Intelligent identification, monitoring and control system for oil-water interface of oil storage settling tank

Abstract

The application provides a storage oil settling tank oil-water interface intelligent identification monitoring and regulating system, and relates to the technical field of intelligent control. The system comprises the following steps: collecting data through a microwave-temperature composite sensor array according to the sensor layout position to obtain measured results, wherein the measured results include pure water layer thickness, emulsion layer thickness and pure oil layer thickness; comparing the measured emulsion layer thickness with the emulsion layer thickness prediction value, and recalibrating the three-dimensional model if the deviation is greater than 10%; calling a demulsifier injection dynamic curve to automatically add chemicals when the emulsion layer thickness is greater than 0; executing a temperature control parameter set to start heating when the pure water layer thickness is greater than 0 and the pure oil layer thickness is greater than 0; discharging the water layer from the primary tank bottom after the temperature control is completed and detecting the water content of the discharged water in real time; and returning to the primary tank for further treatment when the water content of the discharged water is greater than 5% and the predicted recovery time is less than or equal to a threshold value. The application can improve the storage oil treatment efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, and in particular to an intelligent identification, monitoring and control system for the oil-water interface in an oil settling tank. Background Technology

[0002] During crude oil extraction and storage, the crude oil produced from the wellhead is often accompanied by a large amount of formation water, silt, and other impurities, forming an oil-water mixture. To meet the requirements of subsequent refining and pipeline transportation, oil-water separation is necessary through equipment such as settling tanks. Accurate identification of the oil-water interface and efficient control of the separation process are the core links to ensure the quality of crude oil purification, reduce energy consumption, and minimize environmental pollution.

[0003] Existing oil-water separation technologies in oil storage settling tanks primarily rely on the traditional gravity settling principle, supplemented by demulsifier injection and heating. However, the following problems exist in actual operation:

[0004] For example, the accuracy of interface recognition is insufficient. Traditional technologies often use single-point sensors (such as float-type and capacitive type) to monitor the oil-water interface. However, due to the emulsification characteristics of crude oil, a stable emulsion layer often forms in the settling tank. Its thickness and distribution change dynamically, and single-point measurement is difficult to reflect the overall layering state, which can easily lead to misjudgment.

[0005] For example, process control is lagging and inefficient. The dosage of demulsifier, heating temperature, and timing are mostly set manually based on experience, lacking dynamic matching with the real-time separation state. For instance, if the dosage is not adjusted in time when the emulsion layer thickness increases suddenly, it will lead to incomplete demulsification; excessive heating increases energy consumption, while insufficient heating slows down the separation process. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide an intelligent identification, monitoring and control system for the oil-water interface in oil collection and storage tanks, which can improve the efficiency of oil collection and storage processing.

[0007] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0008] The first aspect is a method for intelligent identification, monitoring, and control of the oil-water interface in oil settling tanks, including:

[0009] Step 1: Based on the geometric structure of the target settling tank, the deployment location of the microwave-temperature composite sensor array, and the material characteristic parameters, construct a multi-physics coupled three-dimensional model.

[0010] Step 2: Based on the three-dimensional model, simulate the oil-water separation process using the finite element method to output key parameters, including predicted values ​​for stratification rate, emulsion layer thickness, demulsifier injection dynamic curve, stratification recovery time, and temperature control parameter set.

[0011] Step 3: Based on the sensor deployment locations in Step 1, data is collected using a microwave-temperature composite sensor array to obtain the measured results, which include the thickness of the pure water layer, the thickness of the emulsion layer, and the thickness of the pure oil layer.

[0012] Step 4: Compare the measured emulsion layer thickness in Step 3 with the predicted emulsion layer thickness in Step 2. If the deviation is >10%, recalibrate the 3D model and return to Step 2 to resimulate.

[0013] When the emulsion layer thickness output in step 3 is greater than 0, the demulsifier addition dynamic curve from step 2 is invoked for automatic dosing.

[0014] Step 5: When the output of Step 3 shows that the thickness of the pure water layer is >0 and the thickness of the pure oil layer is >0, execute the temperature control parameter set of Step 2 to start heating;

[0015] Step 6: After temperature control is completed in step 5, drain the water layer from the bottom of the primary tank and monitor the water content of the drain in real time;

[0016] If the moisture content of the wastewater is >5%, and the predicted recovery time is ≤ the threshold, it should be returned to the primary tank for further processing, or

[0017] When the predicted recovery time is greater than the threshold, the crude oil is introduced into the secondary purification tank; if the water content of the wastewater is less than or equal to 5%, the crude oil is transferred to the secondary purification tank and the water is controlled until the water content of the wastewater is less than or equal to 0.5%, and finally purified crude oil is output.

[0018] Secondly, the intelligent identification, monitoring, and control system for the oil-water interface in the oil settling tank includes:

[0019] The module is used to construct a multi-physics coupled three-dimensional model based on the geometry of the target settling tank, the deployment location of the microwave-temperature composite sensor array, and material characteristic parameters.

[0020] The output module is used to simulate the oil-water separation process using the finite element method based on the three-dimensional model, and output key parameters, including the predicted value of the separation rate, the predicted value of the emulsion layer thickness, the dynamic curve of demulsifier injection, the predicted value of the separation recovery time, and the temperature control parameter set.

[0021] The measurement module is used to collect data through a microwave-temperature composite sensor array based on the sensor deployment location to obtain measurement results, including the thickness of the pure water layer, the thickness of the emulsion layer, and the thickness of the pure oil layer.

[0022] The comparison module is used to compare the measured emulsion layer thickness with the predicted emulsion layer thickness. If the deviation is >10%, the 3D model is recalibrated and the simulation is restarted. When the output emulsion layer thickness is >0, the demulsifier injection dynamic curve is invoked for automatic dosing.

[0023] The control module is used to initiate heating by executing the temperature control parameter set when both the output pure water layer thickness and the pure oil layer thickness are greater than 0. After temperature control is completed, the water layer is discharged from the bottom of the primary tank, and the water content of the discharged water is monitored in real time. If the water content of the discharged water is greater than 5%, and the predicted recovery time is less than or equal to the threshold, the system returns to the primary tank for further processing.

[0024] When the predicted recovery time is greater than the threshold, the crude oil is introduced into the secondary purification tank; if the water content of the wastewater is less than or equal to 5%, the crude oil is transferred to the secondary purification tank and the water is controlled until the water content of the wastewater is less than or equal to 0.5%, and finally purified crude oil is output.

[0025] The above-described solution of the present invention has at least the following beneficial effects:

[0026] By acquiring thickness data of the pure water layer, emulsion layer, and pure oil layer using a microwave-temperature composite sensor array, three-dimensional monitoring of the stratification state within the settling tank was achieved, overcoming the limitation of traditional single-point sensors in capturing the dynamic distribution of the emulsion layer. Combining simulation predictions from a multi-physics coupled three-dimensional model with real-time comparison and calibration of measured data (automatic model recalibration when deviation exceeds 10%) further reduces measurement errors caused by material characteristic fluctuations and in-tank flow field interference, providing precise data support for subsequent control.

[0027] Based on real-time monitoring of the emulsion layer thickness, the system automatically calls upon the demulsifier dosing dynamic curve for precise dosing, avoiding issues of insufficient or excessive dosage caused by manual experience-based dosing. This improves demulsification efficiency and reduces reagent consumption. Simultaneously, through the linkage control of temperature control parameters and stratification status (heating is only activated when both pure water and pure oil layers are present), energy is allocated on demand, reducing energy waste caused by ineffective heating.

[0028] During the drainage stage, graded treatment is achieved through real-time monitoring of moisture content: when the drainage moisture content exceeds 5%, the system intelligently selects to return the wastewater to the tank for further treatment or to introduce it into the secondary purification tank based on the stratified recovery time threshold, thus avoiding interference from unqualified media to subsequent processes; when the moisture content meets the standard, a stepped water cut-off control is implemented (transferring crude oil when ≤5%, and terminating water cut-off when ≤0.5%) to ensure that the final output purified crude oil meets high standards, reducing crude oil loss and minimizing environmental risks.

[0029] The entire process requires no manual intervention. From model simulation, parameter acquisition, dosing temperature control to media transfer, closed-loop automatic control is achieved. This not only reduces the intensity of manual labor but also avoids efficiency losses caused by delays or misjudgments due to human operation. Through dynamic model calibration and simulation iteration, the separation process always matches the actual working conditions, shortening the stratification time and improving the continuous operation capability of the equipment. Attached Figure Description

[0030] Figure 1This is a schematic diagram of the intelligent identification, monitoring and control method for the oil-water interface in an oil settling tank provided in an embodiment of the present invention.

[0031] Figure 2 This is a schematic diagram of an intelligent identification, monitoring and control system for the oil-water interface in an oil settling tank provided in an embodiment of the present invention. Detailed Implementation

[0032] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0033] like Figure 1 As shown, embodiments of the present invention propose an intelligent identification, monitoring, and control method for the oil-water interface in oil settling tanks, comprising:

[0034] Step 1: Based on the geometric structure of the target settling tank, the deployment location of the microwave-temperature composite sensor array, and the material characteristic parameters, construct a multi-physics coupled three-dimensional model.

[0035] Step 2: Based on the three-dimensional model, simulate the oil-water separation process using the finite element method to output key parameters, including predicted values ​​for stratification rate, emulsion layer thickness, demulsifier injection dynamic curve, stratification recovery time, and temperature control parameter set.

[0036] Step 3: Based on the sensor deployment locations in Step 1, data is collected using a microwave-temperature composite sensor array to obtain the measured results, which include the thickness of the pure water layer, the thickness of the emulsion layer, and the thickness of the pure oil layer.

[0037] Step 4: Compare the measured emulsion layer thickness in Step 3 with the predicted emulsion layer thickness in Step 2. If the deviation is >10%, recalibrate the 3D model and return to Step 2 to resimulate.

[0038] When the emulsion layer thickness output in step 3 is greater than 0, the demulsifier addition dynamic curve from step 2 is invoked for automatic dosing.

[0039] Step 5: When the output of Step 3 shows that the thickness of the pure water layer is >0 and the thickness of the pure oil layer is >0, execute the temperature control parameter set of Step 2 to start heating;

[0040] Step 6: After temperature control is completed in step 5, drain the water layer from the bottom of the primary tank and monitor the water content of the drain in real time;

[0041] If the moisture content of the wastewater is >5%, and the predicted recovery time is ≤ the threshold, it should be returned to the primary tank for further processing, or

[0042] When the predicted recovery time is greater than the threshold, the crude oil is introduced into the secondary purification tank; if the water content of the wastewater is less than or equal to 5%, the crude oil is transferred to the secondary purification tank and the water is controlled until the water content of the wastewater is less than or equal to 0.5%, and finally purified crude oil is output.

[0043] In this embodiment of the invention, step 1 involves constructing a multi-physics coupled three-dimensional model by integrating the geometry of the settling tank, the location of sensors, and material characteristic parameters, thus overcoming the limitations of traditional single-physics models. This model can comprehensively reflect the microwave propagation characteristics, temperature field distribution, and material flow state during the oil-water separation process within the tank, providing a digital foundation that closely reflects actual working conditions for subsequent simulations. It avoids simulation distortion caused by neglecting the interactive effects of multiple factors and enhances the model's adaptability to different tank types and different crude oil properties.

