Intelligent control method for tension in high-speed winding process of lithium battery

Abstract

The invention relates to an intelligent control method for tension in a high-speed winding process of a lithium battery in the technical field of intelligent manufacturing, which comprises the following steps of: acquiring tension data and material thickness data acquired by a winding machine sensor in real time, extracting a current production speed value and an environment temperature and humidity value from the tension data and the material thickness data, and fusing the data to obtain an initial tension distribution matrix; an environment change compensation factor is obtained from the adjusted tension target value, a fuzzy logic control algorithm is adopted to further fuse the temperature and humidity values, and a compensated tension target value is determined and used for correcting the interlayer lamination quality in real time; according to the optimized tension control instruction, processing real-time feedback data by adopting an adaptive filtering algorithm, and determining a final tension balance state for maintaining the safety and reliability of the battery performance; and an internal defect risk index is extracted from the final tension balance state, if it is judged that the risk index is lower than a preset threshold value, a stable control signal is output to the winding machine, and a real-time balance result under continuous high-speed winding is obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent manufacturing, and particularly relates to an intelligent control method for tension in high-speed winding of lithium batteries. BACKGROUND

[0002] Lithium battery winding control technology is a core pillar in the field of new energy battery manufacturing, directly related to battery performance and safety reliability. In the production process of lithium batteries, the winding link determines the lamination quality between material layers. If not properly controlled, it will seriously affect the service life and safety of the battery. The importance of this field is self-evident, and high-precision technical means are urgently needed to ensure production quality and efficiency.

[0003] Currently, the tension control of lithium battery winding machines generally faces technical bottlenecks. Although existing equipment uses some basic control methods, these methods are not effective in dealing with complex working conditions. Especially when facing different material properties or changes in production speed, the control system often cannot adjust in time, resulting in uneven tension distribution and affecting the stability of the internal structure of the battery.

[0004] In addition, changes in environmental conditions such as humidity and temperature can also interfere with the tension control effect, and existing technologies lack effective response mechanisms to these external factors, making control precision and stability difficult to meet the high requirements of modern manufacturing. The deeper technical difficulty lies in how to achieve real-time balance of tension in the dynamic process. Tension control not only needs to pay attention to the overall stress state of the material during winding, but also must consider the subtle differences of multi-layer materials during high-speed stacking. If the stress distribution of each layer of material cannot be accurately grasped, it will lead to tight or local over-tightness between layers, and then cause internal defects of the battery. For example, when producing square lithium batteries, the separator and electrode sheet will produce subtle misalignment during the winding process due to differences in thickness or elasticity. This misalignment will be amplified at high speed, which may eventually lead to a decline in battery performance and even safety hazards.

[0005] Therefore, how to dynamically adjust the tension in high-speed winding to adapt to the differences in characteristics of multi-layer materials, while taking into account the impact of environmental changes on control effect, has become a key problem that needs to be solved in lithium battery winding technology. The solution to this problem will directly determine whether production efficiency and battery quality can meet the high standards expected by the industry. SUMMARY

[0006] The present application provides an intelligent control method for tension in high-speed winding of lithium batteries, mainly including:

[0007] The tension data and material thickness data collected in real time by the winding machine sensor are acquired, from which the current production speed value and environmental temperature and humidity value are extracted, and an initial tension distribution matrix is obtained by fusing these data; according to the initial tension distribution matrix, an adaptive filtering algorithm is used to process noise interference, and a filtered tension distribution matrix is obtained, wherein the filtering process is weighted and adjusted according to material differences to determine the stress deviation value between multiple layers of materials; if the stress deviation value exceeds a preset threshold value, a fuzzy logic control algorithm is activated to input the filtered tension distribution matrix and the production speed value, a dynamic adjustment coefficient is calculated through a rule base, and an adjusted tension target value is obtained to balance the uneven distribution in high-speed winding; an environmental change compensation factor is obtained from the adjusted tension target value, and a fuzzy logic control algorithm is used to further fuse the temperature and humidity values to determine the compensated tension target value for real-time correction of the interlayer adhesion quality; for the compensated tension target value, the winding machine actuator feedback signal is acquired, and by comparing the feedback signal with the target value, if the difference is greater than a preset threshold value, the dynamic adjustment coefficient is iteratively updated to obtain an optimized tension control instruction to adapt to the subtle differences in materials; according to the optimized tension control instruction, an adaptive filtering algorithm is used to process real-time feedback data to determine the final tension balance state for maintaining the safety and reliability of the battery performance; from the final tension balance state, the internal defect risk index is extracted, and if the risk index is lower than a preset threshold value, a stable control signal is output to the winding machine to obtain the real-time balance result under continuous high-speed winding.

[0008] The technical scheme provided by the embodiment of the present application can include the following beneficial effects:

[0009] The present application discloses an intelligent control method for tension in the high-speed winding process of lithium batteries, aiming to solve the problem of tension imbalance caused by material differences, environmental changes and production speed fluctuations in battery production. The present application collects tension, thickness, speed and temperature and humidity data in real time, constructs an initial tension distribution matrix, and uses an adaptive filtering algorithm to eliminate noise interference, and combines material differences for weighted adjustment to accurately calculate the stress deviation of multiple layers of materials. When the deviation or risk index exceeds the threshold value, the fuzzy logic control algorithm is activated, the production speed and environmental factors are fused, the adjustment coefficient and the compensated tension target value are dynamically generated, and through iterative optimization and actuator feedback comparison, the tension control instruction is ensured to adapt to subtle changes, and finally the tension balance state is realized. The present application significantly improves the stability of high-speed winding and the reliability of battery performance, reduces the risk of internal defects, and ensures the quality consistency under continuous production. BRIEF DESCRIPTION OF DRAWINGS

[0010] Fig. 1 The flowchart of the intelligent control method for tension in the high-speed winding process of lithium batteries of the present application.

[0011] Fig. 2A schematic diagram of the intelligent control method for tension in the high-speed winding process of lithium batteries.

[0012] Fig. 3 Another schematic diagram of the intelligent control method for tension in the high-speed winding process of lithium batteries. DETAILED DESCRIPTION

[0013] The technical solutions of the present application will be described clearly and completely below in conjunction with embodiments. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0014] As Figs. 1-3 The intelligent control method for tension in the high-speed winding process of lithium batteries can specifically include the following steps:

[0015] S101, acquiring tension data and material thickness data collected by a winding machine sensor in real time, extracting a current production speed value and an environmental temperature and humidity value therefrom, and obtaining an initial tension distribution matrix by fusing these data.

