Method and system for predicting membrane separation fluid shear stress distribution

Abstract

The present application belongs to the technical field of computer-aided fluid dynamics simulation, and discloses a shell membrane separation fluid shear force distribution prediction method and system, which solves the problems of inaccurate shear force prediction, simulation unable to match process constraints, and poor working condition adaptability of existing equipment. The scheme collects four types of parameters to build a fluid dynamics calculation domain to solve the flow field, generates the spatial distribution of wall shear force and constructs the effective shear force action zone; relies on the measured data to generate an error correction field to correct the boundary conditions, carries out multi-objective collaborative optimization to output the equipment control parameters, and supports dynamic boundary and acoustic-solid coupling correction to restore the membrane peeling dynamic process. The whole scheme forms a closed-loop regulation and control, reduces the risk of residual membrane and membrane damage, and improves the prediction accuracy and working condition adaptability.

Description

Technical Field

[0001] This invention relates to the field of computer-aided fluid dynamics simulation technology, specifically to a method and system for predicting the shear force distribution of fluids in shell-membrane separation. Background Technology

[0002] Eggshell resource utilization can yield calcium carbonate and eggshell membrane materials. The eggshell membrane adheres to the inner wall of the eggshell, and the separation process typically relies on fluid scouring, agitation, pressure changes, or shearing forces generated by the flow guiding structure within the separation chamber. When the fluid shear force is too low, the eggshell membrane is difficult to peel from the inner wall, resulting in decreased separation efficiency; when the fluid shear force is too high, the shear force direction changes abruptly, or the high shear force duration is too long, the eggshell membrane is prone to breakage, reducing the integrity of the membrane product. The amplitude, direction, gradient, duration, and spatial coverage of the shear force on the inner wall of the shell-membrane separation chamber directly affect the shell-membrane separation effect.

[0003] Existing shell-and-membrane separation equipment typically uses manual experience to set stirring speed, flow rate, pressure, and baffle position. While this experience-based approach can yield usable results under specific equipment and material conditions, it fails to accurately reflect changes in the spatial distribution of shear force within the separation chamber when variations occur in the separation chamber size, eggshell batch size, fluid dynamic viscosity, inlet velocity, or temperature.

[0004] While conventional computational fluid dynamics (CFD) simulations can obtain the spatial distribution of velocity, pressure, and wall shear forces, these simulation results typically do not directly incorporate constraints such as the lower limit shear limit for shell-membrane peeling, the upper limit shear limit for membrane integrity, the continuity constraint of shear force direction, the effective duration constraint of shear force, and experimental measurement error correction. Therefore, they are difficult to directly translate into control parameters for shell-membrane separation equipment. Judging the shell-membrane separation effect solely based on the average wall shear force or single-point pressure can easily overlook eggshell membrane residue caused by localized low-shear regions, as well as eggshell membrane damage caused by localized high-shear regions.

[0005] During the gradual loosening and peeling of the eggshell membrane, the local morphology of the boundary between the shell membrane and the wall also changes. Local bulging, partial detachment, and complete detachment of the shell membrane all alter the flow state in the near-wall region. If the boundary between the shell membrane and the wall is always treated as a static rigid boundary, the calculation results cannot reflect the dynamic interaction between fluid shear force and the shell membrane peeling state. Furthermore, using only velocity and pressure measurement data to correct the fluid dynamics calculation domain may still fail to directly reflect the actual location of shell membrane fiber breakage and local peeling. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for predicting the shear force distribution of fluids in shell-membrane separation, which solves the technical problems of existing shell-membrane separation equipment, such as difficulty in accurately predicting the spatial distribution of shear force on the inner wall of the separation chamber, difficulty in correlating computational fluid dynamics simulation results with constraints on shell-membrane peeling and eggshell membrane integrity, difficulty in correcting the fluid dynamics calculation domain based on experimental measurement errors, and difficulty in coordinating and optimizing stirring parameters, flow parameters, pressure parameters, and baffle configuration.

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

[0008] Methods for predicting the shear force distribution of fluids in membrane separation include:

[0009] Obtain structural parameters, fluid properties, shell-membrane separation effect constraint parameters, and experimental measurement data of the shell-membrane separation chamber;

[0010] A fluid dynamics computational domain was established based on the structural parameters and fluid properties of the shell-membrane separation cavity, and the velocity field, pressure field, and near-wall velocity gradient field were obtained by solving.

[0011] The spatial distribution of wall shear force is generated based on the near-field velocity gradient field of the wall.

[0012] Based on the constraint parameters of shell-membrane separation effect, an effective shear force zone is constructed from the spatial distribution of wall shear force;

[0013] An error correction field is generated based on experimental measurement data, velocity field, pressure field, and near-wall velocity gradient field.

[0014] The boundary conditions of the fluid dynamics computational domain are updated using the error correction field, and the corrected spatial distribution of shear force is obtained by resolving the problem. The effective action zone of shear force is then updated based on the corrected spatial distribution of shear force.

[0015] Based on the corrected spatial distribution of shear force and the updated effective shear force zone, the stirring parameters, flow parameters, pressure parameters and baffle configuration are optimized in a coordinated manner to output the control parameters of the shell-and-membrane separation equipment.

[0016] As a preferred embodiment of the present invention, the structural parameters of the shell-membrane separation chamber include the geometric boundary of the separation chamber, the inlet boundary, the outlet boundary, the stirring component boundary, the candidate baffle installation boundary, and the shell-membrane contact wall boundary; the fluid property parameters include density, dynamic viscosity, temperature, and inlet flow velocity; the shell-membrane separation effect constraint parameters include the lower limit shear constraint for shell-membrane peeling, the upper limit shear constraint for membrane integrity, the continuity constraint for shear force direction, and the constraint for the effective duration of shear force action.

[0017] As a preferred technical solution of the present invention, the establishment of the fluid dynamics calculation domain includes: setting a wall layer mesh on the boundary of the shell-film contact wall; setting a velocity inlet on the boundary of the inlet; setting a pressure outlet on the boundary of the outlet; setting a rotational reference region on the boundary of the stirring component; and using the candidate installation boundary of the baffle as an updatable geometric constraint input for the numerical solution process.

[0018] As a preferred technical solution of the present invention, the solution to obtain the velocity field, pressure field and near-wall velocity gradient field includes: solving the fluid continuity control equation and momentum control equation using a finite volume discretization method; generating turbulent viscosity correction through a turbulent closed model; updating the velocity field and pressure field through pressure-velocity coupling iteration; and outputting the velocity field, pressure field and near-wall velocity gradient field within the separation chamber after the calculated residuals and mass flow imbalance satisfy the preset convergence conditions.

[0019] As a preferred embodiment of the present invention, the construction of the effective shear force zone includes: extracting the shear force amplitude, shear force direction, shear force gradient, and shear force residence time from the spatial distribution of wall shear force; marking the positions where the shear force amplitude satisfies the lower limit shear constraint of shell-membrane peeling and does not exceed the upper limit shear constraint of membrane integrity as candidate effective regions; eliminating positions where the shear force direction changes abruptly based on the continuity constraint of shear force direction, and eliminating positions where the shear force residence time is insufficient based on the effective shear force duration constraint, thereby obtaining the effective shear force zone.

[0020] As a preferred embodiment of the present invention, the generation of the error correction field includes: generating a measurement velocity field using particle image velocimetry data, and generating a measurement pressure field using pressure sensing data; mapping the measurement velocity field and the measurement pressure field to the grid nodes of the fluid dynamics computational domain; calculating the local velocity deviation between the calculated velocity field and the measured velocity field, the local pressure deviation between the pressure field and the measured pressure field, and the local velocity gradient deviation between the near-wall velocity gradient field and the corresponding gradient of the measured velocity field; and spatially propagating and smoothing the local velocity deviation, local pressure deviation, and local velocity gradient deviation along the boundary of the shell-film contact wall to form the error correction field.

[0021] As a preferred embodiment of the present invention, the step of updating the boundary conditions of the fluid dynamics computational domain using the error correction field includes: extracting the inlet velocity deviation, outlet pressure deviation, and local velocity gradient deviation from the error correction field; updating the inlet boundary based on the inlet velocity deviation; updating the outlet boundary based on the outlet pressure deviation; updating the wall layer correction coefficient based on the local velocity gradient deviation; resolving the corrected shear force spatial distribution; and reconstructing the effective shear force zone based on the corrected shear force spatial distribution; and updating the baffle candidate installation boundary based on the positional offset and coverage change between the reconstructed effective shear force zone and the original effective shear force zone.

