Pump-controlled lithium battery pole piece rolling mill control method, device, equipment and storage medium

Abstract

The application provides a pump-controlled lithium battery pole piece rolling mill control method, device, equipment and storage medium, and relates to the technical field of active disturbance rejection control. The method comprises the following steps: receiving a desired displacement signal, inputting the desired displacement signal and a first expansion and contraction factor of a current control period into a fuzzy active disturbance rejection controller, and calculating a current target value and a current estimated value of the current control period; calculating a second expansion and contraction factor of the current control period according to the current target value and the current estimated value; calculating a current control amount of the current control period according to the current estimated value, the current target value and the second expansion and contraction factor of the current control period, and controlling a hydraulic cylinder of the pump-controlled lithium battery pole piece rolling mill according to the current control amount. The application can realize high-precision adaptive control of the pump-controlled lithium battery pole piece rolling mill in a full working condition range.

Description

Technical Field

[0001] This invention relates to the field of active disturbance rejection control technology, and in particular to a pump-controlled lithium battery electrode rolling mill control method, apparatus, equipment and storage medium. Background Technology

[0002] The performance of lithium-ion batteries is highly dependent on the manufacturing precision of the electrodes. The compaction density and thickness consistency of the electrodes within a lithium-ion battery are key parameters determining battery capacity, operational stability, and cycle life. The electrode rolling process is a core production step that ensures these indicators meet standards and improves the overall quality of the battery. Currently, battery electrode rolling mills generally use electro-hydraulic servo systems to achieve automatic thickness control. However, as the new energy industry continues to increase its requirements for battery production precision, energy efficiency, and process stability, the rolling mill control system must simultaneously meet the operational needs of high-precision thickness control, rapid dynamic response, and high energy efficiency under complex operating conditions such as sudden changes in rolling force, uneven incoming material thickness, and time-varying hydraulic parameters. This places higher demands on the global robustness and adaptive capability of the control algorithm.

[0003] To address the shortcomings of traditional valve-controlled electro-hydraulic thickness control systems, such as low energy efficiency, complex structure, and sensitivity to oil contamination, electro-hydraulic servo pump control technology, with its advantages of high efficiency, energy saving, and high power density, is gradually being applied to battery electrode rolling mills, effectively reducing equipment installation and operating costs. However, pump-controlled servo systems are typical electromechanical-hydraulic coupled nonlinear systems, exhibiting characteristics such as uncertain model parameters, large load characteristic fluctuations, and multi-source disturbances. Traditional linear control methods are difficult to adapt to high-precision control requirements. Active disturbance rejection control (ADRC) does not rely on a precise mathematical model of the controlled object and can estimate and compensate for the total system disturbance in real time through an extended state observer, exhibiting strong robustness. However, its parameter tuning relies on manual experience, and fixed parameter structures are difficult to adapt to a wide range of operating conditions. Therefore, researchers have combined fuzzy logic with ADRC to form a fuzzy ADRC scheme to achieve parameter self-tuning.

[0004] However, the inventors discovered that the fuzzy active disturbance rejection controllers in the prior art usually use a fixed input-output domain of discourse. When the system operating conditions deviate significantly from the preset range, the fuzzy inference sensitivity decreases, the dynamic response lags, and the control accuracy degrades, making it difficult to meet the control requirements of pump-controlled rolling mills in highly time-varying and wide-condition scenarios. Summary of the Invention

[0005] This invention provides a pump-controlled lithium battery electrode rolling mill control method, apparatus, equipment, and storage medium to solve the problem that existing fuzzy active disturbance rejection controllers typically use fixed input-output domains, resulting in decreased fuzzy inference sensitivity, delayed dynamic response, and degraded control accuracy when the system operating conditions deviate significantly from the preset range.

[0006] In a first aspect, embodiments of the present invention provide a pump-controlled lithium battery electrode rolling mill control method, comprising: Receive the desired displacement signal and input the desired displacement signal and the first scaling factor of the current control cycle into the fuzzy active disturbance rejection controller to calculate the current target value and the current estimated value of the current control cycle; Calculate the second scaling factor for the current control cycle based on the current target value and the current estimated value; Based on the current estimated value, the current target value, and the second scaling factor of the current control cycle, calculate the current control quantity for the current control cycle, and control the hydraulic cylinder of the pump-controlled lithium battery electrode rolling mill according to the current control quantity.

[0007] Secondly, embodiments of the present invention provide a pump-controlled lithium battery electrode rolling mill control device, comprising: The calculation module is used to receive the desired displacement signal and input the desired displacement signal and the first scaling factor of the current control cycle into the fuzzy active disturbance rejection controller to calculate the current target value and the current estimated value of the current control cycle. The scaling module is used to calculate the second scaling factor for the current control cycle based on the current target value and the current estimated value. The control module is used to calculate the current control quantity for the current control cycle based on the current estimated value, the current target value, and the second scaling factor of the current control cycle, and to control the hydraulic cylinder of the pump-controlled lithium battery electrode rolling mill based on the current control quantity.

[0008] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect or any possible implementation thereof.

[0009] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect or any possible implementation thereof.

[0010] In this embodiment of the invention, by receiving the desired displacement signal and inputting it along with the first scaling factor of the current control cycle into the fuzzy active disturbance rejection controller, the tracking differentiator can arrange a smooth transition process for the desired displacement. The extended state observer can determine the current gain within the variable universe framework after the dynamic adjustment of the first scaling factor, thereby calculating the current target value and the current estimated value, providing smooth instructions and accurate states for control decisions. Then, based on the deviation between the current target value and the current estimated value, the second scaling factor of the current control cycle is calculated, enabling the second scaling factor to track the dynamic characteristics of the error signal and the rate of change of error in real time. Furthermore, the second scaling factor is used to adjust the fuzzy universe of the nonlinear state error feedback law, ensuring that the fuzzy inference always operates within the optimal sensitivity range under the current error state. Finally, the current control quantity is calculated based on the current estimated value, the current target value, and the second scaling factor of the current control cycle. By using feedforward compensation of the total disturbance estimate to eliminate internal uncertainties and external disturbances in the system, the accurate generation of the control quantity is achieved, and the hydraulic cylinder is effectively controlled. This realizes high-precision adaptive control of the pump-controlled lithium battery electrode rolling mill across the entire operating range. Attached Figure Description

