Film tension control method for ultra-thin film winding

Abstract

The present application belongs to the technical field of industrial automation, and particularly relates to a film tension control method for ultra-thin film winding, which aims at the tension control problem of micron-level ultra-thin film material in the high-speed winding process, and the core innovation is that: the static winding diameter estimation logic based on thickness integration and the dynamic winding diameter calculation logic based on the ratio of linear velocity / angular velocity are run in parallel, and the optimal estimated winding diameter is generated through the weighted fusion algorithm based on the speed confidence; the time-varying moment of inertia is calculated in real time based on the optimal estimated winding diameter, and the total disturbance of the system is observed in real time; under the structure condition without mechanical floating roller, the precise tension adjustment is realized by using the feedforward compensation and closed-loop feedback cooperation. The present application effectively solves the contradiction between low-speed winding diameter calculation divergence and high-speed cumulative error in the ultra-thin film winding process, eliminates the friction dead zone limitation of the mechanical floating roller, and realizes the high-precision tension control in the full-speed domain and full-winding diameter range.

Description

Technical Field

[0001] This invention belongs to the field of industrial automation technology, and more specifically, relates to a thin film tension control method for ultrathin film winding. Background Technology

[0002] With the rapid development of data storage and the new energy vehicle industry, the demand for ultrathin film materials such as magnetized thin films and lithium-ion battery separators is increasing. The core materials for these products, such as LTO magnetic tape (approximately 6 μm thick), lithium battery separators (5-12 μm thick), and optical thin films (3-10 μm thick), require high-speed winding processes during production. Due to the extremely thin thickness and low tensile strength (typically below 50 MPa), even minute tension fluctuations can lead to tensile deformation, wrinkling, or even breakage. Therefore, ultrathin film winding places extremely high demands on the precision of tension control and dynamic response.

[0003] Existing winding tension control technologies mainly suffer from the following drawbacks: First, there are inherent limitations to mechanical floating roller systems. Traditional closed-loop tension control generally relies on floating rollers to absorb tension fluctuations. However, the floating roller mechanism itself has mass inertia and cylinder / spring friction. For low-tension control scenarios of ultrathin films (e.g., setting the tension to 0.5N-2N), the static friction of the mechanism can be as high as 0.5N-1N, forming a significant "mechanical dead zone." Within the dead zone, the control system cannot sense the actual tension changes, leading to control failure. Furthermore, the mechanical inertia of the floating roller limits the dynamic response bandwidth of the system, making it difficult to meet the requirements of high-speed start-stop conditions.

[0004] Secondly, there is a bottleneck in the accuracy of roll diameter calculation. Roll diameter is a fundamental parameter for calculating torque commands, and its accuracy directly affects the tension control effect. The commonly used linear velocity / angular velocity ratio method (D=v / ω) produces extremely large errors or even numerical singularities in the calculation results during equipment start-up, shutdown, or low-speed crawling phases because the denominator of the angular velocity approaches zero. While the simple thickness integration method is numerically stable, it fails to consider interlayer air entrainment and material thickness tolerances, resulting in increasingly larger cumulative errors as the roll diameter increases.

[0005] Third, dynamic inertia compensation is insufficient. The rotational inertia of the winding roller changes drastically with the fourth power of the roll diameter. During high-speed acceleration and deceleration (e.g., accelerating from 0 to 800 m / min in just 5 seconds), the torque required to overcome inertia is often much greater than the torque required to overcome tension. If the system inertia cannot be identified and compensated for in real time and accurately, it will inevitably lead to excessive tension during acceleration and slack tension during deceleration. Existing technologies typically use fixed inertia models or simple segmented compensation, which cannot adapt to the actual situation of continuously changing roll diameter.

[0006] Fourth, PID control has poor parameter adaptability. The winding process is a typical time-varying and nonlinear process. As the winding diameter changes from empty to full, the controlled object model of the system undergoes a huge change, and the equivalent time constant may change several times. A PID controller with fixed parameters cannot simultaneously meet the steady-state accuracy and dynamic response requirements for both empty and full winding.

[0007] Therefore, the field urgently needs a tension control method that can adapt to high-speed winding of ultrathin films, has extremely high dynamic response characteristics and steady-state accuracy. Summary of the Invention

[0008] In view of the above-mentioned defects or improvement needs of the prior art, the present invention provides a film tension control method in ultra-thin film winding. The purpose is to provide a tension precision control method that can adapt to high-speed winding of ultra-thin films, has extremely high dynamic response characteristics and steady-state accuracy, and overcomes the problems of low control accuracy of ultra-thin films caused by mechanical floating roller dead zone, roll diameter calculation error, insufficient inertia compensation and poor adaptability of PID parameters in the prior art.

