A high-speed roll changing control method and system based on tension prediction

Abstract

The application relates to the technical field of roll production line control, and discloses a high-speed roll changing control method and system based on tension prediction, wherein the method comprises the following steps: initializing control parameters; collecting the roll diameter of a main roll shaft and material tension in real time, and predicting a roll changing trigger time based on a roll diameter change rate; generating a prepared roll shaft acceleration curve based on an electronic cam algorithm, and calculating a tension feedforward compensation amount according to the moment of inertia and angular acceleration; outputting a speed error feedback item, a tension error feedback item and a feedforward compensation item to a servo driver after superposition; triggering a cutter and a compression roller in sequence to complete roll changing after determining a speed synchronization condition; and completing smooth transfer of tension control rights through a coordination gradual change strategy.

Description

Technical Field

[0001] This invention relates to the field of roll production line control technology, and more specifically, to a high-speed roll changing control method and system based on tension prediction. Background Technology

[0002] In high-speed roll-to-roll production lines for paper, film, and foil, the roll changeover process is a critical step in switching the winding task from a full main roll to a pre-wound roll. Currently, production lines generally use feedback control to adjust the material tension during the roll changeover process, that is, by using a tension sensor to detect the actual tension deviation and then driving the servo system to respond.

[0003] However, the aforementioned feedback control method has two shortcomings. First, during the acceleration of the pre-roll reel from rest to synchronize with the production line speed, the driving torque required to overcome its own rotational inertia disturbs the material tension. Pure feedback control can only respond after the tension deviation occurs, resulting in inherent control lag and causing material tension fluctuations during the acceleration phase to exceed the allowable range. Second, after the roll change is completed, if the tension control is suddenly switched from the main reel to the pre-roll reel, it will cause a tension step jump, affecting the consistency of material quality. These two problems are particularly prominent under conditions of high production line speed and high material tension sensitivity, easily leading to material wrinkling, material breakage, or failure to connect materials. Summary of the Invention

[0004] This invention provides a high-speed roll change control method and system based on tension prediction, which solves the technical problems of severe tension fluctuation, insufficient speed synchronization accuracy, and unsmooth transfer of tension control during the roll change process in related technologies.

[0005] This invention discloses a high-speed roll-changing control method based on tension prediction, comprising: real-time acquisition of the current actual tension value of the material and the current roll diameter of the main roll, and calculation of the roll-changing trigger time based on the roll diameter change rate;

[0006] When the current roll diameter of the main roll decreases to the roll diameter triggering threshold, the acceleration curve of the pre-roll roll is generated based on the electronic cam algorithm, and the tension feedforward compensation is calculated based on the rotational inertia and angular acceleration of the pre-roll roll.

[0007] Using the control cycle as the interval, the speed error feedback term, tension error feedback term, and the tension feedforward compensation amount are superimposed and output to the servo driver corresponding to the pre-reel;

[0008] When the speed error condition and phase alignment condition between the pre-roll and the main production line are met simultaneously, the cutter cylinder and the pressure roller cylinder are triggered in sequence to perform the roll changing action.

[0009] After the roll changing operation is completed, the main roll torque output is linearly reduced synchronously and the pre-roll tension closed-loop control gain coefficient is linearly increased during the transition time, thus completing the smooth transfer of tension control.

[0010] Furthermore, the calculation of the roll change trigger time based on the roll diameter change rate includes: performing differential processing on two adjacent roll diameter sampling values ​​to obtain the roll diameter change rate. : ,in, For the first The roll diameter value at the next sampling time. For the first The roll diameter value at the next sampling time. The time interval between two samplings; the roll change trigger time is calculated based on the roll diameter change rate. : ,in, The current scroll diameter of the main scroll. The preset volume size is the volume diameter that triggers the volume change.

[0011] Furthermore, the calculation of the tension feedforward compensation based on the rotational inertia and angular acceleration of the pre-reel includes: calculating the rotational inertia of the pre-reel in each control cycle. : ,in, To prepare for the current total mass of the scroll, The current roll diameter is set for the pre-loaded reel; the pre-loaded reel rotation angle command is output by the electronic cam algorithm. Perform a second-order difference to obtain the current angular acceleration of the pre-roll reel. : ,in, , , The first , , The angle command value for each control cycle. To control the cycle duration; the tension feedforward compensation amount .