[0022] As a preferred technical solution of the present invention, the collaborative optimization of stirring parameters, flow parameters, pressure parameters and baffle configuration includes: establishing a multi-objective optimization model with shear force effective action zone coverage, shear force distribution variability, membrane integrity risk value and energy consumption constraint value as evaluation quantities, and using the corrected shear force spatial distribution and the updated shear force effective action zone as constraint inputs of the multi-objective optimization model.

[0023] As a preferred technical solution of the present invention, the multi-objective optimization model first determines the candidate combinations of continuous variables of stirring parameters, flow parameters and pressure parameters through the response surface model, and then determines the candidate combinations of discrete variables of baffle position and baffle angle through the genetic search model. The candidate combinations of continuous variables, as well as the coverage of the effective action zone of shear force, the variability of shear force distribution, the membrane integrity risk value and the energy consumption constraint value corresponding to the candidate combinations of continuous variables, are input into the genetic search model as fitness calculation constraints for the candidate combinations of discrete variables.

[0024] As a preferred embodiment of the present invention, the method further includes establishing a shell-membrane separation effect mapping model. The shell-membrane separation effect mapping model receives the effective shear force action zone coverage, shear force distribution variability, shear force direction continuity index, membrane integrity risk value, and corrected shear force spatial distribution, and outputs the separation efficiency prediction value and the membrane integrity prediction value.

[0025] As a preferred technical solution of the present invention, the predicted values ​​of separation efficiency and membrane integrity output by the shell-membrane separation effect mapping model are fed back to the multi-objective optimization model; when the difference between the control parameters of the shell-membrane separation equipment output by the previous two times is lower than the control update threshold, the effective shear force action zone coverage reaches the preset coverage threshold, the predicted value of separation efficiency reaches the preset separation efficiency threshold, and the predicted value of membrane integrity reaches the preset integrity threshold, the collaborative optimization is terminated.

[0026] As a preferred embodiment of the present invention, after outputting the control parameters of the shell-membrane separation device, the control parameters are sent to the separation device control node; the separation device control node adjusts the stirring driver, flow regulating valve, pressure regulating valve and baffle actuator according to the control parameters of the shell-membrane separation device; the experimental measurement data collected after the shell-membrane separation device is running is transmitted back to generate the next round of error correction field.

[0027] As a preferred embodiment of the present invention, after generating the spatial distribution of wall shear force based on the near-zone velocity gradient field of the wall and before constructing the effective shear force action zone, the method further includes: extracting the transient pulsating components of the spatial distribution of wall shear force on the time scale, and constructing a membrane local loosening fatigue accumulation function; dynamically dividing the shell-membrane contact wall boundary into an unpeeled boundary region, a transitional peeling boundary region, and a fully peeled boundary region based on the output value of the membrane local loosening fatigue accumulation function; generating a dynamic mesh displacement field for the transitional peeling boundary region, and updating the three-dimensional morphological features of the shell-membrane contact wall boundary based on the dynamic mesh displacement field during the time step iteration in the hydrodynamic computation domain; the membrane local loosening fatigue accumulation function is:

[0028] ;

[0029] in, This represents the cumulative fatigue index of localized membrane loosening, expressed in seconds. Indicates time, in seconds; This represents the integration time variable, in seconds. This represents the amplitude of the wall shear force at the boundary between the shell and membrane contact wall, in Pascals (Pa). This represents the initial threshold for membrane loosening, expressed in Pascals (Pa). This indicates a positive reference shear force parameter, in Pascals (Pa). This represents the fatigue index, which is dimensionless. This indicates the positive part operator. The difference inside the parentheses is taken when it is greater than zero, otherwise it is taken as zero. Indicates time integration operation; This represents the integral time element, with the unit being seconds.

[0030] As a preferred embodiment of the present invention, the step of generating an error correction field based on experimental measurement data, velocity field, pressure field, and near-wall velocity gradient field further includes: acquiring high-frequency acoustic emission signals from the experimental measurement data; extracting the energy value of the high-frequency acoustic emission signal within a preset stripping main frequency band and spatially mapping the energy value to the transition stripping boundary region; calculating the acoustic-structure coupling deviation between the energy value and the displacement change rate of the dynamic grid displacement field within the transition stripping boundary region; spatially propagating and smoothing the acoustic-structure coupling deviation, local velocity deviation, local pressure deviation, and local velocity gradient deviation along the shell-membrane contact wall boundary to synthesize the error correction field; and using the acoustic-structure coupling deviation to perform bounded correction on the mapping relationship between the fatigue index of the membrane local loosening fatigue accumulation function and the displacement change rate of the transition stripping boundary region.

[0031] As a preferred technical solution of the present invention, it includes a processor, a memory, a data acquisition interface, and a control output interface; the memory stores program instructions that can be executed by the processor, and when the program instructions are executed by the processor, the method for predicting the shear force distribution of the membrane separation fluid as described above is implemented.

[0032] The data acquisition interface is used to receive structural parameters of the shell-membrane separation chamber, fluid property parameters, shell-membrane separation effect constraint parameters, and experimental measurement data, and outputs the received data to the processor.

[0033] The processor is used to generate velocity field, pressure field, near-wall velocity gradient field, spatial distribution of wall shear force, effective shear force zone, error correction field, corrected spatial distribution of shear force, and control parameters of the shell-and-film separation equipment. The control output interface is used to output the control parameters of the shell-and-film separation equipment to the stirring actuator, flow regulating valve, pressure regulating valve, and baffle actuator.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] This invention inputs the structural parameters of the shell-membrane separation chamber, fluid property parameters, shell-membrane separation effect constraint parameters, and experimental measurement data into the fluid dynamics prediction process, expanding the prediction object from a general velocity or pressure field to the spatial distribution of wall shear force at the shell-membrane contact wall boundary. By constructing an effective shear force zone, the lower limit shear constraint for shell-membrane peeling, the upper limit shear constraint for membrane integrity, the continuity constraint for shear force direction, and the constraint for the effective duration of shear force are transformed into calculable spatial constraints, reducing the risk of local residual membrane and local membrane rupture caused by evaluating the shell-membrane separation effect solely based on the average value of wall shear force.

[0036] This invention updates the boundary conditions of the fluid dynamics computational domain through an error correction field and updates the effective shear force zone after resolving the problem. It can utilize particle image velocity measurement data and pressure sensor data to correct deviations between the initial boundary conditions and actual operating conditions, improving prediction consistency under different fluid dynamic viscosity, temperature, and inlet flow velocity conditions. Through cascaded synergistic optimization of the response surface model and the genetic search model, it can simultaneously handle continuous control variables such as stirring parameters, flow rate parameters, and pressure parameters, as well as discrete structural variables such as baffle position and baffle angle.

[0037] This invention uses a shell-membrane separation effect mapping model to correlate the corrected spatial distribution of shear force with predicted values ​​of separation efficiency and membrane integrity, forming a closed-loop technical path from prediction, correction, optimization, control to re-measurement. It reflects the influence of gradual loosening, local bulging, and complete peeling of the eggshell membrane on the near-wall flow state through a membrane local loosening fatigue accumulation function and a dynamic grid displacement field. Furthermore, it improves the predictive ability of the dynamic peeling process of the eggshell membrane by generating acoustic-structure interaction deviation through high-frequency acoustic emission signals. Attached Figure Description

[0038] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0039] Figure 1 This is an overall flowchart of the method described in this invention.

[0040] Figure 2 This is a detailed flowchart of the process for generating the error correction field and updating the boundary conditions in this invention.

[0041] Figure 3 This is a detailed flowchart of the collaborative optimization of the present invention.

[0042] Figure 4 This is a block diagram of the architecture of the membrane separation fluid shear force distribution prediction system of the present invention.

[0043] Figure 5 This is a data flow diagram illustrating the closed loop of data acquisition, prediction, optimization, control, and feedback in this invention. Detailed Implementation

[0044] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0045] The following is in conjunction with the appendix Figures 1-5 The embodiments of the present invention will be described in detail below.