[0009] To achieve the above objectives, according to one aspect of the present invention, a film tension control method for ultra-thin film winding is provided, which is applicable to rigid winding structures without floating rollers, wherein the following control steps are performed on the unwinding shaft and the take-up shaft respectively in each control cycle: The current sampling time of the reel is calculated using the static integration method. k diameter : Calculate layer thickness and from sampling time up to the sampling time k The product of the number of already wound layers is used as the roll diameter increment and superimposed onto the [the product]. , obtain the roll diameter The current sampling time of the reel is calculated using the dynamic ratio method. k diameter : Get the current sampling time of the reel k The angular velocity and linear velocity are used. If the absolute value of the angular velocity is greater than a preset threshold, the ratio of the linear velocity to the angular velocity is used as the roll diameter. Otherwise, make the volume diameter Equal to its static volume diameter; for and Weighted fusion is performed to obtain the optimal estimated roll diameter. This is especially important when the winding structure is operating in a low-speed phase or a preset unsteady-state phase. The weight is set to 0 when running in the preset steady-state high-speed phase. The weight is positively correlated with the linear velocity and is greater than The weight, weights and The weights sum to 1; The time-varying moment of inertia of the spool is calculated based on the optimal estimated spool diameter to calculate the dynamic torque; the internal friction force and external load disturbance of the winding structure are observed to generate a total disturbance estimate; the total disturbance estimate and the dynamic torque value are converted into a current feedforward signal. The reference angular velocity is calculated based on the preset main line speed and the optimal estimated roll diameter. A speed correction amount is generated by a variable gain PID controller based on the tension deviation between the current measured tension value and the target set value. The ratio of the empty roll diameter to the optimal estimated roll diameter is used as an adjustment coefficient to dynamically correct the gain coefficient. The sum of the reference angular velocity and the speed correction amount forms the final speed command, which is input to the servo motor speed loop to generate a current command. This command is superimposed with the current feedforward signal and input to the servo motor current loop to achieve coordinated control of feedforward compensation and closed-loop feedback.

[0010] Furthermore, the roll diameter The calculation method is as follows:

[0011] In the formula, Represents pi; Indicates from the sampling time up to the sampling time k The length of the wound film is obtained by integrating the linear velocity. This indicates the effective layer thickness of the film on the roll as it accumulates during the roll's rotation. , Indicates the nominal thickness of the film. This represents the air layer correction factor. , This indicates the preset target tension. Indicates the current sampling time k linear velocity; 、 、α These represent the preset values ​​of the air layer correction parameters. This is an empirical constant related to the surface roughness of thin film materials. To prevent the compensation coefficient from being zero in the denominator, α The speed-affected index.

[0012] further, The weights are determined using a preset curve, preferably an S-curve, and are expressed as follows:

[0013] In the formula, Indicates the current linear velocity Relative stable target value Confidence factor , The preset maximum dynamic weight; The slope factor of the preset S-curve controls the speed of the transition; This is the midpoint of the preset S-curve, i.e., the starting point where the weight accelerates from 0. Indicates based on The determined The weight.

[0014] Furthermore, the value obtained by the dynamic ratio method Multiple roll diameters determined at a predetermined number of sampling times adjacent to each other Sum and average, then update the current sampling time. lower volume This is used for subsequent weighted fusion, where the number of adjacent elements preceding each other is negatively correlated with the linear velocity.

[0015] Furthermore, when performing the weighted fusion, the method also includes: calculating the deviation in real time. ,like If slippage or tape breakage is detected, the optimal estimated roll diameter determined at the previous sampling time is used as the optimal estimated roll diameter at the current sampling time to execute subsequent steps and trigger an alarm signal; where, This indicates the preset deviation threshold.

[0016] Furthermore, dynamic torque ;in, Represents the time-varying moment of inertia. The angular acceleration of the reel is represented by a differential tracker, which is extracted using a nonlinear track-differentiator structure.

[0017] Furthermore, the gain coefficient is corrected as follows: proportional gain Integral gain ; in, , These are the baseline PID parameters for an empty volume. The diameter of the empty roll. This represents the optimal estimated roll diameter.

[0018] Furthermore, an extended state observer is used to generate an estimate of the total disturbance, which is achieved as follows:

[0019]

[0020] In the formula, The preset sampling period, To control the gain, it represents the equivalent torque effect produced by a unit current. , Indicates the current sampling time The measured value of the rotational speed, The data observed by the observer at the previous sampling time is used for the current sampling time. The estimated rotational speed; The data observed by the observer at the current sampling time is used for the next sampling time. The estimated rotational speed; , The observer gain is a preset value, selected according to the bandwidth configuration principle. Let the desired bandwidth of the observer be... Then in the continuous domain , , The preferred value is 3-5 times the preset speed loop bandwidth; The current sampling time The control quantity, namely the current command.

[0021] Furthermore, before executing the control steps, the winding equipment is soft-started in the following manner: At startup, switch all servo motors of the reel to torque control mode and disable the speed loop; Control each reel servo motor to apply a preset proportion of initial preload torque The measured value of film tension Gradually increase to the tension set value The tension state is maintained for a preset duration, and the tension fluctuation amplitude is monitored. When the tension fluctuation amplitude is within a preset range and there is no preset abnormal vibration, the current static roll diameter is locked. As the initial roll diameter value This is used for subsequent calculations and to perform a seamless switching, smoothly transitioning each reel servo drive from torque control mode to speed closed-loop control mode, and executing the control steps to start the normal winding process.

[0022] According to another aspect of the present invention, an ultrathin film winding system is provided, comprising: an unwinding shaft and its servo motor, a take-up shaft and its servo motor, and a controller; wherein the controller is configured to cooperate with each servo motor of the reel to perform the control steps described above.

