Dry-method electrode preparation process control method and system and storage medium

Abstract

The invention relates to the technical field of electrode preparation, and discloses a dry-method electrode preparation process control method and system and a storage medium. The method comprises the steps that in the rolling forming process of a dry-method electrode material, pressure roller gap data and rotation angle data of a driving motor are monitored and collected in real time, and a support vector machine algorithm is used for recognizing the transmission ratio change trend of a pressure roller system; calculating a pitch precision offset caused by wear through Kalman filtering in combination with historical wear data; when the offset exceeds the limit, an adjustment coefficient is generated based on a simulation model, and the control precision is improved through feed-forward compensation and feedback optimization; when the adjustment precision meets the process requirement, a self-adaptive parameter set is determined through dynamic deviation residual analysis, and an actuator is driven to adjust the gap between the compression rollers; and monitoring the thickness uniformity index of the electrode plate produced after adjustment in real time, and finally forming quality closed-loop control by taking the thickness uniformity of the electrode plate as a verification standard. According to the invention, the quality precision of dry-method electrode production is improved.

Description

Technical Field

[0001] This application relates to the field of electrode preparation technology, and in particular to a method, system and storage medium for controlling the dry electrode preparation process. Background Technology

[0002] With the widespread application of new energy batteries, especially the increasing demand in electric vehicles and energy storage devices, battery performance and production efficiency have become key factors. Battery performance largely depends on the quality of electrode materials, and the uniformity of electrode material thickness directly affects the battery's charge-discharge efficiency, lifespan, and safety. The electrode manufacturing process typically involves coating, pressing, and other steps, among which the control of the gap between the pressure rollers is crucial for the uniformity of electrode thickness.

[0003] In traditional dry electrode fabrication technology, the adjustment of the pressure roller gap typically relies on manual experience or fixed preset parameters. However, with prolonged equipment operation, especially wear and tear on mechanical components such as lead screws, the transmission ratio changes. Traditional control methods fail to effectively address these changes, leading to increased fluctuations in electrode thickness and impacting battery performance. This is particularly true in high-precision electrode production, where the accuracy and adaptability of traditional control methods often cannot meet constantly changing production demands. Furthermore, most existing technologies rely on static parameters for adjustment, lacking real-time dynamic adjustment capabilities. When equipment wears down or operating conditions change, the control strategy cannot be corrected in time, resulting in uneven electrode thickness during production. This limitation affects the stability and consistency of battery products, thus hindering the further development of the battery industry.

[0004] Therefore, how to achieve real-time dynamic adjustment of the pressure roller gap and adapt to changes in equipment wear and the production environment is a problem that urgently needs to be solved in current electrode fabrication technology. This application provides an adaptive control method based on real-time data acquisition and state estimation technology, which can dynamically adjust the pressure roller gap according to changes in the transmission ratio and equipment wear, thereby ensuring the uniformity of electrode thickness and overcoming the shortcomings of existing technologies. Summary of the Invention

[0005] To address the aforementioned technical problems, this application provides a dry electrode preparation process control method, system, and storage medium, which improves the control accuracy and stability of the pressure roller gap, and enables the system to adapt to time-varying factors such as equipment wear, effectively extending the equipment maintenance cycle and significantly improving product qualification rate and consistency.

[0006] In a first aspect, this application provides a method for controlling the dry electrode fabrication process, the method comprising: Step S1: During the rolling process of dry electrode material, real-time monitoring and collection of pressure roll gap data and drive motor rotation angle data are performed. The collected data is classified and processed using a support vector machine algorithm to identify the transmission ratio change trend of the pressure roll system. Step S2: Combining the transmission ratio change trend with the historical wear data of the lead screw, the wear state of the lead screw in the current pressure roller system is evaluated by the state estimation algorithm, and the pitch accuracy offset caused by wear is calculated. Step S3: If the pitch accuracy offset exceeds a preset threshold, an adjustment coefficient for correcting the motor displacement command is generated based on the pitch accuracy offset. The motor angle command sequence is updated using the adjustment coefficient, and the updated command sequence is iteratively optimized through a feedback loop to improve the actual adjustment accuracy of the pressure roller gap. Step S4: When the adjustment accuracy reaches the process requirements, obtain the dynamic deviation residual of the system, correct the dynamic deviation residual and determine the final adaptive control parameter set, and drive the actuator to adjust the pressure roller gap based on the adaptive control parameter set. Step S5: Monitor the uniformity of the electrode sheet thickness after adjustment in real time. If the deviation of the thickness uniformity from the target range exceeds the preset range, return to step S1 and restart a new round of optimization cycle until the thickness uniformity is qualified, and output the final control parameters.