[0011] Figure 1 This is an application scenario diagram of the pump-controlled lithium battery electrode rolling mill control method provided in the embodiments of the present invention; Figure 2 This is a flowchart illustrating the implementation of the pump-controlled lithium battery electrode rolling mill control method provided in this embodiment of the invention. Figure 3 This is a flowchart illustrating the implementation of step S210 of the pump-controlled lithium battery electrode rolling mill control method provided in this embodiment of the invention. Figure 4 This is a schematic diagram of the structure of the pump-controlled lithium battery electrode rolling mill control device provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0012] Pump-controlled servo systems are typical electromechanical-hydraulic coupled nonlinear systems, exhibiting inherent characteristics such as model parameter uncertainty, drastic load characteristic variations, and multi-source disturbances. Traditional linear control methods struggle to achieve high-precision control of pump-controlled servo systems across the entire operating range. While active disturbance rejection control (ADRC) can estimate and compensate for total disturbances in real time using an extended state observer, its fixed parameter structure limits its adaptability under wide operating conditions. To address this, fuzzy ADRC utilizes fuzzy rules to achieve parameter self-tuning. However, conventional schemes employ fixed input-output universes of discourse, leading to varying control performance degradation mechanisms under different disturbance types during the variable actual conditions of battery electrode rolling.

[0013] However, the inventors discovered that under rated operating conditions, the pump control servo system is stable near the preset operating point, and both the input error and the rate of change of error fall within the effective coverage range of the fuzzy domain. The quantization factor and the scaling factor are well matched, and the fuzzy inference has sufficient resolution. Therefore, the fuzzy active disturbance rejection controller can obtain ideal steady-state accuracy and dynamic response.

[0014] When the rolling force changes abruptly, the load surges instantaneously, causing the piston displacement error to increase dramatically. The error signal exceeds the upper bound of the fixed domain of discourse, and the fuzzy subset's ability to distinguish the input quantity drops sharply. The inference process enters the saturation region, where errors of different magnitudes are mapped to the same fuzzy level. The fuzzy active disturbance rejection controller cannot output differentiated gain corrections based on the actual degree of deviation, and the pump-controlled servo system exhibits slow response or oscillation.

[0015] When the elastic modulus of the oil decreases due to changes in temperature or gas content, the equivalent stiffness of the pump-controlled servo system decreases, causing the open-loop gain and natural frequency to drift. The estimated bandwidth preset by the fixed-parameter extended state observer is severely mismatched with the actual system bandwidth, resulting in phase lag and amplitude errors in the extended state observer's estimation of the total disturbance. If the fuzzy controller still uses a fixed universe of discourse, its output gain correction cannot effectively compensate for the estimation bias of the extended state observer; instead, it may exacerbate the control lag due to an inappropriate proportional factor, leading to increased thickness deviation.

[0016] When the thickness of the incoming material fluctuates unevenly, the load disturbance exhibits a wide-bandwidth, large-amplitude random characteristic. The error signal frequently crosses the zero point and extreme value region. The fixed domain of discourse has insufficient resolution when the error is small, making it difficult to achieve fine adjustment. However, it is prone to saturation when the error is large, and cannot simultaneously meet the dual requirements of rapid correction of large deviations and steady-state convergence of small deviations.

[0017] Although the impact mechanisms of the four typical operating conditions mentioned above differ, they all stem from the fact that a fixed universe of discourse structure cannot dynamically adjust the quantification scale of the error and the rate of change of error according to changes in operating conditions. This prevents fuzzy inference from always operating within its optimal sensitivity range. If the universe of discourse is fixed, no matter how sophisticated the subsequent fuzzy rules are designed, once the input mapping of the perception layer fails, the optimization effect of the control decision will be significantly weakened. Therefore, it is necessary to introduce a variable universe of discourse mechanism. By dynamically expanding or shrinking the input and output universes of discourse based on the real-time error magnitude through a scaling factor, fuzzy inference can maintain effective identification capability under large deviations and sufficient resolution under small deviations. This allows for adaptive perception layer adjustment to cope with multi-source and multi-form non-standard disturbances, fundamentally improving the controller's global robustness and accuracy adaptability.

[0018] Furthermore, the inventors discovered that even after introducing a variable universe of discourse mechanism, directly designing a uniform scaling factor for the entire controller is still insufficient to solve the problem. This is because the changes in system stiffness and operating conditions during the battery electrode rolling process have fundamentally different effects on the action path and magnitude of the expanded state observer and the NLSEF, leading to drastically different variation patterns in the actual scales of their input and output variables.

[0019] When the elastic modulus of the hydraulic fluid decreases or a sudden change in rolling force causes a significant reduction in the equivalent stiffness of the system, the observation error input to the extended state observer is essentially the difference between the actual displacement and the displacement estimated by the model. Under the same load disturbance, lower stiffness will cause the piston to produce a larger displacement offset, thus amplifying the magnitude of the observation error by a factor of two or even an order of magnitude. This means that the change in stiffness directly and drastically alters the dynamic range of the input signal to the extended state observer. If the scaling factor cannot expand the domain of discourse proportionally, the extended state observer will saturate and become unable to effectively identify the true magnitude of the disturbance.

[0020] For the nonlinear state error feedback process, the input is the tracking error and its derivative arranged through the transient process. Although load disturbances also increase the tracking error, the amplification magnitude depends not only on the system stiffness but also on the damping ratio, command smoothness, and the nonlinear gain of the electro-hydraulic servo system. Its scale variation is not proportional to the observation error of the extended state observer. Especially during the steady-state fine-tuning phase, the nonlinear state error feedback needs to maintain high resolution within a small error range to suppress minute thickness fluctuations. If the large expansion requirement of the extended state observer forces the nonlinear state error feedback's domain to be excessively amplified synchronously, the small error will be compressed into a very small fuzzy subset, resulting in the feedback control law being unable to make fine gain adjustments, and the steady-state accuracy deteriorating drastically.

[0021] In summary, existing pump-controlled lithium battery electrode rolling mill control methods struggle to balance robustness under all operating conditions with high-precision control, necessitating an improved control scheme that can adaptively adjust and effectively address nonlinear and multi-source disturbances.

[0022] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0023] Figure 1 This diagram illustrates an application scenario of the pump-controlled lithium battery electrode rolling mill control method provided in this embodiment of the invention. (Example:) Figure 1 As shown, after improving the fuzzy active disturbance rejection controller using the method provided in the embodiments of the present invention, the improved fuzzy active disturbance rejection controller includes a tracking differentiator, an extended state observer, a nonlinear state error feedback mechanism, and four fuzzy controllers.

