the gas/liquid separation performance of the intake device itself, some improvements have been made to the calculations used:
Some inlet devices are characterized by a simple 90-degree change in flow direction (for example, no inlet device, simple splitter/baffle, half-pipe inlet, and vane inlet [calculation modified]). These inlet devices have been modeled as simple impingement separators, where the trapping efficiency of liquid droplets entrained in the gas is given by Burkholz (1989) in the equation. 6.
2. Correlation is added to estimate the proportion of bulk liquid that is "splashed" into droplets due to impact on vertical surfaces (for example, splitters/baffles), and the correlation of the droplet size distribution of the splashed liquid is also used .
3. For the cyclone inlet device, the approximate relationship between the number of inlet cyclones, the size of the cyclone tube and the removal efficiency of liquid droplets (liquid in gas) is adopted.
Since the focus of this article is oil/water separation, the above areas will not be discussed further.
Although several types of inlet devices may cause a certain degree of oil/water separation, mainly through centrifugal force and/or coalescence of dispersed phase droplets-especially cyclone inlet devices-there is little data to support or quantify these effects . Therefore, this article ignores any oil/water separation and corresponding changes in droplet size distribution that may be attributed to different inlet devices. This is an area of future investigation.
In addition to these calculations, a method similar to the method used to install the feed tube was used to estimate the influence of the inlet area of the inlet device on the turbulence/shear-induced droplet size. These are only estimates of expected behavior, but the results seem reasonable.
The method used to determine the size of the separator is similar in most respects to the method proposed by Grødal and Realff (1999), and the paper should be consulted for details. However, the current method does not use a predefined separable droplet size target as the basis for the calculation, but uses the required outlet fluid quality specification, combined with the estimated inlet droplet size distribution and the associated estimated dispersed phase volume as the phase separation The basis of calculation. The paper by Song et al. 2010 is one of the few other references in the literature that use methods somewhat similar to those discussed here.
Due to the various interrelated variables involved, the required calculations are inherently quite complex and iterative. In addition, for a given separation application, there may be multiple "technically feasible" solutions. Therefore, the calculation is carried out as an optimization exercise, and the objective function is to minimize the total weight of the ship-which should be related to the cost-while achieving the specified separation goals and being bound by definitions. For the gas/liquid calculations described previously, the Excel spreadsheet is used with the Solver add-in.
For a given set of fluid characteristics, operating conditions, and selected separator internals, the main driving factors for container size are as follows:
The amount of water in the separated oil
The amount of oil in the separated water
Separate the amount of liquid in the gas
Minimum distance between alarm and shutdown levels
Minimum control belt height, namely LIL-HIL, LLL-HLL
Minimum time between alarm and shutdown levels
The shortest time to control the volume, namely LIL-HIL, LLL-HLL
The maximum separator% full, namely HHLL/Di
Gas speed limit to prevent liquid re-entrainment
Oil-water axial speed limit
Figure 3 shows the geometry, main components and liquid level settings of a typical horizontal three-phase separator.
NLL-normal liquid (oil) level
For some applications, the minimum distance or time between levels will be controlled, depending on the selected hypothetical constraint value. Typical values are provided in Table 2.
Obviously, these values must be considered because they can drive the container size. The minimum distance requirement can have a significant impact on the diameter of the container, especially for containers less than 5 feet in diameter, and when using optimization procedures to minimize the weight of the container, the minimum residence/surging time requirement often affects the length of the container.
The logic behind these minimum distance/time applications is usually inconsistent. E.g:
What are the reasons for the rise or fall, are they credible, and the estimated frequency?
If left unchecked, what are the potential consequences of different ascent/decrease levels? Sometimes they are serious, sometimes they are not.
When an alert is received, what intervention options are available to the operator (if there is an operator), and how long will it take for these options to be implemented?
The author believes that more attention should be paid to providing sufficient "control volume" (the volume between low and high alarms) to handle intermittent/slug flow and smooth flow through downstream. In most oil and gas applications, the 2-5 minute "operator intervention time" requirement usually recommended in a typical published separator size guide may be unrealistic or unnecessary. Please also note that the application of the constraints given in Table 2 will often result in container dimensions that are inconsistent with the recommended liquid residence time guidelines from various sources. For difficult separation applications (for example, low WIO or OIW specifications, heavy oil [reduced oil/water density difference and high oil viscosity]), separation requirements sometimes exceed the above-mentioned minimum distance/level time limit, although usually increased The separation ability is most effectively achieved by increasing the length of the horizontal separator.