[0044] Step 2 involves simulating the 3D model using the finite element method. This allows for the prediction of key parameters such as delamination rate and emulsion layer thickness, and generates control schemes including dynamic curves for demulsifier addition and temperature control parameter sets. This process achieves a "digital twin" simulation of the separation process, enabling optimization of process parameters before actual operation and reducing trial-and-error costs. Simultaneously, the quantified prediction results provide precise execution basis for subsequent automated control, avoiding the subjectivity and lag of traditional manual experience-based control.

[0045] Step 3: Based on the preset sensor deployment locations, a composite sensor array is used to collect thickness data for each layer, enabling real-time, multi-point, three-dimensional monitoring of the pure water layer, emulsion layer, and pure oil layer. Compared to traditional single-point sensors, this method can more comprehensively capture the spatial distribution characteristics of the layers within the tank, especially for the accurate measurement of the emulsion layer thickness. This solves the problem of traditional techniques struggling to identify the boundaries of the emulsion layer, providing high spatiotemporal resolution measured data support for subsequent model calibration and control decisions, and ensuring the objectivity of the layer status assessment.

[0046] Step 4 involves calibrating the model by comparing the measured emulsion layer thickness with the predicted value (recalibrating the model when the deviation exceeds 10%), forming a closed loop of "simulation-measurement-feedback optimization". This dynamically corrects the deviation between the model and actual working conditions, continuously improving simulation accuracy and avoiding long-term error accumulation caused by model solidification. Simultaneously, by triggering a dynamic demulsifier dosing curve based on the emulsion layer thickness, real-time matching of dosage and emulsification degree is achieved. This solves the problems of "excessive waste" or "insufficient efficiency" in traditional fixed dosing modes, improving demulsification efficiency and reducing reagent consumption.

[0047] Step 5, the temperature control parameter set is started only when both the pure water layer and the pure oil layer exist, achieving on-demand start of the heating process. This design avoids energy waste caused by blind heating when the stratification has not formed, while ensuring that the heating acts on the effective separation stage, improving the thermal energy utilization efficiency, reducing the overall energy consumption, and is especially suitable for the separation scenario of high-viscosity crude oil.

[0048] Step 6, by real-time monitoring of the water content in the drained water and implementing a hierarchical treatment strategy (reprocessing in the tank, transferring to the secondary tank or direct transfer), the dynamic optimization of the separation process is achieved. When the water content in the drained water exceeds the standard, the treatment path is flexibly selected according to the recovery time threshold, avoiding interference of unqualified media on subsequent processes; strictly controlling the water cut to ≤0.5%, ensuring the quality of the final purified crude oil, reducing crude oil loss and environmental protection risks, while improving the treatment efficiency of the secondary purification tank and shortening the overall purification cycle.

[0049] In a preferred embodiment of the present invention, Step 1, based on the geometric structure of the target settling tank, the layout position of the microwave-temperature composite sensor array, and the material characteristic parameters, a three-dimensional multi-physics field coupling model is constructed, including:

[0050] Step 11, based on the actual size of the target settling tank, a basic geometric model including the tank height, diameter, bottom cone angle, and steam coil layout path is generated, including: Step 111, according to the actual height H and diameter D of the target settling tank, a cylindrical tank body main geometric framework is generated; Step 112, based on the cylindrical framework in Step 111, a conical frustum structure with a cone angle θ is added at the bottom to form the complete container contour of the settling tank; Step 113, according to the actual pitch P and starting height h0 of the steam coil, a spiral path model is generated within the container contour in Step 112; Step 114, according to the tank body main geometric framework, the container contour, and the spiral path model, a basic geometric model is generated;

[0051] Step 12, according to the vertical layout spacing and radial insertion depth of the microwave-temperature composite sensor array, the three-dimensional space coordinates of each sensor are marked, and an enhanced geometric model with sensor coordinates is output, including: Step 121, according to the basic geometric model, N measurement planes are divided along the tank height direction at an interval of △L; Step 122, within each measurement plane, according to the radial insertion depth d of the sensor, the distance from the end of the sensor to the tank wall is determined, where d < D / 2; Step 123, based on the plane position in Step 121 and the radial depth in Step 122, the three-dimensional coordinates (xn, yn, zn) of each sensor are marked in the basic geometric model; Step 124, the marked coordinates are written into the basic geometric model, and an enhanced geometric model with a spatial coordinate set is output;

[0052] Step 13: On the enhanced geometric model of step 12, define the material property functions of density-viscosity-dielectric constant of crude oil, water and emulsion mixture as a function of temperature, define the thermal conductivity coefficient of steam coil and the thermal boundary conditions of tank wall, define the microwave emission source location and signal attenuation calculation domain, so as to obtain a multiphysics framework model with three field definitions.

[0053] Step 14: Based on the pre-stored crude oil characteristic database, bind the component data of the target crude oil, namely the water content threshold and demulsifier reaction kinetic parameters, to the fluid domain of the multiphysics framework model in Step 13, and output the material binding simulation model.

[0054] Step 15: Based on the basic geometric model, enhanced geometric model, multiphysics framework model, and material binding simulation model, generate a multiphysics coupled three-dimensional model that can simultaneously solve fluid stratification, heat conduction, and microwave signal attenuation.

[0055] In this embodiment of the invention, by restoring the actual dimensions, bottom cone angle, and steam coil path of the settling tank, the basic geometric model provides a physical boundary that closely matches the real equipment for subsequent simulations, avoiding errors caused by geometric simplification. The enhanced geometric model, by labeling the three-dimensional coordinates of the sensors, ensures a one-to-one correspondence between the microwave-temperature signal acquisition points and the spatial positions of the simulation model, providing a precise spatial benchmark for comparing measured data and simulation results. The relationships between density, viscosity, and dielectric constant with temperature are defined, and combined with the thermal parameters of the steam coil and the microwave attenuation domain, enabling the model to realistically reflect the impact of temperature on oil-water separation and improving the physical consistency of the simulation. Binding the target crude oil's water content threshold and demulsifier kinetic parameters allows the model to specifically simulate the separation characteristics of particular crude oils, avoiding the problem of poor adaptability of general models to different crude oils. Integrating the interaction relationships of fluid, heat conduction, and microwave fields enables real-time feedback of "stratification-temperature-microwave," solving the limitation of traditional single-field simulations that cannot reflect the multi-factor linkage effect; the model's output parameters, such as stratification speed and demulsifier injection scheme, can directly guide the operation and control of the settling tank, reducing experimental costs and improving separation efficiency and stability.

[0056] In this embodiment of the invention, step 111 above generates the main frame of the cylindrical tank.

[0057] Read the actual height H and diameter D of the target settling tank, and generate cylindrical geometric data (including the spatial coordinate set of the tank wall) with a height range of Z=0 to Z=H and a cross-sectional diameter of D in the three-dimensional modeling coordinate system (with the center of the tank bottom as the origin and the vertical direction as the Z axis), forming the framework structure of the main body of the tank.

[0058] Step 112 above: Add a bottom frustum structure

[0059] Based on the cylindrical framework in Step 111, add a frustum of a cone structure with a cone angle of θ at Z = 0 (the bottom of the tank): The top surface of the frustum of the cone is connected to the cylindrical bottom (at Z = 0) (with a diameter of D), and the bottom diameter is calculated and determined according to the cone angle θ and the height of the frustum of the cone (usually a fixed value or according to the actual size), forming a complete container contour of "cylinder + bottom frustum of the cone" to ensure consistency with the actual tank bottom structure.

[0060] The above Step 113: Generate the spiral path model of the steam coil

[0061] Read the actual pitch P of the steam coil (the vertical distance between two adjacent coils) and the starting height h0 (the Z coordinate of the lowest end of the coil). Starting from Z = h0 with the central axis of the tank as the spiral center, generate a spiral path upward according to the pitch P: For every 360° rotation, the Z coordinate increases by P, and the diameter of the spiral line adapts to the internal space of the tank (without touching the tank wall), and output the three-dimensional coordinate sequence of this spiral line as the path model of the steam coil.

[0062] The above Step 114: Integrate and generate the basic geometric model

[0063] Integrate the geometric data of the cylindrical main body in Step 111, the bottom frustum of the cone in Step 112, and the spiral path of the steam coil in Step 113 into the same three-dimensional coordinate system, ensuring that there are no conflicts in the spatial positions of each part (such as the coil not penetrating the tank wall and the seamless connection between the frustum of the cone and the cylinder), and output the basic geometric model data including the complete container structure and the steam coil.

[0064] The above Step 121: Divide the measurement planes

[0065] Based on the basic geometric model in Step 114, along the height direction of the tank (Z-axis), divide N horizontal measurement planes (each plane perpendicular to the Z-axis) at intervals of a preset spacing △L from the bottom of the tank (Z≈0) to the top of the tank (Z≈H), and calculate and record the Z coordinate of each plane (such as Z1 = △L, Z2 = 2△L…ZN = N×△L).

[0066] The above Step 122: Determine the radial position of the sensor

[0067] Read the radial insertion depth d of the sensor (the insertion distance from the tank wall to the center of the tank). Since d < D / 2 (to avoid the sensor touching the opposite tank wall), calculate the position of the end of the sensor in the radial direction: The distance from the tank wall to the central axis is D / 2, so the distance from the end of the sensor to the tank wall is d, and the distance from the central axis is (D / 2 - d), and record the spatial parameters corresponding to this radial distance.

[0068] The above Step 123: Mark the three-dimensional coordinates of the sensor

[0069] Combining the Z-coordinate (Zn) of the measurement plane from step 121 and the radial depth d from step 122, calculate the three-dimensional coordinates (xn, yn, zn) of the end of each sensor within each measurement plane (distributed at a preset angle, such as uniformly distributing 3-4 sensors). Here, zn is the Z-coordinate of the measurement plane; xn and yn are determined in the plane coordinate system based on the radial depth d and the angle within the plane (such as 0°, 90°, 180°, etc.) (ensuring that the sensors are uniformly distributed in the radial section inside the tank). Finally, output the coordinate set of all sensors.

[0070] Step 124: Generate the enhanced geometric model

[0071] The sensor 3D coordinate data from step 123 is integrated with the basic geometric model data from step 114. Sensor coordinate marking information (such as the spatial location point and number of each sensor) is added to the basic model to form an enhanced geometric model that includes the tank, coil, and sensor locations.

[0072] Step 121 above: Divide the measurement plane

[0073] Read the actual height H and diameter D of the target settling tank, and generate cylindrical geometric data (including the spatial coordinate set of the tank wall) with a height range of Z=0 to Z=H and a cross-sectional diameter of D in the three-dimensional modeling coordinate system (with the center of the tank bottom as the origin and the vertical direction as the Z axis), forming the framework structure of the main body of the tank.

[0074] Step 122 above: Add a bottom frustum structure

[0075] Based on the cylindrical frame in step 111, a frustum structure with a cone angle of θ is added at Z=0 (bottom of the tank): the top surface of the frustum connects with the bottom of the cylinder (Z=0) (diameter D), and the diameter of the bottom surface is calculated and determined according to the cone angle θ and the height of the frustum (usually a fixed value or according to the actual size), forming a complete container outline of "cylinder + bottom frustum", ensuring consistency with the actual tank bottom structure.