[0024] See Figure 2The flowchart illustrating the implementation of the pump-controlled lithium battery electrode rolling mill control method provided in this embodiment of the invention is described in detail below: Step S210: Receive the desired displacement signal and input the desired displacement signal and the first scaling factor of the current control cycle into the fuzzy active disturbance rejection controller to calculate the current target value and the current estimated value of the current control cycle.

[0025] In some embodiments, the desired displacement signal is a target displacement command set by the operator or the upper control system according to the process requirements of lithium battery electrode rolling. This command is input to the fuzzy active disturbance rejection controller in the form of analog or digital quantities, reflecting the desired position of the hydraulic cylinder piston rod. The hydraulic cylinder of the pump-controlled lithium battery electrode rolling mill is used to drive the rolls to apply rolling force to the electrode. The piston displacement of the hydraulic cylinder directly determines the amount of roll reduction, which in turn affects the compaction density and thickness consistency of the electrode. Therefore, precise control of the hydraulic cylinder displacement is the core link to achieve high-precision control of electrode thickness. The fuzzy active disturbance rejection controller is the main execution body of the control method of this invention. It integrates three main functional modules: a tracking differentiator, an extended state observer, and a nonlinear state error feedback law. The three modules work together to complete the tracking control of the desired displacement signal.

[0026] It should be noted that the first scaling factor is a key parameter in variable universe fuzzy control, used to adjust the range of the input and output universes of discourse used for fuzzy inference during the determination of the expanded state observer gain within the current control cycle. The value of the first scaling factor is pre-calculated through fuzzy inference based on the observer's estimated deviation state in the previous control cycle, and its magnitude reflects the degree of deviation of the observer's estimated deviation from the preset range in the previous control cycle. When the observation error in the previous control cycle is large, the value of the first scaling factor is increased accordingly to expand the fuzzy universe of discourse in the current cycle, so that large-amplitude observation errors can be effectively mapped to the fuzzy subset; when the observation error is small, the value of the first scaling factor is decreased accordingly to shrink the fuzzy universe of discourse, so that fuzzy inference maintains sufficient sensitivity under small deviations.

[0027] See Figure 3 The specific processing method of the above step S210 includes steps S2101 to S2103, the details of which are as follows: Step S2101: Based on the tracking differentiator in the fuzzy active disturbance rejection controller, the desired displacement signal is discretely calculated to obtain the current target value of the current control cycle.

[0028] In some embodiments, the tracking differentiator functions to provide a smooth transition for the desired displacement signal. By transforming abrupt input signals into a smooth transition, the entire hydraulic servo control system of the pump-controlled lithium battery electrode rolling mill can smoothly track command changes, avoiding overshoot and oscillation caused by sudden command changes. Internally, the tracking differentiator uses a fastest synthesis function to process the desired displacement signal. This function effectively suppresses overshoot while maintaining tracking speed, outputting a softened tracking signal and its derivative signal. The discrete form of the tracking differentiator is as follows: in, For the first The expected displacement signal for each control cycle. For the first Target displacement tracking value for each control cycle For the first The target differential tracking value for each control cycle The integration step size is... For the velocity factor, the function The fastest synthesis function is defined as follows: in, Indicates the threshold parameter. Indicates the current speed at Predicted displacement values ​​within a time period Indicates the total error. , They represent intermediate variables, Represents a comprehensive variable. Represents a symbolic function. Represents a switching function, and when hour ,when hour , This represents the final output value of the fastest synthesis function.

[0029] Through the above calculations, the tracking differentiator transforms the abrupt desired displacement signal into a continuous and smooth tracking signal and its derivative signal, denoted as the target displacement tracking value and the target derivative tracking value, respectively. Together, they constitute the current target value for the current control cycle. The target displacement tracking value represents the ideal displacement position that the system should achieve in the current control cycle, while the target derivative tracking value represents the rate of change of this ideal position. Both provide smooth and forward-looking instruction information for subsequent error calculations.

[0030] Step S2102: Determine the gain of the current control cycle based on the historical estimate, historical displacement, and the first scaling factor of the current control cycle from the previous control cycle.

[0031] In some embodiments, the historical estimate is a collective term for the displacement estimate, velocity estimate, and total disturbance estimate output by the extended state observer in the previous control cycle, reflecting the observer's estimation of the system state in the previous cycle. The historical displacement is the actual displacement of the hydraulic cylinder acquired by the displacement sensor in the previous control cycle, reflecting the true output of the system in the previous cycle. The first scaling factor of the current control cycle is a parameter pre-calculated based on the observation error state at the end of the previous control cycle, used to adjust the domain range of fuzzy inference during the gain determination process of the current cycle.

[0032] It should be noted that the gain of the extended state observer includes , and Three parameters correspond to the observation gains of the displacement estimation, velocity estimation, and total disturbance estimation channels, respectively. These parameters determine the observer's estimation speed and accuracy of the system state and total disturbance. The gain determination process employs fuzzy inference, using historical observation errors and their rate of change as input variables. Within the target fuzzy universe of discourse framework determined by the first scaling factor of the current control cycle, fuzzy inference is performed according to a preset fuzzy rule control table, ultimately outputting the gain value to be adopted in the current control cycle. Through this method, the gain determination considers both the observer's estimation bias in the previous cycle and ensures the sensitivity of fuzzy inference under the adjustment of the variable universe of discourse mechanism, enabling the gain parameters to adaptively adjust with changes in the system state.

[0033] In one possible implementation, step S2102 is specifically processed as follows: the difference between the historical displacement estimate in the historical estimate of the previous control cycle and the historical displacement of the previous control cycle is determined as the historical observation error of the previous control cycle; the historical observation error of the previous control cycle is differentiated to obtain the historical observation error change rate of the previous control cycle; the first scaling factor of the current control cycle is multiplied by the preset fuzzy universe to obtain the target fuzzy universe; wherein, the first scaling factor of the current control cycle is obtained by fuzzy calculation based on the historical observation error, the historical observation error change rate and the first fuzzy rule control table of the previous control cycle; fuzzy calculation is performed based on the historical observation error, the historical observation error change rate and the target fuzzy universe of the previous control cycle to obtain the gain of the current control cycle.