[0023] In summary, compared with the prior art, the technical solutions conceived by this invention have the following beneficial effects: 1. This invention proposes a floating roller tension control method based on dual roll diameter confidence fusion. By parallel operation of the static integral method and the dynamic ratio method, and weighted fusion based on speed confidence, the static integral method is numerically stable in the low-speed range, while the dynamic ratio method has high accuracy in the high-speed range. The two methods complement each other, and an anomaly elimination mechanism is introduced to prevent slippage or belt breakage from causing abrupt changes in roll diameter. This enables the continuous and smooth optimal estimated roll diameter across the entire speed range, effectively suppressing low-speed calculation divergence and high-speed cumulative errors, and keeping the roll diameter estimation accuracy within a small deviation range. A step is introduced to observe and actively compensate for the total system disturbance in real time. The total disturbance includes friction fluctuations, load fluctuations, and other unmodeled dynamics, which are packaged into a total disturbance estimate and actively offset by a compensation term in the control law. This enables real-time suppression of internal system disturbances and external load changes without relying on mechanical floating roller buffers, significantly improving the robustness of tension control. In addition, a cascaded control architecture of "speed inner loop + tension outer loop" is constructed. The input of the tension outer loop is the tension deviation, and the output is the speed deviation. The input of the speed inner loop is the speed deviation, and the output is the current command. A variable gain PID controller and current feedforward are used for coordinated adjustment. By scheduling PID parameters based on the inverse of the optimal estimated roll diameter, the increased lever arm and slower response caused by the increase in roll diameter are compensated. Since the reference angular velocity feedforward undertakes the main synchronous tracking task, the variable gain PID only needs to handle the remaining small deviations. At the same time, the current feedforward compensates for the effects of inertia and disturbances in advance, which can maintain the constant control loop gain within the entire roll diameter variation range, achieve sub-Newton level micro-tension control, and achieve high steady-state tension accuracy. It also completely eliminates the friction dead zone limitation of the mechanical floating roller, effectively overcoming the problems of low control accuracy of ultra-thin films caused by the dead zone of the mechanical floating roller, roll diameter calculation error, insufficient inertia compensation, and poor adaptability of PID parameters in the existing technology. It realizes a tension precision control method that can adapt to high-speed winding of ultra-thin films and has extremely high dynamic response characteristics and steady-state accuracy.

[0024] 2. In the static estimation of roll diameter, this invention introduces an air layer correction coefficient to dynamically compensate for the roll diameter deviation caused by interlayer air entrainment based on linear velocity and tension, thereby ensuring the accuracy of theoretical roll diameter calculation and thus ensuring the control accuracy of ultra-thin film winding.

[0025] 3. This invention introduces an extended state observer to perform real-time observation and active compensation of the total system disturbance. A second-order linear extended state observer is constructed, which is recursively updated by the error feedback between the previous speed estimate and the current sampling time. Since the observer can package friction fluctuations, load fluctuations and other unmodeled dynamics into the total disturbance estimate and actively cancel them out by setting compensation terms in the control law, it can achieve real-time suppression of internal system disturbances and external load changes without relying on mechanical floating roller buffers, significantly improving the robustness of tension control. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating a method for controlling film tension during ultrathin film winding, provided as an embodiment of the present invention.

[0027] Figure 2 This is a schematic diagram of the thin film tension control method provided in an embodiment of the present invention.

[0028] Figure 3 This is a structural diagram of an ultrathin film winding system provided in an embodiment of the present invention.

[0029] In all the accompanying drawings, the same reference numerals are used to denote the same elements or structures, wherein: 1 is the unwinding unit, 2 is the tension roller, 3 is the idler roller, 4 is the drive roller and pressure roller, 5 is the speed roller, 6 is the winding unit, and 7 is the equipment base. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0031] Example 1 A method for controlling film tension in ultrathin film winding, specifically for rigid winding structures without floating rollers, such as... Figure 1 As shown, the following control steps are performed on the unwinding and rewinding shafts respectively in each control cycle: The current sampling time of the reel is calculated using the static integration method. k diameter : Calculate layer thickness and from sampling time up to the sampling time k The product of the number of already wound layers is used as the roll diameter increment and superimposed onto the [the product]. , obtain the roll diameter The current sampling time of the reel is calculated using the dynamic ratio method. k diameter : Get the current sampling time of the reel k The angular velocity and linear velocity are used. If the absolute value of the angular velocity is greater than a preset threshold, the ratio of the linear velocity to the angular velocity is used as the roll diameter. Otherwise, make the volume diameter Equal to its static volume diameter; for and Weighted fusion is performed to obtain the optimal estimated roll diameter. This is especially important when the winding structure is operating in a low-speed phase or a preset unsteady-state phase. The weight is set to 0 when running in the preset steady-state high-speed phase. The weight is positively correlated with the linear velocity and is greater than The weight, weights and The weights sum to 1; The time-varying moment of inertia of the spool is calculated based on the optimal estimated spool diameter to calculate the dynamic torque; the internal friction force and external load disturbance of the winding structure are observed to generate a total disturbance estimate; the total disturbance estimate and the dynamic torque value are converted into a current feedforward signal. The reference angular velocity is calculated based on the preset main line speed and the optimal estimated roll diameter. A speed correction amount is generated by a variable gain PID controller based on the tension deviation between the current measured tension value and the target set value. The ratio of the empty roll diameter to the optimal estimated roll diameter is used as an adjustment coefficient to dynamically correct the gain coefficient. The sum of the reference angular velocity and the speed correction amount forms the final speed command, which is input to the servo motor speed loop to generate a current command. This command is superimposed with the current feedforward signal and input to the servo motor current loop to achieve coordinated control of feedforward compensation and closed-loop feedback.