[0076] In steps 123 above, the actual pitch P (vertical distance between two adjacent coils) and starting height h0 (Z coordinate of the lowest point of the coil) of the steam coil are read. Taking the central axis of the tank as the center of the spiral, a spiral path is generated upwards from Z = h0 according to the pitch P: the Z coordinate increases by P every 360° of rotation. The diameter of the spiral line is adapted to the internal space of the tank (without touching the tank wall). The three-dimensional coordinate sequence of the spiral line is output as the path model of the steam coil.

[0077] Step 124 above integrates the geometric data of the cylindrical body in step 111, the bottom truncated cone in step 112, and the spiral path of the steam coil in step 113 into the same three-dimensional coordinate system, ensuring that there is no conflict in the spatial position of each part (such as the coil not penetrating the tank wall and the truncated cone and cylinder being seamlessly connected), and outputs basic geometric model data containing the complete container structure and steam coil.

[0078] Step 13 above involves adding multiphysics parameter definitions based on the enhanced geometric model from step 12. The specific process is as follows:

[0079] Read the physical property database of crude oil, water, and emulsion mixture (measured values ​​of density, viscosity, and dielectric constant at different temperatures), generate temperature-property correlation data for the three (such as a table of values ​​for density ρ, viscosity μ, and dielectric constant ε corresponding to temperature T), and use it as a material property function;

[0080] In the temperature-property correlation of crude oil, water, and emulsion mixtures, since the relationship between material properties and temperature is affected by multiple factors such as composition, purity, and environment, empirical fitting formulas are used. The following are typical fitting formulas for the key properties (density, viscosity, dielectric constant) of the three types of materials and temperature:

[0081] The density-temperature function ρ(T) formula:

[0082] ρ(T)=ρ0×[1-β×(T-T0)], where ρ(T) represents the density of the material at temperature T (unit: kg / m³). 3 ); ρ0 represents the density at reference temperature T0 (unit: kg / m³). 3 (Measured baseline value); β represents the volumetric expansion coefficient (unit: °C). -1 The density is fitted from experimental data to characterize the rate of change of density with temperature; T represents the current temperature (unit: °C, measured value); T0 represents the reference temperature (unit: °C, usually taken as 20 °C or measured reference temperature).

[0083] Viscosity-temperature function μ(T) formula:

[0084] Where μ(T) represents the dynamic viscosity of the material at temperature T (unit: Pa·s); μ0 represents the viscosity at the reference temperature (unit: Pa·s, measured reference value). R represents the activation energy of viscous flow (unit: J / mol, fitted from experimental data, characterizing the sensitivity of viscosity to temperature); R represents the gas constant (unit: J / (mol·K), value 8.314); T represents the current temperature (unit: °C, measured value); (T+273.15) is the thermodynamic temperature (unit: K).

[0085] Dielectric constant-temperature function ε(T) formula:

[0086] ε(T) = ε0 × [1 - α × (T - T0)], where ε(T) represents the relative permittivity at temperature T (dimensionless, characterizing the material's response to microwaves); ε0 represents the permittivity at reference temperature T0 (dimensionless, measured baseline value); and α represents the temperature coefficient of the permittivity (unit: °C). -1 (Fitted from experimental data); T represents the current temperature (unit: °C, measured value); T0 represents the reference temperature (unit: °C, the reference temperature for the same density formula).

[0087] Read the thermal conductivity parameters of the steam coil material (such as the thermal conductivity λ of metal materials) and set the adiabatic boundary conditions of the tank wall (such as the heat loss coefficient of the outer side of the tank wall, or the constraint of "the outer surface temperature of the tank wall = the ambient temperature").

[0088] From the "enhanced geometric model with sensor coordinates" generated in step 12, read the three-dimensional coordinate dataset of all microwave-temperature composite sensors {(x1, y1, z1), (x2, y2, z2), ..., (xN, yN, zN)}, where N is the total number of sensors;

[0089] Based on the sensor hardware design (the preset positional relationship between the transmitter and the sensor end, such as "the coordinates of the transmitter and the sensor end are consistent"), the three-dimensional coordinates (xn, yn, zn) of each sensor are directly marked as the corresponding spatial position of the microwave emission source.

[0090] In the three-dimensional coordinate system of the enhanced geometric model, an identifier (such as "source number + coordinates") is added to each microwave source to form a "sensor-source" position mapping relationship, ensuring that the subsequent microwave propagation simulation can accurately locate the signal origin.

[0091] Define the microwave signal attenuation calculation range (covering the entire dielectric area inside the tank):

[0092] The calculation range is determined according to the following logic:

[0093] Extract the boundary of the medium region within the tank, obtain the complete contour data of the tank from the enhanced geometric model, and determine the possible spatial boundaries of the medium inside the tank:

[0094] Axial (height direction): The lower limit is the bottom surface of the frustum cone at the bottom of the tank Z=0, and the upper limit is the top of the tank Z=H, covering the entire height range from the bottom of the tank to the top of the tank;

[0095] Radial (horizontal direction), with the lower limit at the tank's central axis r = 0 and the upper limit at the inner surface of the tank wall r = D / 2, covering the entire radial range from the center to the tank wall; within the above boundary range, a three-dimensional mesh is divided according to a preset precision (such as axial step size △z, radial step size △r, circumferential angle step size △θ), and the coordinates (zi, rj, θk) of each mesh unit correspond to a small spatial region inside the tank;

[0096] Mark the critical propagation path, and mark the path between the microwave emission source and the possible signal receiving area (such as the sensor receiver) in the grid to ensure higher grid density in the oil-water interface and emulsion layer distribution area (to improve calculation accuracy), and finally form a microwave signal attenuation calculation domain covering the entire medium area.

[0097] The microwave, temperature field, and fluid characteristic parameters are integrated into the enhanced geometric model through data correlation. The specific process is as follows:

[0098] Correlate microwave parameters with geometric model: Bind the determined microwave emission source location and the defined microwave attenuation calculation grid to the corresponding spatial region of the enhanced geometric model, so that the model can identify "which locations emit microwaves" and "which regions need to calculate signal attenuation";

[0099] Related temperature field parameters: Retrieve the thermal conductivity coefficient (e.g., λ) of the steam coil from the preset parameter library. 盘管 ), and assign them to the geometry of the steam coil in the enhanced geometric model (spiral path model); at the same time, assign the adiabatic boundary conditions of the tank wall (such as "heat exchange coefficient h of the outer surface of the tank wall") to the geometry of the steam coil in the enhanced geometric model (spiral path model); 壁 =0)” or “External temperature of the tank wall T” 外 =T 环境 The boundary constraints for temperature field calculations are defined by binding the geometric region of the tank wall.

[0100] The associated fluid characteristic parameters are linked by binding the temperature-characteristic correlation data (correspondence table of density, viscosity, dielectric constant and temperature) of crude oil, water and emulsion mixtures to the "fluid domain" (i.e. all spatial areas covered by the mesh inside the tank) of the enhanced geometric model, so that the model can automatically call the corresponding fluid characteristic parameters according to the temperature value at different locations.

[0101] By establishing multiphysics interaction relationships and setting association rules in the model, such as "temperature field change → fluid property parameter update → microwave propagation characteristics in the medium (such as attenuation coefficient) change" and "microwave signal attenuation data feedback → correction of fluid stratification state (such as emulsion layer thickness) → affecting temperature field distribution (such as adjustment of areas requiring heating)," the model can be interactively calculated in real time. Through the above integration, the geometric model is enhanced to describe microwave propagation, temperature field distribution, fluid properties and their interactions, ultimately forming a multiphysics framework model.

[0102] Step 14: To generate the material binding simulation model, perform the following operations:

[0103] Database indexing and positioning: Read the identification information of the target crude oil (such as crude oil number, origin, density grade, etc.) and use it as an index to match the corresponding parameter records in the pre-stored crude oil characteristic database (the database is stored in a structured manner according to "crude oil identifier-parameter set").

[0104] Moisture content threshold extraction: Retrieve the moisture content threshold parameter from the matched records. This parameter is a set of critical values ​​for a certain stage (e.g., emulsion layer determination threshold ≥30%, pure oil layer determination threshold ≤5%, drainage qualified threshold ≤0.5%, etc.), and store it in the form of a numerical list (e.g., [30%, 5%, 0.5%]). Mark the layering status corresponding to each threshold (e.g., "≥30% is emulsion layer" "≤5% is pure oil layer").

[0105] Demulsifier reaction kinetic parameters were extracted by retrieving the demulsifier reaction parameters suitable for this crude oil, including:

[0106] Reaction rate data (such as the reaction rate constant k(T) between the demulsifier and the emulsion layer at different temperatures, in min) -1 );

[0107] Concentration-efficiency relationship (e.g., a table showing the correspondence between demulsifier concentration c and emulsion layer breaking efficiency η, where η = f(c), e.g., when c = 100 ppm, η = 80%).

[0108] The activation energy of the reaction (a correction factor for the effect of temperature on the reaction rate, unit: kJ / mol) forms the set of kinetic parameters.

[0109] Parameters are bound to the fluid domain (associated with a multiphysics framework model):

[0110] The above parameters are then bound to the "fluid domain" of the multiphysics framework model in step 13. The specific process is as follows:

[0111] Fluid domain spatial positioning: Extract the three-dimensional mesh data of the "fluid domain" from the multiphysics framework model, that is, the set of coordinates of all mesh cells in the tank that may contain crude oil, water and emulsion layer {(xi, yi, zi)}.

[0112] Moisture content threshold binding: The staged threshold is associated with the "state determination rule" of the grid cell. For example, when the calculated moisture content w of a certain grid cell meets the condition of 30%≤w<95%, the model automatically marks the cell as an "emulsion layer"; when w≤5%, it is marked as a "pure oil layer", so that the model can determine the layering state in real time through the grid moisture content.

[0113] Demulsifier kinetic parameter binding: Parameters such as reaction rate constant and concentration-efficiency relationship are associated with the "emulsion layer mesh unit" of the fluid domain. The rule is set: when a mesh unit in the model is marked as "emulsion layer", the corresponding reaction rate formula (such as reaction rate = k(T)×c) is automatically called, and the emulsion breaking efficiency of the unit is calculated in real time according to the demulsifier concentration to achieve targeted simulation.

[0114] Parameter verification: Check whether the bound data conflict (such as whether the threshold range covers all possible moisture contents, whether the units of kinetic parameters are consistent with the units of model calculation). If there are no errors, generate a material binding simulation model and output integrated data including "fluid domain mesh + layer judgment rules + demulsification reaction parameters".

[0115] Step 15: Detailed Implementation Process of Generating a Multiphysics Coupled 3D Model

[0116] Import the model data from steps 11 to 14 into a unified 3D modeling platform (such as the database module of finite element simulation software). Specific steps are as follows:

[0117] Extract the basic geometric model data and import the tank structure data (three-dimensional coordinates and dimensional parameters of the cylindrical body, bottom cone, and steam coil path) from step 11 as the physical boundary framework of the model.