[0016] The tension data and material thickness data are collected by the winding machine sensor, the current production speed value is extracted from the tension data by analyzing the fluctuation frequency, and the production speed value is obtained. According to the production speed value, the environmental temperature value and the environmental humidity value are monitored from the material thickness data (thickness sensor integrated temperature and humidity module), and the environmental temperature and humidity value is specifically determined by a speed-related model such as T=f(v,th) (T is the temperature and humidity value, v is the production speed, and th is the thickness data). The production speed value and the environmental temperature and humidity value and the tension data are integrated by using a data fusion method, the fusion value of each data point is calculated by weighted average, and an initial tension distribution matrix is constructed. For the initial tension distribution matrix, sensor calibration verification and matrix element correction are performed, and if the matrix element exceeds a preset threshold, the element value is adjusted to obtain a corrected initial tension distribution matrix.

[0017] Specifically, in an embodiment, the winding machine system collects relevant data in real time by integrating sensors.

[0018] Specifically, the winding machine is equipped with tension and thickness sensors, which are installed at key locations along the winding path, such as between the feed and output rollers. The tension sensor monitors the tensile force on the material during winding, outputting real-time tension data, typically in Newtons. The thickness sensor uses laser or ultrasonic principles to measure the material thickness, generating a thickness data sequence. This acquisition process ensures real-time data, supporting subsequent extraction and fusion. Furthermore, the current production speed value is extracted from the acquired tension and thickness data.

[0019] For example, during the operation of a winding machine, tension data often contains speed-related information because tension changes are proportional to the winding speed. By analyzing the fluctuation frequency of the tension data, the production speed can be calculated.

[0020] For example, the interval between tension peaks can be converted into a velocity value. Similarly, ambient temperature and humidity values ​​can be extracted from material thickness data, which may be embedded in sensor outputs or acquired with the assistance of auxiliary environmental sensors to ensure accurate extraction.

[0021] Preferably, the extraction of ambient temperature and humidity values ​​involves a data filtering step.

[0022] Specifically, if the thickness sensor integrates a temperature and humidity module, these values ​​are directly parsed from the data stream; otherwise, a separate temperature and humidity sensor is used to synchronously acquire data and correlate it with the thickness data. This approach is applicable to different wound materials, such as plastic films or metal wires, demonstrating the versatility of the extraction process.

[0023] In one possible implementation, an initial tension distribution matrix is ​​obtained by fusing these data. First, the extracted production speed values ​​and ambient temperature and humidity values ​​are integrated with the original tension and thickness data. The fusion process involves constructing a two-dimensional matrix where rows represent different winding positions and columns represent influencing factors such as speed and temperature.

[0024] Specifically, for each location, the calculated tension value is influenced by a weighted average of velocity and thickness, for example, by generating matrix elements using a linear combination formula. This matrix represents the initial tension distribution and is used for subsequent optimization.

[0025] It should be noted that the core of the fusion step lies in data normalization processing to eliminate the influence of different units.

[0026] For example, tension data is standardized and mapped to a range of 0 to 1; then, velocity values ​​are incorporated as correction factors to adjust the matrix values.

[0027] For example, in high-temperature and high-humidity environments, the matrix elements increase at higher speeds, reflecting the risk of increased tension. This processing ensures that the matrix accurately reflects production conditions. Furthermore, in the production scenario of wound films, this fusion method can be extended to multi-layer winding. By repeatedly collecting and extracting data in cycles, a dynamic matrix is ​​generated to adapt to speed changes. This implementation improves the accuracy of tension control and reduces material breakage. In another embodiment, for wire winding, the effect of humidity on thickness is given priority during fusion.

[0028] Specifically, humidity values ​​are used to correct thickness data, which is then fused with tension to form a matrix. This variant demonstrates the flexibility of the technology, making it suitable for humidity-sensitive materials.

[0029] Understandably, this method achieves uniform tension distribution through data fusion, supporting high-quality winding products.

[0030] For example, in practice, the initial matrix can be used for feedback control of the winding machine motor to adjust the speed and maintain stable tension. This effect manifests as a reduction in defect rate in continuous production.

[0031] In one embodiment, the system can also record historical matrix data for trend analysis, further enhancing process monitoring.

[0032] S102. Based on the initial tension distribution matrix, an adaptive filtering algorithm is used to process noise interference and obtain the filtered tension distribution matrix. The filtering process performs weighted adjustments based on material differences to determine the stress deviation values ​​between multiple layers of materials.

[0033] An adaptive filtering algorithm is used to process noise interference using an initial tension distribution matrix. This algorithm automatically adjusts filtering parameters based on signal changes to obtain a filtered tension distribution matrix. For this filtered matrix, material difference data, including density and elastic modulus differences, is obtained from a multilayer material property database and weighted to determine the initial stress deviation between the multilayer materials. Based on this initial stress deviation, the adhesive strength index is obtained from the multilayer material interface through interface shear testing. This index is then corrected to obtain the corrected stress deviation. If the corrected stress deviation exceeds a preset threshold, the tension distribution is optimized by weighting and fusing the material differences. This fusion is achieved through a weighted average method to determine the deviation distribution pattern. The final stress deviation between the multilayer materials is quantified based on this deviation distribution pattern.

[0034] Specifically, in one implementation, the initial tension distribution matrix is ​​a two-dimensional data structure obtained by stress testing of multilayer composite materials, where each row represents the tension value distribution of a layer of material and each column corresponds to a measurement point at a specific location.

[0035] For example, in the field of composite material manufacturing, the matrix can be collected from a sensor array, reflecting the tension inhomogeneity of the material in the initial state.

[0036] It should be noted that the construction process of the initial tension distribution matrix includes taking the elastic modulus and thickness of each layer of material as input parameters, simulating the tension distribution by finite element analysis method, and thus forming an m x n matrix, where m is the number of material layers and n is the number of sampling points. This matrix helps to set the reference for subsequent noise processing. Further, the use of an adaptive filtering algorithm to process noise interference is the core step of the scheme. The algorithm is a variant of Kalman filtering, with state vector x representing the tension distribution value, input being the noise signal y collected by the sensor, and output being the filtered tension matrix. The gain matrix K is calculated by K = PH^T(HPH^T+R)^-1, where P is the error covariance matrix, H is the observation matrix, and R is the observation noise covariance; the algorithm updates the filtering coefficient K in real time according to the statistical characteristics of the input signal to minimize the influence of noise.

[0037] Specifically, in the tension analysis of multi-layer materials, the algorithm first estimates the power spectral density of the noise, and then iteratively adjusts the filter weights by the least mean square error criterion. For example.