[0012] Furthermore, the step of superimposing the speed error feedback term, the tension error feedback term, and the tension feedforward compensation amount and outputting them to the servo driver corresponding to the pre-reel includes: outputting a comprehensive control quantity to the servo driver corresponding to the pre-reel. : ,in, This is the speed error proportional gain coefficient. Main production line speed, To prepare the current linear velocity of the reel, This is the tension error gain coefficient. Set the target tension value. This represents the current actual tension value of the material. This is the tension feedforward compensation amount; wherein, the current linear velocity of the pre-reel is... Angular velocity fed back by the servo drive encoder Current roll diameter of the pre-roll reel according to The current roll diameter of the prepared roll is calculated and recalculated based on the distance measuring module update in each control cycle.

[0013] Furthermore, the simultaneous satisfaction of the speed error condition and the phase alignment condition includes: the speed error condition being the current linear velocity of the pre-reel. With the main production line speed Between satisfy The phase alignment condition is the angular phase difference between the pre-reel and the main reel. satisfy ,in The preset phase alignment tolerance is used; when both of the above conditions are met and the duration exceeds the set stable confirmation time, the speed synchronization is completed.

[0014] Furthermore, the step of synchronously and linearly decreasing the main reel torque output and linearly increasing the pre-reel tension closed-loop control gain coefficient during the transition time includes: during the transition time Internal, main reel torque output command and the gain coefficient of the pre-roll tension closed-loop control Adjust as follows: , ,in, To switch the initial torque output value of the main reel at the start, To prepare the target gain coefficient for the closed-loop control of the preheated reel tension after the switching is completed. This represents the elapsed time since the switch began. During the transition period, the current actual tension value of the material is continuously monitored. If it exceeds the allowable deviation range, the gradient slope is automatically adjusted to maintain tension stability.

[0015] Furthermore, before generating the acceleration curve for the pre-reel, the process also includes: pre-adding a damping coefficient to the speed control loop of the main reel during the pre-reel acceleration phase. The damping coefficient This is used to apply braking torque to the servo drive to suppress speed overshoot when the main reel experiences a sudden increase in speed due to cut-off and overload. Before starting the acceleration of the pre-reel, it also includes: controlling the pre-reel to perform axial micro-movement, and adjusting the axial position of the pre-reel in a closed-loop manner by reading the feedback value of the shaft end displacement sensor, so that the axial deviation between the edge of the pre-reel end face and the edge of the main reel end face converges to the preset tolerance range.

[0016] Furthermore, in a multi-axis rewinding scenario involving multiple pre-roll reels, the calculation of the tension feedforward compensation also includes: for each pre-roll reel Based on recent The slope of the predicted volume diameter growth rate is calculated by linear regression fitting of each sampling point. Extrapolate the predicted volume size for each future control cycle And calculate the corresponding rotational inertia prediction sequence. and independent feedforward compensation sequence Based on the wrap angle of each guide roller along the material transport path and coefficient of friction Calculate the transmission coefficient of each guide roller. and reflection coefficient And based on the elastic wave propagation speed of the material Calculate the propagation delay time and amplitude attenuation coefficient for each path order to generate the inter-axis multipath propagation feature matrix. The independent feedforward compensation sequence of each axis is convolved in the time domain with the inter-axis multipath propagation feature matrix to obtain the cross-axis coupling effect time domain signal matrix. The peak time and peak amplitude of the total coupling effect signal at each axis are identified. With the constraint that the superimposed peak of the coupling effect at each axis does not exceed the allowable fluctuation range of the target tension, the application time offset and amplitude scaling ratio of the compensation amount of each axis are jointly solved to generate a collaborative peak-shifting compensation sequence. The collaborative peak-shifting compensation sequence is subjected to propagation inverse compensation processing, and the compensation command is issued in advance according to the corresponding propagation time, generating a time-calibrated roll-changing control command sequence, which replaces the tension feedforward compensation amount output to each axis servo driver.

