The internal structure of electric desalting vessels is an important factor affecting the performance of crude oil dehydration and desalination. The flow and distribution of oil–water emulsions are generally not fully considered in the traditional design of a desalter. Computational fluid dynamics (CFD) simulations enable a deep understanding of oil–water two-phase flows, thereby providing a scientific basis for optimizing structural design of desalters. In this paper, CFD simulations are used to study the flow of oil–water two-phase fluid in a desalting vessel with advection squirrel-cage structure in a SINOPEC refinery, revealing the defects in its electric field layout. A structural improvement scheme is then proposed, and a medium-scale bypass test is conducted. The test results show that the single-stage salt removal rate is increased from 85.4% to 92.1%. The water and salt contents in desalted oil and the oil content in brine are also significantly improved, providing a solid foundation for transformation of the related industrial equipment.

Oil extracted from oil fields contains a certain amount of water and salt that will cause unpredictable risks in the process of oil refining.1 In oil refineries, electric desalting is usually the first important process that is used to remove harmful impurities in crude oil, such as inorganic salts, water, and solid particles. It is necessary to mitigate corrosion of the overhead cooling system and scaling of the heat exchangers in crude distillation units, and to prevent catalyst poisoning in residue fluid catalytic cracking units and residue hydroprocessing units. The performance of electric desalters has a direct impact on the long-term operation of the subsequent processing units.

To achieve the ideal desalting performance, a variety of developed different electric desalting technologies have been developed. For example, Petrolite Corporation developed the BILECTRIC desalter, which distributes the emulsion of oil and water between horizontal electrode grids to achieve rapid coalescence of water droplets, which provides a higher throughput than traditional low-speed desalters with the same vessel size.2,3 By changing from an AC electric field to a modulated high-frequency AC electric field, Cameron Inc. has increased rag layer collapse and the coalescence efficiency for droplets of various sizes, and has improved the adaptability of the BILECTRIC desalter to crude oils, especially those with high basic sediment and water (BS&W).4 The NATCO group developed the Electro-Dynamic Desalting technology combining a dual-polarity electrostatic treater with four additional innovations: composite electrodes, a load-responsive controller, a counterflow dilution water process, and an electrodynamic mixing process, which can promote the performance of a single-stage system to two-stage.5–8 Another patented NATCO Dual Frequency desalter uses a microprocessor-based system to generate a programmed voltage waveform with appropriate frequency and amplitude, which can optimize coalescence of water droplets, leading to good adaptability to high-conductivity crudes, higher treating capacity or smaller vessel size, and less chemical usage and power consumption.9 

The structure of the electric desalting vessel is one of the most important factors that affects performance, including the entry and distribution of oil and water, the arrangement of electrode structure and electric field, and the flow of crude oil in the tank. In the traditional design of an electric desalter, factors such as crude oil properties, processing capacity, and cost have usually been taken into account first to determine the basic type, such as AC desalter, AC/DC desalter, or high-speed electric desalter.10–12 Then, parameters such as the residence time of oil and water phases, the action time of the electric field region, and the rise rate of the oil are calculated to obtain the size and structure of the desalter tank. However, consideration of these parameters alone is not sufficient to understand the flow and distribution of the oil–water emulsion.

Numerical simulation offers an efficient method to understand the flow field, the electric field distribution, and the water droplet coalescence process in the desalting vessel and to optimize the inner structure. Sams and Wallace13 introduced computational fluid dynamics (CFD) models, calculations, and testing methods to identify and improve the hydraulic characteristics of vertical fluid in horizontal desalting vessels, and revealed the uniform distribution and collection mechanism of oil/water two-phase fluids. Using a CFD simulation based on a multiphase model coupled with turbulent interactions, Lu et al.14 evaluated the effectiveness of perforated plate baffles for improving the flow and separation performance of a free-water knockout (FWKO) separator. Lee et al.15 adopted a CFD simulation technique to adjust the equipment inside the separator and change the fluid phase level, through which the volume utilization ratio of the overall separator and both the high- and low-pressure separators were increased. Tarantsev and Tarantseva16 investigated the effect of electric field nonuniformity on the creation and destruction of water–oil emulsions during crude oil desalting by simulating both the flow field and the electric field in an apparatus with three stages of emulsion separation.