[0032] This embodiment proposes a floating roller tension control method based on dual roll diameter confidence fusion. It utilizes parallel execution of the static integral method and the dynamic ratio method, weighted and fused based on speed confidence. Since the static integral method is numerically stable in the low-speed range and the dynamic ratio method is highly accurate in the high-speed range, the two methods complement each other. Furthermore, an anomaly rejection mechanism is introduced to prevent roll diameter jumps caused by slippage or belt breakage, enabling continuous and smooth optimal roll diameter estimation across the entire speed range. This effectively suppresses low-speed calculation divergence and high-speed cumulative errors, keeping the roll diameter estimation accuracy within a small deviation range. A real-time observation and active compensation stage for the total system disturbance is introduced. The total disturbance includes friction fluctuations, load fluctuations, and other unmodeled dynamics, packaged into a total disturbance estimate, and actively offset by a compensation term in the control law. This achieves real-time suppression of internal system disturbances and external load changes without relying on mechanical floating roller buffering, significantly improving the robustness of tension control. Additionally, a cascaded control architecture of "speed inner loop + tension outer loop" is constructed, with the input of the tension outer loop... The input of the speed inner loop is the speed deviation, and the output is the current command. A variable gain PID controller and current feedforward are used for coordinated adjustment. By scheduling PID parameters based on the inverse of the optimal estimated roll diameter, the increase in lever arm and slowdown in response caused by the increase in roll diameter is compensated. Since the reference angular velocity feedforward undertakes the main synchronous tracking task, the variable gain PID only needs to handle the remaining small deviations. At the same time, the current feedforward compensates for the effects of inertia and disturbances in advance, which can maintain the constant control loop gain across the entire roll diameter variation range, achieving sub-Newton level micro-tension control with high steady-state tension accuracy. It also completely eliminates the friction dead zone limitation of the mechanical floating roller, effectively overcoming the problems of low control accuracy of ultra-thin films caused by the dead zone of the mechanical floating roller, roll diameter calculation error, insufficient inertia compensation, and poor adaptability of PID parameters in the existing technology. It realizes a tension precision control method that can adapt to high-speed winding of ultra-thin films and has extremely high dynamic response characteristics and steady-state accuracy.

[0033] This embodiment can be described as a method for controlling the winding tension of ultrathin films based on dual roll diameter confidence fusion and perturbation feedforward. Ultrathin films refer to thin film materials with a thickness ≤10μm. For example... Figure 2 As shown, the core innovation of this method lies in the construction of a three-in-one control architecture of "observation-compensation-adjustment", in which the three core modules form a deeply coupled data closed loop through optimal estimation of the roll diameter.

[0034] First, real-time estimation and fusion of dual roll diameters. This embodiment runs two roll diameter calculation logics in parallel: Logic A is a static estimation based on film thickness integration, which uses the linear velocity integration of the main traction roller and the cumulative layer thickness to calculate the theoretical superposition of the roll diameter; Logic B is a dynamic calculation based on the ratio of linear velocity to angular velocity, which uses encoder feedback to calculate the instantaneous physical roll diameter.

[0035] The key innovation lies in the weighted fusion algorithm based on speed confidence: when the system is in an unsteady state (acceleration or deceleration) or running at low speed, the main weights are allocated to the static integral method, which uses its numerical stability to avoid computational noise caused by speed fluctuations; when the system is in a steady state running at high speed, the weights of the dynamic ratio method are gradually increased, and the high precision of the encoder is used to correct the accumulated error.

[0036] Second, inertia compensation and disturbance observation. Based on the optimal estimated roll diameter, the time-varying moment of inertia is calculated in real time and decomposed into three parts: the fixed inertia of the motor, the fixed inertia of the roll shaft, and the material variable inertia that varies with the fourth power of the roll diameter. The time-varying moment of inertia of the roll shaft is calculated based on the optimal estimated roll diameter and used to calculate the dynamic torque. This torque is directly added to the current loop as a feedforward without PID control, thus achieving real-time inertia compensation.

[0037] Simultaneously, the total disturbance of the real-time observation system, including unmodeled dynamics such as frictional torque fluctuations, changes in material elastic modulus, and mechanical transmission clearances, is actively canceled out in the control law. Observation and inertia feedforward complement each other: observation handles slowly varying disturbances and model uncertainties, while inertia feedforward handles deterministic disturbances related to acceleration.

[0038] Third, PID-based closed-loop tension regulation. In a structure without mechanical floating roller buffer, a cascaded control architecture of "inner speed loop + outer tension loop" is established. The outer tension loop's input is the tension deviation, and its output is the speed deviation; the inner speed loop's input is the speed deviation, and its output is the current command. The main speed command calculates the reference angular velocity based on the main linear velocity (the set target linear velocity) and the optimal estimated roll diameter. Tension PID fine-tuning utilizes the deviation between the tension sensor feedback value and the set value, calculating the speed correction amount through a variable gain PID controller. The variable gain strategy, based on the reciprocal of the roll diameter, schedules PID parameters to compensate for the impact of roll diameter changes on the control loop gain, ensuring consistent control performance across the entire roll diameter range. Finally, the speed command is input to the servo motor speed loop, generating a current command. This command is superimposed on the current feedforward signal and input to the servo motor current loop, achieving coordinated optimization of feedforward and feedback.

[0039] The aforementioned observation-compensation-adjustment system is deeply coupled through optimal estimated roll diameter: it simultaneously serves the gain scheduling law of variable gain PID, the rotational inertia calculation model of inertia feedforward, and the control gain configuration for observed disturbances, forming a dedicated control closed loop for "ultra-thin film time-varying large inertial nonlinear systems". This three-in-one architecture design enables synergistic effects among the modules—dual roll diameter correction provides accurate system state estimation, inertia compensation and disturbance observation provide precise feedforward compensation, and velocity closed-loop tension adjustment provides rapid feedback correction. The three work together to achieve control performance that traditional methods cannot achieve.