[0007] Secondly, this application provides a control system for a dry electrode fabrication process, the system comprising: The data acquisition module is used to monitor and acquire data on the gap between the pressure rollers and the rotation angle of the drive motor in real time during the rolling process of dry electrode materials. The acquired data is classified and processed using a support vector machine algorithm to identify the trend of transmission ratio change of the pressure roller system. The condition assessment module is used to combine the transmission ratio change trend with the historical wear data of the lead screw, and to assess the wear state of the lead screw in the current pressure roller system through a condition estimation algorithm, and to calculate the pitch accuracy offset caused by wear. The iterative optimization module is used to generate an adjustment coefficient for correcting the motor displacement command based on the offset if the pitch accuracy offset exceeds a preset threshold, update the motor angle command sequence using the adjustment coefficient, and iteratively optimize the updated command sequence through a feedback loop to improve the actual adjustment accuracy of the pressure roller gap. The correction module is used to obtain the dynamic deviation residual of the system when the adjustment accuracy reaches the process requirements, correct the dynamic deviation residual and determine the final adaptive control parameter set, and drive the actuator to adjust the pressure roller gap based on the adaptive control parameter set. The loop control module is used to monitor the uniformity of the electrode sheet thickness after adjustment in real time. If the deviation of the thickness uniformity from the target range exceeds the preset range, the module returns to the acquisition module and restarts a new round of optimization loop until the thickness uniformity is qualified, and then outputs the final control parameters.

[0008] A third aspect of this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the aforementioned dry electrode fabrication process control method.

[0009] Compared with the prior art, the beneficial effects of the present invention are at least as follows: This invention provides a method, system, and storage medium for controlling the dry electrode fabrication process. By constructing a complete closed-loop control chain from equipment status perception to product quality verification, it achieves intelligent and precise control of the pressure roller gap. The method first utilizes a support vector machine algorithm to identify trends in real-time data on the pressure roller gap and motor rotation angle, enabling early detection of abnormal changes in the transmission ratio caused by lead screw wear, thus gaining valuable time for subsequent compensation control. By combining this with a Kalman filter algorithm to accurately estimate the lead screw wear state, the precise pitch offset is calculated, providing a reliable quantitative basis for compensation control.

[0010] When significant wear is detected, precise command adjustment coefficients are generated based on a physical simulation model. By combining feedforward compensation and feedback optimization, the impact of mechanical wear on control accuracy is effectively offset. Finally, the uniformity of electrode thickness is used as the quality evaluation standard to form a complete closed loop from equipment control to product quality, ensuring that the adjustment of control parameters can be verified in actual product quality. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of an embodiment of a dry electrode fabrication process control method in this application. Figure 2 This is a schematic diagram of an embodiment of a dry electrode fabrication process control system in this application. Detailed Implementation

[0013] This application provides a method, system, and storage medium for controlling a dry electrode fabrication process. The terms "first," "second," "third," "fourth," etc. (if present)," in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0014] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of a dry electrode fabrication process control method in this application includes: Step S1: During the rolling process of dry electrode material, monitor and collect data on the gap between the pressure rollers and the rotation angle of the drive motor in real time. Use the support vector machine algorithm to classify the collected data and identify the trend of transmission ratio change of the pressure roller system.

[0015] Step S1 includes: collecting real-time data on the gap between the pressure rollers using a displacement sensor and real-time data on the rotation angle of the drive motor using an encoder; using the collected data on the gap between the pressure rollers and the rotation angle as input feature vectors; constructing a classification model using a support vector machine algorithm to classify the input feature vectors, obtaining the classification result of the transmission ratio change trend, fitting the classification result to obtain the transmission ratio change trend curve, and calculating the transmission ratio change rate; if the transmission ratio change rate exceeds a preset threshold, the transmission ratio change trend is marked as an abnormal trend, otherwise it is marked as a normal trend; extracting key time points from the data based on the abnormal trend, analyzing the peak and trough values ​​of the transmission ratio change trend through the key time points, and obtaining quantitative indicators of the transmission ratio change trend.

[0016] Specifically, dry electrode manufacturing refers to the process of shaping materials into electrodes using physical means such as pressure and heat without the use of liquid solvents. This process is commonly used in the manufacture of energy devices such as batteries and capacitors. In this process, the pressure roller is a key piece of equipment, controlling the thickness and density of the material through rolling. During the operation of the pressure roller system, its gap and the rotation angle of the drive motor are crucial parameters. The gap between the two pressure rollers determines the degree of compression of the material during rolling, directly affecting the material quality. The rotation angle of the drive motor reflects the motor's operation and indirectly affects the compression force of the pressure roller. Therefore, throughout the manufacturing process, sensors and encoders are needed to monitor these two data points in real time.

[0017] This application first collects the roller gap data in real time using a high-precision displacement sensor, for example, collecting 10 data points per second to form a roller gap sequence. At the same time, it collects the rotation angle data of the drive motor shaft synchronously through an encoder installed on the drive motor shaft to obtain an angle sequence. The gap sequence and the angle sequence are combined into an input feature vector. The real-time input feature vector is fed into a pre-trained classification model, which is generated based on the support vector machine algorithm. The training data consists of historical data accumulated from the long-term operation of the pressure roller system. The samples in this historical data have been labeled with transmission ratio state categories of "stable", "rising", or "falling" trends by expert knowledge or process results. By learning from these labeled samples, the classification model constructs a classification boundary that can distinguish different transmission ratio states. The classification model processes the real-time input feature vector and outputs the classification result of the transmission ratio change trend at the current moment, such as "stable", "rising", or "falling". The classification results at a series of time points are converted into corresponding numerical sequences, for example, mapped to the values ​​0, +1, and -1, respectively. After smoothing and filtering this numerical sequence to suppress instantaneous fluctuations, the least squares method is used to curve fit it, generating a smooth transmission ratio change trend curve. The slope of the fitted curve is calculated to obtain the transmission ratio change rate, which quantifies the severity of the transmission ratio state change.