The Sinopec Tahe Refining and Chemical Company processes crude oil from the Tahe oil field. This oil has a density of more than 0.95 g/cm3, an average salt content of more than 400 mg NaCl/L, and an asphaltene content of more than 15%, making dehydration and desalination very difficult. Although its 2# atmospheric and coking unit has adopted four-stage electric desalters, the average salt content of crude oil after desalination is as high as 6.7 mg NaCl/L, which is far from meeting the conventional requirement of 3 mg NaCl/L salt content of desalinated crude oil. In this paper, a CFD simulation approach is adopted to study the flow state in advective squirrel-cage electric desalting vessel for processing Tahe crude oil, and the reasons for poor dehydration and desalination effects are examined. An optimization scheme is designed and verified by experiments, and good results are obtained.

Tahe crude oil is a heavy crude oil. Its properties are very detrimental to desalination (density 0.95 g/cm3 at 20 °C, API gravity 17, kinematic viscosity 2877 mm2/s at 50 °C and 147.1 mm2/s at 80 °C, water content 0.55%, salt content 300–400 mg NaCl/L, resin content 19.60%, and asphaltene content 16.14%).

A crude distillation unit that processes Tahe crude oil has a processing capacity of 350 Mt/a and has adopted four-stage electrostatic desalters. However the salt content after desalination can only be reduced to about 5 mg NaCl/L, as shown in Table I. The electric desalting vessel of stages 1–3 has an advection squirrel-cage type structure, as shown in Fig. 1. The crude oil–water emulsion enters from the bottom of one end of the vessel, passes through an orifice plate distributor, and then flows horizontally through three stages of squirrel-cage electric field. Finally, the desalted crude oil flows out from the top of the other end. It is designed to improve space utilization and reduce backmixing between the crude oil and the settling water droplets.

TABLE I.

Processing effects of four-stage electric desalting of Tahe crude oil.

Sampling pointSalt content (mg NaCl/L)Single-stage desalination rate (%)Water content (% m/m)Single-stage dehydration rate (%)
Before desalting 446.2 ⋯ 0.32 ⋯ 
After one-stage desalting 84.3 81.1 1.28 −300 
After two-stage desalting 42.4 49.7 1.32 −3.1 
After three-stage desalting 13.2 68.9 1.25 5.3 
After four-stage desalting 5.2 60.6 0.22 82.4 
Sampling pointSalt content (mg NaCl/L)Single-stage desalination rate (%)Water content (% m/m)Single-stage dehydration rate (%)
Before desalting 446.2 ⋯ 0.32 ⋯ 
After one-stage desalting 84.3 81.1 1.28 −300 
After two-stage desalting 42.4 49.7 1.32 −3.1 
After three-stage desalting 13.2 68.9 1.25 5.3 
After four-stage desalting 5.2 60.6 0.22 82.4 
FIG. 1.

Schematic of advection squirrel-cage electric desalting vessel.

FIG. 1.

Schematic of advection squirrel-cage electric desalting vessel.

Close modal

To improve the desalination and dehydration effects, an improved structure is proposed. The main design principles are to make all crude oil flow through the electric field region smoothly and to ensure an electric field that is strong enough that there is no need for the oil to remain in the electric field for a long time. The detailed improvements are as follows:

  • The oil enters from the bottom of the tank and leaves from the top, and the secondary oil inlet distribution trough is set up to ensure a smooth flow of oil through the electric field.

  • The electric field distribution is simplified, leaving only one layer of strong electric field (vertical hanging plates), and the distance between the plates is reduced to increase the electric field strength.

  • The area of the electrode plates is decreased by reducing their height, thereby reducing the desalting current and thus avoiding transformer overload and improving operational stability.

  • The plates installation position is raised to increase the space between the oil–water interface and the plates, thereby maintaining the stability of the strong electric field and improving the drainage water quality.

  • The electrode plates are changed from grid type to solid plates to make the electric field more uniform and reduce the possibility of point discharge.

A test vessel with a capacity of 10 m3/h based on the above improvements was constructed. Figure 2 is a schematic of this test vessel. Figure 3 compares the structures of the new electrode plate and the traditional electrode plate.

FIG. 2.

Schematic of test vessel.

FIG. 2.

Schematic of test vessel.

Close modal
FIG. 3.

Comparison of electrode plate structures: (a) structure of traditional grid-type electrode; (b) structure of new type electrode.

FIG. 3.

Comparison of electrode plate structures: (a) structure of traditional grid-type electrode; (b) structure of new type electrode.