[0040] In this embodiment, by removing the floating roller, the lower limit of tension control is no longer constrained by cylinder friction, enabling sub-Newton level micro-tension control, which is particularly suitable for ultrathin films with low tensile strength. Accurate roll diameter estimation: The dual roll diameter confidence fusion algorithm helps obtain continuous and smooth roll diameter data across the entire speed range, effectively suppressing computational divergence at low speeds and reducing accumulated errors at high speeds. The roll diameter estimation accuracy can be controlled within a small deviation range. High dynamic response: Inertia feedforward compensation based on a physical model, combined with disturbance observation, effectively suppresses tension fluctuations during rapid acceleration and deceleration, achieving millisecond-level fast response. Stable performance across all operating conditions: The variable gain PID strategy helps maintain a relatively constant control loop gain across a range of roll diameter variations, achieving high steady-state tension control accuracy.

[0041] As a preferred embodiment, the roll diameter The calculation method is as follows:

[0042] In the formula, Represents pi; Indicates from the sampling time up to the sampling time k The length of the wound film can be obtained by integrating the linear velocity. This indicates the effective layer thickness of the film on the roll as it accumulates during the roll's rotation. , Indicates the nominal thickness of the film. This represents the air layer correction factor. , This indicates the preset target tension. Indicates the current sampling time k linear velocity; 、 、α These represent the preset values ​​of the air layer correction parameters. This is an empirical constant related to the surface roughness of thin film materials. To prevent the compensation coefficient from being zero in the denominator, α The speed-affected index.

[0043] Each control cycle in the preferred control method ≤1ms. The linear velocity of the main traction roller is monitored at the sampling time in each control cycle. and film thickness As the roll rotates, the roll diameter increases with each layer of film accumulated. After considering the air layer correction, the effective layer thickness is... Among them, the air layer correction factor Based on the current linear velocity and tension setting value Dynamic calculation: An air layer correction factor is introduced in the static estimation to dynamically compensate for the roll diameter deviation caused by interlayer air entrainment based on linear velocity and tension.

[0044] In the dynamic ratio method, it is necessary to read the angular velocity of the reel. (unit: rad / s and linear velocity (unit: m / min It needs to be converted to m / s ).when (Recommendations 1-5) rad / s When the dynamic volume is... .

[0045] To prevent measurement noise, further optimization can be made for... A moving average filter is applied, with the filter window length adaptively adjusted based on speed: the window shortens at high speeds to improve response and lengthens at low speeds to suppress noise. In practice, the filter window length is calculated using a dynamic ratio method. Multiple roll diameters determined at a predetermined number of sampling times adjacent to each other Sum and average, then update the current sampling time. lower volume This is used for subsequent weighted fusion, where the number of adjacent elements preceding each other is negatively correlated with the linear velocity.

[0046] This can be considered a preferred implementation method. The weights are determined using a preset curve, preferably an S-curve, and are expressed as follows:

[0047] In the formula, Indicates the current linear velocity Relative stable target value Confidence factor , The preset maximum dynamic weight; The slope factor of the preset S-curve controls the speed of the transition; This is the midpoint of the preset S-curve, i.e., the starting point where the weight accelerates from 0. Indicates based on The determined The weight.

[0048] This method requires defining a velocity confidence factor, when β<1 When (speed has not reached steady state) or This indicates that the system is in a low-speed or rapidly changing speed state, at which point the dynamic roll diameter is severely affected by noise interference, and the dynamic weights... ;whenβ≥1 and hour, Increasing according to the above S-curve formula, where A value of 0.7 to 0.9 is recommended. Recommended values , A value of 1.2 is recommended, meaning that a noticeable switch begins after the speed exceeds the steady-state threshold by 20%. Optimal estimated roll diameter. .

[0049] The weighted fusion algorithm based on speed confidence proposed in this embodiment can, in specific implementation, first set a low-speed threshold. and steady-state threshold When linear velocity or angular acceleration At that time, the system is in the transient / low-speed region, and the fusion weights are... Optimal estimated volume diameter When linear velocity and angular acceleration At that time, the system is in the steady-state high-speed region, and the dynamic weights... The optimal estimated roll diameter is obtained by increasing the diameter according to a preset curve. Introduce an anomaly removal mechanism: if If slippage or tape breakage occurs, the roll diameter value from the previous moment will be maintained and an alarm will be triggered.

[0050] As a further preferred embodiment, when performing the weighted fusion, the method further includes: calculating the deviation in real time. ,like If slippage or tape breakage is detected, the optimal estimated roll diameter determined at the previous sampling time is used as the optimal estimated roll diameter at the current sampling time to execute subsequent steps and trigger an alarm signal; where, This indicates the preset deviation threshold.

[0051] Based on optimal estimated volume diameter The time-varying moment of inertia is calculated in real time. The time-varying moment of inertia is the sum of the known fixed inertia of the motor and the fixed inertia of the reel, as well as the material's variable inertia, which varies with the fourth power of the optimal estimated reel diameter. In other words, the total moment of inertia is decomposed into: ,in and For fixed inertia, the fixed inertia of the motor and the fixed inertia of the reel are respectively, which can be calibrated through no-load tests; for variable inertia of materials... It varies with the fourth power of the roll diameter. Indicates the material's moment of inertia; Indicates the density of the thin film material; Indicates the width of the film; This represents the fourth power of the optimal estimated roll diameter; This represents the fourth power of the core diameter.

[0052] As a preferred implementation method, dynamic torque ;in, Represents the time-varying moment of inertia. The angular acceleration of the reel is extracted using a differential tracker. This tracker employs a nonlinear track-differentiator structure, enabling it to obtain a high-quality differential signal while suppressing noise. Dynamic torque. Without PID control, it is directly added to the current loop command as a feedforward quantity to achieve real-time inertia compensation.