[0017] Furthermore, it also includes: during the rewinding acceleration period of each pre-rewinding roll combination, performing cross-correlation analysis on the time-calibrated rewinding control command sequence of each roll combination and the actual tension signal at the corresponding sensor position to extract the estimated actual propagation delay and the estimated actual amplitude attenuation; comparing the estimated actual propagation delay and the estimated actual amplitude attenuation with the predicted values ​​of the direct path in the inter-axis multipath propagation feature matrix to calculate the prediction error; and using a recursive identification algorithm to update the propagation velocity and transmission coefficient in the inter-axis multipath propagation feature matrix online. , ,in, and These represent the propagation speed before and after the update. To recursively identify the learning step size for propagation speed, For propagation delay prediction error, For scrolls To the scroll The length of the direct path, and These are the guide rollers before and after the update. The transmission coefficient, The recursive identification learning step size for the transmission coefficient. To correct the amplitude attenuation prediction error, the actual tension deviation at each axis is simultaneously distributed in reverse to the linear regression model of each axis according to the coupling weight, thereby correcting the predicted slope of the roll diameter growth rate.

[0018] This invention discloses a high-speed roll change control system based on tension prediction, used to execute the above-mentioned high-speed roll change control method based on tension prediction, including: a parameter initialization unit, used to read the target tension setpoint, the main production line speed and material characteristic parameters, and to perform calibration processing on the tension detection module and the distance measurement module;

[0019] The data acquisition and roll change prediction unit is used to collect the current actual tension value of the material and the current roll diameter of the main roll in real time, and calculate the roll change trigger time based on the roll diameter change rate.

[0020] The acceleration curve and feedforward compensation generation unit is used to generate the acceleration curve of the pre-roll spool based on the electronic cam algorithm, and to calculate the tension feedforward compensation based on the rotational inertia and angular acceleration of the pre-roll spool.

[0021] The integrated control command output unit is used to continuously output the speed error feedback item, tension error feedback item and the tension feedforward compensation amount to the servo driver corresponding to the pre-reel at control cycle intervals.

[0022] The speed synchronization determination and roll changing execution unit is used to determine whether the speed error condition and phase alignment condition between the pre-roll and the main production line are met simultaneously, and when they are met, the cutter cylinder and pressure roller cylinder are triggered in sequence to perform the roll changing action.

[0023] The tension control gradual transfer unit is used to synchronously and linearly decrease the main reel torque output and linearly increase the pre-reel tension closed-loop control gain coefficient during the transition time, so as to complete the smooth transfer of tension control.

[0024] The present invention has the following beneficial effects.

[0025] First, a tension feedforward compensation strategy based on rotational inertia is introduced during the pre-rolling acceleration phase. This strategy pre-adds the driving torque required to overcome the rotational inertia of the pre-rolling shaft to the control output. Compared to pure feedback control, which relies solely on the tension deviation signal to drive the response, feedforward compensation intervenes before the tension deviation occurs, eliminating the accumulation process of tension deviation caused by control lag. This helps to control the material tension fluctuation amplitude within a smaller range during the acceleration phase.

[0026] Secondly, after the roll change is completed, the torque output command of the main roll shaft and the gain coefficient of the closed-loop control of the tension of the pre-roll shaft are coordinated and linearly gradually changed. The resultant force output of the two shafts during the transition period is theoretically continuous, avoiding tension jump caused by sudden transfer of control, which helps to ensure a smooth transition of material tension after the roll change is completed.

[0027] Third, by adding a damping coefficient in the main shaft speed control loop during the pre-rolling acceleration phase, the main shaft servo system can apply braking torque more quickly when the material is cut off by the cutter. This helps to limit the speed overshoot to a small range and reduce the tension impact amplitude at the moment of roll change.

[0028] Fourth, in the scenario of simultaneous multi-axis roll changing, by establishing an inter-axis multipath propagation feature matrix based on the elastic wave transmission model, the independent feedforward compensation sequence of each axis is convolved with the propagation feature matrix in the time domain to predict the actual arrival waveform of the compensation effect of each axis at other axis sensors. By adjusting the application time and amplitude scaling ratio of the compensation amount of each axis, it is helpful to avoid tension oscillation caused by the temporal overlap of the multi-axis compensation peak at the downstream axis due to propagation delay.

[0029] Fifth, by continuously extracting the actual propagation delay and attenuation characteristics through cross-correlation analysis, the propagation characteristic parameters are updated online in a recursive manner. This enables the system to maintain the consistency between the propagation characteristic parameters and the actual working conditions under changes in operating conditions such as guide roller wear and material batch changes, which helps to maintain the accuracy of tension prediction and the quality of roll changing control. Attached Figure Description

[0030] Figure 1 This is a flowchart of a high-speed roll-changing control method based on tension prediction provided in an embodiment of the present invention. Detailed Implementation

[0031] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, some features described in the examples may be combined in other examples.