Close modal

CFD simulations were carried out on the advective squirrel-cage electric desalting vessel17 (model A) and the improved electric desalting vessel (model B). The geometric model was established on a 1:1 scale and the grid was then divided, as shown in Figs. 4 and 5. The CFD solution procedure was performed using Ansys Fluent software.

FIG. 4.

Schematic of grid division of model A.

FIG. 4.

Schematic of grid division of model A.

Close modal
FIG. 5.

Schematic of grid division of model B.

FIG. 5.

Schematic of grid division of model B.

Close modal

The medium properties were a density of 949.5 kg/m3, a kinematic viscosity of 9.88 × 10−5 m2/s, a dynamic viscosity of 0.0938 Pa·s, and a water content of 5% v/v. The input conditions for model A were a processing capacity of 439 m3/h, an inlet-pipe cross-sectional area of 0.1146 m2, an inlet velocity of 1.065 m/s, and a turbulence intensity of 5.65%, and those of model B were a processing capacity of 10 m3/h, an inlet-pipe cross-sectional area of 0.0036 m2, an inlet velocity of 0.75 m/s, and a turbulence intensity of 7.3%. The simulation adopted inlet velocity and outlet pressure boundary conditions, a pressure-based steady-state solver, a multiphase flow model for the mixture, and the standard k-ε turbulence model.

1. Advection squirrel-cage electric desalting vessel (model A)

To analyze the internal flow field of the electric desalting vessel, three radial sections of each squirrel-cage electrode and an axial section of the whole tank were considered. Figure 6 shows velocity nephograms of the oil phase at each radial section of model A. It is found that the fluid velocity in the lower part of the tank is greater than that in the upper part, which indicates that the flow rate of crude oil in the lower part is greater. As the crude oil flows to the outlet, the unevenness of the velocity gradually decreases. Table II compares the average flow velocities, residence times, and flow rates of the three sections of the electric field region. It can be seen that the closer the location is to the outlet, the greater is the average flow velocity. At the same time, a portion of the crude oil does not flow through the electric field, which indicates a problem in the structural design of the squirrel-cage electric desalting vessel.

FIG. 6.

Oil-phase velocity nephograms of each radial section of model A.

FIG. 6.

Oil-phase velocity nephograms of each radial section of model A.

Close modal
TABLE II.

Average residence times in three sections of the electrode region.

LocationLength of electric field region (m)Average velocity of oil phase (m/s)Residence time (s)Percentage of flow rate in electrode region (%)
Right electric field (near entrance) 6.4 0.002 3–0.005 3 2783–1208 39.4 
Intermediate electric field 6.4 0.003 72–0.006 4 1720–1000 19.6 
Left electric field (near exit) 0.007 68–0.024 7 651–202 10.1 
LocationLength of electric field region (m)Average velocity of oil phase (m/s)Residence time (s)Percentage of flow rate in electrode region (%)
Right electric field (near entrance) 6.4 0.002 3–0.005 3 2783–1208 39.4 
Intermediate electric field 6.4 0.003 72–0.006 4 1720–1000 19.6 
Left electric field (near exit) 0.007 68–0.024 7 651–202 10.1 

Figure 7 shows streamline diagrams of oil–water two-phase flow in the tank at different times. After the oil–water emulsion enters the electric desalting vessel, several large eddies are generated and propagate from the inlet to the outlet. The overall flow regime is in a relatively uniform advection state. However, Fig. 7 clearly shows that some crude oil does not flow through the electric field region, which is one of the most critical reasons for the poor dehydration and desalination effect of the first three-stage electric desalting vessel.

FIG. 7.

Streamline diagrams of axial longitudinal section of model A.

FIG. 7.

Streamline diagrams of axial longitudinal section of model A.

Close modal

The aim of the advection squirrel-cage structure design is to make the fluid flow horizontally, in contrast to a bottom-in top-out structure, so that it does not flow backward with the settling water droplets, thus reducing back-mixing of the water droplets with the crude oil and increasing the settling velocity of the water droplets. However, Fig. 7 shows that owing to the arrangement of the squirrel-cage electric fields in the upper part of the tank, the fluid flowing horizontally in the lower part of the tank does not pass through any electric field for most of the residence time, and it only passes through the two layers of horizontal grid electrodes at the end of the last section of the electric field near the outlet. In addition, the distance between the two layers of the squirrel-cage electrode structure is too large to provide a sufficiently high electric field strength. As a result, the final dehydration and desalination performance is poor.