[0046] Example 1: This example provides a method for predicting the distribution of fluid shear force in shell-membrane separation. The eggshell-membrane separation device includes a separation chamber, an inlet, an outlet, a stirring component, an adjustable baffle, and a shell-membrane contact wall boundary. The shell-membrane contact wall boundary is the area where the fluid shear force directly acts on the eggshell membrane or the attached membrane layer.

[0047] The method for predicting the shear force distribution of shell-membrane separation fluid generates the spatial distribution of wall shear force through computer-aided fluid dynamics simulation, and converts the spatial distribution of wall shear force into control parameters for shell-membrane separation equipment that can guide the operation of the eggshell-membrane separation equipment.

[0048] Specifically as follows:

[0049] Step S1: Obtain the structural parameters of the shell-membrane separation chamber, fluid property parameters, shell-membrane separation effect constraint parameters, and experimental measurement data.

[0050] The structural parameters of the shell-membrane separation chamber are obtained from 3D scanning data, engineering drawings, or 3D modeling data of the equipment, including the geometric boundary of the separation chamber, the inlet boundary, the outlet boundary, the boundary of the stirring component, the candidate installation boundary of the baffle, and the boundary of the shell-membrane contact wall.

[0051] Fluid properties are obtained from density, dynamic viscosity, temperature and inlet velocity by densitometer, viscometer, temperature sensor and flow rate sensor.

[0052] The constraint parameters for shell-membrane separation effect are obtained from historical shell-membrane separation data, bench test data, or process calibration data, including the lower limit shear constraint for shell-membrane peeling, the upper limit shear constraint for membrane integrity, the continuity constraint for shear force direction, and the constraint for the effective duration of shear force action.

[0053] The lower shear limit constraint for shell membrane peeling represents the minimum shear force required for the eggshell membrane to effectively loosen or peel from the inner shell wall. The upper shear limit constraint for membrane integrity represents the maximum shear force the eggshell membrane can withstand without excessive rupture. The shear force direction continuity constraint represents the permissible range of shear force direction variation between adjacent wall locations. The effective duration constraint for shear force represents the minimum duration for which the shear force must remain in effect to reach the lower shear limit constraint for shell membrane peeling.

[0054] The lower limit shear constraint for shell membrane peeling and the effective duration constraint for shear force were calibrated by progressively changing the flow rate or stirring speed and recording the wall shear force and duration at the first stable loosening of the eggshell membrane. The upper limit shear constraint for membrane integrity was calibrated by progressively increasing the flow rate or stirring speed and recording the wall shear force when the eggshell membrane rupture rate began to exceed the process allowable value. The continuity constraint for shear force direction was calibrated by comparing the tearing of the eggshell membrane edge under different shear force direction variations.

[0055] Experimental measurement data includes particle image velocimetry data and pressure sensing data. Particle image velocimetry data is used to form the measurement velocity field within the separation chamber. Pressure sensing data is collected by pressure sensors located near the inlet boundary, outlet boundary, and shell-membrane contact wall boundary, and is used to form the measurement pressure field.

[0056] Step S2: Establish the fluid dynamics calculation domain based on the structural parameters of the shell-membrane separation cavity and the fluid property parameters.

[0057] The fluid dynamics computational domain includes the fluid interior region, the near-wall region, and the baffle candidate update region. A wall layer mesh is set at the shell-film contact wall boundary to capture velocity changes near the shell-film contact wall boundary. The initial height, growth rate, and number of layers of the wall layer mesh are determined based on the near-wall treatment method used in the fluid dynamics solution, the target dimensionless wall distance, and the mesh independence test results.

[0058] A velocity inlet is set at the inlet boundary, receiving inlet flow velocity, fluid density, fluid dynamic viscosity, and fluid temperature. A pressure outlet is set at the outlet boundary, receiving outlet pressure boundary conditions. A rotational reference region is set at the agitator component boundary, receiving the agitator speed setpoint. The candidate baffle installation boundary is used as an updatable geometric constraint input to the numerical solution process, enabling subsequent co-optimization processes to modify the baffle position and angle.

[0059] When performing a mesh independence check on the fluid dynamics computational domain, the mesh density is increased sequentially in the fluid interior region and the near-wall region. When the pressure loss, spatial distribution statistics of wall shear force, and changes in the effective shear force zone coverage corresponding to two consecutive mesh levels are all below the allowable values ​​for mesh independence, the first-level mesh with lower computational cost is selected as the subsequent solution mesh.

[0060] Step S3: Solve for the velocity field, pressure field, and near-wall velocity gradient field.

[0061] The solution process employs a finite volume discretization method, discretizing the fluid continuity and momentum control equations into a grid control volume within the fluid dynamics computational domain. For operating conditions where liquid density changes are negligible, the fluid is treated as incompressible.

[0062] The governing equations for fluid continuity are:

[0063] ;

[0064] in, Represents the spatial differential operator; It represents the dot product operation, and when combined with the spatial differential operator, it represents the divergence operation; This represents the fluid velocity vector, with units of meters per second.

[0065] The momentum governing equation is:

[0066] ;

[0067] in, This indicates the density of the fluid, expressed in kilograms per cubic meter. Represents partial differential operators; This represents the fluid velocity vector, with units of meters per second. Indicates time, in seconds; It represents the dot product operation, and when combined with the spatial differential operator, it represents the divergence operation; Represents the spatial differential operator; Pressure is expressed in Pascals (Pa). This indicates the dynamic viscosity of a fluid, measured in Pascals per second (Pa). Represents the Laplace operator; This represents the equivalent volumetric force density vector introduced by the stirring component and the rotating reference region, in Newtons per cubic meter.

[0068] The turbulent closed model generates turbulent viscosity corrections based on the local flow state. The appropriate turbulent closed model can be selected based on the Reynolds number, rotating flow intensity, and near-wall solution requirements for rotating shear flow. Pressure-velocity coupled iteration updates the velocity and pressure fields in each iteration step or time step.

[0069] The preset convergence conditions include at least the following: the computational residuals of the fluid continuity control equation and the momentum control equation are lower than the residual threshold; the imbalance between the inlet mass flow rate and the outlet mass flow rate is lower than the mass conservation threshold; and the changes in the spatial distribution statistics of the wall shear force corresponding to multiple consecutive iteration steps are lower than the stability threshold. The residual threshold, mass conservation threshold, and stability threshold are determined based on the solver accuracy requirements and the results of the mesh independence test.

[0070] After the preset convergence conditions are met, the output shows the velocity field, pressure field, and near-wall velocity gradient field within the separation cavity.

[0071] Step S4: Generate the spatial distribution of wall shear force based on the near-zone velocity gradient field of the wall.

[0072] The fluid velocity near the boundary between the membrane and the wall changes along the normal direction of the wall. The wall shear force is determined by both the fluid dynamic viscosity and the near-wall velocity gradient. In a local tangential direction, the wall shear force is:

[0073] ;

[0074] in, This represents the amplitude of the wall shear force at the boundary between the shell and membrane contact wall, in Pascals (Pa). This indicates the dynamic viscosity of a fluid, measured in Pascals per second (Pa). Represents partial differential operators; This represents the velocity component along the tangential direction of the shell-membrane contact wall boundary, in meters per second. The normal coordinates are perpendicular to the boundary of the shell-membrane contact wall, and the unit is meters. Indicates the location of the shell-membrane contact wall boundary; This indicates the operation of taking the absolute value.

[0075] For the three-dimensional shell-membrane contact wall boundary, calculate the wall shear force components in two mutually orthogonal local tangential directions, and then determine the wall shear force amplitude and shear force direction based on the two wall shear force components.

[0076] The shear force calculation unit outputs the shear force amplitude, shear force direction, shear force gradient, and shear force residence time at each mesh node on the shell-film contact wall boundary. For transient solutions, the shear force residence time is obtained by accumulating the continuous time during which the shear force at the corresponding wall location meets the specified amplitude range. For steady-state or quasi-steady-state solutions, the shear force residence time is obtained from the near-wall trace tracing results and the residence time of the fluid in the corresponding wall region.

[0077] Step S5: Construct the effective shear force zone based on the shell-membrane separation effect constraint parameters.

[0078] The effective action zone generation unit extracts shear force amplitude, shear force direction, shear force gradient, and shear force residence time from the spatial distribution of wall shear force. Locations where the shear force amplitude reaches the lower limit of shell-membrane peeling shear constraint but does not exceed the upper limit of membrane integrity shear constraint are marked as candidate effective regions.