[0018] The calculated rate of change is compared with a pre-set threshold, for example, 0.05 units / second. If the rate exceeds this threshold, an abnormal trend is identified and an alarm is triggered; otherwise, it is marked as a normal trend. Once a trend is marked as abnormal, key time points are automatically extracted from the original data sequence corresponding to that abnormal trend. These time points are usually the moments when the transmission ratio fluctuates drastically, with peaks and troughs. By analyzing the transmission ratio peaks and troughs corresponding to these key time points, quantitative indicators of the transmission ratio change trend, such as fluctuation amplitude and fluctuation period, are calculated. These precise quantitative indicators provide an indispensable basis for assessing the severity of mechanical wear, diagnosing potential fault sources, and triggering subsequent adaptive compensation control, thus realizing a closed loop from state perception to intelligent diagnosis.

[0019] Step S2: Combining the transmission ratio change trend with the historical wear data of the lead screw, the wear state of the lead screw in the current pressure roller system is evaluated by the state estimation algorithm, and the pitch accuracy offset caused by wear is calculated.

[0020] Step S2 includes: querying historical operating logs based on the transmission ratio change trend, extracting screw wear-related indicators matching the trend, including wear rate and cumulative wear amount; initializing the state vector and covariance matrix of the Kalman filter, using the screw wear-related indicators as observations, performing iterative state estimation using the Kalman filter algorithm, calculating the state prediction value and updating the prediction covariance matrix; executing the measurement update step, calculating the Kalman gain, and fusing the observations and state prediction values ​​to obtain the posterior state estimate value, and determining the final pitch accuracy offset based on the posterior state estimate value.

[0021] Specifically, during the rolling process of dry electrode materials, mechanical components such as pressure rollers and lead screws will wear down as the equipment is used for a long time. This will lead to a decrease in the control accuracy of electrode thickness, which in turn will affect the battery performance and stability. Therefore, it is necessary to estimate the wear of the equipment to provide a basis for subsequent accurate compensation.

[0022] Specifically, based on the identified abnormal transmission ratio trend, the historical operation log database is queried to extract screw wear-related indicators corresponding to past records with similar characteristics to the current abnormal pattern. These indicators mainly include wear rate (wear amount per unit time) and cumulative wear amount. These indicators, as direct observations reflecting the health status of the screw, are input into the state estimation process. Then, the Kalman filter algorithm is used for state estimation. First, the state vector and covariance matrix of the filter are initialized. The state vector contains the screw wear state variables to be estimated, and the covariance matrix quantifies the degree of uncertainty in the initial estimation of this state. Next, the algorithm enters an iterative estimation loop, each loop containing two core steps: prediction and further estimation. In the prediction step, the algorithm extrapolates the current state prediction from the posterior state of the previous moment based on a physical model describing the dynamics of screw wear, such as a linear model considering the average wear rate. Simultaneously, the prediction covariance matrix is ​​updated to reflect the new uncertainties introduced by model imperfections and process noise. Subsequently, in the measurement update step, the algorithm calculates the Kalman gain, a dynamic weight used to optimally balance the confidence between the prediction and the current observation. This gain is used to fuse the latest observations—the wear rate and cumulative wear extracted from historical logs—with the state prediction to generate a new, more accurate posterior state estimate, and the covariance matrix of the posterior estimate is updated accordingly. This "prediction-update" process iterates continuously with the arrival of new data, causing the state estimate to converge to the true value through continuous correction. Finally, based on the optimized posterior state estimate output by the Kalman filter, the pitch accuracy offset caused by the wear of the lead screw is calculated. This offset is a precisely quantified physical quantity that represents the deviation between the actual transmission accuracy of the lead screw and the ideal value, providing crucial input for the subsequent generation of high-precision compensation commands.

[0023] The above technical solution, by combining data-driven historical experience (observations) with a physics-based prediction model, effectively overcomes the limitations of relying solely on a model or observation. It can still achieve high-precision and robust estimation of the wear state of the lead screw under the interference of sensor noise and equipment operation fluctuations, thus laying a solid foundation for ensuring the thickness uniformity of the electrode sheets from the root.

[0024] Step S3: If the pitch accuracy offset exceeds the preset threshold, an adjustment coefficient is generated based on the pitch accuracy offset to correct the motor displacement command. The motor angle command sequence is updated using the adjustment coefficient, and the updated command sequence is iteratively optimized through a feedback loop to improve the actual adjustment accuracy of the pressure roller gap.

