A dynamic modeling system for packaging electrostatic field digital twinning

Abstract

This invention relates to the field of dynamic modeling technology and discloses a dynamic modeling system for digital twin of packaging electrostatic field, comprising: acquiring the feedback signal of the drive frequency actuator, the displacement data of the controlled target object, and the reference field strength signal; extracting the high-frequency ripple component of the current signal to determine the dynamic friction coupling factor, characterizing the micro-fluctuation of contact stress; mapping the conductive transmission component as a virtual topological zero-potential anchor point, and switching the leakage resistance parameter when the controlled target object is displaced to the corresponding coordinate; using the dynamic friction coupling factor to compensate for the charge injection gain, and solving the charge distribution state of the controlled target object through a two-layer evolution architecture; thereby generating a digital twin model of the distributed physical quantity field and correcting the deviation. This invention establishes the correlation logic between drive dynamics and charge evolution, offsets the fluctuation of charging gain caused by mechanical dynamic drift, and improves the modeling accuracy.

Description

Technical Field

[0001] This invention relates to a dynamic modeling system for digital twin of packaging electrostatic field, belonging to the field of dynamic modeling technology. Background Technology

[0002] In current high-speed conveying and multi-stage roller packaging processes for flexible films, electrostatic accumulation not only induces material adsorption shifts but also poses a risk of damage to sensitive electronic components. To monitor the electrostatic state on the production floor, the industry typically employs electrostatic field digital twin modeling, using real-time potential data from a limited number of observation points to extrapolate the overall field strength distribution on the material surface. Current mainstream technologies establish quasi-static lookup table logic between point measurements and preset thresholds, or solve physical field equations through numerical calculations to provide control support for subsequent actuators. As packaging production lines evolve towards ultra-high speed and high integration, the increased material movement speed leads to charge convection effects. Existing solutions treat the electric field evolution logic and mechanical displacement process as independent calculation flows, ignoring displacement increments under high-speed conditions. To compensate for modeling distortion, methods such as increasing the number of sensors or improving processor performance are typically employed, resulting in increased deployment costs and an inability to handle errors caused by mechanical state drift.

[0003] To address the aforementioned challenges, simply increasing the deployment density of non-contact potentiometers not only causes coupling interference between sensing signals but also fails to eliminate prediction lag caused by material displacement at the mechanistic level. Furthermore, using complex numerical simulation software for real-time solutions is limited by the computational resources of industrial processors, resulting in prediction delays. Existing modeling methods struggle to reconcile the real-time consistency between mechanical dynamics and electrostatic evolution logic under limited computing power. Bottlenecks exist in hardware sensing and computing power allocation. Optimizing the real-time performance of field strength reconstruction at the algorithmic logic level is insufficient, as modeling methods often fail to capture the dynamic essence of the physical process. For example, publication number CN11388... Chinese invention patent 7059A discloses a meshless numerical calculation method for electrostatic field based on digital twin electrostatic field MLPG. By optimizing the shape of the trial function of the meshless algorithm to adapt to different influence domains, it improves the numerical accuracy of complex fields. However, such solutions are limited to pure numerical solution optimization and do not understand the coupling mechanism between the dynamic characteristics of the drive system and the charging gain. Under high-speed packaging conditions, roller runout and bearing wear cause contact stress fluctuations, resulting in random drift of the charging gain. The simple numerical algorithm lacks real-time extraction and online compensation of mechanical state characteristics, and cannot solve the contradiction between the computational cost and prediction accuracy under high dynamic conditions, causing the model to deviate during the material speed change stage.

[0004] Therefore, how to achieve real-time alignment between the displacement signal of the packaging material and the charge evolution logic, and how to perform online correction of the model parameters according to the dynamic feedback mechanism, is the technical problem to be solved by this invention. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A dynamic modeling system for digital twin of packaging electrostatic field, the system comprising: The signal acquisition unit is used to acquire the stator current feedback signal of the drive frequency actuator, the real-time displacement data of the controlled target, and the reference field strength signal at key locations in the process transmission path; The dynamic feedback identification module is used to separate high-frequency ripple components with frequencies higher than 500Hz from the stator current feedback signal through a digital high-pass filter, calculate the energy density of the high-frequency ripple components within a preset sampling period, and determine the dynamic friction coupling factor that characterizes the micro-fluctuation of contact stress between the controlled target and the transmission drive components based on the mapping relationship between the energy density and the preset friction torque fluctuation model. The physical boundary logic mapping module is used to map the conductive transmission components in the process transmission path to zero potential anchor points in the virtual logic topology. Combined with real-time displacement data, it simulates the contact charge transfer process by transiently rewriting the leakage resistance parameters of the logic cells when the logic cell corresponding to the controlled target moves to the coordinates of the zero potential anchor point. A two-layer charge evolution calculator is used to dynamically compensate the charge injection gain term of the logic cell by using the dynamic friction coupling factor as a correction term, and to construct a two-layer evolution architecture containing a surface moving charge operator and an inner bound charge operator to solve the charge distribution state of the controlled target during the dynamic displacement process. The twin field strength output module is used to generate a digital twin model of the physical quantity field of the controlled target object based on the charge distribution state, and to perform deviation closed-loop correction on the digital twin model based on the reference field strength signal.