[0079] For adjacent grid nodes within the candidate effective region, the angle change between shear force directions is calculated. When the angle change exceeds the shear force direction continuity constraint, the corresponding location is identified as a shear force direction abrupt change location and removed from the candidate effective region. For the remaining locations, the shear force residence time is compared with the shear force effective action duration constraint, and locations with insufficient shear force residence time are removed to obtain the effective shear force action zone.

[0080] The effective shear force coverage zone is the ratio of its area to the total area of ​​the shell-membrane contact wall boundary. The membrane integrity risk value is determined by the ratio of the area corresponding to the grid node whose shear force amplitude exceeds the upper limit of membrane integrity shear constraint to the total area of ​​the shell-membrane contact wall boundary. The shear force distribution variability is determined by the dispersion of the shear force amplitude within the effective shear force coverage zone. The shear force direction continuity index is determined by the change in shear force direction between adjacent grid nodes within the effective shear force coverage zone; the smaller the change in shear force direction, the higher the shear force direction continuity index.

[0081] Step S6: Generate an error correction field based on experimental measurement data, velocity field, pressure field, and near-wall velocity gradient field.

[0082] The particle image velocimetry data is processed through image registration, particle displacement calculation, and outlier vector removal to generate a measured velocity field. Pressure sensor data is used to generate a measured pressure field based on the pressure sensor's installation location and sampling time.

[0083] The error correction unit maps the measured velocity and pressure fields to the mesh nodes of the fluid dynamics computational domain. Mapping methods can include nearest neighbor interpolation, radial basis interpolation, or finite element projection. When selecting a mapping method, the spatial continuity of the mapped measured velocity and pressure fields is checked to avoid incorrectly mapping measurement points located in solid regions to fluid interior regions.

[0084] The error correction unit calculates the local velocity deviation between the velocity field and the measured velocity field, the local pressure deviation between the pressure field and the measured pressure field, and the local velocity gradient deviation between the velocity gradient field near the wall and the corresponding gradient of the measured velocity field.

[0085] Local velocity deviation, local pressure deviation, and local velocity gradient deviation are assigned corresponding spatial coordinates and sampling times. The error correction unit sets spatial weights based on the distance between the measurement point and the boundary grid nodes of the shell-membrane contact wall, and spatially propagates and smooths the local velocity deviation, local pressure deviation, and local velocity gradient deviation along the boundary of the shell-membrane contact wall to form an error correction field.

[0086] Step S7: Update the boundary conditions of the fluid dynamics computational domain using the error correction field, and re-solve to obtain the corrected spatial distribution of shear force.

[0087] The inlet velocity deviation, outlet pressure deviation, and local velocity gradient deviation are extracted from the error correction field. A positive inlet velocity deviation indicates that the measured inlet velocity is higher than the current inlet boundary setpoint; a negative inlet velocity deviation indicates that the measured inlet velocity is lower than the current inlet boundary setpoint. The inlet boundary update amount is determined based on the inlet velocity deviation and the boundary update step size.

[0088] The outlet pressure deviation is used to update the outlet boundary. The positive or negative direction of the outlet pressure deviation is determined by the difference between the measured outlet pressure and the current outlet boundary setting.

[0089] Local velocity gradient bias is used to update the wall layer correction coefficient. The wall layer correction coefficient is a dimensionless coefficient that corrects the simulated near-wall velocity gradient to the corresponding gradient in the measured velocity field. At a correction position, if the gradient corresponding to the measured velocity field is higher than the near-wall velocity gradient field, the wall layer correction coefficient is increased; if the gradient corresponding to the measured velocity field is lower than the near-wall velocity gradient field, the wall layer correction coefficient is decreased. The amount of wall layer correction coefficient updated in a single round is limited by the update step size and is confined between a preset lower limit and a preset upper limit to avoid non-physical abrupt changes in boundary conditions.

[0090] After updating the inlet boundary, outlet boundary, and wall layer correction coefficients, the velocity field, pressure field, and near-wall velocity gradient field are resolved, and the spatial distribution of wall shear force is regenerated to obtain the corrected spatial distribution of shear force.

[0091] Based on the corrected spatial distribution of shear force, the effective shear force zone is reconstructed. The reconstructed effective shear force zone is then spatially registered with the original effective shear force zone to determine changes in the center position, boundary position, and coverage area. Based on these changes, the positional offset of the effective shear force zone is determined, and the candidate installation boundary for the baffle is updated.

[0092] Step S8: Perform synergistic optimization based on the corrected spatial distribution of shear force and the updated effective shear force zone.

[0093] A multi-objective optimization model is established using the collaborative optimization unit. The multi-objective optimization model aims to improve the coverage of the effective shear force action zone, reduce the variability of shear force distribution, and reduce the risk to membrane integrity, while also satisfying energy consumption constraints.

[0094] The energy consumption constraint is determined based on the input power of the agitator actuator and the pumping power required to transport the fluid. The input power of the agitator actuator is obtained from measured values ​​of the actuator's electrical parameters or the motor power curve. The pumping power is obtained based on flow rate parameters, pressure parameters, and pump efficiency. The energy consumption constraint is set based on the allowable energy consumption per unit throughput or per unit operating time of the shell-and-film separator.

[0095] The stirring parameters, flow rate parameters, and pressure parameters are continuous variables. The baffle position and baffle angle are discrete variables. The multi-objective optimization model uses the corrected spatial distribution of shear force and the updated effective shear force zone as constraint inputs.

[0096] The response surface methodology (RSM) is built based on continuous variable experimental design samples. Each continuous variable experimental design sample includes a set of stirring parameters, flow rate parameters, and pressure parameters, as well as the corresponding effective shear force coverage, shear force distribution variability, membrane integrity risk value, and energy consumption constraint value. The RSM determines candidate combinations of continuous variables based on the continuous variable experimental design samples.

[0097] The genetic search model uses baffle position and angle as discrete genes. It establishes an initial population, calculates the fitness of baffle configurations within this population, and generates subsequent populations through selection, crossover, and mutation. Each baffle configuration is paired with candidate combinations of continuous variables output by the response surface methodology, and a corrected shear force spatial distribution or a validated surrogate model is invoked to calculate the corresponding effective shear force action zone coverage, shear force distribution variability, membrane integrity risk value, and energy consumption constraint value.

[0098] The genetic search model outputs discrete variable candidate combinations when the optimal fitness change is below the fitness stability threshold for several consecutive generations, or when the preset maximum number of iterations is reached. The fitness stability threshold, the preset maximum number of iterations, the population size, the crossover probability, and the mutation probability are determined through collaborative optimization convergence experiments.

[0099] Step S9: Establish a mapping model for shell-membrane separation effect.

[0100] The shell-membrane separation effect mapping model receives the effective shear force coverage, shear force distribution variability, shear force direction continuity index, membrane integrity risk value, and corrected shear force spatial distribution, and outputs the predicted separation efficiency and membrane integrity values.

[0101] The shell-membrane separation effect mapping model can be constructed using fully connected neural networks, gradient boosting trees, support vector regression, or Gaussian process regression. When using one of these models, the input data for the training samples and the input data for the model runtime use the same feature definitions and normalization methods.

[0102] Training samples were derived from historical shell-membrane separation experiments, simulation results corrected for experimental measurement data, and operating records of the shell-membrane separation equipment. The separation efficiency label for the training samples was determined based on the change in residual membrane area or residual membrane mass before and after eggshell treatment. The membrane integrity label for the training samples was determined based on the proportion of recovered eggshell membranes meeting the preset intact size requirements to the total area of ​​recovered eggshell membranes.

[0103] The training samples were divided into training, validation, and test sets according to the batches of membrane separation. Data from the same batch were not simultaneously included in different sets. The model training process aimed to reduce the prediction errors for separation efficiency and membrane integrity. Training was terminated when the prediction error corresponding to the validation set no longer decreased after several consecutive rounds, and the model parameters with the lowest prediction error in the validation set were selected.

[0104] Step S10: Update the multi-objective optimization model based on the output of the shell-membrane separation effect mapping model.

[0105] The predicted separation efficiency and membrane integrity values ​​output by the shell-membrane separation performance mapping model are fed back into the multi-objective optimization model. The multi-objective optimization model adjusts the candidate combinations of continuous and discrete variables based on these predicted values.