[0038] In one possible implementation, for each element in the initial tension distribution matrix, the algorithm applies a variant of Kalman filtering to adaptively track noise patterns such as random fluctuations caused by thermal noise or vibration interference. In this way, the filtering process can effectively suppress external interference and ensure the reliability of the matrix data. In the context of composite material detection, this algorithm can process data collected in real time on the production line, gradually smoothing the tension value curve and avoiding the occurrence of false peaks caused by noise. The detailed implementation of this step includes initializing the filtering state vector, calculating the gain matrix and updating the estimated value, and the entire process is iterated until convergence, with the typical number of iterations controlled within 10 to 50 times, thus obtaining a preliminary denoised matrix.

[0039] Preferably, the filtering process is adjusted with weights for material differences, further improving the accuracy of the scheme. Material differences mainly refer to the differences in density, elastic coefficient and thermal expansion coefficient of each layer in multi-layer composite materials, which can cause tension distribution inhomogeneity. The principle of weighted adjustment is to compensate for these differences by introducing a weight factor.

[0040] For example, higher weights are assigned to high-density layers to emphasize their contribution to the overall tension.

[0041] Specifically, in the filtered matrix, a weighting function is applied to the tension values of each layer, for example, higher weight is assigned to high-density layers to emphasize their contribution to the overall tension. Specifically, in the filtered matrix, first calculate the weight of each element w_ij=(E_ij / E_avg)*(d_ij / d_avg), where E_ij is the elastic modulus of element ij, E_avg is the average elastic modulus of all elements, and d_ij and d_avg are the density values, respectively. Then, multiply the tension value of each element in the matrix by the corresponding weight w_ij to obtain the weighted tension value. This adjustment ensures that the filtering result is more in line with the actual material properties.

[0042] In one embodiment, for a multi-layer structure of carbon fiber reinforced composite material, the algorithm first identifies the interlayer difference, then multiplies the weight matrix element by element to achieve weighted filtering, and finally outputs the adjusted tension distribution matrix. This process not only reduces the bias, but also improves the prediction accuracy under bending load.

[0043] For example, in the production environment of composite material plates, this weighted adjustment can be applied to interlayer analysis of different thicknesses, and the weight is dynamically optimized according to experimental calibration data.

[0044] It can be understood that determining the stress deviation value between the multi-layer materials based on the filtered tension distribution matrix is a subsequent key link. The stress deviation value is defined as a quantitative indicator of the tension difference between adjacent layers, and its calculation process involves matrix difference operation.

[0045] Specifically, first calculate the element difference vector of adjacent rows in the filtered matrix, and then take its norm as the deviation value, such as using the Euclidean norm to obtain the deviation amplitude. Further, to enhance robustness, a threshold judgment can be introduced, and if the deviation value exceeds the preset threshold, it is marked as a high-risk area. In one embodiment, for a multi-layer composite plate of glass fiber and epoxy resin, the deviation value calculated after the calculation can be used to evaluate the interlayer peeling risk, and the filtering effect is quantified by comparing the initial and filtered deviations. The business value of this step is to discover material defects in time, thereby optimizing manufacturing process parameters, such as adjusting the lamination temperature to reduce the deviation. In another embodiment, this scheme can be extended to the tension monitoring of aerospace composites, where the initial matrix is obtained from wind tunnel tests, adaptive filtering is used to process aerodynamic noise, and weighted adjustment considers the difference between carbon fiber and aluminum alloy layers. The final deviation value is used for structural integrity evaluation.

[0046] It should be noted that the logical flow of the entire process forms a closed loop from data acquisition to deviation output, ensuring general application in the field of composite materials.

[0047] S103, if the force deviation value exceeds the preset threshold value, activate the fuzzy logic control algorithm to input the filtered tension distribution matrix and the production speed value, calculate the dynamic adjustment coefficient through the rule base, and obtain the adjusted tension target value to balance the uneven distribution in high-speed winding.

[0048] During the winding process, the force deviation value and the production speed value are collected from the sensor, and the winding thickness data is recorded, wherein the winding thickness data is updated in real time based on the aforementioned material thickness data. If the force deviation value exceeds the preset threshold value, the change characteristics are extracted from the production speed value and combined with the winding thickness data to generate an initial tension distribution matrix. The initial tension distribution matrix is processed by an adaptive filtering algorithm to obtain a filtered tension distribution matrix to smooth noise interference. For the filtered tension distribution matrix and the production speed value, the fuzzy logic control algorithm is activated to input these data, and a dynamic adjustment coefficient is calculated through a preset rule base, wherein the rule base is constructed based on expert experience and contains 50 rules, the input is the tension distribution and the speed value, and the output is the adjustment coefficient, which goes through the processes of fuzzification, reasoning and defuzzification. According to the dynamic adjustment coefficient, the adjusted tension target value is determined and applied to the winding equipment, and through the feedback mechanism, real-time monitoring and iterative adjustment are realized to achieve closed-loop control of balancing uneven distribution.

[0049] Specifically, in an embodiment, the tension control in high-speed winding process first involves the judgment of force deviation value. The force deviation value is calculated by comparing the difference between the actual tension distribution and the ideal uniform distribution.

[0050] For example, on a yarn winding production line, a sensor array collects tension data at multiple points in real time, forming a tension distribution matrix that represents the force on the winding material at different positions. If the calculated deviation value exceeds the preset threshold value, such as 5% relative error, the system activates the fuzzy logic control algorithm. This judgment mechanism ensures that control is only necessary, avoiding unnecessary computational overhead, thereby maintaining stable production efficiency in high-speed winding. Further, the activation of the fuzzy logic control algorithm depends on the preparation of input data. The filtered tension distribution matrix is obtained by applying a low-pass filter to the original matrix, and the filtering process removes high-frequency noise.

[0051] For example, in the thin film winding scenario, the production speed value, representing the speed of the winding machine, is another input, usually measured in meters per minute. These inputs are collectively provided to the algorithm to capture the dynamic characteristics of uneven tension.

[0052] It should be noted that filtering helps to improve the accuracy of the data, ensuring that subsequent calculations are based on reliable foundations.

[0053] Specifically, the core of the fuzzy logic control algorithm lies in the application of the rule base. The rule base is a collection of fuzzy rules defined based on expert knowledge.

[0054] For example, if the tension profile matrix shows a local high deviation and the production speed is above a threshold, the adjustment coefficient should be increased. The algorithm first fuzzifies the inputs, i.e. maps the filtered matrix values and speed values to fuzzy sets like "low deviation" "medium speed" etc. Then, it applies the rule base through an inference engine to perform fuzzy reasoning and calculates the dynamic adjustment coefficient. This coefficient is a value between 0 and 1 that is used to modify the original tension target value.