[0053] Angular acceleration is extracted by a differential tracker, which adopts a nonlinear track-differentiator structure. The input signal is tracked by constructing a nonlinear function and its output is differentiated. Since this structure can actively separate high-frequency noise while tracking rapidly changing signals, it can obtain a smooth angular acceleration signal with small phase delay under the premise of effectively suppressing noise amplification, which serves as the basis for dynamic torque calculation.

[0054] As a preferred implementation, an extended state observer is used to generate an estimate of the total disturbance, which is achieved as follows:

[0055]

[0056] In the formula, The preset sampling period, To control the gain, the equivalent angular acceleration generated per unit current is expressed as (its value is the motor torque constant divided by the total moment of inertia of the system, with dimensions of 1 / 2π). ); , Indicates the current sampling time The measured rotational speed value comes from the encoder. The data observed by the observer at the previous sampling time is used for the current sampling time. The estimated rotational speed; The data observed by the observer at the current sampling time is used for the next sampling time. The estimated rotational speed; , The observer gain is a preset value, selected according to the bandwidth configuration principle. Let the desired bandwidth of the observer be... Then in the continuous domain , , The preferred value is 3-5 times the preset speed loop bandwidth. This is the current control quantity (unit: A, i.e., current command).

[0057] The Extended State Observer (ESO) is used to estimate two key variables in real time: the actual rotational speed of the reel system. Secondly, the total disturbance experienced by the system. (unit: (including the equivalent torque of all unknown disturbances such as friction fluctuations, sudden load changes, and material elasticity changes) in each control cycle. Within this period, the estimated value is updated according to the aforementioned state equation.

[0058] formula This indicates that the estimated rotational speed at the next moment = the estimated current rotational speed + sampling time × (current estimated disturbance). + angular acceleration effect generated by electric current + Correction term for observation error ).

[0059] formula This indicates that the estimated disturbance at the next moment = the current estimated disturbance + sampling time × (another correction term for the observation error). ).

[0060] As the recursion proceeds, the observation error... It will tend to zero, at which point Accurately track the actual rotational speed. It accurately reflects the true total disturbance.

[0061] Observer Gain , Based on the bandwidth configuration principle, let the expected bandwidth of the observer be... Then in the continuous domain , The recommended observer bandwidth is 3-5 times that of the velocity loop bandwidth, thus ensuring that the ESO can quickly estimate the disturbance (faster than the velocity loop response) without excessively amplifying the measurement noise.

[0062] Final control quantity u Calculated according to the control law ,in The current command output by the tension PID controller, This is a disturbance compensation term used to offset the total disturbance estimated by the ESO. In this invention, the dynamic torque compensation amount (from the inertia compensation module) and the total disturbance estimate are... Multiply by the current time-varying moment of inertia The resulting disturbance torques are added together and divided by the torque constant. This is converted into a total current feedforward signal, which is then superimposed on the speed loop output and fed into the current loop.

[0063] This preferred embodiment introduces an extended state observer to perform real-time observation and active compensation of the total system disturbance. A second-order linear extended state observer is constructed, which is recursively updated by the error feedback between the previous speed estimate and the current sampling time. Since the observer can package friction fluctuations, load fluctuations and other unmodeled dynamics into the total disturbance estimate and actively cancel them out by setting compensation terms in the control law, it can achieve real-time suppression of internal system disturbances and external load changes without relying on mechanical floating roller buffers, significantly improving the robustness of tension control.

[0064] As a preferred implementation method, the gain coefficient can be corrected by: proportional gain Integral gain ;in, , These are the baseline PID parameters for an empty volume. The diameter of the empty roll. This represents the optimal estimated roll diameter.

[0065] In a structure without mechanical floating rollers, a cascaded control architecture of "inner speed loop + outer tension loop" is established. Main speed command. The calculation is divided into two parts: velocity feedforward. ,in Main linear velocity (unit: m / min). For optimal estimated roll diameter (unit: mm); speed correction The variable gain PID controller calculates the tension deviation.

[0066] The gain scheduling strategy of the variable gain PID controller is based on the inverse of the optimal estimated convolution diameter: proportional gain. Integral gain ,in , These are the baseline PID parameters for an empty volume. The empty roll diameter is used as the scheduling strategy. The physical meaning of this strategy is that as the roll diameter increases, the system lever arm increases and the response slows down. By increasing the controller gain, the attenuation of the loop gain is compensated, ensuring consistent control performance across the entire roll diameter range.

[0067] Final speed command The speed loop is directly output to the servo driver. Simultaneously, the dynamic torque corresponding to the time-varying moment of inertia is... and the total disturbance estimated by ESO Convert to current feedforward signal: ,in This represents the motor torque constant. Current feedforward and speed closed-loop work synergistically: the speed loop handles rapid tracking, while the current feedforward provides precise compensation; their combination enables high-precision tension control during ultrathin film winding.

[0068] Before implementing the method of this invention, a hardware architecture is established and initialized. For example, this embodiment uses an FPGA-based real-time motion controller as the core computing unit PLC, connected to a high-performance servo driver Servo via an EtherCAT isochronous bus. The PLC and Servo communicate via the ECAT bus, and the scan time (i.e., the value is equivalent to the control cycle) is set to 1ms. The servo driver uses an AC servo system with a current loop bandwidth ≥2kHz and a speed loop bandwidth ≥1kHz to ensure rapid response to control commands. The traction shaft is equipped with a high-resolution absolute encoder (e.g., 23-bit) responsible for establishing the main speed reference for the entire line. The winding shaft operates in speed closed-loop mode, receiving speed commands from the motion controller.