[0025] In step S3, generating adjustment coefficients for correcting motor displacement commands based on offset includes: if the pitch accuracy offset exceeds a preset threshold, inputting the pitch accuracy offset into the simulation model, calculating the meshing fluctuation amplitude caused by the pitch accuracy offset based on the simulation model, including the fluctuation frequency and peak amplitude; fitting a displacement correspondence curve based on the meshing fluctuation amplitude, determining the reciprocal of the curve slope as the initial value of the adjustment coefficient, calibrating the initial value using an optimization algorithm to obtain a refined adjustment coefficient; verifying the accuracy of the displacement correspondence through the refined adjustment coefficient, and outputting the adjustment coefficient if the accuracy is higher than the preset threshold.

[0026] Specifically, when the calculated pitch accuracy deviation exceeds the preset safety threshold such as 0.01mm, it indicates that the wear of the lead screw has had a substantial impact on the transmission accuracy, and compensation control must be initiated. In order to solve the problem of inaccurate control of the pressure roller gap caused by mechanical wear, this application generates accurate command correction coefficients by establishing a method that combines physical simulation and optimization algorithm, so as to realize feedforward compensation of motor displacement commands.

[0027] Specifically, a simulation model characterizing the dynamic characteristics of the pressure roller transmission system is constructed. This model can be simplified as a spring-mass-damping system, where the spring stiffness is equivalent to the transmission system stiffness, the mass block is equivalent to the pressure roller inertia, and the damping coefficient is equivalent to the system friction characteristics. This model simulates the system dynamic response under different wear conditions by adjusting its physical parameters. The calculated pitch accuracy offset is input into the simulation model as a key input parameter. By running the model, the dynamic response of meshing fluctuation caused by a specific offset is calculated. The output includes two key indicators: fluctuation frequency and peak amplitude. The fluctuation frequency reflects the periodic characteristics of transmission instability, while the peak amplitude directly quantifies the maximum deviation of the gap fluctuation. Subsequently, based on the data point set output by the simulation model, a mapping relationship from theoretical displacement command to peak fluctuation amplitude is established. The least squares method is used to fit the displacement correspondence curve, and the slope of this curve within the current process-defined working range is calculated. The reciprocal of the curve slope is used as the initial value of the adjustment coefficient, which is essentially a compensation factor for pre-scaling the motor displacement command. Next, a gradient descent optimization algorithm is used to calibrate the initial value of the adjustment coefficient, obtaining a refined adjustment coefficient. The specific calibration process will be explained in detail later. Finally, the refined adjustment coefficient is used for reverse verification: the corrected displacement command is re-inputted into the simulation model. If the calculated peak fluctuation amplitude is reduced by more than a preset threshold compared to before optimization, the accuracy of the displacement correspondence is determined to meet the preset requirements. Finally, the refined adjustment coefficient is output to the motor control system.

[0028] After the refining adjustment coefficient is output to the motor control system, the system immediately initiates the instruction update and closed-loop optimization process to ensure continuous and accurate control of the pressure roller gap. Specifically, the refining adjustment coefficient is multiplied by each instruction value in the original motor angle instruction sequence to generate a pre-compensated updated motor angle instruction sequence. Subsequently, the updated instruction sequence is input into a real-time feedback control loop. This loop first drives the servo actuator to act according to the instruction, and simultaneously collects the actual pressure roller gap data through a high-precision displacement sensor. The loop immediately calculates the real-time deviation between the actual output value and the expected output value set by the process. This deviation signal comprehensively reflects the residual error of feedforward compensation and any unmodeled dynamic interference. Based on this deviation, an incremental PID control algorithm is used to iteratively fine-tune the instruction sequence parameters online, correcting the residual error and unmodeled interference cycle by cycle. When the gap deviation of several consecutive cycles is stable within the preset tolerance range, the system is considered to have converged. This process, through the integration of feedforward compensation and feedback optimization, significantly improves the actual adjustment accuracy and robustness of the pressure roller gap.

[0029] The above technical solution achieves a precise conversion from wear quantification to instruction compensation by combining physical modeling and optimization algorithms. This method not only overcomes the shortcomings of poor adaptability of traditional fixed parameter compensation, but also generates the optimal compensation strategy according to the specific wear conditions, providing key technical support for improving the actual adjustment accuracy of the pressure roller gap.

[0030] The process of calibrating the initial value using an optimization algorithm to obtain the refined adjustment coefficients includes: starting with the initial value as the initial point, taking minimizing the meshing fluctuation amplitude as the objective function, and using the gradient descent method for iterative search. In each iteration, the gradient of the objective function with respect to the adjustment coefficients is calculated and the adjustment coefficients are updated in the opposite direction of the gradient. When the change in the objective function is less than the convergence threshold or the maximum number of iterations is reached, the iteration is stopped, and the optimal solution at this time is taken as the refined adjustment coefficients.