[0106] The control update threshold is determined based on the control resolution of the agitator actuator, flow control valve, pressure control valve, and baffle actuator, as well as the allowable fluctuation range of the shell-and-film separation process. The preset coverage threshold, preset separation efficiency threshold, and preset integrity threshold are determined based on product quality requirements and the calibration results of the shell-and-film separation process.

[0107] Collaborative optimization terminates when the difference between the control parameters of the membrane separation device output in two consecutive iterations is lower than the control update threshold, the effective shear force coverage reaches the preset coverage threshold, the predicted separation efficiency reaches the preset separation efficiency threshold, and the predicted membrane integrity reaches the preset integrity threshold.

[0108] The control parameters of the shell-and-membrane separation equipment include the stirring speed setting, flow rate setting, pressure setting, baffle position setting, and baffle angle setting.

[0109] Step S11: Send the control parameters of the membrane separation device to the separation device control node.

[0110] The separation equipment control node adjusts the stirring driver according to the stirring speed setting, the flow regulating valve according to the flow rate setting, the pressure regulating valve according to the pressure setting, and the baffle actuator according to the baffle position setting and the baffle angle setting.

[0111] After the eggshell membrane separation equipment is operational, the data acquisition interface continues to collect particle image velocity data and pressure sensor data. The newly acquired experimental measurement data is transmitted back to generate the next round of error correction field, forming a closed loop of prediction, correction, optimization, control, and remeasurement.

[0112] In this embodiment, the near-zone velocity gradient field of the wall is input into the shear force calculation process; the spatial distribution of the wall shear force is input into the shear force effective action zone construction process; the shear force effective action zone is input into the multi-objective optimization model; experimental measurement data is input into the error correction field generation process; the error correction field is input into the fluid dynamics computational domain boundary condition update process; the corrected shear force spatial distribution is again input into the shear force effective action zone construction process; the multi-objective optimization model outputs the control parameters of the shell-membrane separation device; and the control parameters of the shell-membrane separation device are applied to the eggshell-membrane separation device to generate new experimental measurement data.

[0113] Example 2: This example provides a shell-and-membrane separation fluid shear force distribution prediction system based on Example 1. The shell-and-membrane separation fluid shear force distribution prediction system includes a data acquisition interface, a processor, a memory, a control output interface, and a human-machine display interface.

[0114] The memory stores program instructions, a fluid dynamics solution program, an error correction program, a multi-objective optimization program, and a shell-film separation effect mapping program. The processor executes the program instructions to implement the shell-film separation fluid shear force distribution prediction method described in Example 1.

[0115] The data acquisition interface receives structural parameters of the shell-membrane separation chamber, fluid property parameters, shell-membrane separation effect constraint parameters, and experimental measurement data. The data acquisition interface connects to a 3D scanning device, a densitometer, a viscometer, a temperature sensor, a flow rate sensor, a particle image velocimeter, a pressure sensor, and the operating controller of the eggshell-membrane separation device. The data acquisition interface writes the received data to the memory and sends data update signals to the processor.

[0116] The processor includes a computational domain generation unit, a fluid dynamics solution unit, a shear force calculation unit, an effective action zone generation unit, an error correction unit, a collaborative optimization unit, and an effect mapping unit.

[0117] The computational domain generation unit establishes the fluid dynamics computational domain based on the shell-membrane separation chamber structural parameters. The fluid dynamics solution unit solves for the velocity field, pressure field, and near-wall velocity gradient field. The shear force calculation unit generates the spatial distribution of wall shear force based on the near-wall velocity gradient field. The effective action zone generation unit generates the effective shear force action zone based on the shell-membrane separation effect constraint parameters. The error correction unit generates an error correction field based on experimental measurement data, the velocity field, pressure field, and near-wall velocity gradient field. The collaborative optimization unit outputs the shell-membrane separation equipment control parameters based on the corrected shear force spatial distribution and the updated effective shear force action zone. The effect mapping unit outputs predicted separation efficiency and membrane integrity values ​​and feeds these values ​​back to the collaborative optimization unit.

[0118] The control output interface connects to the agitator driver, flow control valve, pressure control valve, and baffle actuator. The control output interface converts the control parameters of the membrane separation equipment output by the processor into corresponding equipment control commands.

[0119] The human-machine interface displays velocity field cloud map, pressure field cloud map, spatial distribution map of wall shear force, effective shear force zone map, error correction field map, predicted separation efficiency, predicted membrane integrity, and control parameters of the shell-membrane separation equipment.

[0120] When the shell-membrane separation fluid shear force distribution prediction system is running, the data acquisition interface first receives the structural parameters of the shell-membrane separation chamber and the fluid property parameters. The processor establishes the fluid dynamics calculation domain and performs fluid dynamics solutions. The shear force calculation unit generates the spatial distribution of wall shear force. The effective action zone generation unit converts the spatial distribution of wall shear force into the effective action zone of shear force. The error correction unit uses experimental measurement data to generate an error correction field and updates the boundary conditions of the fluid dynamics calculation domain. The collaborative optimization unit generates the control parameters of the shell-membrane separation equipment based on the corrected spatial distribution of shear force and the updated effective action zone of shear force. The control output interface outputs the control parameters of the shell-membrane separation equipment to the stirring actuator, flow regulating valve, pressure regulating valve, and baffle actuator. New experimental measurement data generated after the eggshell membrane separation equipment is running is returned to the shell-membrane separation fluid shear force distribution prediction system through the data acquisition interface.

[0121] The shell-membrane separation fluid shear force distribution prediction system can be deployed in the control cabinet, edge computing node, or factory server of an eggshell-membrane separation production line. When deployed using an edge computing node, the fluid dynamics proxy model and control parameter update program are deployed close to the eggshell-membrane separation equipment. When deployed using a factory server, multiple eggshell-membrane separation devices share computing resources and separately store the shell-membrane separation chamber structural parameters, fluid property parameters, experimental measurement data, error correction field, and control parameters of each device.

[0122] Example 3: This example further illustrates the computational relationship between the effective shear force zone, the error correction field, and the collaborative optimization model. In this example, the eggshell membrane separation device includes a cylindrical separation chamber. The inlet is located at one end of the cylindrical separation chamber, and the outlet is located at the other end. A stirring component and an adjustable baffle are installed inside the cylindrical separation chamber.

[0123] After receiving the structural parameters of the shell-membrane separation cavity from the data acquisition interface, the computational domain generation unit sets the wall layer mesh on the shell-membrane contact wall boundary and marks the candidate installation boundary of the baffle as an updatable geometric boundary.

[0124] After receiving the fluid property parameters, the processor solves for the velocity field, pressure field, and near-wall velocity gradient field using the fluid dynamics solution unit. The tangential velocity at a certain mesh node at the boundary of the shell-film contact wall changes along the wall normal distance. The shear force calculation unit calculates the wall shear force amplitude at the corresponding location based on the fluid dynamic viscosity and the wall normal gradient of the tangential velocity.

[0125] When the shear force direction change at the boundary grid node of the adjacent shell-membrane contact wall satisfies the shear force direction continuity constraint, the shear force amplitude is between the lower limit shear constraint of shell-membrane peeling and the upper limit shear constraint of membrane integrity, and the shear force residence time satisfies the shear force effective action duration constraint, the effective action zone generation unit will assign the corresponding grid node to the effective action zone of shear force.

[0126] If the shear force amplitude in a certain region meets the amplitude requirement but the shear force direction changes abruptly, the effective action zone generation unit will remove the corresponding region from the candidate effective regions. If both the shear force amplitude and shear force direction in a certain region meet the requirements, but the shear force residence time is less than the effective action duration constraint, the effective action zone generation unit will also remove the corresponding region from the candidate effective regions.

[0127] The particle image velocimetry device acquires a local velocity field, and the pressure sensor acquires a local pressure field. The error correction unit maps the measured velocity and pressure fields to the mesh nodes of the fluid dynamics computational domain.

[0128] When the measured velocity near the inlet is lower than the velocity at the corresponding location in the velocity field, the error correction unit generates a negative inlet velocity deviation. When the measured velocity gradient near the boundary of the shell-film contact wall is lower than the velocity gradient at the corresponding location in the near-zone velocity gradient field of the wall, the error correction unit generates a negative local velocity gradient deviation.