[0006] Preferably, when determining the dynamic friction coupling factor, the dynamic feedback identification module is configured to execute the following processing rules: using a preset digital high-pass filter to remove the low-frequency fundamental wave from the stator current feedback signal, and extracting the high-frequency ripple component with a frequency higher than 500Hz; calculating the energy density component of the high-frequency ripple component within the current sampling period. Based on energy density components Based on the preset friction torque fluctuation model, a dynamic friction coupling factor α is determined to reflect the microscopic fluctuation of contact stress between the controlled target and the transmission component; wherein, the dynamic friction coupling factor α is used to nonlinearly offset the charge injection gain term of the logic cell when it contacts the transmission component.

[0007] Preferably, the physical boundary logic mapping module performs the following steps when simulating the contact charge transfer process: Step S1, establish a virtual coordinate system consistent with the mechanical layout of the process transmission path, and calibrate the geometric center coordinates and effective contact length of the transmission component in the virtual coordinate system; Step S2, establish the logic coverage range of the zero potential anchor point around the geometric center coordinates based on the effective contact length; Step S3, monitor the movement trajectory of the logic cell carrying the leakage resistance parameter in real time. If the movement trajectory enters the logic coverage range, a logic instruction is triggered to switch the leakage resistance parameter from a preset high resistance state value to a preset low resistance state value to derive the charge amount in the logic cell.

[0008] Preferably, the dual-layer charge evolution solver includes: a surface evolution module for executing a surface moving charge operator to calculate the flowing charge density generated on the surface of the controlled target object due to the influence of the dynamic friction coupling factor; a polarization memory module for executing an inner bound charge operator, establishing a virtual polarization layer operator, and parsing the pumping logic of bound charges inside the controlled target object to the surface; and a state coupling unit for superimposing the calculation results of the surface moving charge operator and the inner bound charge operator to determine the charge distribution state.

[0009] Preferably, the system further includes a nonlinear discharge correction module, which is used to: monitor the local field strength generated by the charge distribution state in real time, and determine whether the local field strength exceeds the preset air breakdown threshold; if it exceeds the air breakdown threshold, the nonlinear shunting logic triggered by the state variable is activated to simulate the charge loss process caused by local ionization.

[0010] Preferably, when the nonlinear discharge correction module executes the nonlinear shunt logic, it follows the formula below: ,in, The nonlinear charge loss is simulated, and k is a constant factor characterizing the ionization intensity. These are the calculated values ​​for the local field strength. The preset air breakdown threshold is used; the dual-layer charge calculation unit performs real-time subtraction of the charge distribution state based on the nonlinear charge loss.

[0011] Preferably, the signal acquisition unit includes: a current monitoring branch for acquiring stator current feedback signals from the frequency converter controller of the drive frequency actuator; a displacement feedback branch for acquiring real-time displacement data from the displacement encoder of the transmission drive component; and a field strength calibration branch for acquiring reference field strength signals.

[0012] Preferably, the dual-layer charge calculation unit further includes an environmental compensation unit, which is used to acquire real-time temperature data and real-time humidity data of the process transport path; based on the real-time temperature data and real-time humidity data, it queries a preset material property correction matrix to obtain the surface resistivity correction coefficient for the controlled target; and uses the surface resistivity correction coefficient to adjust the charge migration rate of the logic cell in real time.

[0013] Preferably, the twin field strength output module includes: a mesh generation unit, used to divide the modeling space corresponding to the process transmission path into a three-dimensional discrete mesh; a potential function solving unit, used to obtain the spatial potential distribution of each node in the three-dimensional discrete mesh by solving the Poisson equation, using the charge distribution state as the source term and combining the boundary constraints of the zero potential anchor point; and a field strength mapping unit, used to calculate the negative gradient of the spatial potential distribution and generate a vector field distribution model of the physical quantity field of the controlled target object.

[0014] Preferably, the system also includes an early warning management module, which is used to calculate the maximum field strength value of the preset monitoring area in the digital twin model based on the charge distribution state and its evolution trend; if the maximum field strength value continues to exceed the preset safety threshold, it outputs a physical status early warning command for the coordinate position corresponding to the process transmission path.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In the dynamic modeling of digital twins of packaging electrostatic fields, a forced synchronization rule is established between the mechanical displacement signal of packaging and the evolution process of logic cells. The continuous physical space electric field distribution is discretized into a cell flow model controlled in real time by the driving source pulse. Since the step migration of logic cells directly depends on the feedback pulse of the mechanical encoder rather than simple time iteration, the modeling process can automatically eliminate the phase lag generated during the acceleration and deceleration switching of packaging materials. This displacement-triggered modeling method enables the digital twin to maintain precise alignment with the physical entity charge position under high-speed dynamic conditions. It eliminates the boundary drift and phase distortion commonly found in the solution of traditional continuous field equations from the mechanism level, ensuring that the model output value has extremely high real-time tracking accuracy in the multi-process path.