[0069] The tension detection device is installed on a fixed guide roller before the winding station, employing a high-sensitivity tension sensor (range 0-50N, accuracy 0.1%FS). The sensor's installation wrap angle is designed to be constant (recommended 120°-150°) to ensure that the tension detection accuracy is not affected by the lateral offset of the roll material. A key design feature is that no mechanical floating rollers or rocker arms loaded by cylinders, springs, or counterweights are installed along the unwinding to rewinding path; the tension sensor is directly fixed to the frame, forming a rigid tension detection structure without floating rollers.

[0070] The system initialization is as follows: First, configure the basic parameters of the thin film material, including the nominal thickness. (Unit: μm), Width (Unit: mm), density (Unit: kg / m³), Elastic modulus (Unit: GPa) and core diameter (Unit: mm). Then set the control parameters: low speed threshold. (Recommended speed: 10-30 m / min) Steady-state threshold (Recommended speed: 100-200 m / min), angular acceleration threshold (Recommended 50-100 rad / s²), safety deviation threshold (5-10mm recommended). Finally, configure the relevant parameters for the air layer correction factor. , , α These parameters can be obtained through offline calibration experiments.

[0071] As a preferred implementation, the winding equipment can be soft-started before executing the control steps, in the following manner: At startup, switch all servo motors of the reel to torque control mode and disable the speed loop; Control the servo motors of each reel to apply the initial preload torque The measured value of film tension Gradually increase to the tension set value The system maintains the tension at a preset ratio for a preset duration, monitors tension fluctuations, and stops the machine to troubleshoot any abnormalities if the fluctuations are significant. If the tension fluctuations are within a preset range and there are no preset abnormal vibrations, the current static roll diameter is locked. As the initial roll diameter value This is used for subsequent calculations and to perform a seamless switching, smoothly transitioning each reel servo drive from torque control mode to speed closed-loop control mode, and executing the control steps to start the normal winding process.

[0072] In response to the characteristics of low tensile strength and easy breakage of ultrathin films, this embodiment designs a special soft-start program with high startup safety: the soft-start program effectively avoids tension overshoot at startup and reduces the risk of ultrathin film breakage through torque pre-tightening and smooth mode switching.

[0073] In practice, the soft boot procedure can be exemplified by the following steps: Step 1: At the moment of system startup, switch the servo motor to torque control mode and disable the speed loop; Step 2: Set the initial preload torque This makes the tension sensor feedback value Gradually increase to the tension set value The rise time should be controlled within 0.5 to 1 second, and the rise rate should be 30% to 50%. Step 3: Maintain this tension for 0.5 to 1 second, monitor the tension fluctuation amplitude, and confirm that the tension is stable (fluctuation < ±5%) and there is no abnormal vibration; Step 4: Lock the current static volume size As the initial roll diameter value This is used for subsequent roll diameter calculations; Step 5: Perform a seamless switch to smoothly transition the servo drive from torque control mode to speed closed-loop control mode and start the normal winding process.

[0074] The soft-start procedure effectively prevents tension overshoot during startup, making it particularly suitable for ultrathin films with tensile strengths below 50 MPa. The seamless switching technology ensures continuous torque command during mode switching, avoiding any impact.

[0075] The following example, taking a complete process cycle from startup to stable operation and then to shutdown, illustrates the actual control timing of the method of this invention: (1) Initialization phase (t=0-1s): After the system is powered on, material parameters and control parameters are loaded, friction identification program is executed (optional), and initial roll diameter estimation is established.

[0076] (2) Soft start phase (t=1-10s): The above soft start procedure is executed, and the tension gradually increases from 0 to 30%~50% of the set value, and then switches to speed closed loop mode without disturbance.

[0077] (3) Acceleration phase (t=10-30s): The linear velocity accelerates from 0 to the target velocity (e.g., 800m / min), lasting 5 seconds. During this phase, the dual convolution diameter estimation primarily uses the static integration method. <0.3), inertia feedforward provides the main accelerating torque, and ESO observes and compensates for frictional disturbances. Tension fluctuations are controlled within ±5%.

[0078] (4) Steady-state operation stage (t>30s): The linear velocity stabilizes at the target value, and the dual roll diameter estimation gradually transitions to the dynamic ratio method as the dominant method. The encoder feedback is used to correct accumulated errors. The variable gain PID adjusts parameters in real time according to the current roll diameter, maintaining a tension steady-state accuracy of better than ±2%.

[0079] (5) Deceleration and stopping stage: The linear speed is reduced from the target value to 0, and the inertia feedforward provides braking torque to prevent tension slack. After deceleration is completed, you can choose to perform winding or roll changing operations.

[0080] Example 2 An ultrathin film winding system includes: an unwinding shaft and its servo motor, a take-up shaft and its servo motor, and a controller; wherein the controller is used to coordinate with the servo motors of each shaft to execute the control steps as described in Embodiment 1. In a conventional configuration, the winding system may also include idler rollers, power rollers and their servo motors, pressure rollers, tension rollers and their servo motors, and speed measuring rollers, such as... Figure 3 As shown.

[0081] The relevant technical solutions are the same as above, and will not be repeated here.