2. Structurally improved tank (model B)

Figure 8 shows the velocity nephograms of the oil phase at each axial cross section of model B. The flow velocity in the region near the tank wall and outlet pipe is larger, and the velocity distribution in the rest of the space is more uniform. Compared with the squirrel-cage structure, the bottom-in top-out structure has a much larger cross-sectional area perpendicular to the fluid flow direction, and so the fluid velocity is very low, which makes the bulk velocity distribution more uniform in the tank. According to the simulation result, the average rise velocity of the oil phase in the electric field is about 0.425 mm/s. Together with the streamline diagram shown in Fig. 10(a) below, this indicates that the flow state in model B is as expected, with most of the crude oil passing smoothly through the electric field. The situation that arises with model A in which some crude oil does not flow through the electric field does not occur.

FIG. 8.

Oil-phase velocity nephograms of axial cross section of model B.

FIG. 8.

Oil-phase velocity nephograms of axial cross section of model B.

Close modal

To obtain a more ideal flow state and oil–water distribution, the oil inlet distributor was redesigned as shown in Fig. 9. The fluid first enters from a perforated inlet distribution pipe, and then passes into the electric desalination tank through a secondary distributor that is an inverted perforated trough. Figures 10 and 11 present comparisons of streamline diagrams and oil-phase volume fraction nephograms, respectively, of the fluid in the vessel with and without the secondary distribution trough. Figure 10 shows that the addition of the secondary distribution trough eliminates the large eddies at both ends of the vessel, thereby avoiding oil–water interface turbulence and reducing oil–water remixing. Figure 11 shows that the secondary distributor makes the oil–water interface more distinct, which is conducive to dehydration.

FIG. 9.

Structure of oil inlet distributor.

FIG. 9.

Structure of oil inlet distributor.

Close modal
FIG. 10.

Streamline diagrams of axial longitudinal section of model B: (a) without a secondary oil inlet distribution trough; (b) with a secondary oil inlet distribution trough.

FIG. 10.

Streamline diagrams of axial longitudinal section of model B: (a) without a secondary oil inlet distribution trough; (b) with a secondary oil inlet distribution trough.

Close modal
FIG. 11.

Oil-phase volume fraction nephograms of axial longitudinal section of model B: (a) without a secondary oil inlet distribution trough; (b) with a secondary oil inlet distribution trough.

FIG. 11.

Oil-phase volume fraction nephograms of axial longitudinal section of model B: (a) without a secondary oil inlet distribution trough; (b) with a secondary oil inlet distribution trough.

Close modal

The distribution revealed by the volume fraction nephograms of the oil phase in the tank shown in Fig. 11 does not take into account the effect of the electric field on the aggregation of water droplets, but is caused by the flow state in the tank and natural sedimentation. Owing to the density difference between oil and water, the water content is higher in the lower part of the tank, and the highest water content occurs in the inlet distribution pipe. However, in the absence of the secondary distribution trough, no clear oil–water interface can be seen in Fig. 11(a). The water content of crude oil in the distribution trough is high, and there is obvious oil–water stratification outside the trough, with a clear oil–water interface, as shown in Fig. 11(b).

The above simulation results clearly show that there is a severe problem with model A, resulting in poor dehydration and desalination effects. By contrast, the flow state in model B is in line with expectations, representing a significant improvement.

The test vessel based on model B was run in parallel with the first-stage electrical desalting vessel of the industrial unit, with feedstock undesalted Tahe crude oil that had been mixed with water and demulsifier. Thus, the main process variables, such as the operating temperature, the demulsifier injection rate, the water injection rate, and the mixing pressure drop were consistent with those for model A.

After the system was running smoothly, the salt content of the crude oil (measured by the Coulomb method), the water content of the crude oil (measured by the distillation method), and the oil content in the drainage (measured by the ultraviolet fluorescence method) were sampled and tested twice a day. The test lasted for 10 days, and the data for the model B test tank and for model A were compared and analyzed. Figure 12 shows a photograph of the test equipment.

FIG. 12.

In-field photograph of test equipment.

FIG. 12.

In-field photograph of test equipment.