[0016] 2. By utilizing the current ripple characteristics of the drive power system, the gain of charge generation within the model is corrected in real time. By extracting the high-frequency fluctuation component in the stator current of the packaging drive motor and mapping it to a friction coupling factor reflecting the contact stress between the material and the roller, the system transforms the electrical byproducts of the drive motor into parameter inputs describing the electrostatic physical mechanism. This enables the modeling system to sense mechanical fatigue and operational state deviations. Without adding sensing hardware, it achieves endogenous hedging against electrostatic gain fluctuations caused by roller runout, bearing wear, and microscopic slippage of the material. This cross-dimensional information reuse logic not only reduces the overall system construction cost but also effectively solves the trend prediction deviations generated by static modeling methods under long-term operation in industrial settings.

[0017] 3. By mapping key mechanical structures such as metal rollers inside the packaging machine to zero-potential anchor points in the logical topology, a boundary impedance injection mechanism tightly coupled with the mechanical spatial layout is constructed. When the logical cell moves to the corresponding virtual grounding point coordinates with the displacement signal, the leakage resistance parameter of the cell is transiently rewritten to simulate the real contact discharge process. This method of directly internalizing mechanical structure information into model logic constraints avoids the accumulation error of charge state caused by long-distance transmission in the model. It not only reproduces the potential reset phenomenon after the material comes into contact with the metal component, but also ensures that the field strength output by the digital model maintains a high degree of consistency with the real process state when the material crosses key physical nodes. Attached Figure Description

[0018] Figure 1 This is a diagram showing the overall system architecture and data interaction between modules of the present invention; Figure 2 This is a flowchart illustrating the system operation logic and key parameter evolution and correction process of the present invention. Detailed Implementation

[0019] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments; the following embodiments are intended to explain the present invention and are not intended to limit the scope of protection of the present invention.

[0020] A dynamic modeling system for digital twin of packaging electrostatic field includes a signal acquisition unit, a dynamic feedback identification module, a physical boundary logic mapping module, a double-layer charge evolution calculator, and a twin field strength output module. The signal acquisition unit acquires the stator current feedback signal of the driving frequency actuator, the real-time displacement data of the controlled target, and the reference field strength signal at key locations along the process transmission path. The dynamic feedback identification module, connected to the signal acquisition unit, extracts the fluctuation characteristics of the current signal. The physical boundary logic mapping module establishes the spatial mapping relationship between mechanical components and the logic model. The double-layer charge evolution calculator receives displacement data and dynamic feedback results to perform evolutionary deduction of the charge distribution state. The twin field strength output module generates a digital twin model of the distributed physical quantity field and corrects deviations based on the reference field strength signal. During the operation of the packaging machine, the flexible film is in a high-speed variable-speed conveying state. The charge distribution on the material surface undergoes dynamic convection with displacement. To address the prediction lag issue, the system employs a displacement-triggered synchronization procedure. The displacement feedback branch in the signal acquisition unit acquires the displacement increment Δs of the controlled target object through a displacement encoder. The dual-layer charge evolution calculator divides the packaging path into several logical cells based on the displacement increment Δs, with each logical cell representing a microscopic segment of the packaging material. The processing logic for each sampling period is as follows: the input is the displacement increment Δs and the charge state vector Q(k) from the previous moment. When the accumulated displacement increment Δs reaches the preset cell step size L, the charge state quantity within the logical cell is triggered to migrate to the next node in the material flow direction, thus updating the state. The cell step size L is determined based on the system's spatial resolution requirements and is set to 10mm in this embodiment. This synchronizes the model state evolution with the mechanical displacement, eliminating the prediction phase lag caused by the fixed time step.

[0021] Due to contact stress fluctuations caused by micro-vibration of the roller shaft or wear of the bearing, the triboelectric gain term deviates; the dynamic feedback identification module separates the high-frequency ripple component δI with a frequency higher than 500Hz from the stator current feedback signal through a digital high-pass filter and calculates its energy density. The processing logic is as follows: the input is the stator current feedback signal. The result is obtained through a fourth-order high-pass filter. ,calculate ,in, Where N is the energy density and N is the total number of samples within a single sampling window. Let be the high-frequency ripple current value at the j-th sampling point; according to The dynamic friction coupling factor α is determined by the mapping relationship with the preset friction torque fluctuation model. The mapping relationship of the dynamic feedback identification module is determined based on the speed gradient benchmark calibration during the initial debugging stage. After the first run or material change, the system executes a set of automatic calibration sequences, driving the frequency actuator to drive the controlled target object to run stably for 10 seconds at five speed steps: 50 m / min, 100 m / min, 150 m / min, 200 m / min, and 250 m / min. The average energy density of the high-frequency ripple component at each speed segment is collected simultaneously and linearly interpolated with the offline preset unit contact stress charge gain benchmark to generate a 5-row, 2-column mapping relationship lookup table stored in the system's non-volatile memory. Using the offline calibration platform, a stepping mechanical load is applied to the surface of the controlled target object, and the stator current feedback signal under different speed conditions is collected to separate the high-frequency ripple component and calculate the energy density. The cross-correlation coefficient between the measured charge increment of the non-contact probe and the energy density was established. A linear mapping equation with the dynamic friction coupling factor α as the dependent variable and the independent variable being α, satisfies the following condition: The proportionality coefficient η is determined based on the surface roughness level of the contact surface and through least-squares fitting of the sampled dataset. The background ripple energy density is given by the system under no-load constant-speed operation. The dynamic friction coupling factor α is used to perform nonlinear compensation for the charge injection gain term of the logic cell. The compensation rule is as follows: ,in, For the corrected charge injection gain, The base gain coefficient is preset based on the material properties, α is the dynamic friction coupling factor, and λ is the preset sensitivity weighting coefficient. The power-on parameters are corrected using the electrical signal characteristics of the drive system to offset the errors caused by the drift of the mechanical dynamic state.