[0031] Specifically, after obtaining the initial adjustment coefficients based on the physical model, to further improve the compensation accuracy and overcome the errors caused by model simplification, this application employs a gradient descent-based optimization algorithm to refine the coefficients. This process uses the initial adjustment coefficients as the starting point for optimization and systematically iteratively searches for the refined coefficient values ​​that can most effectively suppress meshing fluctuations. Specifically, regarding the gradient descent optimization process, the initial adjustment coefficients are used as the initial point for the iterative search. In each iteration, the algorithm calculates the objective function value corresponding to the current adjustment coefficient through forward simulation, and then calculates the gradient of the objective function with respect to the adjustment coefficients using automatic differentiation or finite difference methods. This gradient vector indicates the steepest growth direction of the objective function in the parameter space. To minimize the objective function, the algorithm updates the adjustment coefficients along the opposite direction of the gradient, with the update step size controlled by a preset learning rate parameter. This iterative process continues until the optimization terminates when the change in the objective function value between two consecutive iterations is less than a preset convergence threshold (e.g., 0.001), or the number of iterations reaches the maximum limit. The adjustment coefficients obtained at this point are the refined adjustment coefficients.

[0032] In step S3, iterative optimization of the updated command sequence through a feedback loop includes: multiplying the adjustment coefficient by the original motor angle command sequence to generate an updated motor angle command sequence; inputting the updated command sequence into the feedback loop to drive the actuator and collect the actual pressure roller gap as the actual output; calculating the deviation between the actual output and the expected output; and iteratively adjusting the parameters in the motor angle command sequence based on the deviation until the deviation converges to within a preset range.

[0033] Specifically, after obtaining the initial adjustment coefficients based on the simulation model, in order to overcome the influence of model simplification and parameter errors, this application adopts a closed-loop feedback mechanism to perform online iterative optimization of the motor angle command sequence. This process combines feedforward compensation and feedback correction. By continuously comparing the actual output with the expected target, the control command is dynamically corrected to ensure that the system converges to the optimal state.

[0034] Specifically, the refining adjustment coefficient is multiplied by the original motor angle command sequence to generate a pre-compensated update command sequence. This sequence controls the actuator through a servo driver to complete the initial adjustment of the pressure roller gap. At the same time, the system initiates a feedback optimization process, using displacement sensors installed on the pressure roller assembly to collect actual gap data in real time at a sampling frequency of 10kHz. This data is then digitally filtered and used as the actual output value of the system. This actual output is compared in real time with the expected target value set by the process to calculate the deviation signal reflecting the control accuracy. This deviation value not only includes the residual error of feedforward compensation but also covers the influence of various unmodeled disturbances.

[0035] Based on this deviation signal, the system employs an improved incremental PID control algorithm to correct the command sequence online. The control mechanism of this algorithm is as follows: the proportional term rapidly responds to deviation changes, the integral term eliminates steady-state error, and the derivative term suppresses overshoot. The synergistic effect of these three factors causes the output deviation to gradually converge along the negative gradient direction of the error surface. To ensure the stability and robustness of the control system, this implementation also introduces a parameter adaptive mechanism. This mechanism dynamically adjusts the PID control parameters by monitoring the system response characteristics in real time. For example, it automatically reduces the proportional gain when the response is too fast and appropriately increases the integral coefficient when a persistent deviation occurs. Simultaneously, the system also sets up a safety guarantee strategy, automatically switching to a conservative control mode when abnormal fluctuations or excessive overshoot are detected. The convergence state is determined by monitoring the statistical characteristics of the output deviation in real time. When the output deviation remains within the preset range for multiple consecutive control cycles and the root mean square value of the deviation is less than the preset threshold, the system is determined to have reached stable convergence. This closed-loop optimization mechanism of "instruction execution - output measurement - deviation calculation - parameter correction" enables the control system to have online self-correction and adaptive adjustment capabilities, and ultimately realizes real-time precise control of the pressure roller gap of the current production task, so that the adjustment accuracy can stably meet the process requirements.

[0036] Step S4: When the adjustment accuracy meets the process requirements, obtain the dynamic deviation residual of the system, correct the dynamic deviation residual and determine the final adaptive control parameter set, and drive the actuator to adjust the pressure roller gap based on the adaptive control parameter set.

[0037] The step S4, which involves correcting the dynamic deviation residual and determining the final adaptive control parameter set, includes: obtaining the dynamic deviation residual value from the system output when the adjustment accuracy meets the process requirements; constructing a residual correction model through a compensation mechanism, inputting the dynamic deviation residual value into the correction model for correction calculation, and outputting the corrected residual value; obtaining the adaptive control parameter set based on the corrected residual value, including the gain and time constant of the control system; verifying the stability of the adaptive control parameter set, and if the stability index is higher than a preset threshold, confirming the effectiveness of the adaptive control parameter set, and using it as the final adaptive control parameter set output.