[0034] In some embodiments, the historical displacement estimate is the estimated value of the actual system displacement output by the extended state observer in the previous control cycle, and the historical displacement is the actual system displacement value collected by sensors in the previous control cycle. The historical observation error reflects the estimation deviation of the extended state observer of the actual system displacement in the previous control cycle. When there is a deviation between the observer's estimate and the actual value, it indicates that the observer's state estimation is not accurate enough, and the gain needs to be adjusted to improve the estimation accuracy. When the deviation is small, it indicates that the observer's estimation is relatively accurate, and the current gain setting is reasonable. The magnitude of the observation error directly reflects the degree of change in the system state and the degree of matching of the observer parameters, and is an important basis for adjusting the observer gain.

[0035] It should be noted that the rate of change of observation error reflects the trend of estimation bias over time. When the observation error increases, the rate of change is positive, indicating that the estimation bias is widening and the observer's tracking ability may be insufficient, requiring an increase in gain to improve response speed. When the observation error decreases, the rate of change is negative, indicating that the estimation bias is narrowing and the observer's estimation is becoming more accurate, allowing for a reduction in gain to avoid noise amplification. When the observation error remains constant, the rate of change is zero, indicating that the estimation bias is in a stable state and the current gain setting is appropriate. The rate of change of observation error, together with the observation error itself, constitutes two dimensions describing the state of the observer's estimation bias. They comprehensively reflect the observer's estimation quality from the aspects of amplitude and trend, respectively, providing complete input information for fuzzy inference of gain.

[0036] In some embodiments, the preset fuzzy universe of discourse is a pre-defined range of the basic universes of discourse for the input and output variables. This range is determined through theoretical analysis or experimental debugging under the rated operating conditions of the system, reflecting the normal value range of each input and output variable under the rated operating conditions. The first scaling factor is obtained by fuzzy calculation based on the historical observation error, the rate of change of the historical observation error, and the first fuzzy rule control table of the previous control cycle. Its value reflects the degree of deviation of the observer estimation bias from the preset range in the previous control cycle. The first fuzzy rule control table is shown in Table 1: Table 1 First Fuzzy Rule Control Table Multiplying the first scaling factor by the preset fuzzy universe of discourse is equivalent to scaling the basic universe of discourse proportionally: when the first scaling factor is greater than 1, the universe of discourse is expanded, the mapping range of the input variables increases, and large-amplitude observation errors can be effectively mapped to the fuzzy subset; when the first scaling factor is less than 1, the universe of discourse is shrunk, the mapping range of the input variables decreases, and small-amplitude observation errors can still obtain sufficient discrimination in the fuzzy domain.

[0037] By using historical observation errors and their rate of change as input variables for fuzzy inference, fuzzification is performed within the framework of the target fuzzy domain. Inference is then conducted according to a pre-defined fuzzy rule control table, and finally, the gain value for the current control cycle is obtained through defuzzification. This gain value is the observation gain parameter used by the extended state observer in the current control cycle. , and Through the above processing, the observation error and its rate of change from the previous control cycle are used as the basis for determining the gain of the current cycle. This provides historical information to support the gain adjustment and allows it to reflect the continuous changing trend of the system state. The dynamic adjustment of the fuzzy domain by the first scaling factor ensures that the fuzzy inference process of the gain is always within the optimal sensitivity range, avoiding the saturation distortion problem that occurs when the observation error deviates significantly from the preset range in a fixed domain. The information from both the observation error and the rate of change of the observation error participates in the fuzzy inference, making the gain adjustment more comprehensive and accurate.

[0038] In one possible implementation, the specific processing method for "performing fuzzy calculations based on the historical observation error, the rate of change of historical observation error, and the target fuzzy universe of discourse to obtain the gain of the current control cycle" is as follows: using trigonometric functions as membership functions, the historical observation error and the rate of change of historical observation error of the previous control cycle are input into a preset fuzzy rule control table to obtain at least one fuzzy subset and the membership degree corresponding to each fuzzy subset; based on the target fuzzy universe of discourse, the fuzzy sub-interval corresponding to each fuzzy subset is determined; based on each fuzzy sub-interval, the membership degree of each fuzzy sub-interval, and a preset defuzzification algorithm, defuzzification is performed to obtain the gain of the current control cycle.

[0039] In some embodiments, the membership function is a function used in a fuzzy set to describe the degree to which an element belongs to the set, and its value ranges from 0 to 1. Trigonometric functions, as a commonly used form of membership function, have the advantages of simple calculation and easy shape adjustment. In this scheme, seven fuzzy subsets are defined for both historical observation error and historical observation error change rate: negative large (NB), negative medium (NM), negative small (NS), zero (ZO), positive small (PS), positive medium (PM), and positive large (PB). By substituting the specific values ​​of historical observation error and historical observation error change rate into the trigonometric membership function, their membership degrees to each fuzzy subset can be calculated. The preset fuzzy rule control table is a pre-defined set of rules describing the gain values ​​under different combinations of observation errors and observation error change rates. This table is pre-defined based on expert experience and system characteristics. The preset fuzzy rule control table includes three fuzzy rule control tables, as shown in Table 2. The fuzzy rule control table, Table 3 is The fuzzy rule control table, Table 4 is The fuzzy rule control table, where e is the historical observation error and ec is the rate of change of the historical observation error: Table 2 Fuzzy rule control table Table 3 Fuzzy rule control table Table 4 Fuzzy rule control table It should be noted that the target fuzzy universe is obtained by multiplying the first scaling factor by the preset fuzzy universe, and its range may expand or shrink compared to the preset fuzzy universe. After the target fuzzy universe is determined, the specific numerical range corresponding to each fuzzy subset is also determined. For example, when the preset fuzzy universe is [-6, 6] and the first scaling factor is 1.5, the target fuzzy universe becomes [-9, 9], and the numerical range corresponding to each fuzzy subset expands accordingly. This is determined by the preset fuzzy rule control table. , and Next, these three values ​​need to be added to the corresponding reference gain to obtain the gain.