[0022] Conductive components such as metal rollers in the packaging path constitute an electrostatic discharge path; the physical boundary logic mapping module marks the conductive transmission components as zero-potential anchor points in the virtual logic topology; the specific procedures are as follows: the system establishes a virtual coordinate system consistent with the mechanical layout of the process transmission path, and determines the geometric center coordinates of each conductive transmission component in the virtual coordinate system. And the effective contact length W; real-time monitoring of the movement trajectory of the logic cell carrying the leakage resistance parameter, when the logic cell moves to the coordinate range When the leakage resistance parameter is reset, a leakage resistance parameter reset instruction is triggered, switching the leakage resistance parameter of the corresponding logic cell from a preset high resistance state value to a preset low resistance state value. The preset high resistance state value is calibrated to 1,000,000,000,000Ω, used to simulate the charge retention state of insulating materials. The preset low resistance state value is calibrated to 1,000,000Ω, used to simulate the grounding discharge resistance at the moment the material comes into contact with the metal roller. The switching instruction is triggered by a hardware interrupt when the center coordinate of the logic cell enters within 5mm before and after the center point of the zero potential anchor point coverage area. The low resistance state value drives the charge in the logic cell to perform step-by-step discharge, used to simulate the potential reset process when the material crosses the physical grounding boundary.

[0023] To simulate the polarization charge characteristics within multi-layer composite packaging materials, a dual-layer charge evolution calculator constructs a dual-layer evolution architecture comprising a surface-layer moving charge operator and an inner-layer bound charge operator. The polarization memory module, by executing the inner-layer bound charge operator, establishes a virtual polarization layer operator to analyze the pumping logic of bound charges from the controlled target material to the surface. The processing logic defines the cross-layer coupling relationship between the surface-layer moving charge and the virtual polarization layer operator. When the system detects that the packaging speed signal has returned to zero, it triggers the evolution operator to enter polarization release mode, using the potential energy difference to drive the internal charge to continuously compensate for the surface charge state. When processing controlled targets with multi-layer interface polarization characteristics, the polarization memory module executes the parameter initialization procedure of the virtual polarization layer operator to simulate the relaxation behavior of bound charges. The system collects the surface resistivity of the material before the packaging material enters the process transport path. With volume resistivity Inner bound charge The initial value is based on the formula It is confirmed that, among them, This refers to the amount of bound charge in the inner layer. Let β be the total surface charge density of the material measured by a non-contact potentiometer, and let β be based on... The system determines the value of the scaling factor based on the pre-set charge distribution mapping table of the material. When the system detects that the ambient temperature change exceeds 2°C, it automatically triggers the correction command of the time constant τ. The Arrhenius equation is used to perform nonlinear compensation on the charge pumping rate. This procedure eliminates the initial convergence deviation of the model caused by the unobservable polarization depth by physically tracing the initial charge state, and realizes the engineering simulation of the electrostatic rebound phenomenon on the surface of the packaging film.

[0024] Under conditions such as high-speed stripping, the local electric field strength may exceed the air breakdown threshold; the nonlinear discharge correction module monitors the local electric field strength generated by the charge distribution in real time. And determine the local field strength Does it exceed the preset air breakdown threshold? ;like The system loads nonlinear shunt logic triggered by state variables to calculate nonlinear charge loss. Loss calculation follows the formula below: ,in, The nonlinear charge loss is simulated, and k is a constant factor characterizing the ionization intensity. These are the calculated values ​​for the local field strength. The system uses a preset air breakdown threshold; it subtracts the charge distribution state in real time based on the nonlinear charge loss to correct the prediction overshoot generated by the linear RC model under extreme conditions; the twin field strength output module uses the charge distribution state of each logic cell as the source term, solves the Poisson equation in combination with the boundary constraints of the zero potential anchor point, generates the three-dimensional spatial potential distribution corresponding to the packaging path, and generates a vector field distribution model; the state assimilation module receives the reference field strength signal, calculates the residual between the model prediction value and the measured value, and corrects the gain coefficient of the charge transfer operator in real time based on the residual; the polarization parameters of the double-layer charge evolution calculator are initialized based on the dielectric properties of the material. Before the controlled target enters the process transmission path, the electrometer collects the surface potential evolution characteristics of the controlled target within the preset discharge time window, and uses nonlinear fitting to extract the charge relaxation time constant τ that reflects the release rate of the bound charge. The value is written into the polarization memory module as the reference for the time constant of the inner bound charge operator to correct the measurement residual caused by the sensor installation posture. The state assimilation module reads the posture offset data of the laser ranging component and calculates the residual vector between the theoretical field strength and the reference field strength signal. Utilizing Kalman gain The state weights of the surface moving charge operator and the inner bound charge operator are adjusted to ensure that the output space potential distribution of the digital twin model converges steadily within the displacement range. This closed-loop correction process ensures that the output value of the digital twin model continuously converges to the true electrostatic field strength of the physical entity. The system also has an early warning management module. When the maximum field strength value of the monitored area in the digital twin model continuously exceeds the preset safety threshold, the system outputs a physical state early warning command at the corresponding coordinate position.