[0129] The inlet velocity deviation is used to update the inlet boundary. The local velocity gradient deviation is used to update the wall layer correction factor. After the fluid dynamics solution element is resolved, the corrected shear force spatial distribution is generated.

[0130] The collaborative optimization unit first calls the response surface model to search for combinations of stirring parameters, flow rate parameters, and pressure parameters. When a set of candidate combinations of continuous variables increases the coverage of the effective shear force zone but also increases the risk value of membrane integrity, the collaborative optimization unit inputs the corresponding candidate combinations of continuous variables and the corresponding evaluation values ​​into the genetic search model, and reduces the local high-shear region by changing the baffle position and baffle angle.

[0131] The genetic search model generates multiple candidate combinations of discrete variables. Each candidate combination invokes the calculation process for the effective shear force zone. When the baffle configuration transforms a locally high-shear region into an effective region with continuous shear force direction and shear force amplitude within the constraints, the fitness increases. Conversely, when the baffle configuration leads to an expansion of the low-velocity region and a decrease in the coverage of the effective shear force zone, the fitness decreases.

[0132] The shell-membrane separation effect mapping model receives the effective shear force action zone coverage, shear force distribution variability, shear force direction continuity index, membrane integrity risk value, and corrected shear force spatial distribution, and outputs the separation efficiency prediction value and membrane integrity prediction value.

[0133] When the predicted separation efficiency increases but the predicted membrane integrity decreases, the co-optimization unit reduces pressure parameters, reduces stirring parameters, or adjusts the baffle angle, and recalculates the effective shear force coverage, shear force distribution variability, and membrane integrity risk value.

[0134] When the predicted separation efficiency reaches the preset separation efficiency threshold, the predicted membrane integrity reaches the preset integrity threshold, the effective shear force coverage reaches the preset coverage threshold, and the difference between the control parameters of the shell-membrane separation equipment in two consecutive tests is less than the control update threshold, the collaborative optimization unit terminates the search and outputs the control parameters of the shell-membrane separation equipment.

[0135] The control parameters of the shell-membrane separation equipment are sent to the stirring driver, flow regulating valve, pressure regulating valve, and baffle actuator via the control output interface. After the eggshell-membrane separation equipment is running, the data acquisition interface continues to receive particle image velocity measurement data and pressure sensing data. The newly acquired experimental measurement data is used to generate the next round of error correction field.

[0136] In this way, the shell-membrane separation fluid shear force distribution prediction system can update the prediction results when the batch of eggshells changes, the fluid dynamic viscosity changes, or material accumulation occurs locally in the separation chamber, thereby reducing the fluctuation of shell-membrane separation effect caused by long-term operation of fixed empirical parameters.

[0137] Example 4: Based on Examples 1 to 3, this example provides a method for predicting the shear force distribution of the shell-membrane separation fluid by introducing a membrane local loosening fatigue accumulation function, a dynamic grid displacement field, and acoustic-structure interaction deviation, targeting the dynamic interaction process at the interface between the fluid and the eggshell membrane.

[0138] The fluid dynamics solution unit employs a transient solution method capable of analyzing unsteady turbulent vortex structures.

[0139] In one implementation, the fluid dynamics solution element employs separated eddy simulation, using Reynolds-averaged solution near the shell-film contact wall boundary, and large eddy simulation solution in the fluid interior region far from the shell-film contact wall boundary where the mesh size meets the requirements.

[0140] The processor obtains the unsteady turbulent vortex structure generated by the rotation of the stirring component and the wake of the baffle through time-step iteration, and extracts transient pulsation components from the spatial distribution of wall shear force corresponding to the unsteady turbulent vortex structure. The transient pulsation components are used to characterize the periodic or non-periodic changes of wall shear force over time.

[0141] The shear force calculation unit receives transient pulsating components and constructs a membrane local loosening fatigue accumulation function. The membrane local loosening fatigue accumulation function is:

[0142] ;

[0143] in, This represents the cumulative fatigue index of localized membrane loosening, expressed in seconds. Indicates time, in seconds; This represents the integration time variable, in seconds. This represents the amplitude of the wall shear force at the boundary between the shell and membrane contact wall, in Pascals (Pa). This represents the initial threshold for membrane loosening, expressed in Pascals (Pa). This represents a positive reference shear force parameter, in Pascals (Pa). This represents the fatigue index, which is dimensionless. This indicates the positive part operator. The difference inside the parentheses is taken when it is greater than zero, otherwise it is taken as zero. Indicates time integration operation; This represents the integral time element, with the unit being seconds.

[0144] The positive part operator ensures that fatigue accumulation does not occur when the amplitude of the wall shear force is lower than the initial threshold for shell loosening, thus avoiding non-integer exponentiation when the difference between the amplitude of the wall shear force and the initial threshold for shell loosening is negative.

[0145] The reference shear force parameter is determined based on the baseline value of the wall shear force acting as a stabilizing agent in the shell membrane peeling bench test. The fatigue index is calibrated based on cyclic shear tests or biaxial tensile fatigue tests of the eggshell membrane material. During calibration, the time required for the first local bulge and stable peeling of the eggshell membrane is recorded under multiple known transient wall shear force conditions, and the fatigue index that minimizes the deviation between the predicted result of the membrane local loosening fatigue accumulation function and the experimental result is selected.

[0146] The computational domain generation unit dynamically divides the boundary of the shell-membrane contact wall based on the output value of the membrane local loosening fatigue accumulation function. The first fatigue threshold corresponds to the output value of the membrane local loosening fatigue accumulation function when a measurable local bulge first appears on the eggshell membrane. The second fatigue threshold corresponds to the output value of the membrane local loosening fatigue accumulation function when the eggshell membrane and the inner shell wall form a stable separation. The first and second fatigue thresholds are calibrated through shell-membrane peeling bench tests, and the second fatigue threshold is greater than the first fatigue threshold.

[0147] When the output value of the membrane local loosening fatigue accumulation function is lower than the first fatigue threshold, the corresponding shell-membrane contact wall boundary mesh node is marked as an unstripped boundary region. When the output value of the membrane local loosening fatigue accumulation function reaches the first fatigue threshold but is lower than the second fatigue threshold, the corresponding shell-membrane contact wall boundary mesh node is marked as a transitionally stripped boundary region. When the output value of the membrane local loosening fatigue accumulation function reaches the second fatigue threshold, the corresponding shell-membrane contact wall boundary mesh node is marked as a completely stripped boundary region.

[0148] The unpeeled boundary region uses wall boundary conditions corresponding to the state where the eggshell membrane is fully attached to the inner wall of the eggshell. The transitionally peeled boundary region generates a dynamic mesh displacement field based on the output value of the fatigue accumulation function for local membrane loosening. The fully peeled boundary region uses wall boundary conditions corresponding to the state where the fluid is in direct contact with the inner wall of the eggshell.

[0149] The displacement rate of the transitional peeling boundary region is determined based on the output value of the membrane local loosening fatigue accumulation function. During calibration, the local bulge height or velocity of the eggshell membrane is mapped to the synchronously calculated output value of the membrane local loosening fatigue accumulation function, forming a displacement rate mapping table. During operation, the displacement rate of the transitional peeling boundary region is determined by interpolating the current output value of the membrane local loosening fatigue accumulation function into the displacement rate mapping table.

[0150] The dynamic mesh displacement field is constructed using the arbitrary Lagrange-Eulerian method. The mesh nodes in the transition peeling boundary region move along the normal direction of the shell membrane contact wall boundary according to the displacement change rate, to simulate the wall morphology changes caused by local bulging of the eggshell membrane.

[0151] The dynamic mesh displacement field is obtained by solving the following Laplace smoothing equation:

[0152] ;

[0153] in, Represents the spatial differential operator; It represents the dot product operation, and when combined with the spatial differential operator, it represents the divergence operation; Represents the dynamic grid boundary diffusion coefficient, which is dimensionless; This represents the dynamic mesh displacement vector, in meters.

[0154] The dynamic mesh boundary diffusion coefficient is set based on the distance between the mesh nodes and the transition stripping boundary region. A larger dynamic mesh boundary diffusion coefficient is used near the transition stripping boundary region, so that the mesh near the moving boundary moves in coordination with the boundary; a smaller dynamic mesh boundary diffusion coefficient is used far away from the transition stripping boundary region, so that the boundary displacement gradually decays into the fluid interior region.