[0040] When the elastic modulus of the hydraulic fluid decreases due to increased temperature or changes in gas content, the equivalent stiffness of the system decreases, and the hydraulic natural frequency drifts. A mismatch occurs between the estimated bandwidth preset by the fixed-parameter extended state observer and the actual system bandwidth, resulting in phase lag and amplitude errors in the observer's estimation of the total disturbance. The first scaling factor is determined based on the dynamic characteristics of historical observation errors and their rate of change. By expanding or contracting the fuzzy domain, the gain is equivalently adjusted, enabling the observation bandwidth of the extended state observer to adaptively reconstruct with changes in system stiffness. When the observation error increases sharply, the first scaling factor expands the domain to accelerate gain adjustment and quickly approximate the optimal observation parameters under the new operating conditions. When the observation error tends to stabilize, the first scaling factor contracts the domain to finely adjust the gain and suppress amplified observation noise.

[0041] In some embodiments, defuzzification is the process of converting fuzzy inference results into specific numerical values. This invention employs the centroid method for defuzzification, which calculates the geometric centroid of the fuzzy output set and uses the value corresponding to this centroid as the final gain value. The centroid method comprehensively considers all activated fuzzy rules and their membership degrees, resulting in a smoother and more reasonable output.

[0042] By employing the above methods, the input variables are fuzzified using trigonometric membership functions, ensuring that even minute changes in the input are smoothly reflected in the fuzzy domain. A pre-defined fuzzy rule control table solidifies expert experience and system characteristics in the form of rules, providing a decision-making basis for fuzzy inference. Dynamic adjustment of the target fuzzy domain allows the partitioning of fuzzy subsets to adapt to changes in system state. The centroid method for defuzzification guarantees the continuity and rationality of the output results. These various steps work together to ensure that the gain determination process is both guided by expert experience and adaptable to changes in operating conditions.

[0043] Step S2103: Calculate the current estimated value of the current control cycle based on the real-time displacement of the hydraulic cylinder, the current target value, and the gain of the current control cycle.

[0044] In some embodiments, the real-time displacement is the actual displacement signal acquired in real time by a displacement sensor mounted on the hydraulic cylinder. This signal serves as the feedback input of the extended state observer, reflecting the true output of the system. The target displacement tracking value in the current target value serves as the reference input of the extended state observer, reflecting the ideal state that the system should achieve in the current control cycle. Under the influence of the gain of the current control cycle, the extended state observer recursively calculates the displacement estimate, velocity estimate, and total disturbance estimate of the current control cycle based on the deviation between the real-time displacement and the historical estimate, using the state equation. The continuous-form state equation of the extended state observer is: in, For real-time displacement, This is the nominal value of the control gain. This is the displacement estimate. This is a speed estimate. This is the estimated total disturbance. , , The gain for the current control cycle. For the filter factor, and For tracking factors (take respectively) < (to enhance the ability to estimate the total disturbance). For nonlinear functions, when > When, nonlinear function The definition of is: when ≤ When, nonlinear function The definition of is: The extended state observer treats unmodeled dynamics, nonlinear friction, parameter perturbations, and external load disturbances as a total disturbance and expands them into new state variables. Perform real-time estimation.

[0045] The displacement estimate is the observed actual displacement of the system, the velocity estimate is the observed velocity of the system, and the total disturbance estimate is a comprehensive estimate of all uncertainties, including unmodeled dynamics within the system, parameter perturbations, and external load disturbances. The displacement and velocity estimates are used for subsequent error calculations, while the total disturbance estimate is used for feedforward compensation of the control input. Together, the displacement, velocity, and total disturbance estimates constitute the current estimate for the current control cycle.

[0046] Step S220: Calculate the second scaling factor for the current control cycle based on the current target value and the current estimated value.

[0047] In some embodiments, the current target value includes the target displacement tracking value and the target derivative tracking value output by the tracking differentiator, and the current estimated value includes the displacement estimate and velocity estimate output by the extended state observer. The second scaling factor is a parameter used for adjusting the fuzzy derivation domain of the nonlinear state error feedback law in the current control cycle, and its function is similar to that of the first scaling factor. The first scaling factor is used in the gain determination process of the extended state observer, while the second scaling factor is used in the parameter determination process of the nonlinear state error feedback law.

[0048] In one possible implementation, the specific processing method of the above step S220 includes: calculating the error signal and error change rate of the current control cycle based on the current target value and the current estimated value; and performing fuzzy calculation based on the error signal, error change rate and the second fuzzy rule control table of the current control cycle to obtain the second scaling factor.

[0049] In some embodiments, the target displacement tracking value is subtracted from the displacement estimate, and the difference is the error signal for the current control cycle. The error signal reflects the current position tracking deviation of the system. When the error signal is positive, it indicates that the actual displacement has not yet reached the expected displacement; when the error signal is negative, it indicates that the actual displacement has exceeded the expected displacement; when the error signal is zero, it indicates that the actual displacement is exactly equal to the expected displacement. The target differential tracking value is subtracted from the velocity estimate, and the difference is the error change rate for the current control cycle. The error change rate reflects the trend of the tracking deviation. When the error change rate is positive, it indicates that the deviation is increasing; when the error change rate is negative, it indicates that the deviation is decreasing; when the error change rate is zero, it indicates that the deviation remains constant.

[0050] In some embodiments, the second fuzzy rule control table is a pre-defined set of rules describing the values ​​that the scaling factor should take under different combinations of error signals and error change rates. The second fuzzy rule control table is shown in Table 5: Table 5 Second Fuzzy Control Rule Table Specifically, the error signal and error rate of change are used as input variables for fuzzy inference. After fuzzification, they are transformed into fuzzy subsets and their membership degrees. Then, fuzzy inference is performed according to the second fuzzy rule control table. Finally, the specific value of the second scaling factor is obtained through defuzzification. This second scaling factor will be used to adjust the fuzzy input-output universe of discourse of the nonlinear state error feedback law in the current control cycle.

[0051] In this way, the error signal and the rate of change of error comprehensively describe the control deviation state of the system from both amplitude and trend dimensions, providing complete state information for subsequent fuzzy inference and control quantity calculation. The error signal reflects the magnitude of the current deviation, determining the strength of the required control action; the rate of change of error reflects the direction of deviation change, determining the direction of control action adjustment. The two work together to enable control decisions to respond to the current deviation amplitude and to anticipate the deviation's changing trend for early adjustments. Simultaneously, the calculation of the observation error and the rate of change of observation error provides input information for gain adjustment in the next cycle, realizing continuous recursion of the control cycle.

[0052] Step S230: Calculate the current control quantity for the current control cycle based on the current estimated value, the current target value, and the second scaling factor for the current control cycle, and control the hydraulic cylinder of the pump-controlled lithium battery electrode rolling mill based on the current control quantity.