[0118] Enhanced geometric model data extraction: Import the sensor coordinate set ((xn, yn, zn)) and measurement plane division data from step 12, mark the sensor positions in the basic geometric framework, and provide spatial anchor points for the microwave signal propagation path;

[0119] Data extraction of the multiphysics framework model is performed, and the multiphysics parameters (temperature field boundary conditions, microwave emission source location and attenuation calculation domain, material property functions) from step 13 are imported and associated with the corresponding regions of the basic geometry (such as the thermal conductivity coefficient of the steam coil and the insulation conditions of the tank wall).

[0120] Material binding simulation model data extraction, import the crude oil-specific parameters (water content threshold, demulsification kinetic parameters) from step 14, and associate them with the fluid domain mesh to ensure that the model can call the stratification determination rules and reaction characteristics of the crude oil.

[0121] Three-field coupled data association (establishing real-time interaction rules):

[0122] By setting up data association logic, real-time interaction of "fluid stratification - heat conduction - microwave attenuation" is achieved. Specific rules are as follows:

[0123] Interaction between temperature field and fluid stratification:

[0124] When the steam coil is heated (the temperature field distribution T(x, y, z) is calculated by the heat conduction model), the model automatically calls the "temperature-material properties" function (step 13) to update the density ρ(T) and viscosity μ(T) of each grid cell.

[0125] Hydrodynamic calculations and interface position correction processes triggered by density / viscosity changes:

[0126] Real-time updates of density and viscosity parameters:

[0127] When the temperature field changes (e.g., the temperature inside the tank rises due to steam coil heating), the system first calls the "temperature-characteristic" correlation data in step 13 based on the real-time temperature of each grid cell: for each grid cell, the density value of the medium (crude oil, water, emulsion mixture) in that cell is updated by querying the "temperature-density" correspondence table; similarly, the viscosity value of the medium is updated by querying the "temperature-viscosity" correspondence table (e.g., when the temperature rises, the viscosity of crude oil decreases significantly, and the viscosity of water decreases slightly).

[0128] Dynamic calculation of layering velocity (based on density difference and viscosity): The layering velocity of each mesh cell is calculated according to the following logic:

[0129] Extract the density difference. For each grid cell, calculate the density difference between the oil phase (or emulsion phase) and the water phase (e.g., if the oil phase density is 850 kg / m³ and the water phase density is 1000 kg / m³, then the density difference is 150 kg / m³).

[0130] The effect of viscosity is related to the resistance to stratification. The lower the viscosity, the better the fluidity of the medium, the less resistance the oil or water droplets experience, and the faster the stratification speed (for example, when the viscosity of crude oil drops from 500 mPa·s to 100 mPa·s, the stratification speed may increase by 3-5 times).

[0131] By comprehensively calculating the velocity, and taking into account the direction of the gravity field inside the tank (vertically downward) and the volume fraction of the medium (e.g., oil phase accounts for 60%), the actual velocity (distance moved per unit time) of the oil phase rising or the water phase sinking within the grid cell is calculated.

[0132] Real-time correction of the oil-water / emulsion layer interface position:

[0133] Based on the layered speed, update the interface position using the following steps:

[0134] Initial interface localization: Obtain the three-dimensional coordinates of the oil-water interface, oil-emulsion layer interface, and emulsion-water interface at the current moment from the fluid stratification model (e.g., the oil-water interface is at a height Z = 3.5m).

[0135] The time step is discrete, and the calculation time interval is set (e.g., updated every 10 seconds). Within each time step, the distance the interface moves within that time period is calculated based on the layering speed of each grid (e.g., if the speed is 0.01m / s, it moves 0.1m within 10 seconds).

[0136] Multi-region collaborative correction involves calculating the interfaces of different regions separately. If the upper oil phase has low viscosity and rises quickly, the oil-emulsion interface moves upward; if the lower water phase has high density and sinks quickly, the emulsion-water interface moves downward.

[0137] Boundary constraint verification ensures that the corrected interface position does not exceed the physical boundary (e.g., not lower than the bottom of the tank, not higher than the top of the tank) and smoothly transitions with the interface position of adjacent meshes (avoiding abrupt changes).

[0138] When the viscosity or density of the emulsion layer mesh changes, the system additionally performs the following:

[0139] Emulsion layer thickness correction: The change in emulsion layer thickness is calculated based on the movement distance of the oil-emulsion layer and the emulsion layer-water interface (e.g., if the two interfaces move by +0.05m and -0.03m respectively, the emulsion layer thickness increases by 0.08m).

[0140] Flow state judgment: If the viscosity of the emulsion layer drops to a certain threshold (e.g., ≤200mPa·s), it is determined that its fluidity is enhanced, which may accelerate demulsification. At this time, the reaction rate of the demulsifier is adjusted (e.g., the reaction efficiency is increased by 10%).

[0141] If the emulsion layer thickness increases and the delamination rate is lower than expected, a signal (such as "heating power needs to be increased") is sent to the temperature field model to further reduce viscosity and accelerate delamination by increasing the temperature.

[0142] Through the above process, the system achieves a dynamic closed loop of "density / viscosity change → stratification speed adjustment → interface position update → emulsion layer state correction", ensuring that the interface position output by the fluid stratification model always matches the real-time operating conditions.

[0143] Interaction between temperature field and microwave attenuation:

[0144] The change in temperature T updates the dielectric constant ε(T) of each grid cell through the "temperature-dielectric constant" function (step 13);

[0145] The change in ε(T) corrects the microwave propagation attenuation coefficient in this grid (if ε increases, the attenuation is enhanced), the microwave model recalculates the attenuation of the signal propagation path, and outputs new microwave detection data.

[0146] Microwave attenuation and the feedback from fluid stratification:

[0147] The attenuation data output by the microwave model (such as the signal strength received by different sensors) is compared with the measured data (step 3 in the following step) or the preset threshold. If the deviation exceeds the set range (such as ±15%), the interface determination of the fluid stratification model is corrected in reverse (such as adjusting the calculated value of the emulsion layer thickness).

[0148] The corrected stratification state is further fed back to the temperature field model (e.g., when the emulsion layer thickens, it triggers the adjustment of the steam coil heating power).

[0149] Coupled model validation and output, perform data consistency verification on the integrated model:

[0150] Check whether the units of each physical field parameter are consistent (e.g., temperature in °C, density in kg / m³); test whether the interaction logic of the three fields is closed-loop (e.g., whether the cycle of temperature change → characteristic update → layer adjustment → microwave correction → temperature readjustment is smooth); verify the stability of the model under extreme conditions (e.g., whether the parameters overflow at high temperature and high emulsification). After verification, output a three-dimensional model that can simultaneously solve the coupling effect of the three fields. This model includes complete geometric structure, sensor location, physical parameters, interaction rules and crude oil-specific characteristics.

[0151] In a preferred embodiment of the present invention, step 2 involves simulating the oil-water separation process using the finite element method based on the three-dimensional model to output key parameters. These key parameters include predicted values ​​for stratification rate, emulsion layer thickness, demulsifier injection dynamic curve, stratification recovery time, and a set of temperature control parameters, including:

[0152] Step 21: Based on the multi-physics coupled three-dimensional model, the fluid motion state during the oil-water separation process is simulated using the finite element method, and the predicted value of the oil-water interface stratification velocity is output.

[0153] Step 22: Based on the predicted stratification rate from Step 21 and the viscosity-temperature relationship in the material property function, calculate the dynamic change of the emulsion layer thickness over time and output the predicted emulsion layer thickness. This includes: Step 221: Based on the predicted stratification rate output in Step 21 and the viscosity-temperature relationship in the material property function, calculate the oil-water emulsion phase separation efficiency coefficient; Step 222: Based on the separation efficiency coefficient and the predicted stratification rate, dynamically correct the rate of change of the emulsion layer thickness over time; Step 223: Integrate the corrected rate of change of the emulsion layer thickness over time and output the dynamic sequence of the predicted emulsion layer thickness over time.

[0154] Step 23: Based on the predicted emulsion layer thickness from Step 22 and the pre-stored demulsifier reaction kinetic parameters, generate a dynamic curve showing the relationship between demulsifier injection amount and time. This includes: Step 231: Extracting the maximum value of the emulsion layer thickness change rate output from Step 222, and combining it with the critical emulsion strength threshold in the pre-stored demulsifier reaction kinetic parameters to determine the initial demulsifier injection rate; Step 232: Based on the dynamic sequence decay trend of the predicted emulsion layer thickness, and according to the concentration-reaction rate relationship in the reaction kinetic parameters, generating segmented injection rate decay curves; Step 233: Superimposing the initial injection rate and the segmented decay curves to generate a dynamic curve showing the continuous change of demulsifier injection amount over time.

[0155] Step 24: Based on the predicted stratification rate from Step 21 and the predicted emulsion layer thickness from Step 22, simulate the time required for the oil-water interface to recover stability after disturbance, and output the predicted stratification recovery time. This includes: Step 241: Calculate the theoretical recovery time of the oil-water interface without emulsion layer disturbance based on the predicted stratification rate from Step 21; Step 242: Extract the peak thickness and duration based on the dynamic sequence of the predicted emulsion layer thickness output in Step 223; Step 243: Calculate the hindrance correction coefficient for interface recovery of the emulsion layer by combining the peak emulsion layer thickness with the dielectric constant-viscosity correlation in the material property function; Step 244: Multiply the theoretical recovery time by the hindrance correction coefficient to output the predicted stratification recovery time after disturbance.

[0156] Step 25: Combining the predicted value of the stratified recovery time from Step 24 with the thermal conductivity coefficient of the steam coil, optimize the combination of heating power and duration to generate a set of temperature control parameters.

[0157] In this embodiment of the invention, the finite element method is used to simulate the fluid motion state, accurately outputting the predicted value of the stratification rate at the oil-water interface, making the separation dynamic process, which was originally difficult to observe directly, predictable. Combining the stratification rate with the viscosity-temperature relationship, the real-time change in emulsion layer thickness is dynamically calculated. By correcting the rate of change and integrating, a sequence of emulsion layer thickness over time is obtained, clearly showing the formation and dissipation patterns of the emulsion layer, facilitating early identification of key separation nodes. Based on the changes in emulsion layer thickness and the reaction characteristics of the demulsifier, the initial injection rate is first determined, and then a segmented decay curve is generated according to the emulsion layer decay trend, ultimately forming a continuous injection dynamic curve. This ensures precise matching of the demulsifier dosage and timing, avoiding excessive waste or insufficient dosage that could affect the separation effect. The peak values ​​of stratification rate and emulsion layer thickness are used to calculate the stratification recovery time after disturbance, taking into account both the theoretical recovery time without an emulsion layer and the hindering effect of the emulsion layer. This allows operators to know the system's anti-interference capability in advance, facilitating the development of emergency control strategies. The heating power and duration are optimized by combining the recovery time and steam coil characteristics, so that temperature control is coordinated with the separation process and demulsification reaction, reducing energy consumption while ensuring separation efficiency and achieving a balance between energy saving and separation effect. The key parameters output from each step (stratification rate, emulsion layer thickness, filling curve, etc.) are interconnected and dynamically adapted to form a complete separation process prediction system. This provides comprehensive and accurate decision-making basis for the efficient operation of the settling tank and parameter adjustment, reducing experimental costs and operational risks.