[0032] Example 1

[0033] This invention discloses a high-speed roll changing control method based on tension prediction, see [link to relevant documentation]. Figure 1 ,include:

[0034] Step 1: Initialize control parameters.

[0035] The central control module reads the target tension setpoint for the current production task from the host computer or local storage medium. (Unit: N) Main production line linear speed (Unit: m / s), Material elastic modulus (Unit: Pa) and mass per unit area of ​​material (Unit: kg / m²), and the above parameters are written into the control register as reference values ​​for subsequent tension closed-loop control and speed synchronization control. Simultaneously, zero-point calibration and range calibration are performed on the output signals of the tension detection module and the distance measuring module to ensure the accuracy and consistency of the output data from each sensor before entering formal operation.

[0036] Step 2: Collect operational data in real time and predict the roll change trigger time.

[0037] After the system completes initialization, the tension detection module continuously acquires the current actual tension value of the material at a sampling frequency of no less than 1 kHz. (Unit: N), the raw signal is low-pass filtered before being transmitted to the central control module. The ranging module synchronously acquires the current roll diameter of the main roll in real time. (Unit: m), update cycle not exceeding 100ms.

[0038] The central control module determines the current roll diameter of the main roll. With the main production line speed Calculate the current angular velocity of the main scroll. (Unit: rad / s):

[0039]

[0040] in, Main production line linear speed (unit: m / s). Current roll diameter of the main roll (unit: m).

[0041] The central control module performs differential processing on two consecutive roll diameter sampling values ​​to calculate the roll diameter change rate. (Unit: m / s):

[0042]

[0043] in, For the first The roll diameter at the next sampling time (unit: m). For the first The roll diameter at the next sampling time (unit: m). This is the current sampling sequence number. The time interval between two samplings (in seconds).

[0044] Based on the roll diameter change rate, the central control module continuously calculates the roll change trigger time. (Unit: s):

[0045]

[0046] in, The preset roll change trigger threshold is the roll diameter (in meters). The roll change trigger time is... The dynamic calculation results provide a basis for the timing planning of the accelerated start of the pre-roll, enabling the system to leave enough time window before the roll change trigger to complete the accelerated synchronization of the pre-roll.

[0047] It should be noted that the above-mentioned roll diameter change rate This refers to the instantaneous rate of change of the main roll diameter within the current sampling period. Since the roll diameter monotonically decreases as material is continuously wound, The value is usually negative, therefore it is used in calculating the volume change trigger time. The absolute value is taken to ensure that the predicted time is positive. When the production line speed or material thickness changes, The volume change trigger time will change accordingly. The prediction results are also dynamically updated to adapt to the roll change sequence requirements under different working conditions.

[0048] Step 3: Generate the pre-roll acceleration curve and calculate the tension feedforward compensation.

[0049] When the current roll diameter of the main roll Reduce to the volume size that triggers the volume change threshold At this point, the central control module enters the roll-changing preparation stage. The central control module generates the acceleration curve of the pre-roll reel based on an electronic cam algorithm. The input to the electronic cam algorithm is the real-time rotation angle of the main roll reel. Based on the current roll diameter data, the output is the pre-set roll rotation angle command. The electronic cam algorithm establishes the pre-set scroll rotation angle. With the main scroll angle The nonlinear mapping relationship between them is dynamically updated based on real-time roll diameter data within each 1ms control cycle to ensure the speed of the prepared roll axis. Continuously follow the main production line speed Instead of using a linear acceleration curve with a fixed slope, the acceleration curve changes.

[0050] During each control cycle, the central control module calculates the moment of inertia of the pre-reel in real time. (Unit: kg·m²):

[0051]

[0052] in, The current total mass of the roll (unit: kg). Preset the current roll diameter (in meters) for the prepared roll.

[0053] Preparing the current angular acceleration of the reel (Unit: rad / s²) Preparatory reel rotation angle command output by the electronic cam algorithm The second difference over time yields:

[0054]

[0055] in, For the first Preparatory reel rotation angle command value for each control cycle (unit: rad). For the first Preparatory reel rotation angle command value for each control cycle (unit: rad). For the first Preparatory reel rotation angle command value for each control cycle (unit: rad). This is the current control cycle number. To control the cycle duration (unit: s, i.e. 1ms).