Close modal

Table III shows a comparison between the results with the test vessel and those with the industrial vessel. Overall, by improving the structure of the electric desalting vessel, the electric desalting effect has been significantly improved. The water content of crude oil after desalting is reduced to a more ideal level, and the salt content after desalting and the oil content of the drainage are also significantly reduced. From the water content after desalting, it can be seen that the pilot test desalination vessel completely eliminates the defects of the original squirrel-cage structure, with a water content of crude oil of 0.22% after the test vessel, compared with over 1% after the industrial squirrel-cage vessel. With regard to the salt content after desalting, the pilot test vessel gives 32.4 mg NaCl/L with a desalination rate of 92.1%, which is significantly better than that of the squirrel-cage vessel under the same operating conditions, with a salt content of 59.5 mg NaCl/L and a desalination rate of 85.4%. The single-stage desalination rate of the test vessel has basically reached that of two-stage squirrel-cage vessels. In addition, compared with the industrial squirrel-cage vessel, the oil content of the drainage water of the experimental plant was significantly reduced from 220 to 78 mg/L on average.

TABLE III.

Comparison of results between test equipment and squirrel-cage equipment for processing of crude oil.

Salt contentSingle-stageTotal desalinationWater contentOil content of
Sampling point(mg NaCl/L)desalination rate (%)rate (%)(% m/m)drainage (mg/L)
Ave.RangeAve.Ave.Ave.RangeAve.Range
Before desalting409353–4680.230.2–0.5
Industrial After one-stage desalting 59.5 55.0–65.6 85.4 85.4 1.01 0.9–1.2 220 35.1–558 
squirrel-cage After two-stage desalting 29.6 26.9–36.5 50.2 92.8 1.36 0.7–1.5 
vessel outlet After three-stage desalting 13.2 11.9–14.6 55.4 96.8 1.26 0.8–1.6 
Pilot test vessel outlet 32.4 21.5–42.6 92.1 92.1 0.22 0.1–0.3 78 2.68–524 
Salt contentSingle-stageTotal desalinationWater contentOil content of
Sampling point(mg NaCl/L)desalination rate (%)rate (%)(% m/m)drainage (mg/L)
Ave.RangeAve.Ave.Ave.RangeAve.Range
Before desalting409353–4680.230.2–0.5
Industrial After one-stage desalting 59.5 55.0–65.6 85.4 85.4 1.01 0.9–1.2 220 35.1–558 
squirrel-cage After two-stage desalting 29.6 26.9–36.5 50.2 92.8 1.36 0.7–1.5 
vessel outlet After three-stage desalting 13.2 11.9–14.6 55.4 96.8 1.26 0.8–1.6 
Pilot test vessel outlet 32.4 21.5–42.6 92.1 92.1 0.22 0.1–0.3 78 2.68–524 

The test results have verified the CFD simulation and once again have shown that the dehydration and desalination effects of the squirrel-cage desalter are poor, with significant room for improvement. On the basis of the results of the CFD simulation and the pilot test, we believe that the performance of the advection squirrel-cage desalter can be improved by modifying the fluid distribution and the geometry and layout of the electrode structure.

The conclusions of this study can be summarized as follows:

  1. CFD simulations have been used to study the flow state in processing of Tahe crude oil using an advection squirrel-cage electric desalting vessel and it has been found that the design of the vessel’s internal structure is defective. Part of the crude oil does not pass through the electric field, which is the key reason for the poor dehydration and desalination effects.

  2. An improvement to the structure of the electric desalting vessel has been proposed. The oil inlet and outlet and the electrode structure have been redesigned. CFD simulation results show that the new structure can ensure the smooth flow of crude oil through the electric field and completely overcome the defects of the squirrel-cage structure.

  3. On the basis of the new design, a 10 m3/h test vessel has been constructed and a bypass test has been performed. The test results show that the single-stage desalination rate is increased from 85.4% to 92.1%, and the water and salt contents after desalting and the oil content of the drainage are significantly reduced, thus providing a solid basis for the improvement of the industrial equipment.

This work is supported by the SINOPEC research project “Research and development of electric desalting technology and equipment for heavy crude oil” (Project No. 318020-9).

The authors have no conflicts to disclose.

Lei Han: Writing – original draft (equal). Yanling Zhang: Writing – review & editing (supporting). Luna Niu: Project administration (supporting). Weiwei Xu: Software (equal); Writing – review & editing (supporting).

The data that support the findings of this study are available on request from the authors.