[0053] In some embodiments, the current estimate includes displacement estimate, velocity estimate, and total disturbance estimate; the current target value includes target displacement tracking value and target derivative tracking value; and the current control quantity is the control signal ultimately output to the controlled object (i.e., the hydraulic cylinder). This signal drives a fixed displacement pump via a servo motor, thereby controlling the piston movement of the hydraulic cylinder. The calculation of the current control quantity requires comprehensive consideration of information from three aspects: system tracking error, state estimation, and disturbance compensation.

[0054] In one possible implementation, the specific processing method of step S230 above includes: performing fuzzy calculation based on the error signal, the error change rate, and the second scaling factor of the current control cycle to obtain the damping factor and control quantity gain of the current control cycle; determining the control quantity calculation formula of the current control cycle according to the damping factor and control quantity gain; inputting the target displacement tracking value, the target differential tracking value, the displacement estimate, and the velocity estimate into the control quantity calculation formula to obtain the initial control quantity of the current control cycle; and determining the difference between the initial control quantity and the preset ratio as the current control quantity; wherein the preset ratio is the ratio of the total disturbance estimate to the preset compensation coefficient.

[0055] In some embodiments, fuzzy calculations are performed based on the error signal, the error rate of change, and a second scaling factor for the current control cycle to obtain the damping factor and control gain for the current control cycle. The error signal is the difference between the target displacement tracking value and the displacement estimate, and the error rate of change is the difference between the target differential tracking value and the velocity estimate. The second scaling factor is used to adjust the input-output domain of the fuzzy inference in the current control cycle. Specifically, the error signal and the error rate of change are used as input variables for fuzzy inference. Fuzzification is performed within the target fuzzy domain framework adjusted by the second scaling factor. Inference is performed according to the fuzzy rule control table of the damping factor and control gain. Finally, the correction values ​​of the damping factor and control gain are obtained through defuzzification. The correction values ​​are then added to their respective reference values ​​to obtain the damping factor and control gain for the current control cycle. The fuzzy rule control tables of the damping factor and control gain are shown in Tables 6 and 7. Table 6. Fuzzy Rule Table for Damping Factor Table 7. Fuzzy Rule Table for Control Variable Gain Based on damping factor and control quantity gain The formula for calculating the control quantity in the current control cycle is determined. Specifically, the damping factor and control quantity gain are substituted into the expression of the nonlinear state error feedback law. The nonlinear state error feedback law adopts the form of the fastest synthesis function, which takes the error signal and error rate of change as inputs, and the damping factor and control quantity gain as adjustable parameters, and outputs the initial control quantity. Target displacement tracking value Target differential tracking value Displacement estimates and speed estimates Inputting this into the control quantity calculation formula yields the initial control quantity for the current control cycle. The initial control input is a nonlinear combination result based on the state error, reflecting the control action required by the system to eliminate the current deviation. Subtracting the preset ratio from the initial control input yields the current control input. The preset ratio is the total disturbance estimate. With preset compensation coefficient The ratio. Total disturbance estimate. It is a comprehensive estimate of the system's internal uncertainties and external disturbances output by the extended state observer, and the preset compensation coefficient is the nominal value of the system's control gain. The specific calculation formula is as follows: in, Let be the step size of the nonlinear feedback law. This is the fastest synthesis function.

[0056] The target displacement tracking value, target derivative tracking value, displacement estimate, and velocity estimate are input into the control quantity calculation formula to obtain the initial control quantity for the current control cycle. The initial control quantity is a nonlinear combination result based on the state error, reflecting the control action required by the system to eliminate the current deviation. By subtracting the ratio of the total disturbance estimate to the compensation coefficient from the initial control quantity, feedforward compensation of the total system disturbance is achieved, making the controlled object approximately compensated as a standard integral-cascaded object.

[0057] When the rolling force changes abruptly or the incoming material thickness fluctuates significantly, the magnitude of the piston displacement error may instantly jump beyond the preset fixed domain boundary. At this point, the second scaling factor dynamically expands the input domain based on the real-time detected error signal and the magnitude of the error rate of change, renormalizing the large-scale error to within the effective fuzzy level mapping range. This avoids saturation distortion caused by the input exceeding the upper bound of the domain, ensuring that the fuzzy rule can still output correct and differentiated gain corrections even under large deviations. Simultaneously, the output domain is adjusted synchronously using the second scaling factor, appropriately changing the damping factor and control gain. When the error converges to a smaller range, the second scaling factor shrinks the input domain in real time, maintaining sufficient sensitivity of fuzzy inference to small error changes and enhancing the fine-tuning capability in the steady-state phase.

[0058] After obtaining the current control input, the control input is output to the hydraulic cylinder drive unit of the pump-controlled lithium battery electrode rolling mill, driving the servo motor to operate in the direction and amplitude indicated by the control input, thereby controlling the piston of the hydraulic cylinder to move to the desired position. At the same time, the system enters the next control cycle and repeats the above steps S210 to S230 to achieve real-time closed-loop control of the system.

[0059] By dynamically adjusting the fuzzy universe of discourse using first and second scaling factors, fuzzy inference always operates within its optimal sensitivity range. Under rated operating conditions, the universe of discourse remains stable, without introducing additional fluctuations. When the rolling force changes abruptly, the universe of discourse expands rapidly to prevent large deviations from falling into the saturation region. When the oil modulus degrades, the universe of discourse contracts to enhance the recognition capability of weak signals. When the incoming material is uneven, the dual-channel independent mapping addresses the inverse amplitude characteristics of the error and the rate of change. This mechanism enables the system to maintain high-precision control under four operating conditions, solving the defect of fixed universe of discourse degradation under non-rated operating conditions, and achieving a leap from single-point optimization to global robustness.

[0060] The controller gain is dynamically and adaptively adjusted in real time by using a variable universe of discourse fuzzy module. It increases when the rolling force changes abruptly. Strengthen bias correction and adjust damping to suppress overshoot. Reconstruct the observation bandwidth of the expansion state observer when the oil modulus degrades. , , To match stiffness variations, high-frequency fine-tuning of parameters tracks random fluctuations when there is uneven material flow. Under rated operating conditions, the nominal optimal parameters are maintained. This mechanism ensures that the control parameters always match the system dynamics under four operating conditions, overcoming the limitations of insufficient adaptability of fixed-parameter structures.