[0155] After each time step completes the dynamic mesh displacement field update, the volume, torsion, and minimum angle of the mesh elements are checked. If the mesh quality does not meet the preset mesh quality conditions, the local hydrodynamic computational domain is re-meshed, and the velocity field, pressure field, and turbulent state variables before re-meshing are conserved and mapped to the re-meshed mesh.

[0156] After the three-dimensional morphology of the shell-membrane contact wall boundary is updated, the location of local bulges alters the near-wall flow state, forming new local backflow or wake regions. In the next time step, the fluid dynamics solution unit resolves the velocity field, pressure field, and near-wall velocity gradient field, and regenerates the spatial distribution of wall shear force. This regenerated spatial distribution of wall shear force is then input into the membrane local loosening fatigue accumulation function, forming a bidirectional coupling process between fluid shear force, shell membrane loosening state, dynamic boundary morphology, and wall shear force redistribution.

[0157] During the generation of the error correction field, the data acquisition interface also receives high-frequency acoustic emission signals from high-frequency acoustic emission sensors mounted on the wall of the separation cavity. The number and installation location of the high-frequency acoustic emission sensors are determined based on the boundary area of ​​the diaphragm contact wall, the sound source localization accuracy, and the structure of the separation cavity.

[0158] The error correction unit performs bandpass filtering and spectral transformation on the high-frequency acoustic emission signal to extract the energy value within the preset stripping main frequency band. The preset stripping main frequency band is determined through eggshell membrane stripping bench tests.

[0159] In one implementation, background signals are first acquired under an unloaded state without shell membrane peeling. Then, high-frequency acoustic emission signals are acquired under conditions of localized shell membrane bulging and fiber breakage. The frequency range with stable energy increments compared to the background signal and capable of characterizing the shell membrane peeling event is selected as the preset peeling main frequency band. In specific equipment, the preset peeling main frequency band can be located in the range of 100 kHz to 300 kHz, and the actual frequency range is determined based on the response characteristics of the high-frequency acoustic emission sensor and bench test results.

[0160] The error correction unit uses the time difference of arrival between multiple high-frequency acoustic emission sensors to locate the sound source and maps the energy value space within the preset stripping main frequency band to the transition stripping boundary region.

[0161] The energy value of the high-frequency acoustic emission signal characterizes the relative intensity of the actual eggshell membrane fiber fracture event. The displacement change rate of the dynamic mesh displacement field characterizes the simulated predicted local loosening rate of the eggshell membrane. The error correction unit calculates the acoustic-structure coupling deviation between the two.

[0162] ;

[0163] in, This represents acoustic-structure coupling deviation, and is dimensionless. This represents a scaling factor, which is dimensionless. This represents the energy value obtained by integrating the square within a preset stripped main frequency band of a high-frequency acoustic emission signal, expressed in volts square seconds. This represents the operation to find the maximum value; This indicates the spatial extent of the transition stripping boundary region currently undergoing correction; Indicates the current error correction time window; Represents vector modulo operation; This represents the vector of the rate of change of dynamic grid displacement, in meters per second. This indicates the operation of taking the absolute value.

[0164] When the maximum energy value or the maximum displacement rate of change is zero, the corresponding normalization term is treated as zero to avoid the normalization denominator being zero.

[0165] The error correction unit combines acoustic-structure interaction deviation, local velocity deviation, local pressure deviation, and local velocity gradient deviation into deviation data with spatial location attributes. The deviation data is spatially propagated and smoothed along the shell-membrane contact wall boundary using a Gaussian weighted kernel to synthesize an error correction field covering the shell-membrane contact wall boundary.

[0166] The acoustic-structure interaction (ASI) deviation is used to correct the mapping relationship between the fatigue index and displacement rate of the fatigue accumulation function for local membrane loosening. The correction process employs a bounded parameter search method, generating candidate parameters within pre-defined allowable intervals for both the fatigue index and displacement rate, and calculating the corresponding ASI deviations for each candidate parameter. Candidate parameters that reduce the ASI deviation within the current error correction time window are selected as update parameters, avoiding directly pre-setting the direction of increase or decrease in the fatigue index based on the range of the normalized amplitude of the wall shear force.

[0167] The corrected spatial distribution of shear force, incorporating dynamic boundary information and acoustic-structure interaction deviation information, is input into the effective action zone generation unit. This unit updates the effective action zone of shear force based on the corrected spatial distribution of shear force, shell-membrane separation effect constraint parameters, and boundary region division results.

[0168] The collaborative optimization unit receives the corrected spatial distribution of shear force, the updated effective shear force zone, the coverage of the effective shear force zone, the variability of the shear force distribution, and the membrane integrity risk value. The genetic search model searches within the discrete variable space of the baffle position and baffle angle.

[0169] As a preferred approach, the fitness calculation also penalizes candidate baffle configurations where the pressure pulsation frequency of the baffle wake falls into the preset peeling frequency band and the membrane integrity risk value increases, in order to reduce eggshell membrane rupture caused by periodic high shear forces.

[0170] Under the same fluid properties and separation chamber structural parameters, the boundary configurations of the static rigid shell membrane contact wall and the dynamic deformable shell membrane contact wall were compared. The comparative experiment included 12 batches of eggshell treatment, with each batch processing 200 kg of raw eggshell fragments.

[0171] The actual residual membrane rate was determined using image analysis. Images of the inner wall of eggshell fragments after shell membrane separation were acquired, the residual shell membrane area was segmented, and the actual residual membrane rate was determined based on the proportion of the residual shell membrane area to the total area of ​​the eggshell inner wall.

[0172] The predicted residual membrane rate is obtained by converting the predicted separation efficiency value output by the shell-membrane separation effect mapping model, that is, by subtracting the percentage corresponding to the predicted separation efficiency value from 100%.

[0173] Eggshell membrane integrity was analyzed using images of recovered eggshell membranes. The proportion of membrane area meeting the preset integrity requirements to the total area of ​​recovered eggshell membranes was used as the eggshell membrane integrity. The membrane integrity improvement rate was the relative increase in eggshell membrane integrity corresponding to the dynamically deformed shell membrane contact wall boundary configuration compared to the eggshell membrane integrity corresponding to the static rigid shell membrane contact wall boundary configuration.

[0174] Table 1 shows the computational domain of the hydrodynamics and the error correction strategy for the two prediction configurations.

[0175] Table 1. Computational domain of the predicted configuration and error correction strategy;

[0176] Configuration A Static rigid shell membrane contact wall boundary The fatigue accumulation function for local membrane loosening is not used. Particle image velocimetry data and pressure sensing data Local velocity deviation, local pressure deviation, and local velocity gradient deviation Configuration B Dynamically deformable shell-film contact wall boundary Employing the membrane local loosening fatigue accumulation function Particle image velocimetry data, pressure sensing data, and high-frequency acoustic emission signals Local velocity deviation, local pressure deviation, local velocity gradient deviation, and acoustic-structure interaction deviation

[0177] Each row in Table 2 represents the average result for the corresponding three batches of eggshell treatment.

[0178] Table 2. Predicted residual film rate and actual equipment operation results;

[0179] Batch 01 to Batch 03 4.50% 8.21% 3.10% 3.25% 14.2% Batch 04 to Batch 06 4.25% 7.95% 2.85% 3.01% 16.5% Batch 07 to Batch 09 4.60% 9.10% 3.05% 2.90% 13.8% Batch 10 to Batch 12 4.80% 8.85% 3.15% 3.22% 15.1%

[0180] The predicted residual film rate of configuration A is lower than the corresponding actual residual film rate, indicating that the static rigid shell membrane contact wall boundary is difficult to reflect the changes in flow state caused by local loosening of the eggshell membrane and changes in boundary morphology.

[0181] The absolute differences between the predicted and actual residual film rates for configuration B were 0.15, 0.16, 0.15, and 0.07 percentage points, respectively, all not exceeding 0.2 percentage points. The average improvement rate of membrane integrity for the four groups was 14.9%.

[0182] The comparative results show that after introducing the membrane local loosening fatigue accumulation function, dynamic grid displacement field and acoustic-structure interaction deviation, the corrected shear force spatial distribution can more closely approximate the actual peeling process of the eggshell membrane, and can provide constraint inputs containing boundary dynamic change information for the multi-objective optimization model.