[0055] In one possible implementation, for high-speed paper winding, the rule base can contain 20 rules covering different speed and deviation scenarios.

[0056] Exemplarily, in the embodiment of yarn winding, assuming the production speed is 500 meters per minute and the filtered tension profile matrix shows a central area deviation of 8%, which exceeds the threshold of 5%. After the algorithm is activated, it calculates the dynamic adjustment coefficient to be 0.75 through the rule base, which is multiplied by the original target value to obtain the adjusted tension target value, e.g. from 100 Newton to 75 Newton. This adjustment balances the uneven distribution and reduces the risk of material stretch unevenness.

[0057] Preferably, the rule base can be dynamically updated according to historical data to adapt to different winding materials. In another implementation, for metal cable winding, the deviation judgment can be combined with multi-sensor data. The stress deviation value calculation process involves the standard deviation of the matrix elements.

[0058] For example, the tension profile matrix is considered as a two-dimensional array, and its overall deviation is calculated. If it exceeds the threshold, the production speed value input to the fuzzy logic algorithm can be obtained from the frequency converter. The rule base emphasizes the amplification effect of speed on tension in this scenario, and when calculating the dynamic adjustment coefficient, it gives priority to rules at high speed, such as "if speed is high and deviation is medium, then coefficient is medium". After obtaining the adjusted tension target value, the system outputs to the actuator to achieve real-time balancing. Further, the calculation process of the dynamic adjustment coefficient can be divided into fuzzification and defuzzification steps. Fuzzification converts the input into membership function values.

[0059] For example, the deviation degree of the tension profile matrix is divided into three fuzzy sets: low, medium, and high, and the production speed value is similarly divided into slow, medium, and fast. The rule base is organized in the form of "if-then", such as "if deviation is high and speed is fast, then adjustment coefficient is high". After reasoning, the precise coefficient is obtained through the center of gravity method. This method effectively balances uneven distribution in high-speed winding because it handles input data with strong uncertainty.

[0060] It can be understood that the universality of the control algorithm is reflected in the adaptation to different winding speeds. In low-speed winding such as textile production at 200 meters per minute, the threshold can be set to 3%, and after activation, the rule base uses simplified rules to calculate the coefficient to fine-tune the tension target value. In high-speed plastic film winding such as 800 meters per minute, the rule base is expanded to include more speed-related rules to ensure that the adjusted target value quickly responds to changes in distribution, thereby maintaining winding quality.

[0061] In an embodiment, the system integrates a feedback loop to further optimize the adjustment process. After obtaining the adjusted tension target value, the actuator such as the motor controller applies the value, and the sensor monitors the effect. If it is still unbalanced, the algorithm can be iteratively activated. This loop improves robustness in the winding production line and is suitable for continuous production environments.

[0062] Preferably, in actual deployment, fuzzy logic control can be used in combination with a PID controller, relying on PID when the deviation does not exceed the threshold, and switching to fuzzy mode when it exceeds. This hybrid strategy achieves smooth transition in high-speed winding, balancing computational complexity and control accuracy.

[0063] S104, obtain an environmental change compensation factor from the adjusted tension target value, further fuse the temperature and humidity values using a fuzzy logic control algorithm to determine a compensated tension target value for real-time correction of interlayer bonding quality.

[0064] Obtain an environmental change compensation factor from the adjusted tension target value. Further fuse the temperature and humidity values using a fuzzy logic control algorithm. Determine a compensated tension target value. Real-time correct interlayer bonding quality through the compensated tension target value. Obtain an environmental change compensation factor from the adjusted tension target value. Further fuse the temperature and humidity values using a fuzzy logic control algorithm. Determine a compensated tension target value. Real-time correct interlayer bonding quality through the compensated tension target value. Obtain an environmental change compensation factor through the adjusted tension target value, use a preset threshold to determine whether the factor is out of range, if it is out of range, fuse the temperature and humidity values using a fuzzy logic control algorithm to obtain a preliminary compensation tension. According to the preliminary compensation tension, determine a compensated tension target value. Through the compensated tension target value, determine the real-time correction parameter of the interlayer bonding quality.

[0065] Specifically, in certain embodiments, obtaining the environmental change compensation factor from the adjusted tension target value first requires an understanding of the concept of the environmental change compensation factor. It refers to the correction coefficient of the tension target based on external environmental variables such as temperature and humidity. In the production line of interlayer lamination, for example, in the process of film material compounding, the initial tension target value may have been set to a certain value, such as 20 N / m, through preliminary calibration, but environmental fluctuations will affect the material viscosity, resulting in uneven lamination. The process of obtaining the compensation factor includes comparing the deviation of the current environmental data from the standard conditions.

[0066] Specifically, real-time temperature and humidity data are collected by sensors, and then the deviation value is calculated, for example, temperature deviation ΔT = current temperature - standard temperature (standard temperature is 25°C), humidity deviation ΔH = current humidity - standard humidity (standard humidity is 50%). ΔT typically ranges from -10 to 10°C, and ΔH typically ranges from -20 to 20%. The compensation factor can be obtained by a preset formula such as F = 1 + k1*ΔT + k2*ΔH, where k1 and k2 are empirical coefficients (for example, k1 = 0.01, k2 = 0.005), which ensure that the factor reflects the impact of the environment on the tension. For example, if the current temperature is 30°C and the humidity is 60%, then ΔT = 5 and ΔH = 10, and F = 1 + 0.01*5 + 0.005*10 = 1.1. This factor is used for subsequent fusion to avoid direct adjustment leading to system instability. Further, a fuzzy logic control algorithm is used to further fuse the temperature and humidity values. This algorithm is a control method based on fuzzy set theory, which can handle uncertainty and nonlinearity. In interlayer lamination quality control, the core of the fuzzy logic algorithm is to fuzz the temperature and humidity, establish a rule base and perform reasoning, and finally de-fuzz the output.

[0067] For example, first, the temperature and humidity are fuzzed, which are divided into fuzzy sets such as "low", "medium", and "high", for example, the temperature range 0-50°C is divided into low (0-20), medium (15-35), and high (30-50), and the membership degree is determined using a triangular membership function. Next, a rule base is established, such as "if the temperature is high and the humidity is high, then the compensation factor increases", these rules are based on material lamination experience, and a total of 20-30 rules can be set to cover various environmental combinations. Then, through a fuzzy reasoning engine such as the Mamdani method, the rule output fuzzy set is aggregated. Finally, de-fuzzing uses the barycenter method to calculate the crisp value, i.e., the fused compensation factor. This process ensures the robustness of the algorithm to environmental changes, and in actual film lamination, it can adapt to tension drift in high-humidity environments such as summer.