[0025] Example 1: In the operation of a flexible film packaging production line with high-frequency start-stop characteristics, the drive device guides the material to perform accelerated motion. At this time, the displacement of the material relative to the roller shaft exhibits nonlinear changes. The signal acquisition unit reads the displacement increment Δs generated by the encoder installed on the main drive shaft of the packaging machine. The dual-layer charge algorithm divides the packaging path into several logic cells of equal length. The length of each logic cell is the cell step size L. The system executes a synchronization procedure based on displacement locking. Its processing logic is as follows: The input is the displacement increment Δs and the charge state vector Q(k) of the previous sampling period. When the cumulative amount of displacement increment Δs is equal to the cell step size L, the step-by-step migration of the charge distribution state is triggered.

[0026] The dynamic feedback identification module synchronously reads the current stator current feedback signal. The high-frequency ripple component δI with a frequency higher than 500Hz is separated by a digital high-pass filter, according to the formula... Calculate the energy density to determine the dynamic friction coupling factor α, where, Let N be the energy density, N be the number of sampling points, δI be the high-frequency ripple component, and α be the dynamic friction coupling factor; the system is based on the formula... Correcting charge injection gain, where For the corrected charge injection gain, The base gain is λ, which is the sensitivity weighting coefficient. Since the evolution logic of the logic cell is triggered by the displacement increment Δs, the charge state evolution in the digital twin system is synchronously mapped with the material displacement. The system compensates for the charge gain deviation caused by contact stress fluctuations through the dynamic friction coupling factor α. The overlap between the spatial potential distribution output by the digital twin model and the measured value of the field sensor reaches 98%, realizing the real-time tracking of the physical entity's operating logic by the model state.

[0027] Example 2: The test platform includes a multi-stage roller conveyor system to simulate the dynamic electrostatic field evolution process of a high-speed packaging production line. The platform is equipped with an electrostatic potential sensor with a range of 0 to 50 kV and a resolution of 0.01 kV, as well as a current acquisition unit with a sampling frequency of 10 kHz. The test data originates from real-time measurements on the physical experimental platform. Gaussian white noise with a signal-to-noise ratio of 20 dB and power frequency interference harmonics with a frequency of 50 Hz are actively superimposed at the signal input of the drive frequency actuator to verify signal processing performance under complex electromagnetic environments. The core parameter is the sampling period. The setting depends on the spectral characteristic bandwidth of the monitored signal. Under the experimental conditions, the spectral characteristic bandwidth of the material conveying speed is 2kHz. To satisfy the sampling theorem and take into account the computational load of the solver, the sampling period is set to... The time interval was set to 0.1ms. During the test start-up phase, the signal acquisition unit acquired the original displacement signal of the controlled target under variable speed conditions. The comparison sample group adopted a continuous field modeling method based on a fixed time step, updating the charge state every 10ms. Under the acceleration condition where the material speed increased from 200m / min to 800m / min, the charge distribution curve output by the comparison sample group showed a phase shift with the actual position of the physical entity, with an average spatial position deviation of 15.6mm. The test group executed a displacement-triggered modeling procedure, setting the cell step size L to 10mm. The test group read the displacement increment Δs, and when the cumulative amount of the displacement increment Δs reached the cell step size L, the state transition of the logic cell was triggered. The measurement results showed that under the same acceleration condition, the spatial prediction deviation of the test group was reduced to 1.2mm, verifying the suppression effect of the displacement synchronization logic on phase lag.

[0028] The test group further implemented kinetic feedback identification logic to correct for energization deviation caused by contact stress fluctuations; the kinetic feedback identification module received feedback signals from the stator current, which contained noise. Extract the high-frequency ripple component δI and calculate the energy density. The processed data stream shows that the power frequency interference component was reduced by 35dB; energy density The calculation formula is as follows: ,in, The energy density of the current signal, in units of N is the number of sampling points. Let J be the high-frequency ripple current amplitude at the j-th sampling point; experimental observations show that the triboelectric gain term exhibits nonlinear characteristics, when When it is in the range of 0.05 to 0.15, the dynamic friction coupling factor α and There is a linear positive correlation; when mechanical vibration intensifies, it leads to... Beyond 0.45, the charge injection gain The growth rate slowed down and stabilized, entering the physical saturation zone; the experimental group, according to the formula... Adjusting the gain reduced the model field strength residual from 145.2V to 42.8V, confirming the synergistic effect of the dynamic feedback mechanism and displacement synchronization logic under variable speed conditions. Ultimately, the twin field strength output module generated a three-dimensional spatial vector field distribution model. The state assimilation module received the reference field strength signal from the field sensors and calculated the real-time residual between the model prediction and the measured value. The data flow characteristics of the experimental group showed that after 12.5ms of system operation, the model prediction residual stably converged to within the preset threshold of 50V. The system calibrated the power-on gain drift online and maintained the consistency between the twin model and the physical entity under nonlinear conditions. Its modeling accuracy was improved by more than 85% compared with the traditional scheme, providing data support for the technical characteristics of dynamic friction compensation and displacement-driven evolution.