[0183] This invention transforms the spatial distribution of wall shear force into a spatial constraint for the shell-membrane separation process by using the effective shear force action zone. It feeds back experimental measurement data to the boundary conditions of the fluid dynamics computational domain through an error correction field, and forms a collaborative optimization closed loop through a response surface model, a genetic search model, and a shell-membrane separation effect mapping model.

[0184] In a further embodiment, the present invention reflects the interaction process between fluid shear force and the dynamic peeling state of eggshell membrane through the membrane local loosening fatigue accumulation function, dynamic grid displacement field and acoustic-structure interaction deviation, so that the predicted results of the fluid shear force distribution in shell-membrane separation can be used for the generation and closed-loop update of control parameters of shell-membrane separation equipment.

[0185] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0186] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the shear force distribution of fluids in membrane separation, characterized in that, include: Obtain structural parameters, fluid properties, shell-membrane separation effect constraint parameters, and experimental measurement data of the shell-membrane separation chamber; A fluid dynamics computational domain was established based on the structural parameters and fluid properties of the shell-membrane separation cavity, and the velocity field, pressure field, and near-wall velocity gradient field were obtained by solving. The spatial distribution of wall shear force is generated based on the near-field velocity gradient field of the wall. Based on the constraint parameters of shell-membrane separation effect, an effective shear force zone is constructed from the spatial distribution of wall shear force; An error correction field is generated based on experimental measurement data, velocity field, pressure field, and near-wall velocity gradient field. The boundary conditions of the fluid dynamics computational domain are updated using the error correction field, and the corrected spatial distribution of shear force is obtained by resolving the problem. The effective action zone of shear force is then updated based on the corrected spatial distribution of shear force. Based on the corrected spatial distribution of shear force and the updated effective shear force zone, the stirring parameters, flow parameters, pressure parameters and baffle configuration are synergistically optimized to output the control parameters of the shell-and-film separator.

2. The method for predicting the shear force distribution of membrane separation fluid according to claim 1, characterized in that, The structural parameters of the shell-membrane separation chamber include the geometric boundary of the separation chamber, the inlet boundary, the outlet boundary, the stirring component boundary, the candidate baffle installation boundary, and the shell-membrane contact wall boundary; the fluid properties include density, dynamic viscosity, temperature, and inlet flow velocity; the shell-membrane separation effect constraint parameters include the lower limit shear constraint for shell-membrane peeling, the upper limit shear constraint for membrane integrity, the continuity constraint for shear force direction, and the constraint for the effective duration of shear force action.

3. The method for predicting the shear force distribution of membrane separation fluid according to claim 1, characterized in that, The establishment of the fluid dynamics computational domain includes: setting a wall layer mesh on the shell-film contact wall boundary; setting a velocity inlet on the inlet boundary; setting a pressure outlet on the outlet boundary; setting a rotational reference region on the stirring component boundary; and using the candidate installation boundary of the baffle as an updatable geometric constraint input for the numerical solution process.

4. The method for predicting the shear force distribution of membrane separation fluid according to claim 1, characterized in that, The solution to obtain the velocity field, pressure field, and near-wall velocity gradient field includes: solving the fluid continuity control equation and momentum control equation using a finite volume discretization method; generating turbulent viscosity correction through a turbulent closed model; updating the velocity field and pressure field through pressure-velocity coupling iteration; and outputting the velocity field, pressure field, and near-wall velocity gradient field within the separation chamber after the calculated residuals and mass flow imbalance satisfy the preset convergence conditions.

5. The method for predicting the shear force distribution of membrane separation fluid according to claim 1, characterized in that, The construction of the effective shear force zone includes: extracting the shear force amplitude, shear force direction, shear force gradient, and shear force residence time from the spatial distribution of wall shear force; marking the positions where the shear force amplitude satisfies the lower limit shear constraint of shell-membrane peeling and does not exceed the upper limit shear constraint of membrane integrity as candidate effective regions; eliminating positions where the shear force direction changes abruptly based on the continuity constraint of shear force direction, and eliminating positions where the shear force residence time is insufficient based on the effective shear force duration constraint, thus obtaining the effective shear force zone.

6. The method for predicting the shear force distribution of membrane separation fluid according to claim 1, characterized in that, The generation of the error correction field includes: generating a measurement velocity field using particle image velocimetry data and a measurement pressure field using pressure sensing data; mapping the measurement velocity field and the measurement pressure field to the grid nodes of the fluid dynamics computational domain; local velocity deviation between the computational velocity field and the measurement velocity field, local pressure deviation between the pressure field and the measurement pressure field, and local velocity gradient deviation between the near-wall velocity gradient field and the corresponding gradient of the measurement velocity field; and spatially propagating and smoothing the local velocity deviation, local pressure deviation, and local velocity gradient deviation along the boundary of the shell-film contact wall to form the error correction field.

7. The method for predicting the shear force distribution of membrane separation fluid according to claim 1, characterized in that, The process of updating the boundary conditions of the fluid dynamics computational domain using the error correction field includes: extracting the inlet velocity deviation, outlet pressure deviation, and local velocity gradient deviation from the error correction field; updating the inlet boundary based on the inlet velocity deviation; updating the outlet boundary based on the outlet pressure deviation; updating the wall layer correction coefficient based on the local velocity gradient deviation; resolving the corrected shear force spatial distribution; reconstructing the effective shear force zone based on the corrected shear force spatial distribution; and updating the baffle candidate installation boundary based on the positional offset and coverage change between the reconstructed effective shear force zone and the original effective shear force zone.

8. The method for predicting the shear force distribution of membrane separation fluid according to claim 1, characterized in that, The collaborative optimization of stirring parameters, flow rate parameters, pressure parameters, and baffle configuration includes: establishing a multi-objective optimization model with shear force effective action zone coverage, shear force distribution variability, membrane integrity risk value, and energy consumption constraint value as evaluation metrics, and using the corrected shear force spatial distribution and the updated shear force effective action zone as constraint inputs to the multi-objective optimization model.

9. The method for predicting the shear force distribution of membrane separation fluid according to claim 1, characterized in that, After generating the spatial distribution of wall shear force based on the near-zone velocity gradient field of the wall and before constructing the effective shear force zone, the method further includes: extracting the transient pulsating components of the spatial distribution of wall shear force on the time scale, and constructing a membrane local loosening fatigue accumulation function; dynamically dividing the shell-membrane contact wall boundary into an unstripped boundary region, a transitionally stripped boundary region, and a completely stripped boundary region based on the output value of the membrane local loosening fatigue accumulation function; generating a dynamic mesh displacement field for the transitionally stripped boundary region, and updating the three-dimensional morphological features of the shell-membrane contact wall boundary based on the dynamic mesh displacement field during the time step iteration in the hydrodynamic computation domain; the membrane local loosening fatigue accumulation function is: ； in, This represents the cumulative fatigue index of localized membrane loosening, expressed in seconds. Indicates time, in seconds; This represents the integration time variable, in seconds. This represents the amplitude of the wall shear force at the boundary between the shell and membrane contact wall, in Pascals (Pa). This represents the initial threshold for membrane loosening, expressed in Pascals (Pa). This represents a positive reference shear force parameter, in Pascals (Pa). This represents the fatigue index, which is dimensionless. This indicates the positive part operator. The difference inside the parentheses is taken when it is greater than zero, otherwise it is taken as zero. Indicates time integration operation; This represents the integral time element, with the unit being seconds.

10. A shell-membrane separation fluid shear force distribution prediction system, used in eggshell membrane separation equipment, characterized in that, It includes a processor, a memory, a data acquisition interface, and a control output interface; the memory stores program instructions that can be executed by the processor, and when the program instructions are executed by the processor, they implement the method for predicting the shear force distribution of the membrane separation fluid as described in any one of claims 1 to 9; The data acquisition interface is used to receive structural parameters of the shell-membrane separation chamber, fluid property parameters, shell-membrane separation effect constraint parameters, and experimental measurement data, and outputs the received data to the processor. The processor is used to generate velocity field, pressure field, near-wall velocity gradient field, spatial distribution of wall shear force, effective shear force zone, error correction field, corrected spatial distribution of shear force, and control parameters of the shell-membrane separation equipment. The control output interface is used to output the control parameters of the shell-membrane separation equipment to the stirring actuator, flow regulating valve, pressure regulating valve, and baffle actuator.

Images