[0068] Specifically, the obtained environmental change compensation factor is multiplied by the adjusted tension target value to obtain the final value.

[0069] For example, if the adjusted target value is 25 N / m and the compensation factor is 1.05, the compensated value is 26.25 N / m. This determination process can be integrated in the control system and executed by the PLC controller to ensure that the output of the tension regulator is accurately matched.

[0070] In one possible implementation, for real-time correction of interlayer bonding quality, the compensated tension target value is fed back to the roll drive system of the bonding machine.

[0071] For example, in the context of paper interlayer bonding, if the humidity increases causing the paper to swell, the compensation value increases the tension to maintain the interlayer tightness and avoid bubbles or separation. The system achieves quality correction through closed-loop control, updating the tension every second.

[0072] Preferably, in another embodiment, the fuzzy logic algorithm can be extended to include more environmental variables, such as wind speed, but is still limited to the interlayer bonding field.

[0073] For example, on a plastic film compounding line, when the fusion wind speed is considered, the rule base adds "if the wind speed is high, then the compensation factor is fine-tuned", which enhances the flexibility of the algorithm without changing the core fusion mechanism.

[0074] It should be noted that this technical solution is universal at different bonding speeds, such as emphasizing precision at low speed (5 m / min) and focusing on stability at high speed (20 m / min). Through compensation, the bonding defect rate can be reduced, ensuring production efficiency.

[0075] For example, in a continuous web bonding process, compensation factor acquisition can be optimized in combination with historical data, and the algorithm fusion uses adaptive rules to adjust the membership function to adapt to seasonal changes. Further, the overall process forms a closed loop from data acquisition to output, ensuring real-time performance. In implementation, the accuracy of the sensor directly affects the accuracy of the compensation, so a high-precision temperature and humidity sensor is selected.

[0076] In one embodiment, for multi-layer bonding, the compensation value is applied in layers, where the inner layer tension target value is defined as the tension target value close to the substrate side, and the outer layer tension target value is defined as the tension target value away from the substrate side, such as the inner layer tension target value multiplied by 1.02 and the outer layer tension target value multiplied by 1.08, to optimize the quality. It can be understood that these embodiments support the broad application of the claims, and the effect is reflected in the improvement of bonding uniformity.

[0077] S105, for the compensated tension target value, obtain the winding machine actuator feedback signal, compare the feedback signal with the target value, if the difference is greater than a preset threshold, iteratively update the dynamic adjustment coefficient, and obtain the optimized tension control instruction to adapt to the subtle differences in materials.

[0078] Based on the compensated tension target value of S104, the winding machine actuator feedback signal, winding speed monitoring data and thickness sensor data are obtained, and by comparing the feedback signal with the tension target value, it is determined that if the difference is greater than a preset threshold, the initial dynamic adjustment coefficient is determined. By fusing the thickness sensor data and the initial dynamic adjustment coefficient, the dynamic adjustment coefficient is iteratively updated to obtain an intermediate optimization coefficient to handle material fine differences. The winding machine execution response corresponding to the intermediate optimization coefficient is obtained, and by comparing the execution response with the compensated tension target value, it is determined that if the response deviation is greater than another preset threshold, the dynamic adjustment coefficient is further iteratively updated. Through the further iteratively updated dynamic adjustment coefficient, an optimized tension control instruction is obtained to adapt to the material fine differences.

[0079] Specifically, in an embodiment, for the compensated tension target value, the winding machine actuator feedback signal needs to be obtained first. This feedback signal usually comes from sensors installed on the winding machine, such as tension sensors or speed sensors, which monitor the actual tension state in real time during winding. The compensated tension target value is the ideal value obtained by preliminary calculation, aiming to maintain the uniformity of material winding.

[0080] Specifically, the target value may be initially set based on material type and winding speed, for example, in textile fiber winding, the target value is set to a certain newton force per meter of material. By obtaining the feedback signal, the actual running state can be monitored to ensure the accuracy of the subsequent comparison step. This way helps to capture deviations in time in industrial production environment. Further, by comparing the feedback signal with the target value to evaluate the system performance.

[0081] Illustratively, the feedback signal represents the actual tension, while the target value is the preset reference. If the actual tension deviates from the target value, it may be due to the resistance change caused by the fine differences in material thickness or elasticity.

[0082] In a possible implementation, the comparison process uses numerical difference calculation, for example, subtracting the target value from the feedback signal to get the difference value. This comparison is not limited to a single moment, but can be extended to a continuous time sequence to observe the trend of tension fluctuations, providing a more comprehensive data basis for judgment. In the film winding scenario, this comparison helps to identify fine deviations caused by temperature changes, ensuring the pertinence of control instructions.

[0083] It should be noted that if the difference is greater than the preset threshold, the iteration update stage is entered. The preset threshold is an empirical value, for example, set to 5% of the target value, to avoid excessive sensitivity or ignore small fluctuations.

[0084] Specifically, if the difference exceeds a threshold, the system triggers an adjustment mechanism. This judgment process embodies the essence of closed-loop control, deciding whether to optimize through real-time evaluation. In the operation of the winding machine, the threshold can be adjusted according to the material type, for example, a smaller threshold is used for materials that are easily deformed to increase sensitivity. This design enables the control system to adapt to the dynamic changes of the production line rather than static settings.

[0085] In one embodiment, the iterative update of the dynamic adjustment coefficient is a core step for generating optimized tension control instructions. The initial value of the dynamic adjustment coefficient is 1, which is gradually modified through an iterative loop. The specific process includes: first, calculate the proportion of the current difference, then update the coefficient according to the proportion multiplied by a learning rate (such as 0.1), for example, new coefficient = old coefficient + learning rate × difference proportion. Repeat this process until the difference is less than the threshold. This iteration is similar to the gradient descent method, but simplifies the application to industrial control to ensure convergence. For example, in paper winding, if the material humidity causes the tension to increase, the iteration will gradually increase the coefficient to compensate for the resistance, thereby outputting the adjusted instruction to control the actuator such as the motor speed. This updating mechanism can effectively adapt to subtle differences in materials, such as thickness unevenness or surface roughness changes. Through multiple iterations, usually within 5 to 10 times, the optimized instructions are obtained, allowing the tension to stabilize around the target value. This detailed iterative process not only improves control accuracy, but also reduces material waste, resulting in higher finished product qualification rates in actual production.

[0086] Preferably, the obtained optimized tension control instructions are directly applied to the winding machine actuator.