[0158] In this embodiment of the invention, step 21 above, which uses the finite element method to simulate the fluid motion state based on a multi-physics coupled three-dimensional model, is as follows:

[0159] First, the internal space of the 3D model is divided into a large number of tiny 3D mesh units (finite element units) with a preset precision (such as 0.1 meters axially and 0.05 meters radially), with each unit corresponding to a specific spatial location inside the tank.

[0160] Each grid cell is assigned initial physical parameters. Based on the material property functions in the multiphysics model, initial temperature, density, and viscosity are set (e.g., crude oil initial temperature 25℃, density 860 kg / m³). 3 Viscosity 300 mPa·s; initial water temperature 25℃; density 1000 kg / m³ 3 Viscosity 1 mPa·s).

[0161] Set boundary conditions: the tank wall is a fixed boundary (fluid cannot penetrate), the steam coil area is a heat conduction boundary (heated at a preset initial temperature), and the tank top and bottom are free flow boundaries (allowing fluid to enter and exit).

[0162] The finite element simulation calculation is initiated, and the specific process of calculating the relative motion between the oil and water phases using the finite element simulation is as follows:

[0163] The three-dimensional space of the settling tank is divided into millions of independent grid cells (similar to small squares in a Rubik's Cube) according to fixed dimensions (such as 0.05 meters axially and 0.05 meters radially). Each cell has a unique number and three-dimensional coordinates (such as (x = 0.1m, y = 0.1m, z = 0.5m)).

[0164] Bind initial physical properties to each unit:

[0165] If the unit is located at the top of the tank (e.g., z≥3m), it is marked as an "oil phase unit" and assigned a density of 850kg / m³. 3 Viscosity 0.3 Pa·s, initial velocity 0 m / s; if the unit is located at the bottom of the tank (e.g., z < 3 m), it is marked as "aqueous phase unit", with a density of 1000 kg / m³, viscosity of 0.001 Pa·s, and initial velocity of 0 m / s; the unit at the interface between the oil and water phases (e.g., 2.9 m ≤ z < 3.1 m) is marked as "emulsion layer unit", and its density and viscosity are calculated by weighting "70% oil + 30% water" (e.g., density 850 × 0.7 + 1000 × 0.3 = 895 kg / m³). 3 Record the adjacent units of each unit (such as the six directions of front, back, left, right, up, and down) to establish the data interaction relationship between units.

[0166] Mass conservation calculation, dynamic mass balance within a cell: For each grid cell, the mass change is calculated according to the following steps:

[0167] Inflow mass calculation involves counting the mass of fluid flowing into the current cell from adjacent cells through the shared surface. For example, if the upper cell moves upward, it will "push" some fluid into the current cell through the bottom shared surface. Inflow mass = density of adjacent cells × flow velocity × area of ​​shared surface × time step (e.g., 0.01 seconds).

[0168] Outflow mass calculation: The fluid mass flowing out of the current cell through the shared surface to the adjacent cells is counted. For example, if the current cell moves downward, it will "release" fluid to the cells below through the bottom shared surface. Outflow mass = current cell density × its own flow velocity × shared surface area × time step.

[0169] Density Update: Adjust the density of the current cell based on the difference between "inflow mass" and "outflow mass".

[0170] If the inflow mass is greater than the outflow mass, the mass within the cell increases, and the density = (original mass + difference) ÷ cell volume; if the inflow mass is less than the outflow mass, the mass within the cell decreases, and the density = (original mass - difference) ÷ cell volume; this ensures that the mass change of each cell conforms to the conservation principle that "mass neither arises from nothing nor disappears from nothing".

[0171] For each mesh element, the velocity variation is determined through force analysis. Specific steps include:

[0172] Gravity and buoyancy calculations:

[0173] The gravity acting on the element = element density × element volume × gravitational acceleration (9.81 m / s²) 2 );

[0174] Buoyancy force on a unit = density of the adjacent unit below × unit volume × gravitational acceleration;

[0175] If the density of the unit cell (e.g., 850 kg / m³ for oil phase) is less than the density of the unit cell below (e.g., 1000 kg / m³ for water phase) 3 When buoyancy is greater than gravity, the resultant force is upward (pushing the oil phase to float); conversely, when buoyancy is less than gravity, the resultant force is downward (pushing the water phase to sink).

[0176] Calculation of pressure difference force:

[0177] Compare the pressure values ​​of the unit with those of the adjacent units (pressure is determined by unit density and height, with lower units typically having higher pressure than upper units), and calculate the pressure difference as (adjacent unit pressure - current unit pressure) × shared surface area.

[0178] If the pressure in the right-hand unit is higher than that in the current unit, the pressure difference will push the fluid in the current unit to the right, and the speed will increase as the pressure difference increases.

[0179] Calculation of viscous resistance:

[0180] The higher the viscosity of a unit, the stronger the flow resistance to adjacent units. For example, the viscosity of an emulsion layer unit (e.g., 0.5 Pa·s) is 1.7 times that of a pure oil layer (0.3 Pa·s), which will reduce the rising speed of adjacent oil phase units by 30% (resistance = viscosity × velocity difference × shared surface area).

[0181] Speed ​​updates:

[0182] Total resultant force = gravity + buoyancy + pressure difference force + viscous resistance;

[0183] Acceleration = Total resultant force ÷ (Unit density × Unit volume);

[0184] New velocity = current velocity + acceleration × time step (e.g., 0.01 seconds), ensuring that the velocity change is proportional to the force applied.

[0185] After each cell completes its calculation, its density and velocity data are synchronized to adjacent cells (through a shared surface): for example, the "upward velocity" of the upper oil phase cell will be used as the "inflow velocity" of the lower cell, affecting the mass and momentum calculation of the lower cell.

[0186] The elements at the tank wall are set to "no slip boundary": the velocity of the tank wall elements is forced to be 0, and the velocity of adjacent elements will be reduced by 50% due to the viscous resistance of the tank wall (simulating the phenomenon that fluid flow slows down when it adheres to the wall in reality).

[0187] At the oil-water interface, the unit (emulsion layer) adds an extra layer of "surface tension resistance": the mutual attraction between the interfacial units slows down the rate at which oil droplets rise and water droplets sink (e.g., making the rate of emulsion layer units 20% lower than that of pure oil layer).

[0188] Time step iteration and interface tracking:

[0189] Set a fixed time step (e.g., 0.01 seconds), and advance one time step after all units have been calculated.

[0190] First time step (0-0.01 seconds): All elements complete the initial mass and velocity update; Second time step (0.01-0.02 seconds): Based on the results of the previous step, recalculate the force and velocity of each element; and so on, until the preset number of time steps is completed (e.g., 1000 steps, i.e., a 10-second simulation).

[0191] After each time step, the interface position is tracked using the cell's "oil phase volume percentage":

[0192] Units with an oil phase volume percentage ≥ 95% are labeled as "pure oil layer"; units with an oil phase volume percentage ≤ 5% are labeled as "pure water layer"; units with an oil phase volume percentage 5%-95% are labeled as "emulsion layer"; record the interface height between the pure oil layer and the emulsion layer (oil-emulsion interface), and the interface height between the emulsion layer and the pure water layer (emulsion-water interface).

[0193] Layered motion simulation and velocity calculation:

[0194] As time progresses, the pure oil layer units, continuously subjected to an upward resultant force, experience a continuous accumulation of velocity (e.g., increasing from 0.001 m / s to 0.01 m / s), causing them to move upwards as a whole, and the oil-emulsion interface gradually rises. The pure water layer units, continuously subjected to a downward resultant force, experience a continuous accumulation of velocity (e.g., increasing from 0.001 m / s to 0.008 m / s), causing them to move downwards as a whole, and the emulsion-water interface gradually decreases. The emulsion layer units, subjected to bidirectional forces and viscous resistance, have a slower velocity (e.g., 0.002 m / s), and their thickness gradually decreases as oil and water separate (the number of pure oil and pure water units increases, while the number of emulsion units decreases).

[0195] Calculate the stratification rate at each time step:

[0196] Oil-emulsion interface stratification rate = (current interface height - previous interface height) ÷ time step; Emulsion-water interface stratification rate = (previous interface height - current interface height) ÷ time step; Take the average velocity at different radial positions throughout the tank as the "predicted stratification rate" at that time point.

[0197] Simulation termination and result output:

[0198] When the stratification velocity is less than 0.0001 m / s for 100 consecutive time steps (i.e., the interface hardly moves anymore), the oil-water separation is considered to have reached a stable state, and the simulation is stopped. The recorded data of all time steps are sorted out and output as follows: the curve of the rising velocity of the pure oil layer over time (e.g., from 0 to 0.01 m / s in 0-1 seconds, and gradually decreasing to 0.0001 m / s in 1-10 seconds); the curve of the sinking velocity of the pure water layer over time; and the spatial distribution of the oil phase, emulsion layer, and water phase at different times (three-dimensional visualization results).

[0199] Through the above-mentioned detailed calculations unit by unit and time step by time, finite element simulation can realistically reproduce the relative motion of the oil and water phases under the action of gravity and buoyancy, and finally output accurate predicted values ​​of the separation velocity, providing data support for the subsequent optimization of the separation process.

[0200] The system records the movement speed of the oil-water interface in each grid cell in real time, calculates the time difference of the interface position change between adjacent grid cells (e.g., the difference between the interface positions at two adjacent moments divided by the time interval), and obtains the stratification speed at each position. The system integrates the stratification speed data of all grids, takes the average value of the cross section at the same height as the predicted value of the stratification speed at that height, and finally outputs the spatial distribution of the oil-water interface stratification speed and the prediction results of its change over time within the entire tank.

[0201] In step 221 above, based on the predicted stratification rate output in step 21 (e.g., an average stratification rate of 0.02 m / s at a certain moment), and combined with the correspondence between "viscosity and temperature" in the material property function (e.g., crude oil viscosity is 200 mPa·s at 30℃ and 100 mPa·s at 40℃), the separation efficiency coefficient is calculated according to the following logic:

[0202] First, determine the viscosity value at the current simulation temperature (e.g., find the viscosity at 35℃ to be 150 mPa·s from the "Temperature-Viscosity" table). Then, based on the negative correlation between viscosity and separation efficiency (higher viscosity means more difficult separation and lower efficiency coefficient), compare it with the standard state (e.g., efficiency coefficient is 1.0 at a viscosity of 100 mPa·s) to calculate the efficiency coefficient at the current viscosity (e.g., efficiency coefficient = 100 / 150 = 0.67 at 150 mPa·s). Finally, multiply this efficiency coefficient by the predicted stratification rate to obtain the corrected effective stratification rate (e.g., 0.02 m / s × 0.67 = 0.0134 m / s), which serves as the core parameter for the separation efficiency coefficient.

[0203] In step 222 above, the initial rate of change of the emulsion layer thickness is determined by the effective separation speed (e.g., an effective separation speed of 0.0134 m / s represents the reduction in thickness of the emulsion layer due to separation per unit time).