[0029] Example 3: This example combines Figures 1 to 2 This paper describes a dynamic modeling system for digital twins of packaging electrostatic fields, such as... Figure 1As shown, the system consists of a signal acquisition unit, a dynamic feedback identification module, a dual-layer charge evolution calculator, a twin field strength output module, and a physical boundary logic mapping module, coupled through signal and data flows. The signal acquisition unit, as the input front end, is responsible for acquiring the stator current feedback signal of the driving frequency actuator, the displacement data of the controlled target, and the reference field strength signal. The stator current feedback signal is unidirectionally transmitted to the dynamic feedback identification module, which extracts high-frequency ripple and determines the dynamic friction coupling factor, and inputs it to the dual-layer charge evolution calculator. The displacement data of the controlled target is transmitted to the physical boundary logic mapping module, which maps a zero-potential anchor point and switches the leakage resistance parameter when the controlled target moves to the corresponding coordinate, and provides a virtual logic topology to the twin field strength output module. The charge distribution state calculated by the dual-layer charge evolution calculator is transmitted to the twin field strength output module, which combines the reference field strength signal to generate a digital twin model and correct the deviation. The corrected state is returned to the dual-layer charge evolution calculator through a feedback loop, forming a closed-loop evolution.

[0030] like Figure 2 As shown, the system's operational logic begins with the data input from the hardware components. The output of the drive frequency actuator is processed to extract and separate current signals with frequencies higher than 500Hz to obtain high-frequency ripple components, thereby determining the dynamic friction coupling factor characterizing the micro-fluctuations of contact stress. The physical properties of the transmission conductive components are used to map zero-potential anchor points to establish a virtual logic topology. When a logic cell moves to the anchor point position, it triggers the switching of leakage resistance parameters. Data collected by the laser ranging and field strength probe array is used to execute the laser pose calibration procedure to correct the virtual coordinate system deviation. Based on this, the system calculates the real-time residual vector between the theoretical and measured field strengths. The Kalman gain is calculated based on the noise covariance matrix R. The gain is used to correct the charge transfer operator and adjust the state weight allocation; the system flow also includes a cell step size adaptive adjustment stage, which is achieved by calculating the spatial cross-correlation coefficient. The step size parameter is optimized, and all the corrected parameters and factors are incorporated into the two-layer charge evolution calculation step to compensate for the charge injection gain, and finally output a high-precision digital twin field strength model.

[0031] Example 4: In the initial configuration process for a precision electronic product packaging line, the system executes an initial state definition procedure to establish a modeling benchmark. The signal acquisition unit is equipped with a current sensor with a sampling frequency of 100kHz and a current measurement resolution better than 1mA, and a displacement encoder with a range of 0 to 5m / s and a linearity deviation of less than 0.1%. The system acquires the stator current signal of the packaging line under no-load constant speed operation as the benchmark signal, and uses the dynamic feedback identification module to establish the zero-point reference point of the friction torque fluctuation model. To eliminate the mapping black box of the dynamic friction coupling factor, the system adopts a linear calibration procedure based on the least squares method. During material transfer, the system determines the calculation path of the dynamic friction coupling factor based on the residual between the observed current energy density at different speed gradients and the reference field strength signal. The specific calculation formula is as follows: Where α is the dynamic friction coupling factor and η is the proportionality coefficient, which is set to 0.85 in this embodiment based on the surface roughness of the material. To measure the energy density, As the baseline energy density under no-load conditions, and considering the cross-layer coupling logic in the two-layer evolution architecture, the two-layer charge evolution calculator executes a charge pumping operator based on the potential energy gradient. Within each sampling period, the system calculates the cross-layer transferred charge per unit step based on the difference between the amount of surface-moving charge and the amount of inner-layer bound charge. The processing logic follows the following discretization equation: ,in, γ represents the amount of charge transferred across layers, and γ is the cross-layer coupling coefficient. This refers to the amount of bound charge in the inner layer. denoted as , where t is the cumulative time elapsed since the cell was created, and τ is the material-specific charge relaxation time constant.

[0032] To address the environmental adaptability issue of discharge correction intensity, the nonlinear discharge correction module incorporates adaptive calibration logic based on ambient humidity. The system receives the relative humidity signal RH from a humidity sensor installed above the process transmission path, and the constant factor k is dynamically calibrated according to the following relationship: Where k is a constant factor characterizing the ionization intensity, ζ is the basic ionization constant under standard atmospheric conditions, ζ is the humidity influence weighting factor and is preset to 0.12, and RH is the real-time relative humidity percentage. By executing parameter calibration and coupling algorithm transparency procedures, the system maintains the rapid convergence of the twin model in scenarios where the contact characteristics change drastically due to material batch changes. Observational data show that after introducing the proportional coefficient η and humidity correction logic, the prediction residual of the model under extreme high humidity conditions is reduced from 120V to less than 25V, improving the system's contribution to reproducibility in variable industrial environments.