[0087] For example, the instructions can adjust the torque or speed of the winding drum to achieve tension balance. In one scenario, for metal foil winding, the instructions will fine-tune the driving motor according to the updated coefficient to ensure that the material does not wrinkle. This adaptability makes the technical solution applicable to a variety of winding materials while remaining within the same industrial field. Further, in practical applications, multiple sets of sensor data can be combined to enhance the accuracy of the iteration.

[0088] For example, using position sensors for auxiliary feedback enriches the basis for comparison. This extension does not change the core logic but improves robustness.

[0089] It can be understood that the logical sequence of the entire process from feedback acquisition to instruction output ensures the coherence of the system. Through the above steps, the control instructions can respond to material variations in real time, achieving stable winding results.

[0090] S106, according to the optimized tension control instructions, using an adaptive filtering algorithm to process real-time feedback data to determine the final tension balance state, which is used to maintain the safety and reliability of the battery performance.

[0091] The real-time feedback data is obtained from the response of the battery electrode material, and the real-time feedback data is processed by using an adaptive filtering algorithm, wherein the adaptive filtering algorithm suppresses noise by adjusting filtering coefficients to obtain filtered data. For the filtered data, instruction optimization processing is performed in combination with the tension control instruction, wherein the instruction optimization processing fuses the fluctuation index in the filtered data and the target value in the tension control instruction to determine a safety threshold setting. According to the safety threshold setting, the charging current adjustment is monitored, and if the charging current adjustment exceeds the safety threshold setting, feedback loop adjustment is performed, wherein the feedback loop adjustment iteratively corrects the deviation to obtain an adjusted current value. The electrolyte flow monitoring is performed by using the adjusted current value, wherein the electrolyte flow monitoring tracks the correspondence between the flow rate and the adjusted current value to determine the final tension balance. In the final tension balance, the battery performance safety reliability is maintained, wherein the maintenance process integrates the state index of the final tension balance to maintain the balance.

[0092] Specifically, in an embodiment, the technical solution is applied to a tension control system in the battery production process, and the device parameters are adjusted by optimizing the tension control instruction.

[0093] Specifically, the system first collects real-time tension data of the battery winding link, for example, uses a sensor to monitor the tensile force value of the electrode material. These data are used as input to generate an initial control instruction.

[0094] It should be noted that the tension control instruction refers to a set of adjustment signals for the winding machine or the coating machine, for example, adjusting the speed or pressure, to prevent material deformation.

[0095] In an embodiment, in the lithium battery electrode coating scenario, the system optimizes the preset threshold value instruction, for example, automatically reduces the winding speed when the tension exceeds the safety range, thereby achieving preliminary tension stability. This optimization process emphasizes real-time performance, ensuring that the instruction adapts to the production needs of different battery specifications. Further, the adaptive filtering algorithm is a core step for processing real-time feedback data. The adaptive filtering algorithm is a signal processing method that dynamically adjusts parameters, which can automatically optimize filtering coefficients according to changes in input data, thereby removing noise and extracting valid information. In the field of battery performance maintenance, this algorithm is applied to process the tension signal feedback by the sensor, such as noise caused by vibration or temperature interference.

[0096] Specifically, the algorithm first initializes the filter parameters, such as setting the initial gain value, and then updates the coefficients by iteration.

[0097] In one possible implementation, after receiving real-time feedback data, the system adjusts the filter using the least mean square error criterion, for example, performing a weighted average of consecutive tension readings to gradually approach the true value. The advantage of this adaptive mechanism is that it can cope with uncertainties in battery production, such as minor changes in material thickness, thereby providing a more accurate data basis. Further expanding, on a lithium battery assembly line, when the feedback data contains burst noise, the algorithm increases the filter window width, for example, from 10 sampling points to 20, to enhance robustness. Through this processing, the system can extract a stable tension trend from the raw data for subsequent balance judgment. The detailed implementation of this step ensures the reliability of the technology, for example, in high-difficulty scenarios such as high-speed winding processes, the algorithm needs to process hundreds of data points per second, gradually updating the filter parameters: first calculate the current error, then adjust the coefficient according to the error gradient until the error is minimized. This iterative process involves multiple loops, for example, the initial loop uses historical data as a reference, and the subsequent loop incorporates new feedback, thereby achieving adaptive effect. In actual battery production, this helps to avoid misjudgment due to noise and ensures high precision of tension data. Based on the above processing, the final tension balance state is further determined.

[0098] In one embodiment, the balance state refers to the condition that the tension value is stable within a preset range, for example, for a specific battery type, the balance threshold is set to plus or minus 5% deviation. The system determines by comparing the processed data with the threshold, for example, if the tension fluctuation of multiple consecutive periods is less than the threshold, it is confirmed that the balance is reached.

[0099] Preferably, in the maintenance of battery performance safety and reliability, this determination process combines multi-sensor data fusion, for example, integrating tension and temperature feedback, to comprehensively evaluate the state.

[0100] For example, in the implementation scenario of a lithium battery winding device, the system first performs adaptive filtering on real-time data, and then generates optimization instructions based on the filtering results, such as adjusting the motor torque.

[0101] In one embodiment, when the tension imbalance is detected, the algorithm calculates the balance deviation value and iteratively optimizes until the deviation converges. This method demonstrates the versatility of the technology, which is applicable to different battery capacity specifications.

[0102] It can be understood that this scheme is used to maintain the performance, safety and reliability of the battery, and prevent battery short circuit or capacity attenuation caused by tension imbalance through the above steps.

[0103] In one possible implementation, after determining the balance state, the system outputs an alarm or automatically adjusts the parameters, for example, stopping the equipment in real time on the production line to avoid failure. This application ensures the long-term stability of the battery, for example, in an electric vehicle battery pack, maintaining tension balance can improve cycle life.

[0104] Specifically, the implementation details of the adaptive filtering algorithm are further described. The core of the algorithm lies in the adaptive updating mechanism, for example, using recursive least squares to adjust the filtering coefficients. In real-time feedback of the battery, the process includes: collecting the original data sequence, calculating the difference between the current estimated value and the actual value, and then updating the weight according to the difference.

[0105] Exemplarily.

[0106] In one embodiment, for the tension data sequence, the algorithm is initialized as a simple moving average, and then an adaptive factor is gradually introduced, such as dynamically changing the learning rate according to the noise level. This detailed process ensures that the algorithm is effective in complex battery environments, for example, when the noise increases under high temperature conditions, the algorithm will reduce the learning rate to stabilize the output. Through multiple iterations, such as 10 to 50 cycles, the system gradually determines the filtered data for tension balance judgment. The technical effect of this step lies in improving the accuracy of the data, thereby supporting reliable maintenance of battery performance. In another implementation, for the application of the battery test phase, the system optimizes the tension control instruction by incorporating a feedback loop.