[0056] Based on the rotational inertia of the pre-rolled reel and the current angular acceleration of the pre-roll reel The central control module calculates the tension feedforward compensation amount. (Unit: N·m):

[0057]

[0058] Tension feedforward compensation This indicates the additional driving torque required for the pre-loaded reel to overcome its own moment of inertia under the current acceleration state.

[0059] In this embodiment, in order to actively suppress the sudden increase in rotational speed caused by subsequent cut-off and loss of load on the main reel during the acceleration phase, the central control module also pre-adds a damping coefficient to the speed control loop of the main reel during the pre-reel acceleration phase. Damping coefficient Its function is to apply braking torque more quickly when the main spindle servo drive detects a sudden increase in speed, thus controlling the overshoot of the main spindle speed within the allowable range. Damping coefficient. The values ​​are calibrated during the system debugging phase based on the dynamic response characteristics of the main scroll servo system.

[0060] Step 4: Output integrated control commands to the pre-loaded reel servo driver.

[0061] The central control module continuously outputs integrated control signals to the servo driver corresponding to the pre-loaded reel with a control cycle of 1ms. (Unit: N·m):

[0062]

[0063] in, The speed error proportional gain coefficient reflects the controller's response sensitivity to speed deviations, and its value is adjusted based on the system identification results. Preparing the current linear velocity of the reel (unit: m / s); This is the tension error gain coefficient, reflecting the controller's correction strength for tension deviation; Set the target tension value (unit: N); The material's current actual tension value (unit: N); This is the tension feedforward compensation amount (unit: N·m).

[0064] It should be noted that the speed error term in the above formula... The unit is m / s, tension error term The unit is N, and the feedforward compensation term The units are N·m, and the dimensions of the three terms are different. Velocity error proportional gain coefficient. and tension error gain coefficient During the system identification phase, calibration is performed to unify the dimensions of each output quantity to N·m, ensuring the consistency of the dimensions after the superposition of the three sub-items and matching the torque input dimension of the servo drive.

[0065] Comprehensive control quantity It consists of three superimposed parts: a speed error feedback term, a tension error feedback term, and a feedforward compensation term. The feedforward compensation term pre-adds the torque required to overcome rotational inertia to the control output before tension deviation occurs, thereby actively suppressing tension fluctuations during acceleration.

[0066] It should be noted that the current linear velocity of the aforementioned pre-roll reel This refers to the tangential velocity of the prepared reel surface, which is calculated as the angular velocity fed back by the servo drive encoder. Current roll diameter of the pre-roll reel Half of the product, that is Due to the current roll diameter of the pre-roll reel. As the material rolls up, the central control module continuously increases the diameter of the prepared roll using the distance measuring module updated in each control cycle. Value recalculated for the current linear velocity of the pre-roll reel This is to ensure the accuracy of speed feedback.

[0067] In this embodiment, to further improve the axial alignment accuracy before the pre-roll reel acceleration, before starting the pre-roll reel acceleration in step 3, the central control module also controls the pre-roll reel to perform axial micro-movement. By reading the feedback value from the displacement sensor installed at the shaft end, the axial position of the pre-roll reel is adjusted in a closed-loop manner, so that the axial deviation between the edge of the pre-roll reel end face and the edge of the main reel end face converges to within a preset tolerance range. After this process is completed, the central control module outputs an axial alignment completion signal before the pre-roll reel acceleration process can be started.

[0068] Step 5: Determine the speed synchronization condition and perform the roll change operation.

[0069] The central control module continuously monitors the current linear velocity of the pre-loaded reel with a control cycle of 1ms. With the main production line speed To determine the speed error between the two speeds, we need to check whether the following two synchronization conditions are met simultaneously:

[0070] Condition 1 (Speed ​​Error Condition):

[0071]

[0072] Condition 2 (Phase Alignment Condition):

[0073]

[0074] in, The angular phase difference (in rad) between the pre-reel and the main reel. The preset phase alignment tolerance (unit: rad) is determined based on the material splicing overlap length requirements.