[0033] Example 5: In the case of spatial electric field reconstruction for a long-span roller conveyor path, the twin field strength output module divides the space containing the material and the physical entity of the roller into non-uniformly spaced grids and establishes a discretized three-dimensional grid model. The zero-potential anchor points calibrated by the physical boundary logic mapping module are mapped to the Dirichlet boundary conditions of the Poisson equation, that is, the potential is always 0V when the grid node coordinates are at the physical boundary of the roller. The system uses the super-relaxation iterative method to solve the discretized Poisson equation. Where Φ is the space potential distribution in V, and ρ is the space charge density determined based on the logical cell migration state in V. , The relaxation factor ω is the vacuum permittivity. During the solution process, the relaxation factor ω is dynamically adjusted in steps within the range of 1.2 to 1.8 based on the residual convergence rate of the computational domain.

[0034] When the system faces initial uncertainty in model parameters due to material switching, the state assimilation module initiates the sensor pose calibration procedure and uses the laser ranging component to correct the position deviation vector of the reference field strength sensor in the virtual coordinate system, calculating the real-time residual vector between the theoretical field strength and the measured field strength at the reference position. The system is based on the preset sensor measurement noise covariance matrix R and prediction error covariance matrix. Real-time calculation of Kalman gain matrix The solution logic follows the formula ,in, For Kalman gain, Let H be the prediction error covariance matrix, H be the state observation matrix, and R be the sensor measurement noise covariance matrix. The system utilizes the real-time residual vector. The state weight allocation of the charge transfer operator is modified so that the field strength output value of the digital twin model converges to the preset residual threshold within the first 5 cell steps of material movement.

[0035] Example 6: In the spatial resolution calibration of a semiconductor component packaging line, the system selects the cell step size L based on the rate of change of the spatial electric field gradient along the transmission path. The signal acquisition unit arranges a probe array with a sampling interval of 5.2 mm at the stripping start point of the transmission path. The dual-layer charge algorithm gradually decreases the value of the cell step size L by adjusting the operator, and calculates the spatial correlation coefficient between the predicted field strength and the measured field strength of the digital twin model at each length. When the spatial interrelationship number When the growth rate ΔC / ΔL is below the gain threshold of 0.05% per millimeter, the system determines the current value as the optimal value for the cell step size L. Under this condition, the system calibrates the cell step size L to 8.5 mm to ensure that the digital twin system captures electrostatic charge convection characteristics while controlling the single calculation time within the pulse width modulation period of the drive frequency actuator. Under the condition of spatial accuracy discretization calibration for the anti-static packaging path of electronic components, the system executes the spatial discretization density optimization procedure to determine the calculation boundary for solving the Poisson equation. The physical boundary logic mapping module determines the local densification zone based on the radius of curvature r of the physical entity of the roller. When the normal curvature 1 / r of the roller surface is greater than 0.2 per millimeter, the twin field strength output module sets the grid step size h of this region to 1 mm, while setting the grid step size of the material suspension transmission zone to 5 mm. The system starts the convergence verification process by calculating the global potential norm residual under adjacent grid densities. To evaluate the reconstruction accuracy, the processing logic follows the following criteria: when the field strength variation rate caused by doubling the number of grid nodes is less than 0.1%, the system terminates grid refinement and locks the current discretized topology parameters. Under this condition, the system divides the three-dimensional solution domain into 256,000 hexahedral elements to control the single calculation time within a drive control cycle of less than 15ms while ensuring the resolution of the spatial electric field gradient abrupt change region.

[0036] When the system is applied to composite thin film materials with different conductivity properties, the state assimilation module determines the parameter benchmark of the Kalman gain matrix through the sensor noise statistical procedure. The system collects 2048 discrete voltage samples from the reference field strength sensor under unloaded conditions and with electromagnetic interference at background levels on the packaging line, and calculates their variance, using these as the diagonal elements of the sensor measurement noise covariance matrix R. Simultaneously, the physical boundary logic mapping module injects detached samples with known charge at the material entry point. The system calculates the cumulative prediction error of the solver over 10 displacement cycles and determines the prediction error covariance matrix accordingly. The initial value is used to establish the Kalman gain through this calibration process. The weight allocation logic is used to eliminate prediction oscillations caused by sensor temperature drift or environmental background charge fluctuations, and to achieve field strength convergence in the early stage of material transfer.

[0037] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A dynamic modeling system for digital twin of packaging electrostatic field, characterized in that, The system includes: The signal acquisition unit is used to acquire the stator current feedback signal of the drive frequency actuator, the real-time displacement data of the controlled target, and the reference field strength signal at key locations in the process transmission path; The dynamic feedback identification module is used to separate high-frequency ripple components with frequencies higher than 500Hz from the stator current feedback signal through a digital high-pass filter, calculate the energy density of the high-frequency ripple components within a preset sampling period, and determine the dynamic friction coupling factor that characterizes the micro-fluctuation of contact stress between the controlled target and the transmission drive components based on the mapping relationship between the energy density and the preset friction torque fluctuation model. The physical boundary logic mapping module is used to map the conductive transmission components in the process transmission path to zero potential anchor points in the virtual logic topology. Combined with real-time displacement data, it simulates the contact charge transfer process by transiently rewriting the leakage resistance parameters of the logic cells when the logic cell corresponding to the controlled target moves to the coordinates of the zero potential anchor point. A two-layer charge evolution calculator is used to dynamically compensate the charge injection gain term of the logic cell by using the dynamic friction coupling factor as a correction term, and to construct a two-layer evolution architecture containing a surface moving charge operator and an inner bound charge operator to solve the charge distribution state of the controlled target during the dynamic displacement process. The twin field strength output module is used to generate a digital twin model of the physical quantity field of the controlled target object based on the charge distribution state, and to perform deviation closed-loop correction on the digital twin model based on the reference field strength signal.