[0107] For example, after real-time data is adaptively filtered, if the balance state is not reached, a new instruction is generated in a loop until it is stable. This loop mechanism enhances the flexibility of the scheme.

[0108] It should be noted that determining the final tension balance state can be achieved by a state machine model, for example, defining multiple states such as "unstable", "transition", and "balance", and transitioning according to the filtered data.

[0109] In one embodiment, in the battery aging test, this model helps monitor long-term tension changes to ensure safety. Further, the implementation of this technical solution in battery recycling processing is also applicable, for example, controlling tension to avoid damaging components during disassembly. By processing feedback data of recycling equipment through an adaptive algorithm, the next operation is performed after balancing.

[0110] For example, in a complete process of a lithium battery production line, the system starts from data acquisition, goes through filtering processing and instruction optimization, and finally outputs a balance state report. This integrated approach demonstrates the comprehensiveness of the scheme, which can maintain the safety and reliability of battery performance while improving production efficiency.

[0111] S107, extracting an internal defect risk indicator from the final tension balance state, and if the risk indicator is lower than a preset threshold, outputting a stable control signal to the winding machine to obtain a real-time balance result under continuous high-speed winding.

[0112] The internal defect data is obtained from the tension balance state, the defect density distribution, i.e. the statistical distribution of the defect data, is calculated, and the risk index value is extracted through the distribution. The winding speed monitoring data is obtained first, and the risk index value is adjusted through the data to determine the winding tension distribution. For the comparison between the risk index value and the preset threshold value, i.e. 0.5, if it is lower than the threshold value, the material uniformity parameter is generated by calculating the material density standard deviation through the uniformity evaluation. According to the material uniformity parameter, a stable control signal is output to the winding machine equipment, and the winding process stability is obtained through signal feedback. The winding parameters are adjusted through the winding process stability to obtain the real-time balance result under continuous high-speed winding.

[0113] Specifically, in an embodiment, the winding machine system first obtains the final tension balance state, which is determined by monitoring the tension sensor data during the winding process.

[0114] Specifically, the tension balance state refers to the overall mechanical distribution of the winding material when the tension value tends to be stable under high-speed rotation, for example, the average value and fluctuation range of the tension data collected in real time by the tension sensor are calculated to form a quantitative description of the balance state. The calculation process is as follows: n tension data points T1 to Tn are collected, the average value μ = (T1 + T2 +... + Tn) / n, where μ is the average tension and n is the number of data points; the fluctuation range R = max(Ti) - min(Ti), where R is the fluctuation range, Ti is a single data point, and max and min are the maximum and minimum values, respectively. The acquisition of such state helps to extract the internal defect risk index subsequently, ensuring the continuity of the winding process. Further, the process of extracting the internal defect risk index from the final tension balance state involves in-depth analysis of the tension data. The internal defect risk index is a quantitative index for evaluating the possible cracks or unevenness in the winding material, which may cause tension instability.

[0115] Exemplarily, the extraction process first performs Fourier transform spectral analysis on the tension fluctuation curve in the balance state, the input is the tension fluctuation curve data, and the output is the abnormal frequency peak value in the frequency spectrum, which corresponds to the potential defects in the material. Then, by calculating the deviation value of the fluctuation amplitude from the standard balance curve, the risk index is generated, for example, when the deviation value exceeds a certain proportion, it indicates high risk. This analysis ensures the accuracy of the index and is suitable for the processing of different materials such as metal wires or plastic films in the winding field.

[0116] It is necessary to point out that the extraction of internal defect risk indicators is based on the principle of tension balance: under the state of tension balance, internal defects in the material will cause local stress concentration, resulting in slight disturbance of the tension signal. By capturing these disturbances and quantifying, for example, converting the disturbance amplitude into a risk indicator, the system can identify problems early. This principle is embodied in the chain from sensor data to indicator generation in the business process, ensuring that the winding machine avoids sudden failure when running at high speed, thereby maintaining production efficiency.

[0117] In one possible implementation, determining whether the risk indicator is lower than the preset threshold is achieved by a comparison algorithm.

[0118] Specifically, the preset threshold is set according to historical winding data, for example, the threshold is 0.5, indicating that the risk score is considered safe when it is lower than this value. The system compares the extracted risk indicator with the threshold value, and if it is lower than the threshold value, it is confirmed that the internal defect risk of the material is controllable. This judgment step covers various scenarios in the winding business, such as setting a lower threshold when winding metal wire at high speed to ensure accuracy, or setting a slightly higher threshold when winding flexible materials to accommodate fluctuations.

[0119] Preferably, if the risk indicator is lower than the preset threshold, a stable control signal is output to the winding machine. This signal is a digital instruction for adjusting the speed of the winding machine or the tension regulator, for example, by sending a pulse signal through the controller to make the winding machine maintain the current parameters. This output process ensures the timeliness of the system's response, which in actual business can bring about an improvement in the stability of the winding process and avoid interruptions caused by defects.

[0120] For example, in the scenario of winding plastic film, after the system extracts the risk indicator and determines that it is lower than the threshold, it outputs a signal to maintain high-speed winding, resulting in real-time balancing. This result is manifested as uniform distribution of tension in the wound product, reducing the defect-induced scrap rate.

[0121] It can be understood that this technical solution exhibits generality in the winding field, for example, when applied to cable winding, it also achieves balancing control through the above steps. Further, the real-time balancing result under continuous high-speed winding is obtained by integrating the aforementioned steps.

[0122] Specifically, the balancing result includes the running log of the winding machine and the tension curve graph, showing that the defect risk has been controlled within a safe range. This result can achieve the goal of continuous production in business.

[0123] In one embodiment, the entire process focuses on a single winding machine, and after the indicator is extracted, the signal is judged and output to ensure the balancing of the winding.

[0124] The above merely describes a specific implementation of the present specification, and those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be described here. It should be understood that the protection scope of the present specification is not limited to this, and any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed in the present specification, and these modifications or replacements should be covered within the protection scope of the present specification.