For horizontal three-phase separators, there are basically four configurations commonly used for liquid handling: overflow weirs; submerged weirs; buckets and weirs; and heavy (water) phase guidance.
Generally speaking, these have their pros and cons. The pilot design is usually only used for relatively small weight/water volumes, such as <10% of the total liquid.
In this article, the submerged weir configuration is assumed. From the point of view of operability, the author believes that it has advantages over other configurations. In future articles, various liquid handling configurations will be compared in more detail.
For liquid/liquid separation, in addition to the actual gravitational separation calculation (such as Stokes's Law or Intermediate Law), the following parameters are the key:
The relationship between the droplet size and volume of the dispersed phase (WIO, OIW) at the entrance of the gravity separation section
The effective "release point" characteristics of the dispersed phase at the entrance of the gravity separation section
Effective average continuous phase velocity distribution along the length of the gravity separation section
Item 1 has been discussed earlier in this article. Regarding item 2, for the oil phase and the water phase, it is assumed that the dispersed phase is uniformly "released" at the vertical height of each phase at the centerline of the container at the entrance of the gravity separation section. For the water phase, this height is the normal oil/water interface level (NIL), and for the oil phase, this height is the normal oil level (NLL) minus the normal oil/water interface level (NIL). Compared with the commonly used assumptions, the dispersed phase is released as a "point source" at the worst-case release point, that is, for the oil phase, the dispersed WIO is released in gas/oil. This is less conservative and more realistic in terms of effective gravity. Contact at the entrance of the separation section, while for the water phase, the dispersed OIW is released at the bottom of the container at the entrance of the effective gravity separation section. The effective average continuous phase velocity for a given flow rate is a function of several factors:
The axial cross-sectional area occupied by phases such as oil or water, based on the level used in item 2
Effective reselection zone length/diameter
Whether the perforated plate/baffle is used for rectification purposes.
According to the sedimentation trajectory shown in Figure 4, calculate the minimum droplet size separated with 100% efficiency.
These restricted trajectories are equivalent to equating the vertical settling time of a given dispersed phase with the horizontal transport time, or
H = the height of the continuous phase layer (the centerline of the vessel)
Le = effective separation zone length of continuous layer
Vt,100 = the sedimentation velocity of the smallest droplet removed at 100%
Vh = effective horizontal velocity of continuous phase
Equation 7 can be rearranged to give the minimum droplet settling/rising velocity corresponding to the limit trajectory:
This velocity can then be inserted into the appropriate sedimentation law expression (Stokes, middle) and solved for the smallest separable droplet diameter (100% removal) or D100. For example, Stokes's law is as follows:
Vt = droplet terminal settling/rising velocity
g = acceleration due to gravity
ρc = continuous phase density
µc = viscosity of the continuous phase
Assume that all droplets larger than this size are completely removed. Droplets smaller than this size will be partially removed.
In view of the assumption in item 2 above that the dispersed phase is uniformly released at the height of the vertical continuous phase, it can be proved that the removal efficiency of droplets smaller than d100 is given by the following formula:
= Separation efficiency of droplets smaller than d100
Vt = Settling velocity of a given droplet size smaller than d100
Vt,100 = the sedimentation velocity of the smallest droplet removed at 100%
Using the d100 value calculated for each stage (WIO, OIW), the removal efficiency of the smaller droplet size can be calculated from Equation 2. 10. Then apply these removal efficiencies to the relationship between the droplet size of each phase and the volume of the dispersed phase (item 1) to obtain overall droplet removal performance. The unremoved (smaller) droplets remain in the separated continuous phase leaving the separator and must be less than or equal to the specified requirements, such as WIO <5% v/v, OIW <2,000 ppmv. This "idealized" droplet separation model will be modified due to turbulence effects, as described below.
There are some empirical data available that indicate that the simple vector addition (Vt and Vh) that produces the tilted synthetic stable trajectory depicted in Figure 4 may not be suitable for all conditions. The main explanation for this is that turbulence in the continuous phase and related vertical velocity fluctuations hinder the gravitational settling of droplets, especially small droplets with low settling velocity.
The recommended operating limits usually take the following form:
Vh/Vt <15, the maximum Vh is 0.05–0.10 ft/s.
Where Vt is the settling velocity of the target separable droplet size.
The size determination method discussed in this article does not use the "target separable droplet size", so the Vh/Vt <15 limit does not apply.
Both of these constraints may be restrictive, but the limit of Vh,max = 0.05–0.10 ft/sec in particular may be a heavy specification that can push the separator diameter (axial flow area) upward.