[0204] Based on the separation efficiency coefficient from step 221, the initial rate of change is corrected. If the separation efficiency coefficient is lower than 0.8 (indicating greater separation difficulty), the rate of change is multiplied by 0.9 (further reducing the rate to reflect the characteristic that the emulsion layer is not easy to separate); if the efficiency coefficient is higher than 0.8, the initial rate of change remains unchanged.

[0205] Meanwhile, considering the stability of the emulsion layer (e.g., the greater the thickness, the stronger the internal disturbance and the slower the separation), an additional thickness correction factor is introduced (e.g., when the current emulsion layer thickness is 1.2m, the correction factor = 1 / 1.2≈0.83). This factor is multiplied by the above-mentioned corrected rate of change to obtain the final emulsion layer thickness change rate (e.g., 0.0134m / s × 0.9 × 0.83 ≈ 0.0101m / s).

[0206] Step 223 above, starting from the initial emulsion layer thickness (e.g., 1.5m at model startup), accumulates the corrected rate of change at time intervals (e.g., every minute):

[0207] Thickness at the end of the first minute = initial thickness - rate of change × 60 seconds (e.g., 1.5m - 0.0101m / s × 60s ≈ 1.5m - 0.606m ≈ 0.894m).

[0208] Thickness at the end of the 2nd minute = Thickness at the end of the 1st minute - Current rate of change × 60 seconds (If the rate of change becomes 0.012 m / s due to the improvement in separation efficiency, then 0.894 m - 0.012 × 60 ≈ 0.894 m - 0.72 m ≈ 0.174 m).

[0209] Repeat the above calculations until the emulsion layer thickness stabilizes (e.g., thickness ≤ 0.1m). Arrange the thickness values ​​at each moment in chronological order to form a dynamic sequence of the predicted emulsion layer thickness over time (e.g., [1.5m, 0.894m, 0.174m, ...]).

[0210] In step 231 above, the maximum value of the emulsion layer thickness change rate is extracted from the output of step 222 (e.g., 0.02 m / s, which represents the moment when the emulsion layer separates the fastest, and at this time the strongest demulsification effect is required).

[0211] Consult the pre-stored demulsifier reaction kinetic parameters to obtain the critical emulsion strength threshold (e.g., when the rate of change is ≥0.015m / s, the maximum initial injection needs to be started); compare the maximum value with the threshold. If the maximum value (0.02m / s) > the threshold (0.015m / s), then use the "maximum effective injection rate" (e.g., 20L / h) in the kinetic parameters as the initial injection rate; if the maximum value ≤ the threshold, then set the initial rate at 80% of the maximum rate (e.g., 16L / h).

[0212] In step 232 above, the dynamic sequence of the predicted emulsion layer thickness output in step 223 is analyzed to determine the attenuation stage:

[0213] Phase 1 (0-10 minutes): The emulsion layer thickness decreases from 1.5m to 0.8m, with a relatively rapid decay (change rate 0.01-0.015m / s). According to the "high concentration-high reaction rate" relationship in the kinetic parameters, the injection rate is set to decrease by 10% every 2 minutes (e.g., from 20L / h → 18L / h → 16.2L / h...).

[0214] The second stage (10-20 minutes): The thickness decreases from 0.8m to 0.2m, and the decay slows down (change rate 0.005-0.01m / s). According to the "medium concentration-medium reaction rate" relationship, the injection rate is set to decrease by 5% every 5 minutes (e.g., 16.2L / h→15.4L / h→14.6L / h...).

[0215] The third stage (after 20 minutes): the thickness is <0.2m, the decay is gradual (change rate <0.005m / s), and the injection rate is kept constant at 14.6L / h according to the "low concentration-low reaction rate" relationship until the emulsion layer thickness approaches 0.

[0216] The initial injection rate (20 L / h) in step 231 is used as the starting point of the curve, and is continuously spliced ​​with the segmented decay curve in step 232:

[0217] The first stage of decay data was used for 0-10 minutes (20→18→16.2...), the second stage data was used for 10-20 minutes (16.2→15.4→14.6...), and the flow rate was maintained at 14.6 L / h after 20 minutes.

[0218] Integrating the injection rate at each time point (rate × time) yields the cumulative injection volume (e.g., cumulative injection volume from 0 to 10 minutes = (20 + 18 + 16.2) ÷ 6 × 10 ≈ 8.7 L), ultimately forming a dynamic curve relating "time - injection rate - cumulative injection volume" to ensure that the injection volume changes continuously over time and matches the emulsion layer decay trend.

[0219] Step 241 above, based on the predicted stratification velocity value from step 21 (e.g., stratification velocity of 0.015 m / s in a stable state), simulates a "disturbance" scenario (e.g., sudden drainage causing fluctuations in the oil-water interface, deviating from the stable position by 0.5 m); the theoretical recovery time = disturbance distance ÷ predicted stratification velocity value (e.g., 0.5 m ÷ 0.015 m / s ≈ 33.3 seconds), represents the time it takes for the interface to return to a stable position by its own stratification motion when there is no emulsion layer obstruction.

[0220] In step 242 above, the maximum value of the emulsion layer thickness after disturbance is found from the dynamic sequence of emulsion layer thickness in step 223 (e.g., the thickness increases from 0.8m to 1.1m after disturbance, and the peak value is 1.1m); the time of the peak value occurrence (e.g., the 2nd second after disturbance) and the duration of the peak value (e.g., from the 2nd second to the 8th second, lasting 6 seconds) are recorded as the main stage of the emulsion layer's influence on interface recovery.

[0221] In step 243 above, based on the relationship between "dielectric constant and viscosity" in the material property function (the larger the dielectric constant, the higher the viscosity, and the stronger the hindrance effect), the dielectric constant (e.g., 35) and viscosity (e.g., 300 mPa·s) of the emulsion layer corresponding to a peak thickness of 1.1 m are found; the hindrance coefficient corresponding to the reference viscosity (e.g., 100 mPa·s) is set to 1.0. The current viscosity of 300 mPa·s is 3 times the reference value, so the basic hindrance coefficient = 3.0; combined with the peak duration of 6 seconds (accounting for 18% of the theoretical recovery time of 33.3 seconds), the final hindrance correction coefficient = 1 + basic hindrance coefficient × 18% ≈ 1 + 3.0 × 0.18 ≈ 1.54.

[0222] In step 244 above, the theoretical recovery time in step 241 is multiplied by the stagnation correction coefficient in step 243 (33.3 seconds × 1.54 ≈ 51.3 seconds) to obtain the actual recovery time after considering the stagnation effect of the emulsion layer. The above calculation is repeated for different disturbance intensities (such as deviations from the stable position of 0.3m and 0.7m), and the maximum value is taken as the final predicted value of the layered recovery time (such as 51.3 seconds).

[0223] Step 25 above, combined with the predicted stratification recovery time (51.3 seconds) from step 24 and the thermal conductivity of the steam coil (e.g., 45 W / (m·℃) for metal coils), optimizes the combination of heating power and duration:

[0224] First stage (rapid heating): To shorten the recovery time, high-power heating (e.g., 50kW) is used in the first 20 seconds. The temperature rise rate inside the tank is calculated based on the thermal conductivity (e.g., 2°C every 5 seconds) to ensure that the emulsion temperature rises from 25°C to 40°C (to reduce viscosity).

[0225] Second stage (heat preservation): After 20 seconds, if the stratification recovery time is 31.3 seconds, use low power heating (20kW) to maintain the temperature at 40℃±2℃ to avoid overheating and increased energy consumption;

[0226] The third stage (stop heating): When the predicted recovery time is only 5 seconds, turn off the heating and use the residual heat to complete the final stage of stratified recovery;

[0227] Simultaneously, the heating power (50kW, 20kW, 0kW), corresponding duration (20 seconds, 26.3 seconds, 5 seconds), and target temperature (40℃) of each stage are recorded to form a set of temperature control parameters including "power-duration-temperature" to ensure efficient heating in the shortest recovery time.

[0228] In a preferred embodiment of the present invention, step 3 involves collecting data using a microwave-temperature composite sensor array based on the sensor placement locations from step 1 to obtain measured results. These measured results include the thickness of the pure water layer, the thickness of the emulsion layer, and the thickness of the pure oil layer, including:

[0229] Step 31: Based on the enhanced geometric model with spatial coordinate set output in step 124, activate the microwave-temperature composite sensor array according to the sensor's three-dimensional coordinates, and synchronously collect the microwave signal attenuation rate and temperature value at each spatial point.

[0230] Step 32: Based on the material property function of dielectric constant changing with temperature defined in Step 13, perform temperature compensation calibration on the attenuation rate of the microwave signal acquired in Step 31.

[0231] Step 33: Identify the height of the oil-emulsion layer interface and the height of the emulsion layer-water layer interface based on the abrupt change point of the calibrated microwave signal attenuation rate.

[0232] Step 34: Combining the dual-interface height and tank geometry identified in Step 33, calculate the thickness of the pure water layer, the emulsion layer, and the pure oil layer, and output the measured results.

[0233] In this embodiment of the invention, by activating and enhancing the sensor array corresponding to the geometric model coordinates, synchronous acquisition of microwave signal attenuation rate and temperature is achieved, ensuring strict matching between the data source and the model's spatial location, providing a reliable measured benchmark for subsequent simulation result verification. Microwave attenuation rate compensation is performed based on the dielectric constant-temperature characteristic function, eliminating the interference of temperature changes on the signal, making attenuation data at different times and locations comparable, and improving the accuracy of interface identification. The oil-emulsion layer and emulsion-water layer dual interfaces are directly located through attenuation rate mutation points, replacing traditional manual observation or sampling analysis, reducing human error, and achieving real-time, automated identification of layered interfaces. The thickness of each layer is calculated by combining the dual interface height and tank geometric parameters, ensuring that the thickness data of the pure oil layer, emulsion layer, and pure water layer are consistent with actual working conditions, providing a quantitative basis for evaluating the separation effect. The output measured results can be directly compared with the simulated predicted values ​​to correct model parameters (such as material characteristic functions and reaction kinetic parameters), forming a closed loop of "simulation-measurement-optimization," continuously improving the model's prediction accuracy. The real-time output of each layer thickness data can dynamically reflect the oil-water separation process, providing immediate feedback for adjusting operations such as demulsifier dosage and heating power, and helping the settling tank to operate efficiently and stably.

[0234] In this embodiment of the invention, step 31 involves retrieving the three-dimensional coordinates of all microwave-temperature composite sensors (e.g., (x1, y1, z1) = (0.5m, 0m, 1m), (x2, y2, z2) = (0.5m, 0m, 1.5m), etc.) from the enhanced geometric model of step 124, locating the corresponding sensors in the physical tank according to the coordinates, and sending an activation command to the sensor array to enable all sensors to start data acquisition at the same time point (e.g., at second 0), with the acquisition frequency set to once per second (to ensure data time synchronization).

[0235] Signal acquisition content:

[0236] Each sensor collects two sets of data simultaneously:

[0237] Microwave signal attenuation rate: The proportion of signal strength loss from the transmitter to the receiver (e.g., if the transmitter strength is 1000mV and the receiver strength is 600mV, the attenuation rate = (1000-600) / 1000 = 40%).