Claims

1. A method for intelligent tension control during high-speed winding of lithium batteries, characterized in that, include: Tension and material thickness data are obtained from the winding machine's sensors to determine the production speed and ambient temperature and humidity values. The production speed value, the ambient temperature and humidity value, and the tension data are integrated using a data fusion method to construct an initial tension distribution matrix; An adaptive filtering algorithm is used to process the noise interference in the initial tension distribution matrix to obtain the filtered tension distribution matrix. Based on the filtered tension distribution matrix, the stress deviation values ​​between the multilayer materials are determined; If the force deviation value exceeds the preset threshold, the dynamic adjustment coefficient is calculated by the fuzzy logic control algorithm to obtain the adjusted tension target value; Based on the adjusted tension target value and the environmental change compensation factor, the compensated tension target value is determined. By comparing the feedback signal from the winding machine actuator with the compensated tension target value, the dynamic adjustment coefficient is iteratively updated to obtain the optimized tension control command. The real-time feedback data is processed according to the optimized tension control command to determine the final tension balance state.

2. The intelligent tension control method for high-speed winding of lithium batteries as described in claim 1, characterized in that, The process of acquiring tension data and material thickness data from the winding machine sensors to determine the production speed and ambient temperature and humidity values ​​includes: The tension data is collected in real time by the winding machine sensor, and the current production speed value is extracted from it. Based on the production speed value, the ambient temperature and humidity values ​​are monitored from the material thickness data, and the ambient temperature and humidity values ​​are integrated to obtain the ambient temperature and humidity values. Data preprocessing is performed on the tension data, the production speed value, and the ambient temperature and humidity value to remove outliers; The tension data, production speed value, and ambient temperature and humidity value are initially fused using a preset weighting method to generate fused data. Based on the fused data, the data consistency is verified. If the consistency is lower than the preset standard, the acquisition parameters are adjusted and the data is acquired again. The initial tension distribution matrix is ​​constructed using the verified fused data.

3. The intelligent tension control method for high-speed winding of lithium batteries as described in claim 1, characterized in that, The step of using an adaptive filtering algorithm to process noise interference in the initial tension distribution matrix to obtain a filtered tension distribution matrix includes: The initial tension distribution matrix is ​​processed by the adaptive filtering algorithm to automatically adjust the filtering parameters to suppress noise interference and generate a preliminary filtering matrix. For the aforementioned preliminary filtering matrix, material difference data are obtained from a multi-layer material property database; Based on the material difference data, the preliminary filtering matrix is ​​weighted and adjusted to generate the filtered tension distribution matrix; The tension distribution characteristics of each layer of material are extracted using the filtered tension distribution matrix. Based on the aforementioned tension distribution characteristics, calculate the preliminary force deviation value; The initial stress deviation value is corrected by combining the interface adhesion strength index to obtain the final stress deviation value.

4. The intelligent tension control method for high-speed winding of lithium batteries as described in claim 1, characterized in that, If the force deviation value exceeds a preset threshold, a dynamic adjustment coefficient is calculated using a fuzzy logic control algorithm to obtain the adjusted tension target value, including: Extract deviation distribution characteristics from the force deviation values ​​and combine them with the production speed values ​​to generate input data; The input data is processed by the fuzzy logic control algorithm, and the dynamic adjustment coefficient is calculated based on the preset rule base. Based on the dynamic adjustment coefficient, the tension values ​​of the filtered tension distribution matrix are adjusted to generate a preliminary adjustment matrix; Based on the initial adjustment matrix, distribution optimization is performed using winding thickness data to obtain the adjusted tension target value; The uniformity of tension distribution is verified using the adjusted tension target value; If the uniformity is lower than the preset standard, the dynamic adjustment coefficient is adjusted and the calculation is repeated.

5. The intelligent tension control method for high-speed winding of lithium batteries as described in claim 1, characterized in that, The step of determining the compensated tension target value by integrating the adjusted tension target value with an environmental change compensation factor includes: The environmental change compensation factor is extracted from the adjusted tension target value; The fuzzy logic control algorithm is used to fuse the ambient temperature and humidity values ​​with the environmental change compensation factor to generate preliminary compensation tension. Based on the initial compensation tension, obtain material deformation monitoring data and calculate the deformation rate to determine the compensation adjustment value under the influence of deformation; By combining the compensation adjustment value with the adhesive strength evaluation index, the compensated tension target value is generated. Verify the interlayer bonding quality parameters for the compensated tension target value; If the interlayer bonding quality parameters are lower than the preset standard, the compensation adjustment value is adjusted and then recalculated.

6. The intelligent tension control method for high-speed winding of lithium batteries as described in claim 1, characterized in that, The step of comparing the feedback signal from the winding machine actuator with the compensated tension target value and iteratively updating the dynamic adjustment coefficient to obtain an optimized tension control command includes: Acquire the feedback signal from the winding machine actuator and the winding speed monitoring data; By comparing the feedback signal with the compensated tension target value, it is determined whether the difference is greater than a preset threshold. If the difference is greater than the preset threshold, then the initial dynamic adjustment coefficient is determined; By fusing thickness sensor data with the initial dynamic adjustment coefficients, intermediate optimization coefficients are obtained through iterative updates. Based on the intermediate optimization coefficient, obtain the execution response data and determine whether the response deviation is greater than another preset threshold. If the response deviation is greater than another preset threshold, the dynamic adjustment coefficient is further iterated and updated to generate the optimized tension control command.

7. The intelligent tension control method for high-speed winding of lithium batteries as described in claim 1, characterized in that, The step of processing real-time feedback data according to the optimized tension control command to determine the final tension balance state includes: The real-time feedback data is obtained from the response of the battery electrode material; The real-time feedback data is processed by the adaptive filtering algorithm, the filtering coefficients are adjusted to suppress noise, and filtered data is generated. The filtered data is then optimized in conjunction with the optimized tension control command, and the fluctuation index and target value are fused to determine the safety threshold setting. Based on the safety threshold setting, monitor the charging current adjustment and determine whether it exceeds the safety threshold setting. If the current exceeds the set safety threshold, a feedback loop adjustment is performed to generate an adjusted current value. By incorporating the adjusted current value into the electrolyte flow monitoring, the final tension balance state is determined.

8. The intelligent tension control method for high-speed winding of lithium batteries as described in claim 1, characterized in that, After determining the final tension equilibrium state, the internal defect risk indicators are further processed, including: Extract internal defect data from the final tension equilibrium state; Risk index values ​​are generated by calculating the defect density distribution. The risk index value is adjusted based on the winding speed monitoring data to determine the winding tension distribution; The winding tension distribution is compared with a preset threshold to determine whether it is lower than the preset threshold. If the value is below the preset threshold, a material uniformity parameter is generated through uniformity evaluation. Based on the material uniformity parameters, a stable control signal is output to the winding machine to obtain stability data of the winding process and generate real-time balance results under continuous high-speed winding.

Images