[0038] Specifically, after feedback control enables the system to achieve the required adjustment accuracy, in order to optimize the control performance of subsequent production tasks, improve the system convergence speed, and establish long-term adaptive capability, this application initiates an offline optimization process based on dynamic deviation residual analysis and parameter self-learning. Specifically, a dynamic deviation residual acquisition and analysis system is established. When the control accuracy of the pressure roller gap is detected to be continuously stable within the required range, it indicates that the system has entered the quasi-steady-state working stage. At this time, the dynamic deviation data extracted from the output of the control system refers to the sequence of small differences between the actual output value and the predicted output value obtained based on the ideal mathematical model of the control system when the control accuracy of the pressure roller gap has reached the required process and the system is in quasi-steady-state operation. It contains dynamic characteristics and external disturbance information not represented by the model, such as periodic micro-vibrations caused by wear of the transmission chain or low-frequency fluctuations caused by changes in material properties.

[0039] The specific implementation methods for residual correction and parameter optimization are as follows: First, a residual correction model based on modern control theory is constructed. This model establishes a mapping relationship between the residual sequence and the system's dynamic characteristics, enabling the identification and separation of systematic deviation components and random noise components in the residuals. After inputting the collected dynamic deviation residual values ​​into the model, a recursive least squares algorithm is used for real-time calculation, effectively filtering out measurement noise and extracting the corrected residual values ​​that reflect the true dynamic characteristics of the system. Based on the corrected residual values, the model reference adaptive method is used for controller parameter tuning. The core idea of ​​this step is to use the corrected residual information to infer the actual dynamic characteristics of the system, thereby optimizing the controller parameters. By solving the Lyapunov equations, control parameters that make the actual system's dynamic characteristics approach the ideal reference model are calculated, including gain parameters that determine the system's response speed and time constant parameters that affect system stability. These parameters together constitute the adaptive control parameter set.

[0040] Rigorous stability verification is performed immediately after obtaining the adaptive control parameter set. This step is crucial because inappropriate parameters can lead to system instability. The verification process is based on the Nyquist stability criterion. By analyzing the open-loop frequency response characteristics of the control system, the phase margin and gain margin of the system are evaluated. When the verification results show that the stability margin of the system is higher than the preset safety threshold, it indicates that the parameter set can guarantee the stability of the system under all expected operating conditions. At this point, the adaptive control parameter set is confirmed to be effective and is used as the final parameter output. Finally, the system downloads the verified adaptive control parameter set to the actuator drive system. This step completes the transformation from parameter optimization to practical application. The new parameter set, by adjusting the core algorithm parameters of the servo controller, achieves precise adjustment of the pressure roller gap, enabling the control system to automatically adapt to changes in equipment characteristics and maintain long-term stable control performance.

[0041] Step S5: Monitor the uniformity of the electrode sheet thickness after adjustment in real time. If the deviation of the thickness uniformity from the target range exceeds the preset range, return to step S1 and restart a new round of optimization cycle until the thickness uniformity index is qualified, and output the final control parameters.

[0042] The step S5, which involves real-time monitoring of the uniformity index of the electrode sheet thickness after adjustment, includes: generating a drive signal based on the adaptive control parameter set, inputting it into the actuator to adjust the gap between the pressure rollers, collecting the adjusted gap data, processing the collected gap data using filtering technology, calculating the uniformity index of the electrode sheet thickness based on the processed data, including the uniformity variance and the average thickness, and if the variance of the uniformity index is lower than the preset threshold, then the adjustment of the pressure roller gap is confirmed to be effective, and the uniformity index of the thickness is output.

[0043] Specifically, after optimizing the control system parameters, in order to establish a complete quality closed-loop control, this application constructs a final verification and feedback mechanism based on the actual quality indicators of the product. In specific implementation, the system first executes production control based on optimized parameters: generating precise drive signals according to the determined adaptive control parameter set, and driving the actuator through the servo system to make final adjustments to the gap between the pressure rollers. This step transforms the results of all the aforementioned technical optimizations into specific mechanical actions, realizing precise control of the gap between the pressure rollers. During the adjustment process, the thickness data of the electrode sheet is collected in real time through an online thickness gauge. The sampling frequency setting needs to meet the quality monitoring requirements of the production process to ensure that the true changes in the thickness of the electrode sheet can be fully reflected.

[0044] The following technical solution is adopted for the accurate calculation of thickness uniformity index: First, the collected raw thickness data is filtered in real time. An adaptive Kalman filter algorithm is used to effectively separate the actual thickness variation and measurement noise in the signal. This processing is crucial for subsequent accurate calculation. Based on the high-quality filtered data, key quality indicators are calculated through the following steps: The average thickness index is calculated by statistically averaging the thickness values ​​of all sampling points within a set evaluation period. This index reflects the overall thickness level of the electrode sheet and is a core parameter for evaluating whether the product meets basic specifications. The uniformity variance index is calculated by using an unbiased estimation method in statistics to weighted average the squared deviations of the thickness of each sampling point from the average thickness. This index quantitatively characterizes the uniformity of thickness distribution in spatial and temporal dimensions and is directly related to the consistency and performance stability of battery products. The calculated thickness uniformity index is rigorously compared with the process standard. When the uniformity variance is lower than the preset qualified threshold and the average thickness meets the product specification requirements, the system determines that the current control parameters are valid, outputs a thickness uniformity qualified report, and locks the current control parameters as the standardized parameters for this production process. This determination is based on rigorous statistical process control principles to ensure the scientific nature and reliability of the decision.