[0061] By deeply integrating the disturbance estimation compensation of the extended state observer with the variable universe of discourse gain adjustment, a dual disturbance rejection mechanism is formed. When the rolling force changes abruptly, the extended state observer quickly estimates the impact load, while the variable universe of discourse synchronously expands the universe of discourse and adjusts the gain, ensuring coordinated output of disturbance compensation and gain correction for rapid recovery to steady state. When the oil modulus degrades, the variable universe of discourse reconstructs the observation bandwidth of the extended state observer to restore phase matching, avoiding estimation lead and compensation mismatch of the fixed observer when stiffness decreases. When the incoming material is uneven, the dual-channel independent expansion and contraction ensure continuous and accurate estimation of the extended state observer under wideband random disturbances, while the variable universe of discourse synchronously adjusts the control gain at high frequency to match the fluctuation frequency. The synergy of these three mechanisms enables the system to maintain stable control quality under various disturbances, overcoming the response lag and accuracy degradation problems of conventional fuzzy active disturbance rejection controllers.

[0062] Independent first and second scaling factors are configured for error, error rate of change, observation error, and observation error rate of change, respectively, and these factors are independently inferred and adjusted in parallel. Independent adjustment and adaptive tuning are performed for four operating conditions. This independent scaling structure ensures that both input dimensions maintain optimal mapping under each operating condition, improving the system's adaptability to time-varying and highly nonlinear operating conditions.

[0063] In summary, this invention achieves adaptive adjustment of controller parameters and real-time compensation for system disturbances by deeply integrating variable universe fuzzy control and fuzzy active disturbance rejection controller, without the need for a precise mathematical model, thus obtaining high-precision and robust control performance across the entire operating range.

[0064] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0065] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0066] Figure 4 A schematic diagram of the pump-controlled lithium battery electrode rolling mill control device provided in an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown, and are described in detail below: like Figure 4 As shown, the pump-controlled lithium battery electrode rolling mill control device 4 includes: Calculation module 41 is used to receive the desired displacement signal and input the desired displacement signal and the first scaling factor of the current control cycle into the fuzzy active disturbance rejection controller to calculate the current target value and the current estimated value of the current control cycle. The scaling module 42 is used to calculate the second scaling factor for the current control cycle based on the current target value and the current estimated value. The control module 43 is used to calculate the current control quantity of the current control cycle based on the current estimated value, the current target value and the second scaling factor of the current control cycle, and to control the hydraulic cylinder of the pump-controlled lithium battery electrode rolling mill based on the current control quantity.

[0067] In one possible implementation, the calculation module 41 is specifically used to: perform discrete calculations on the desired displacement signal based on the tracking differentiator in the fuzzy active disturbance rejection controller to obtain the current target value of the current control cycle; determine the gain of the current control cycle based on the historical estimate value, historical displacement, and the first scaling factor of the current control cycle of the previous control cycle; and calculate the current estimate value of the current control cycle based on the real-time displacement of the hydraulic cylinder, the current target value, and the gain of the current control cycle.

[0068] In one possible implementation, the calculation module 41 is further configured to: determine the difference between the historical displacement estimate in the historical estimate of the previous control cycle and the historical displacement of the previous control cycle as the historical observation error of the previous control cycle; differentiate the historical observation error of the previous control cycle to obtain the historical observation error change rate of the previous control cycle; multiply the first scaling factor of the current control cycle with a preset fuzzy universe of discourse to obtain the target fuzzy universe of discourse; wherein, the first scaling factor of the current control cycle is obtained by fuzzy calculation based on the historical observation error, the historical observation error change rate and the first fuzzy rule control table of the previous control cycle; and perform fuzzy calculation based on the historical observation error, the historical observation error change rate and the target fuzzy universe of discourse of the previous control cycle to obtain the gain of the current control cycle.

[0069] In one possible implementation, the calculation module 41 is further configured to: input the historical observation error and the rate of change of historical observation error of the previous control cycle into a preset fuzzy rule control table using trigonometric functions as membership functions, to obtain at least one fuzzy subset and the membership degree corresponding to each fuzzy subset; determine the fuzzy sub-interval corresponding to each fuzzy subset according to the target fuzzy universe of discourse; and perform defuzzification based on each fuzzy sub-interval, the membership degree of each fuzzy sub-interval, and a preset defuzzification algorithm to obtain the gain of the current control cycle.

[0070] In one possible implementation, the scaling module 42 is specifically used to: calculate the error signal and error rate of change of the current control cycle based on the current target value and the current estimated value; and perform fuzzy calculations based on the error signal, error rate of change of the current control cycle, and the second fuzzy rule control table to obtain the second scaling factor.

[0071] In one possible implementation, the current target value includes the target displacement tracking value and the target differential tracking value; the current estimated value includes the displacement estimated value and the velocity estimated value; the scaling module 42 is further configured to: determine the difference between the target displacement tracking value and the displacement estimated value as the error signal of the current control cycle; and determine the difference between the target differential tracking value and the velocity estimated value as the error change rate of the current control cycle.

[0072] In one possible implementation, the current estimate also includes the total disturbance estimate; the control module 43 is specifically used for: performing fuzzy calculations based on the error signal, the error rate of change, and the second scaling factor of the current control cycle to obtain the damping factor and control gain of the current control cycle; determining the control quantity calculation formula for the current control cycle based on the damping factor and control gain; inputting the target displacement tracking value, the target differential tracking value, the displacement estimate, and the velocity estimate into the control quantity calculation formula to obtain the initial control quantity of the current control cycle; and determining the difference between the initial control quantity and the preset ratio as the current control quantity; wherein the preset ratio is the ratio of the total disturbance estimate to the preset compensation coefficient.

[0073] Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Figure 5 As shown, the electronic device 5 of this embodiment includes a processor 50 and a memory 51. The memory 51 stores a computer program 52. When the processor 50 executes the computer program 52, it implements the steps in the various method embodiments described above. Alternatively, when the processor 50 executes the computer program 52, it implements the functions of each module / unit in the various device embodiments described above.

[0074] For example, computer program 52 may be divided into one or more modules / units, which are stored in memory 51 and executed by processor 50 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 52 in electronic device 5.