1.
L.
Vafajoo
,
K.
Ganjian
, and
M.
Fattahi
, “
Influence of key parameters on crude oil desalting: An experimental and theoretical study
,”
J. Pet. Sci. Eng.
90–91
,
107
111
(
2012
).
2.
R. B.
Martin
, “
Plural parallel stage desalting and dehydration
,” U.S. patent 4209374 (
24 June 1980
).
3.
G. W.
Sams
and
S. P. B.
Mandewalkar
, “
Petroleum desalting utilizing voltage modulation
,” U.S. patent 9764336B2 (
19 September 2017
).
4.
K. W.
Warren
, “
New tools for heavy oil dehydration
,” in paper presented at the
SPE International Thermal Operations and Heavy Oil Symposium and International Horizontal Well Technology Conference
,
Calgary, AB, Canada
,
4–7 November 2002
, SPE/Petroleum Society of CIM/CHOA 78944.
5.
C.
Noïk
,
J.
Chen
, and
C.
Dalmazzone
, “
Electrostatic demulsification on crude oil: A state-of-the-art review
,” in paper presented at the
SPE International Oil and Gas Conference and Exhibition
,
Beijing, China
,
5–7 December 2006
, Paper No. SPE 103808.
6.
E.
Sellman
,
G.
Sams
, and
S.
Pavan Mandewalkar
, “
Benefits of using advanced electrostatic fields in crude oil dehydrators and desalters
,” in
paper presented at the SPE International Production and Operations Conference and Exhibition,
Doha, Qatar
,
14
16
May 2012, Paper No. SPE 156299
.
7.
F. L.
Prestridge
and
B. C.
Johnson
, “
Distributed charge composition electrodes and desalting system
,” U.S. patent 4702815 (
27 October 1987
).
8.
F. L.
Prestridge
and
B. C.
Johnson
, “
Electrostatic mixer/separator
,”
U.S. patent 4606801
(
19 August 1986
).
9.
G. W.
Sams
, “
Separating multiple components of a stream
,” U.S. patent 7758738B2 (
20 July 2010
).
10.
E.
Sellman
,
P.
Mandewalkar
, and
G.
Sams
, “
Improved desalting of challenging crude slates
,” in paper presented at the
SPE Kuwait Oil and Gas Show and Conference
,
Mishref, Kuwait
,
7–10 October 2013
, Paper No. SPE 167376.
11.
E.
Sellman
,
G.
Sams
, and
S.
Pavan Mandewalkar
, “
Improved dehydration and desalting of mature crude oil fields
,” in paper presented at the
SPE Middle East Oil and Gas Show and Conference
,
Manama, Bahrain
,
10–13 March 2013
, Paper No. SPE 164289.
12.
E.
Sellman
and
P.
Mandewalkar
, “
Highly efficient dehydration and desalting of crude oil from mature fields in the middle east
,” in paper presented at the
SPE Annual Technical Conference and Exhibition
,
Dubai, UAE
,
26–28 September 2016
, Paper No. SPE-181724-MS.
13.
G. W.
Sams
and
H. G.
Wallace
, “
Improving process efficiencies by optimizing fluid hydraulics electrostatic oil dehydrators
,” in
paper presented at the Offshore Technology Conference,
Houston, TX
,
2001
,
Paper No. OTC-13216-MS
.
14.
Y.
Lu
,
J. M.
Lee
,
D.
Phelps
, and
R.
Chase
, “
Effect of internal baffles on volumetric utilization of an FWKO – A CFD evaluation
,” in
SPE Annual Technical Conference and Exhibition
(
Society of Petroleum Engineers
,
Anaheim, CA
,
2007
), Paper No. SPE-109944-MS.
15.
J. M.
Lee
,
R. I.
Khan
, and
D. W.
Phelps
, “
Debottlenecking and computational-fluid-dynamics studies of high- and low-pressure production separators
,”
SPE Proj., Facil. Constr.
4
(
4
),
124
131
(
2009
).
16.
K. V.
Tarantsev
and
K. R.
Tarantseva
, “
Effect of electric field nonuniformity on the processes of creation and destruction of water-oil emulsions during crude oil desalting
,”
Chem. Pet. Eng.
53
(
11–12
),
707
710
(
2018
).
17.
S.
Lou
,
P.
Jia
,
O.
Sha
et al, “
Electric dewcatering and desalting appliance
,” China patent CN1140611C (
3 March 2004
).