[0045] When the system detects that the thickness uniformity index exceeds the preset range, it will automatically trigger the quality feedback mechanism, record the current quality deviation data, and further analyze the spatiotemporal distribution characteristics of the deviation to identify abnormal patterns. This complete quality information will serve as the input for a new round of optimization cycle and be fed back to the transmission ratio trend analysis stage in step S1, restarting the complete process from state perception to parameter optimization. This closed-loop control mechanism based on the final product quality ensures that the system can continuously self-optimize and gradually approach the optimal control state.

[0046] The above describes a dry electrode fabrication process control method according to an embodiment of this application. The following describes a dry electrode fabrication process control system according to an embodiment of this application. Please refer to [link / reference]. Figure 2 One embodiment of a dry electrode fabrication process control system in this application includes: The data acquisition module is used to monitor and acquire data on the gap between the pressure rollers and the rotation angle of the drive motor in real time during the rolling process of dry electrode materials. The acquired data is classified and processed using a support vector machine algorithm to identify the trend of transmission ratio change of the pressure roller system. The condition assessment module is used to combine the transmission ratio change trend with the historical wear data of the lead screw, and to assess the wear condition of the lead screw in the current pressure roller system through the condition estimation algorithm, and calculate the pitch accuracy offset caused by wear. The iterative optimization module is used to generate an adjustment coefficient for correcting the motor displacement command based on the offset if the pitch accuracy offset exceeds a preset threshold. The adjustment coefficient is used to update the motor angle command sequence, and the updated command sequence is iteratively optimized through a feedback loop to improve the actual adjustment accuracy of the pressure roller gap. The correction module is used to obtain the dynamic deviation residual of the system when the adjustment accuracy meets the process requirements, correct the dynamic deviation residual and determine the final adaptive control parameter set, and drive the actuator to adjust the pressure roller gap based on the adaptive control parameter set. The loop control module is used to monitor the uniformity of the electrode sheet thickness after adjustment in real time. If the deviation of the thickness uniformity from the target range exceeds the preset range, it returns to the acquisition module and restarts a new round of optimization loop until the thickness uniformity index is qualified, and then outputs the final control parameters.

[0047] This application also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of a dry electrode preparation process control method.

[0048] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0049] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0050] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for controlling the dry electrode fabrication process, characterized in that, The method includes: Step S1: During the rolling process of dry electrode material, real-time monitoring and collection of pressure roll gap data and drive motor rotation angle data are performed. The collected data is classified and processed using a support vector machine algorithm to identify the transmission ratio change trend of the pressure roll system. Step S2: Combining the transmission ratio change trend with the historical wear data of the lead screw, the wear state of the lead screw in the current pressure roller system is evaluated by the state estimation algorithm, and the pitch accuracy offset caused by wear is calculated. Step S3: If the pitch accuracy offset exceeds a preset threshold, an adjustment coefficient for correcting the motor displacement command is generated based on the pitch accuracy offset. The motor angle command sequence is updated using the adjustment coefficient, and the updated command sequence is iteratively optimized through a feedback loop to improve the actual adjustment accuracy of the pressure roller gap. Step S4: When the adjustment accuracy reaches the process requirements, obtain the dynamic deviation residual of the system, correct the dynamic deviation residual and determine the final adaptive control parameter set, and drive the actuator to adjust the pressure roller gap based on the adaptive control parameter set. Step S5: Monitor the uniformity of the electrode sheet thickness after adjustment in real time. If the deviation of the thickness uniformity from the target range exceeds the preset range, return to step S1 and restart a new round of optimization cycle until the thickness uniformity is qualified, and output the final control parameters.

2. The method according to claim 1, characterized in that, Step S1 includes: The pressure roller gap data is collected in real time by a displacement sensor, and the rotation angle data of the drive motor is collected in real time by an encoder. The collected pressure roller gap data and rotation angle data are used as input feature vectors. A classification model is constructed using the support vector machine algorithm to classify the input feature vectors, obtain the classification result of the transmission ratio change trend, and fit the classification result to obtain the transmission ratio change trend curve, and calculate the transmission ratio change rate. If the rate of change of the transmission ratio exceeds a preset threshold, the transmission ratio change trend is marked as an abnormal trend; otherwise, it is marked as a normal trend. Based on the abnormal trend, key time points are extracted from the data, and the peak and trough values ​​of the transmission ratio change trend are analyzed through the key time points to obtain a quantitative index of the transmission ratio change trend.