[0075] Electronic device 5 may include, but is not limited to, processor 50 and memory 51. Those skilled in the art will understand that... Figure 5This is merely an example of electronic device 5 and does not constitute a limitation on electronic device 5. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device 5 may also include input / output devices, network access devices, buses, etc.

[0076] The processor 50 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0077] The memory 51 can be an internal storage unit of the electronic device 5, such as a hard disk or RAM. The memory 51 can also be an external storage device of the electronic device 5, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory 51 can include both internal and external storage units of the electronic device 5. The memory 51 is used to store the computer program 52 and other programs and data required by the electronic device 5. The memory 51 can also be used to temporarily store data that has been output or will be output.

[0078] For the sake of simplicity and clarity, only the above-described functional modules / units are used as examples. In practical applications, the functions described above can be assigned to different functional modules / units as needed. These modules / units can be implemented in hardware, software, or a combination of both.

[0079] This invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the methods described in the above-described method embodiments.

[0080] This invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the methods described in the above-described method embodiments.

[0081] Computer programs include computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. Computer-readable media can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0082] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0083] 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 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 the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A pump-controlled lithium battery electrode rolling mill control method, characterized in that, include: Receive the desired displacement signal and input the desired displacement signal and the first scaling factor of the current control cycle into the fuzzy active disturbance rejection controller to calculate the current target value and the current estimated value of the current control cycle; Calculate the second scaling factor for the current control cycle based on the current target value and the current estimated value; Based on the current estimated value, the current target value, and the second scaling factor of the current control cycle, the current control quantity of the current control cycle is calculated, and the hydraulic cylinder of the pump-controlled lithium battery electrode rolling mill is controlled according to the current control quantity.

2. The pump-controlled lithium battery electrode rolling mill control method according to claim 1, characterized in that, The step of inputting the desired displacement signal and the first scaling factor of the current control cycle into the fuzzy active disturbance rejection controller to calculate the current target value and the current estimated value of the current control cycle includes: The target value for the current control cycle is obtained by discretely calculating the desired displacement signal based on the tracking differentiator in the fuzzy active disturbance rejection controller. The gain of the current control cycle is determined based on the historical estimate, historical displacement, and the first scaling factor of the current control cycle from the previous control cycle. The current estimated value of the current control cycle is calculated based on the real-time displacement of the hydraulic cylinder in the current control cycle, the current target value, and the gain of the current control cycle.

3. The pump-controlled lithium battery electrode rolling mill control method according to claim 2, characterized in that, The process of determining the gain of the current control cycle based on the historical estimate, historical displacement, and the first scaling factor of the current control cycle includes: The difference between the historical displacement estimate in the historical estimate of the previous control period and the historical displacement of the previous control period is determined as the historical observation error of the previous control period. Differentiating the historical observation error of the previous control cycle yields the rate of change of the historical observation error of the previous control cycle; The target fuzzy universe is obtained by multiplying the first scaling factor of the current control cycle with the preset fuzzy universe of discourse; wherein, the first scaling factor of the current control cycle is obtained by fuzzy calculation based on the historical observation error, the rate of change of historical observation error and the first fuzzy rule control table of the previous control cycle; Fuzzy calculations are performed based on the historical observation error, the rate of change of historical observation error, and the target fuzzy universe of discourse of the previous control cycle to obtain the gain of the current control cycle.

4. The pump-controlled lithium battery electrode rolling mill control method according to claim 3, characterized in that, The gain for the current control cycle is obtained by performing fuzzy calculations based on the historical observation error, the rate of change of historical observation error, and the target fuzzy universe of discourse from the previous control cycle, including: Using trigonometric functions as membership functions, the historical observation error and the rate of change of historical observation error from the previous control cycle are input into a preset fuzzy rule control table to obtain at least one fuzzy subset and the membership degree corresponding to each fuzzy subset. Based on the target fuzzy universe of discourse, determine the fuzzy sub-interval corresponding to each fuzzy subset; Based on each fuzzy sub-interval, the membership degree of each fuzzy sub-interval, and the preset defuzzification algorithm, defuzzification is performed to obtain the gain of the current control cycle.

5. The pump-controlled lithium battery electrode rolling mill control method according to any one of claims 1-4, characterized in that, The step of calculating the second scaling factor for the current control cycle based on the current target value and the current estimated value includes: Based on the current target value and the current estimated value, calculate the error signal and error rate of change for the current control cycle; The second scaling factor is obtained by performing fuzzy calculations based on the error signal of the current control cycle, the error change rate, and the second fuzzy rule control table.

6. The pump-controlled lithium battery electrode rolling mill control method according to claim 5, characterized in that, The current target value includes the target displacement tracking value and the target differential tracking value; the current estimated value includes the displacement estimate and the velocity estimate; The step of calculating the error signal and error rate of change for the current control cycle based on the current target value and the current estimated value includes: The difference between the target displacement tracking value and the displacement estimation value is determined as the error signal for the current control cycle; The difference between the target differential tracking value and the velocity estimate is determined as the error change rate of the current control cycle.

7. The pump-controlled lithium battery electrode rolling mill control method according to claim 6, characterized in that, The current estimate also includes a total disturbance estimate; the calculation of the current control quantity for the current control period based on the current estimate, the current target value, and the second scaling factor for the current control period includes: Based on the error signal, the error rate of change, and the second scaling factor of the current control cycle, fuzzy calculation is performed to obtain the damping factor and control quantity gain of the current control cycle. Based on the damping factor and the control quantity gain, determine the control quantity calculation formula for the current control cycle; The target displacement tracking value, target differential tracking value, displacement estimate, and velocity estimate are input into the control quantity calculation formula to obtain the initial control quantity for the current control cycle. The difference between the initial control quantity and the preset ratio is determined as the current control quantity; wherein the preset ratio is the ratio of the total disturbance estimate to the preset compensation coefficient.

8. A pump-controlled lithium battery electrode rolling mill control device, characterized in that, include: The calculation module is used to receive the desired displacement signal and input the desired displacement signal and the first scaling factor of the current control cycle into the fuzzy active disturbance rejection controller to calculate the current target value and the current estimated value of the current control cycle. The scaling module is used to calculate the second scaling factor for the current control cycle based on the current target value and the current estimated value. The control module is used to calculate the current control quantity for the current control cycle based on the current estimated value, the current target value, and the second scaling factor of the current control cycle, and to control the hydraulic cylinder of the pump-controlled lithium battery electrode rolling mill based on the current control quantity.

9. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images