Some historically published partition sizing methods include these restrictions, while others do not. The data available to justify these limits is quite limited, at least in the public domain, and there is some evidence that these horizontal speed limits are conservative.
For the current work, an axial velocity limit of 0.2 ft/sec is used for both liquid phases. If turbulence effects are correctly quantified, these restrictions should not be strictly necessary. The Vh,max = 0.2 ft/sec limit generally does not limit the situation studied in this article, although sometimes applying this limit to the water phase requires an increase in the oil/water interface level to provide more axial flow area and to compensate for the curvature of the lower part of the vessel. The effect of turbulence on droplet settlement used here is based on the method of Miedema and Vlasblom (1996), which involves settlement calculations used in the dredging industry. Although this method is generally not used for oil/water separation, the results seem reasonable.
Part 2 of the series provides a basic case example with specified parameters to illustrate the methods and calculations discussed in Part 1.
Bothamley, M. 2013. "Gas-Liquid Separator-Quantitative Separation Performance Part 1-3", SPE Oil and Gas Facilities, August to December.
Zhang, H.-Q., Sarica, C. 2005. "Unified Modeling of Gas/Oil/Water Pipe Flow-Basic Method and Preliminary Verification", SPE 95749, SPE ATCE, Dallas, October 9-12.
Keskin, C., Zhang, H.-Q., Sarica, C. 2007. "Recognition and Classification of New Three-Phase Gas/Oil/Water Flow Patterns", SPE 110221, SPE ATCE, Anaheim, California, November 11-14.
Taitel, Y., Barnea, D., Brill, JP 1995. "Layered Three-Phase Flow in Pipelines", International Journal of Multiphase Flow 21 (1), 53–60.
Heydari, N., Sadeghi, MT 2011. "Analysis and modeling of three-phase stratified flow in pipelines using mechanical methods", CHEMECA 2011, Sydney, Australia, September 18-21.
Hinze, OJ 1955. "The basis of the hydrodynamic mechanism of splitting in the dispersion process", AIChE J. 1:289-295.
Karabelas, AJ 1978. "Spectrum of droplet size produced in turbulent pipe flow of diluted liquid/liquid dispersion", AIChE J. 24: 170–180.
Simmons, MJH, Azzopardi, BJ 2001. "Dispersion-liquid droplet size distribution in pipe flow", International Journal of Multiphase Flow 27: 843–859.
Heijckers, C. 2012. "The Effect of Flow Regulation on Separation", SPE Webinar, September 12th.
Burkholz, A. 1989. Droplet separation, published by VCH.
Grodal, EO, Realff, MJ, "Optimal design of two-phase and three-phase separators: mathematical programming formulas", SPE 56645, SPE ATCE, Houston, Texas, October 3-6, 1999.
Song, JH, Jeong, BE, Kim, HJ and Gil, SS, "Three-phase separator size using droplet size distribution", OTC 20558, Offshore Technology Conference, Houston, Texas, May 3, 2010 Day to 6th.
Miedema, SA, Vlasblom, WJ 1996. "Hopper Deposition Theory", 29th Annual Texas A&M Dredging Symposium, New Orleans, June.
Mark Bothamley is an independent consultant specializing in oil and gas processing and related facilities, and is the President of Mark Bothamley Consulting LLC. For the past few years, he has been a senior technical consultant for PetroSkills, a large technical training company that supports the oil and gas industry. Prior to this, he was chief engineer at John M. Campbell and Company in Norman, Oklahoma for 8 years. Before joining JMC, he worked at BP/Amoco for 24 years, working in multiple locations around the world. His experience covers the design, operation, troubleshooting and optimization of offshore and onshore oil and gas production and processing facilities. Bothamley is the former chairman of the SPE Facilities Subcommittee and a former member of the GPSA Engineering Data Manual Editorial Review Committee. He received a bachelor's degree in chemical engineering from Lakehead University in Thunder Bay, Ontario, Canada, and a diploma in natural gas and petroleum technology from the British Columbia Institute of Technology in Vancouver, British Columbia, Canada.
Get JPT articles in your LinkedIn feed and stay up to date with oil and gas news and technology.
Journal of Petroleum Technology is the flagship magazine of the Society of Petroleum Engineers, providing authoritative briefings and topics about the advancement of exploration and production technology, oil and gas industry issues, and news about SPE and its members.
ISSN: 1944-978X (online) ISSN: 0149-2136 (print)