3. The method according to claim 1, characterized in that, Step S2 includes: Based on the transmission ratio change trend, query the historical operation log and extract the lead screw wear-related indicators that match the trend, including wear rate and cumulative wear amount; Initialize the state vector and covariance matrix of the Kalman filter, take the screw wear-related index as the observed value, use the Kalman filter algorithm to perform iterative state estimation, calculate the state prediction value and update the prediction covariance matrix; Perform a measurement update step, calculate the Kalman gain, and fuse the observed value with the state prediction value to obtain the posterior state estimate. Based on the posterior state estimate, determine the final pitch accuracy offset.

4. The method according to claim 1, characterized in that, Step S3, which generates adjustment coefficients for correcting motor displacement commands based on the offset, includes: If the pitch accuracy offset exceeds a preset threshold, the pitch accuracy offset is input into the simulation model, and the meshing fluctuation amplitude caused by the pitch accuracy offset is calculated based on the simulation model, including the fluctuation frequency and amplitude peak value. Based on the displacement correspondence curve fitted to the meshing fluctuation amplitude, the reciprocal of the curve slope is determined as the initial value of the adjustment coefficient. The initial value is then calibrated using an optimization algorithm to obtain the refined adjustment coefficient. The accuracy of the displacement correspondence is verified by refining the adjustment coefficient. If the accuracy is higher than a preset threshold, the adjustment coefficient is output.

5. The method according to claim 4, characterized in that, The preliminary values ​​are calibrated using an optimization algorithm to obtain refined adjustment coefficients, including: Using the initial value as the starting point and minimizing the meshing fluctuation amplitude as the objective function, an iterative search is performed using the gradient descent method; In each iteration, the gradient of the objective function with respect to the adjustment coefficients is calculated and the adjustment coefficients are updated in the opposite direction of the gradient. When the change in the objective function is less than the convergence threshold or the maximum number of iterations is reached, the iteration stops and the optimal solution at this time is taken as the refined adjustment coefficients.

6. The method according to claim 5, characterized in that, Step S3 involves iteratively optimizing the updated instruction sequence using a feedback loop, including: The adjustment coefficient is multiplied by the original motor angle command sequence to generate an updated motor angle command sequence. The updated command sequence is then input into the feedback loop to drive the actuator and collect the actual pressure roller gap as the actual output. Calculate the deviation between the actual output and the expected output, and iteratively adjust the parameters in the motor angle command sequence based on the deviation until the deviation converges to within a preset range.

7. The method according to claim 1, characterized in that, Step S4, which involves correcting the dynamic deviation residual and determining the final adaptive control parameter set, includes: When the adjustment accuracy meets the process requirements, the dynamic deviation residual value is obtained from the system output; A residual correction model is constructed through a compensation mechanism. The dynamic deviation residual value is input into the correction model for correction calculation, and the corrected residual value is output. The adaptive control parameter set, including the gain and time constant of the control system, is obtained based on the corrected residual values. The stability of the adaptive control parameter set is verified. If the stability index is higher than the preset threshold, the adaptive control parameter set is confirmed to be valid and is used as the final adaptive control parameter set output.

8. The method according to claim 1, characterized in that, Step S5 involves real-time monitoring of the uniformity of electrode thickness after adjustment, including the following indicators: A drive signal is generated based on the adaptive control parameter set, input to the actuator to adjust the gap between the pressure rollers, and the adjusted gap data is collected. The collected gap data is processed using filtering technology, and the thickness uniformity index of the electrode sheet is calculated based on the processed data, including uniformity variance and average thickness. If the variance of the uniformity index is lower than the preset threshold, the adjustment of the pressure roller gap is confirmed to be effective, and the thickness uniformity index is output.

9. A dry electrode fabrication process control system, used to implement the dry electrode fabrication process control method as described in any one of claims 1-8, characterized in that, The system includes: The data acquisition module is used to monitor and acquire data on the gap between the pressure rollers and the rotation angle of the drive motor in real time during the rolling process of dry electrode materials. The acquired data is classified and processed using a support vector machine algorithm to identify the trend of transmission ratio change of the pressure roller system. The condition assessment module is used to combine the transmission ratio change trend with the historical wear data of the lead screw, and to assess the wear state of the lead screw in the current pressure roller system through a condition estimation algorithm, and to calculate the pitch accuracy offset caused by wear. The iterative optimization module is used to generate an adjustment coefficient for correcting the motor displacement command based on the offset if the pitch accuracy offset exceeds a preset threshold, update the motor angle command sequence using the adjustment coefficient, and iteratively optimize the updated command sequence through a feedback loop to improve the actual adjustment accuracy of the pressure roller gap. The correction module is used to obtain the dynamic deviation residual of the system when the adjustment accuracy reaches the process requirements, correct the dynamic deviation residual and determine the final adaptive control parameter set, and drive the actuator to adjust the pressure roller gap based on the adaptive control parameter set. The loop control module is used to monitor the uniformity of the electrode sheet thickness after adjustment in real time. If the deviation of the thickness uniformity from the target range exceeds the preset range, the module returns to the acquisition module and restarts a new round of optimization loop until the thickness uniformity is qualified, and then outputs the final control parameters.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement a dry electrode fabrication process control method as described in any one of claims 1-8.

Images