[0082] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for controlling film tension in ultrathin film winding, characterized in that, For a rigid winding structure without floating rollers, the following control steps are performed on the unwinding shaft in each control cycle: The current sampling time of the reel is calculated using the static integration method. k diameter : Calculate layer thickness and from sampling time up to the sampling time k The product of the number of already wound layers is used as the roll diameter increment and added to... , obtain the roll diameter The current sampling time of the reel is calculated using the dynamic ratio method. k diameter : Get the current sampling time of the reel k The angular velocity and linear velocity are used. If the absolute value of the angular velocity is greater than a preset threshold, the ratio of the linear velocity to the angular velocity is used as the roll diameter. Otherwise, make the volume diameter Equal to its static volume diameter; for and Weighted fusion is performed to obtain the optimal estimated roll diameter. This is especially important when the winding structure is operating in a low-speed phase or a preset unsteady-state phase. The weight is set to 0 when running in the preset steady-state high-speed phase. The weight is positively correlated with the linear velocity and is greater than The weight, weights and The weights sum to 1; The time-varying moment of inertia of the spool is calculated based on the optimal estimated spool diameter to calculate the dynamic torque; the internal friction force and external load disturbance of the winding structure are observed to generate a total disturbance estimate; the total disturbance estimate and the dynamic torque value are converted into a current feedforward signal. The reference angular velocity is calculated based on the preset main line speed and the optimal estimated roll diameter. A speed correction amount is generated by a variable gain PID controller based on the tension deviation between the current measured tension value and the target set value. The ratio of the empty roll diameter to the optimal estimated roll diameter is used as an adjustment coefficient to dynamically correct the gain coefficient. The sum of the reference angular velocity and the speed correction amount forms the final speed command, which is input to the servo motor speed loop to generate a current command. This command is superimposed with the current feedforward signal and input to the servo motor current loop to achieve coordinated control of feedforward compensation and closed-loop feedback.

2. The thin film tension control method as described in claim 1, characterized in that, roll diameter The calculation method is as follows: In the formula, Represents pi; Indicates from the sampling time up to the sampling time k The length of the wound film is obtained by integrating the linear velocity. This indicates the effective layer thickness of the film on the roll as it accumulates during the roll's rotation. , Indicates the nominal thickness of the film. This represents the air layer correction factor. , This indicates the preset target tension. Indicates the current sampling time k linear velocity; 、 、α These represent the preset values ​​of the air layer correction parameters. This is an empirical constant related to the surface roughness of thin film materials. To prevent the compensation coefficient from being zero in the denominator, α The speed-affected index.

3. The thin film tension control method as described in claim 1, characterized in that, The weights are determined using a preset curve, preferably an S-curve, and are expressed as follows: In the formula, Indicates the current linear velocity Relative stable target value Confidence factor , The preset maximum dynamic weight; The slope factor of the preset S-curve controls the speed of the transition; This is the midpoint of the preset S-curve, i.e., the starting point where the weight accelerates from 0. Indicates based on The determined The weight.

4. The thin film tension control method as described in claim 1, characterized in that, The result calculated using the dynamic ratio method Multiple roll diameters determined at a predetermined number of sampling times adjacent to each other Sum and average, then update the current sampling time. lower volume This is used for subsequent weighted fusion, where the number of adjacent elements preceding each other is negatively correlated with the linear velocity.

5. The thin film tension control method as described in claim 1, characterized in that, When performing the weighted fusion, the method further includes: calculating the deviation in real time. ,like If slippage or tape breakage is detected, the optimal estimated roll diameter determined at the previous sampling time is used as the optimal estimated roll diameter at the current sampling time to execute subsequent steps and trigger an alarm signal; where, This indicates the preset deviation threshold.

6. The thin film tension control method as described in claim 1, characterized in that, Dynamic torque ;in, Represents the time-varying moment of inertia. The angular acceleration of the reel is represented by a differential tracker, which is extracted using a nonlinear track-differentiator structure.

7. The thin film tension control method as described in claim 1, characterized in that, The method for correcting the gain coefficient is: proportional gain Integral gain ; in, , These are the baseline PID parameters for an empty volume. The diameter of the empty roll. This represents the optimal estimated roll diameter.

8. The thin film tension control method as described in claim 1, characterized in that, The total disturbance estimate is generated using an extended state observer, implemented as follows: In the formula, The preset sampling period, To control the gain, it represents the equivalent torque effect produced by a unit current. , Indicates the current sampling time The measured value of the rotational speed. The data observed by the observer at the previous sampling time is used for the current sampling time. The estimated rotational speed; The observations of the observer at the current sampling time are used for the next sampling time. The estimated rotational speed; , The observer gain is a preset value, selected according to the bandwidth configuration principle. Let the desired bandwidth of the observer be... In the continuous domain , , The preferred value is 3-5 times the preset speed loop bandwidth; The current sampling time The control quantity, namely the current command.

9. The thin film tension control method as described in claim 1, characterized in that, Before performing the control steps, the winding equipment is soft-started in the following manner: At startup, switch all servo motors of each reel to torque control mode and disable the speed loop; Control the servo motors of each reel to apply the initial preload torque The measured value of film tension Gradually increase to the tension set value The system maintains the tension at a preset ratio for a preset duration, monitors the tension fluctuation amplitude, and locks the current static roll diameter when the tension fluctuation amplitude is within a preset range and there is no preset abnormal vibration. As the initial roll diameter value This is used for subsequent calculations and to perform a seamless switching, smoothly transitioning each reel servo drive from torque control mode to speed closed-loop control mode, and executing the control steps to start the normal winding process.

10. An ultrathin film winding system, characterized in that, include: An unwinding spool and its servo motor, a rewinding spool and its servo motor, and a controller; wherein the controller is used to coordinate with each spool servo motor to perform the control steps as described in any one of claims 1 to 9.

Images