[0238] Temperature value: The temperature of the medium at the location of the sensor probe (e.g., the temperature at a height of 1.5m is 45℃).

[0239] All data are stored in the format of “sensor number-time-attenuation rate-temperature” (e.g., sensor 1-5th second-35%-42℃).

[0240] In step 32 above, the dielectric constant-temperature characteristic function is retrieved from the multiphysics framework model in step 13 (e.g., the dielectric constant of crude oil decreases by 2 for every 10°C increase in temperature). This function describes the ability of the medium to absorb microwave signals at different temperatures.

[0241] Compensation calculation:

[0242] The attenuation rate of the microwave signal acquired by each sensor is calibrated:

[0243] For example, a sensor collects an attenuation rate of 35% and a temperature of 42℃ at the 5th second, while the dielectric constant corresponding to the standard temperature (e.g., 25℃) at the sensor location is higher. According to the characteristic function, 42℃ is 17℃ higher than 25℃, and the dielectric constant decreases by 3.4 (17℃×0.2 / ℃), resulting in a microwave attenuation rate that is 5% lower than that at the standard temperature (the dielectric constant and attenuation rate are positively correlated). The calibrated attenuation rate = original attenuation rate + compensation value = 35% + 5% = 40%, ensuring that the attenuation rate at different temperatures can be compared horizontally.

[0244] Data correction:

[0245] Repeat the above steps to perform temperature compensation on the attenuation rate of all sensors at each time point, generating a "calibrated attenuation rate dataset".

[0246] In step 33 above, based on the sensor's height coordinates (e.g., sensor heights from the bottom to the top of the tank are 0.5m, 1m, 1.5m...5m), the calibrated attenuation rates at the same time point are sorted by height, and a "height-attenuation rate" curve is plotted (horizontal axis is height, vertical axis is attenuation rate; analyze the signal abrupt change characteristics in the curve (attenuation rates differ significantly between different media)).

[0247] Pure oil layers have low dielectric constants (e.g., 2-3) and low microwave attenuation rates (e.g., 10%-20%).

[0248] The emulsion layer has a moderate dielectric constant (e.g., 10-20) and a moderate attenuation rate (e.g., 30%-50%).

[0249] Pure water layers have high dielectric constants (e.g., 70-80) and high attenuation rates (e.g., 60%-80%).

[0250] When the curve suddenly jumps from a low decay rate (20%) to a medium decay rate (40%), it corresponds to the "oil-emulsion layer interface" (e.g., at a height of 3m).

[0251] When the curve suddenly jumps from a medium decay rate (40%) to a high decay rate (70%), it corresponds to the "emulsion layer-water layer interface" (e.g., at a height of 1.5m).

[0252] Interface height confirmed:

[0253] Compare the abrupt change points of the curves at three consecutive time points. If the abrupt change characteristics at the same height are consistent (e.g., a sudden increase in decay rate occurs for 3 consecutive seconds at 3m), then the height is confirmed as the stable interface height.

[0254] In step 34 above, the total height of the tank (e.g., 5m) and the height of the bottom truncated cone (e.g., 0.5m) are obtained from the basic geometric model in step 11, and the effective measurement height range is determined to be from 0.5m (top of the truncated cone) to 5m (top of the tank).

[0255] Thickness calculation for each layer:

[0256] Pure oil layer thickness = tank top height - oil-emulsion layer interface height (e.g., 5m - 3m = 2m);

[0257] Emulsion layer thickness = oil-emulsion layer interface height - emulsion layer-water layer interface height (e.g., 3m - 1.5m = 1.5m);

[0258] Pure water layer thickness = emulsion layer - water layer interface height - truncated cone top height (e.g., 1.5m - 0.5m = 1m).

[0259] Data validation and output:

[0260] Check that the sum of the thicknesses of the three layers equals the effective measured height (2m + 1.5m + 1m = 4.5m, consistent with 5m - 0.5m) to ensure there is no calculation error;

[0261] The thicknesses of the pure oil layer, emulsion layer, and pure water layer are arranged in chronological order (e.g., a set of data is recorded every 10 seconds) to form a sequence of measured results (e.g., at the 10th second: pure oil layer 2m, emulsion layer 1.5m, pure water layer 1m).

[0262] In the above steps, step 4 extracts the current emulsion layer thickness (e.g., 1.2m) from the measured results of step 3, and at the same time retrieves the predicted emulsion layer thickness (e.g., 1.4m) output from step 2.

[0263] Calculate the deviation rate: (measured value - predicted value) ÷ measured value × 100% = (1.2 - 1.4) ÷ 1.2 × 100% ≈ -16.7%, the absolute value is > 10%, triggering model calibration;

[0264] When recalibrating the 3D model, the focus is on correcting the material property functions in step 13 (such as adjusting the slope of the viscosity-temperature curve of the emulsion mixture to make the predicted values ​​closer to the actual measurements).

[0265] After calibration, return to step 2 to re-execute the finite element simulation and generate new key parameters (layering speed, filling curve, etc.).

[0266] If the deviation rate is ≤10%, the model prediction is considered valid, no calibration is required, and the subsequent steps can be continued.

[0267] Automatic demulsifier dispensing trigger:

[0268] Real-time monitoring of the emulsion layer thickness output from step 3:

[0269] When the thickness is greater than 0.1m (the set minimum effective thickness threshold), the system automatically calls the demulsifier injection dynamic curve generated in step 23; for example, if the curve shows that the injection rate needs to be 15L / h at the 5th minute, then a command is sent to the injection pump to adjust the flow rate to 15L / h, and the injection rate is continuously updated according to the curve (such as reducing it to 10L / h at the 10th minute); if the emulsion layer thickness is ≤0.1m, injection is paused to avoid wasting demulsifier.

[0270] In step 5 above, continuously monitor the thickness of the pure water layer and the pure oil layer output from step 3:

[0271] When the thickness of the pure water layer is >0.2m (a distinct water layer can be detected) and the thickness of the pure oil layer is >0.5m (a distinct oil layer can be detected), it is determined that the oil and water have been initially separated and the heating conditions are met. The heating scheme is retrieved from the temperature control parameters set in step 25 (e.g., 50kW heating for 20 seconds in the first stage, and 20kW heating for 26.3 seconds in the second stage). Instructions are sent to the temperature control valve of the steam coil to adjust the steam flow according to the scheme (e.g., 80% valve opening for 50kW, and 30% opening for 20kW). The temperature inside the tank is collected in real time (via the temperature sensor in step 31) to ensure that the deviation between the actual temperature and the target temperature in the parameter set (e.g., 40℃) is ≤±2℃.

[0272] If the thickness requirements for the pure water layer and pure oil layer are not met, heating will be delayed, and the test will be repeated every 30 seconds.

[0273] Step 6 above: Drainage of the primary tank and moisture content testing:

[0274] After temperature control is completed in step 5 (e.g., the temperature stabilizes at 40℃ and is maintained for 5 minutes), open the drain valve at the bottom of the primary tank (initially 30%), and simultaneously install an online moisture content analyzer on the drain pipe to detect the moisture content of the drained water once per second.

[0275] When the moisture content of the drainage is >5%:

[0276] Retrieve the predicted stratified recovery time (e.g., 60 minutes) output from step 24 and compare it with a preset threshold (e.g., 30 minutes):

[0277] If the recovery time is ≤30 minutes (the system can recover quickly), close the drain valve and pump the discharged water back to the primary tank for continued settling (at the same time, increase the demulsifier dosage by 10% and extend the heating time by 5 minutes); if the recovery time is >30 minutes (the system recovers slowly), stop returning water to the primary tank and introduce the high water content wastewater into the pretreatment area of ​​the secondary purification tank (specifically for treating oily wastewater).

[0278] When the drainage moisture content is ≤5%:

[0279] Triggering the crude oil transfer process:

[0280] Turn on the crude oil transfer pump from the primary tank to the secondary purification tank to transfer pure oil layer crude oil (water content ≤5%) to the secondary purification tank; during the transfer process, keep the drain valve slightly open (10%), continuously monitor the water content of the drained oil, and control the drainage volume by adjusting the valve opening (close it less when the water content increases and open it more when it decreases); until the water content of the drained oil drops to ≤0.5% (meeting the standard), completely close the drain valve; after receiving the crude oil, the secondary purification tank performs further dehydration treatment (such as heating and sedimentation), and finally outputs purified crude oil with a water content ≤0.5%, completing the entire separation process.

[0281] Through the above steps, the entire process from model calibration, reagent injection, temperature control to final crude oil purification is automated, ensuring that separation efficiency and crude oil quality meet the standards.

[0282] like Figure 2 As shown, the intelligent identification, monitoring, and control system for the oil-water interface in the oil settling tank includes:

[0283] The module is used to construct a multi-physics coupled three-dimensional model based on the geometry of the target settling tank, the deployment location of the microwave-temperature composite sensor array, and material characteristic parameters.

[0284] The output module is used to simulate the oil-water separation process using the finite element method based on the three-dimensional model, and output key parameters, including the predicted value of the separation rate, the predicted value of the emulsion layer thickness, the dynamic curve of demulsifier injection, the predicted value of the separation recovery time, and the temperature control parameter set.

[0285] The measurement module is used to collect data through a microwave-temperature composite sensor array based on the sensor deployment location to obtain measurement results, including the thickness of the pure water layer, the thickness of the emulsion layer, and the thickness of the pure oil layer.

[0286] The comparison module is used to compare the measured emulsion layer thickness with the predicted emulsion layer thickness. If the deviation is >10%, the 3D model is recalibrated and the simulation is restarted. When the output emulsion layer thickness is >0, the demulsifier injection dynamic curve is invoked for automatic dosing.

[0287] The control module is used to initiate heating by executing the temperature control parameter set when both the output pure water layer thickness and the pure oil layer thickness are greater than 0. After temperature control is completed, the water layer is discharged from the bottom of the primary tank, and the water content of the discharged water is monitored in real time. If the water content of the discharged water is greater than 5%, and the predicted recovery time is less than or equal to the threshold, the system returns to the primary tank for further processing.

[0288] When the predicted recovery time is greater than the threshold, the crude oil is introduced into the secondary purification tank; if the water content of the wastewater is less than or equal to 5%, the crude oil is transferred to the secondary purification tank and the water is controlled until the water content of the wastewater is less than or equal to 0.5%, and finally purified crude oil is output.