2. The dynamic modeling system for digital twin of packaging electrostatic field according to claim 1, characterized in that, When determining the dynamic friction coupling factor, the dynamic feedback identification module is configured to execute the following processing rules: remove the low-frequency fundamental wave from the stator current feedback signal using a preset digital high-pass filter, and extract the high-frequency ripple component with a frequency higher than 500Hz; calculate the energy density component of the high-frequency ripple component within the current sampling period. Based on energy density components Based on the preset friction torque fluctuation model, a dynamic friction coupling factor α is determined to reflect the microscopic fluctuation of contact stress between the controlled target and the transmission component; wherein, the dynamic friction coupling factor α is used to nonlinearly offset the charge injection gain term of the logic cell when it contacts the transmission component.

3. The dynamic modeling system for digital twin of packaging electrostatic field according to claim 1, characterized in that, When simulating the contact charge transfer process, the physical boundary logic mapping module performs the following steps: Step S1, establish a virtual coordinate system consistent with the mechanical layout of the process transmission path, and calibrate the geometric center coordinates and effective contact length of the transmission component in the virtual coordinate system; Step S2, establish the logic coverage range of the zero potential anchor point around the geometric center coordinates based on the effective contact length; Step S3, monitor the movement trajectory of the logic cell carrying the leakage resistance parameter in real time. If the movement trajectory enters the logic coverage range, a logic instruction is triggered to switch the leakage resistance parameter from the preset high resistance state value to the preset low resistance state value to derive the charge amount in the logic cell.

4. The dynamic modeling system for digital twin of packaging electrostatic field according to claim 1, characterized in that, The two-layer charge evolution solver includes: a surface evolution module, used to execute the surface moving charge operator to calculate the flowing charge density generated on the surface of the controlled target object under the influence of the dynamic friction coupling factor; a polarization memory module, used to execute the inner bound charge operator, establish a virtual polarization layer operator, and analyze the pumping logic of bound charges inside the controlled target object to the surface; and a state coupling unit, used to superimpose the calculation results of the surface moving charge operator and the inner bound charge operator to determine the charge distribution state.

5. The dynamic modeling system for digital twin of packaging electrostatic field according to claim 1, characterized in that, The system also includes a nonlinear discharge correction module, which is used to: monitor the local field strength generated by the charge distribution in real time, and determine whether the local field strength exceeds the preset air breakdown threshold. If the air breakdown threshold is exceeded, a nonlinear shunt logic based on state variables is activated to simulate the charge loss process caused by local ionization.

6. A dynamic modeling system for digital twin of packaging electrostatic field according to claim 5, characterized in that, When the nonlinear discharge correction module executes nonlinear shunt logic, it follows the formula: ,in, The nonlinear charge loss is simulated, and k is a constant factor characterizing the ionization intensity. These are the calculated values ​​for the local field strength. The preset air breakdown threshold is used; the dual-layer charge calculation unit performs real-time subtraction of the charge distribution state based on the nonlinear charge loss.

7. The dynamic modeling system for digital twin of packaging electrostatic field according to claim 1, characterized in that, The signal acquisition unit includes: a current monitoring branch for obtaining stator current feedback signals from the frequency converter controller of the drive frequency actuator; a displacement feedback branch for obtaining real-time displacement data from the displacement encoder of the transmission drive component; and a field strength calibration branch for obtaining reference field strength signals.

8. The dynamic modeling system for digital twin of packaging electrostatic field according to claim 1, characterized in that, The double-layer charge calculation unit also includes an environmental compensation unit, which is used to acquire real-time temperature data and real-time humidity data of the process transport path; Based on real-time temperature and humidity data, a preset material property correction matrix is ​​queried to obtain the surface resistivity correction coefficient for the controlled target object. The charge migration rate of logic cells is adjusted in real time using a surface resistivity correction coefficient.

9. A dynamic modeling system for digital twin of packaging electrostatic field according to claim 1, characterized in that, The twin field strength output module includes: a mesh generation unit, used to divide the modeling space corresponding to the process transmission path into a three-dimensional discrete mesh; a potential function solving unit, used to obtain the spatial potential distribution of each node in the three-dimensional discrete mesh by solving the Poisson equation, using the charge distribution state as the source term and combining the boundary constraints of the zero potential anchor point; and a field strength mapping unit, used to calculate the negative gradient of the spatial potential distribution and generate a vector field distribution model of the physical quantity field of the controlled target object.

10. A dynamic modeling system for digital twin of packaging electrostatic field according to claim 1, characterized in that, The system also includes an early warning management module, which is used to calculate the maximum field strength value of the preset monitoring area in the digital twin model based on the charge distribution state and its evolution trend. If the maximum field strength value continues to exceed the preset safety threshold, a physical status warning command will be output for the coordinate position corresponding to the process transmission path.

Images