[0289] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for intelligent identification, monitoring, and control of the oil-water interface in an oil settling tank, characterized in that, include: Step 1: Based on the geometric structure of the target settling tank, the deployment location of the microwave-temperature composite sensor array, and the material characteristic parameters, construct a multi-physics coupled three-dimensional model. Step 2: Based on the three-dimensional model, simulate the oil-water separation process using the finite element method to output key parameters. Key parameters include predicted values ​​of stratification rate, predicted values ​​of emulsion layer thickness, dynamic curve of demulsifier injection, predicted values ​​of stratification recovery time, and temperature control parameter set. This includes: Step 21: Based on the multi-physics coupled three-dimensional model, simulate the fluid motion state during the oil-water separation process using the finite element method to output predicted values ​​of oil-water interface stratification rate. Step 22: Based on the predicted stratification rate from Step 21 and the viscosity-temperature relationship in the material property function, calculate the dynamic change of the emulsion layer thickness over time and output the predicted emulsion layer thickness. Step 23: Based on the predicted emulsion layer thickness from Step 22 and the pre-stored demulsifier reaction kinetic parameters, generate a dynamic curve showing the relationship between demulsifier dosage and time. Step 24: Based on the predicted stratification rate from Step 21 and the predicted emulsion layer thickness from Step 22, simulate the time required for the oil-water interface to recover stability after disturbance and output the predicted stratification recovery time. Step 25: Combining the predicted stratification recovery time from Step 24 and the thermal conductivity coefficient of the steam coil, optimize the combination of heating power and duration to generate a set of temperature control parameters. Step 3: Based on the sensor deployment locations in Step 1, data is collected using a microwave-temperature composite sensor array to obtain the measured results, which include the thickness of the pure water layer, the thickness of the emulsion layer, and the thickness of the pure oil layer. Step 4: Compare the measured emulsion layer thickness in Step 3 with the predicted emulsion layer thickness in Step 2. If the deviation is >10%, recalibrate the 3D model and return to Step 2 to resimulate. When the emulsion layer thickness output in step 3 is greater than 0, the demulsifier addition dynamic curve from step 2 is invoked for automatic dosing. Step 5: When the output of Step 3 shows that the thickness of the pure water layer is >0 and the thickness of the pure oil layer is >0, execute the temperature control parameter set of Step 2 to start heating; Step 6: After temperature control is completed in step 5, drain the water layer from the bottom of the primary tank and monitor the water content of the drain in real time; If the moisture content of the wastewater is >5%, and the predicted recovery time is ≤ the threshold, it should be returned to the primary tank for further processing, or When the predicted recovery time is greater than the threshold, the crude oil is introduced into the secondary purification tank; if the water content of the wastewater is less than or equal to 5%, the crude oil is transferred to the secondary purification tank and the water is controlled until the water content of the wastewater is less than or equal to 0.5%, and finally purified crude oil is output.

2. The intelligent identification, monitoring, and control method for the oil-water interface in an oil settling tank according to claim 1, characterized in that, Step 1: Based on the geometric structure of the target settling tank, the deployment location of the microwave-temperature composite sensor array, and material characteristic parameters, construct a multi-physics coupled three-dimensional model, including: Step 11: Based on the actual dimensions of the target settling tank, generate a basic geometric model including the tank height, diameter, bottom cone angle, and steam coil layout path; Step 12: Based on the vertical spacing and radial insertion depth of the microwave-temperature composite sensor array, mark the three-dimensional spatial coordinates of each sensor and output an enhanced geometric model with sensor coordinates. Step 13: On the enhanced geometric model of step 12, define the material property functions of density-viscosity-dielectric constant of crude oil, water and emulsion mixture as a function of temperature, define the thermal conductivity coefficient of steam coil and the thermal boundary conditions of tank wall, define the microwave emission source location and signal attenuation calculation domain, so as to obtain a multiphysics framework model with three field definitions. Step 14: Based on the pre-stored crude oil characteristic database, bind the component data of the target crude oil, namely the water content threshold and demulsifier reaction kinetic parameters, to the fluid domain of the multiphysics framework model in Step 13, and output the material binding simulation model. Step 15: Based on the basic geometric model, enhanced geometric model, multiphysics framework model, and material binding simulation model, generate a multiphysics coupled three-dimensional model that can simultaneously solve fluid stratification, heat conduction, and microwave signal attenuation.

3. The intelligent identification, monitoring, and control method for the oil-water interface in an oil settling tank according to claim 2, characterized in that, Step 11: Based on the actual dimensions of the target settling tank, generate a basic geometric model including the tank height, diameter, bottom cone angle, and steam coil layout path, including: Step 111: Generate the main geometric frame of the cylindrical tank body based on the actual height H and diameter D of the target settling tank; Step 112: Based on the cylindrical frame of Step 111, add a frustum-shaped structure with a cone angle θ at the bottom to form the complete container outline of the settling tank. Step 113: Based on the actual pitch P and initial height h0 of the steam coil, generate a spiral path model within the container outline from step 112. Step 114: Generate the basic geometric model based on the main geometric framework of the tank, the container outline, and the spiral path model.

4. The intelligent identification, monitoring, and control method for the oil-water interface in an oil settling tank according to claim 3, characterized in that, Step 12: Based on the vertical spacing and radial insertion depth of the microwave-temperature composite sensor array, mark the three-dimensional spatial coordinates of each sensor, and output an enhanced geometric model with sensor coordinates, including: Step 121: Based on the basic geometric model, divide the tank into N measurement planes with a spacing of ΔL along the height direction; Step 122: In each measurement plane, determine the distance from the sensor tip to the tank wall based on the sensor's radial insertion depth d, where d <D / 2; Step 123: Based on the planar position in step 121 and the radial depth in step 122, mark the three-dimensional coordinates (xn, yn, zn) of each sensor in the basic geometric model; Step 124: Write the labeled coordinates into the basic geometric model and output the enhanced geometric model with a spatial coordinate set.

5. The intelligent identification, monitoring, and control method for the oil-water interface in an oil settling tank according to claim 4, characterized in that, Step 22: Based on the predicted stratification rate from Step 21, and combining it with the viscosity-temperature relationship in the material property function, calculate the dynamic change of the emulsion layer thickness over time, and output the predicted emulsion layer thickness, including: Step 221: Based on the predicted stratification rate output in Step 21, and combined with the viscosity-temperature relationship in the material property function, calculate the oil-water emulsion phase separation efficiency coefficient. Step 222: Dynamically correct the rate of change of emulsion layer thickness over time based on the separation efficiency coefficient and the predicted value of stratification rate; Step 223: Integrate the corrected emulsion layer thickness change rate over time and output the dynamic sequence of the predicted emulsion layer thickness over time.

6. The intelligent identification, monitoring, and control method for the oil-water interface in an oil settling tank according to claim 5, characterized in that, Step 23: Based on the predicted emulsion layer thickness from Step 22 and the pre-stored demulsifier reaction kinetic parameters, generate a dynamic curve showing the relationship between demulsifier injection volume and time, including: Step 231: Extract the maximum value of the emulsion layer thickness change rate output in step 222, and combine it with the critical emulsion strength threshold in the pre-stored demulsifier reaction kinetic parameters to determine the initial demulsifier injection rate; Step 232: Based on the dynamic sequence decay trend of the predicted emulsion layer thickness, generate the injection rate decay curve in segments according to the concentration-reaction rate relationship in the reaction kinetic parameters. Step 233: Superimpose the initial injection rate and the segmented decay curve to generate a dynamic curve showing the continuous change of demulsifier injection amount over time.

7. The intelligent identification, monitoring, and control method for the oil-water interface in an oil settling tank according to claim 6, characterized in that, Step 24: Based on the predicted stratification rate from Step 21 and the predicted emulsion layer thickness from Step 22, simulate the time required for the oil-water interface to recover stability after disturbance, and output the predicted stratification recovery time, including: Step 241: Based on the predicted stratification rate from Step 21, calculate the theoretical recovery time of the oil-water interface without emulsion layer disturbance. Step 242: Based on the dynamic sequence of predicted emulsion layer thickness output in step 223, extract the thickness peak value and duration. Step 243: Combine the peak thickness of the emulsion layer with the dielectric constant-viscosity relationship in the material property function to calculate the hindrance correction coefficient of the emulsion layer to interface recovery. Step 244: Multiply the theoretical recovery time by the stagnation correction coefficient to output the predicted value of the stratified recovery time after the disturbance.

8. The intelligent identification, monitoring, and control method for the oil-water interface in an oil settling tank according to claim 7, characterized in that, Step 3: Based on the sensor deployment locations in Step 1, data is collected using a microwave-temperature composite sensor array to obtain the measured results. These results include the thickness of the pure water layer, the emulsion layer, and the pure oil layer, including: Step 31: Based on the enhanced geometric model with spatial coordinate set output in step 124, activate the microwave-temperature composite sensor array according to the sensor's three-dimensional coordinates, and synchronously collect the microwave signal attenuation rate and temperature value at each spatial point. Step 32: Based on the material property function of dielectric constant changing with temperature defined in Step 13, perform temperature compensation calibration on the attenuation rate of the microwave signal acquired in Step 31. Step 33: Identify the height of the oil-emulsion layer interface and the height of the emulsion layer-water layer interface based on the abrupt change point of the calibrated microwave signal attenuation rate. Step 34: Combining the dual-interface height and tank geometry identified in Step 33, calculate the thickness of the pure water layer, the emulsion layer, and the pure oil layer, and output the measured results.

9. A smart identification, monitoring, and control system for the oil-water interface in an oil settling tank, characterized in that, The system is used to perform the method according to any one of claims 1 to 8, comprising: The module is used to construct a multi-physics coupled three-dimensional model based on the geometry of the target settling tank, the deployment location of the microwave-temperature composite sensor array, and material characteristic parameters. The output module is used to simulate the oil-water separation process using the finite element method based on the 3D model, and output key parameters, including predicted values ​​for stratification rate, emulsion layer thickness, demulsifier injection dynamic curve, stratification recovery time, and a set of temperature control parameters. This includes: simulating the fluid motion state during oil-water separation using the finite element method based on the multi-physics coupled 3D model, and outputting the predicted value for the oil-water interface stratification rate; calculating the dynamic change of the emulsion layer thickness over time based on the predicted stratification rate and the viscosity-temperature relationship in the material property function, and outputting the predicted emulsion layer thickness; generating a dynamic curve of the demulsifier injection amount versus time based on the predicted emulsion layer thickness and pre-stored demulsifier reaction kinetic parameters; simulating the time required for the oil-water interface to recover stability after disturbance based on the predicted stratification rate and emulsion layer thickness, and outputting the predicted stratification recovery time; and optimizing the combination of heating power and duration based on the predicted stratification recovery time and the heat transfer coefficient of the steam coil to generate a set of temperature control parameters. The measurement module is used to collect data through a microwave-temperature composite sensor array based on the sensor deployment location to obtain measurement results, including the thickness of the pure water layer, the thickness of the emulsion layer, and the thickness of the pure oil layer. The comparison module is used to compare the measured emulsion layer thickness with the predicted emulsion layer thickness. If the deviation is >10%, the 3D model is recalibrated and the simulation is restarted. When the output emulsion layer thickness is >0, the demulsifier injection dynamic curve is invoked for automatic dosing. The control module is used to initiate heating by executing the temperature control parameter set when both the output pure water layer thickness and the pure oil layer thickness are greater than 0. After temperature control is completed, the water layer is discharged from the bottom of the primary tank, and the water content of the discharged water is monitored in real time. If the water content of the discharged water is greater than 5%, and the predicted recovery time is less than or equal to the threshold, the system returns to the primary tank for further processing. When the predicted recovery time is greater than the threshold, the crude oil is introduced into the secondary purification tank; if the water content of the wastewater is less than or equal to 5%, the crude oil is transferred to the secondary purification tank and the water is controlled until the water content of the wastewater is less than or equal to 0.5%, and finally purified crude oil is output.

Images