[0075] Once both of the above conditions are met and the duration exceeds the set stabilization confirmation time (which can be adjusted within the range of 5ms to 20ms depending on the operating conditions), the central control module determines that speed synchronization is complete and executes the following roll changing actions in sequence:

[0076] Action 1: The central control module sends a trigger command to the solenoid valve corresponding to the cutter cylinder. The piston rod of the cutter cylinder extends rapidly, cutting the material on the main roll and physically separating the new and old rolls.

[0077] Action 2: Within the set delay after the cutter action is completed, the central control module sends an action command to the solenoid valve corresponding to the pressure roller cylinder. The piston rod of the pressure roller cylinder extends, and the pressure roller presses the material onto the surface of the prepared roll at a set pressure, completing the physical bonding of the new and old roll materials.

[0078] It should be noted that the time interval between the cutting action and the pressure roller action mentioned above refers to the time difference between the energization of the solenoid valve of the cutting cylinder and the energization of the solenoid valve of the pressure roller cylinder. This time interval is calibrated during the system debugging phase based on material characteristics, production line speed, and equipment mechanical response characteristics to ensure that the timing of the cutting and pressing actions meets the material receiving process requirements.

[0079] Step 6: Coordinate the gradual transition to complete the transfer of tension control.

[0080] After the roll changing operation is completed, the central control module follows a coordinated gradual change strategy during the transition time. (Unit: seconds) Synchronous adjustment of main reel torque output command and the gain coefficient of the pre-roll tension closed-loop control :

[0081]

[0082]

[0083] in, The initial torque output value of the main shaft at the start of the switch (unit: N·m); The target gain coefficient for the pre-loaded reel tension closed-loop control after the switching is completed; transition time. The value can be determined within the range of 0.5s to 3s based on the material's tension sensitivity and the production line speed; The elapsed time since the switch started (in seconds). .

[0084] Throughout the transition, the central control module continuously monitors the current actual tension value of the material. This ensures that the combined force output from the main reel and the pre-reel is always maintained at the target tension setting value. Within the nearby allowable deviation range. If the material's current actual tension value... If the deviation exceeds the allowable range, the central control module automatically adjusts the gradient to prioritize tension stability.

[0085] when Reaching the transition time At that time, the main reel torque output command Reduced to zero, preparatory reel tension closed-loop control gain coefficient Increase to the target gain factor Tension control is completely transferred to the pre-reel, and the transition process ends. Afterward, the pre-reel acts as the new main take-up shaft, and its control logic is the same as that of the main reel in step 2, allowing the system to return to normal operation.

[0086] It should be noted that the above-mentioned main reel torque output command With the pre-reel tension closed-loop control gain coefficient The linear gradual change relationship refers to the synchronous linear change of both axes in opposite directions along the time axis, ensuring that the sum of the output torques of the two axes remains constant under ideal conditions at any given time, thus avoiding tension jumps caused by sudden transfers of control over a single axis. In actual operation, since material tension is affected by various factors, the central control module continuously monitors the current actual tension value of the material. It also dynamically corrects the gradient slope and provides real-time compensation for the linear gradient strategy.

[0087] In this embodiment, to reset the system state before the next roll change cycle begins, in addition to step 6, the central control module sends a deceleration command to the original master roll servo driver. After the original master roll smoothly decelerates to zero speed according to a preset deceleration curve, the central control module waits to receive a confirmation signal from the operator confirming the completion of the winding process. Upon receiving the confirmation signal, the central control module resets the original master roll to a new preparatory roll and sets the initial roll diameter of the original master roll. (i.e., the outer diameter of the hollow shaft, unit: m) is written to the control register to complete the system parameter reset and wait for the trigger of the next roll change cycle.

[0088] In steps 3 and 4 of this invention, a tension feedforward compensation strategy based on rotational inertia is introduced to compensate for the driving torque tension feedforward required to overcome the rotational inertia of the pre-reel. The tension feedforward compensation strategy pre-adds to the control output. Compared to pure feedback control, which relies solely on the tension deviation signal to drive the response, the tension feedforward compensation strategy eliminates the tension deviation accumulation process caused by control lag, thereby controlling the material tension fluctuation amplitude within a smaller range during the pre-rolling acceleration stage.

[0089] In step 6, by coordinating and linearly varying the gain coefficient of the main shaft torque output command and the pre-spindle tension closed-loop control, the resultant force output of the two shafts during the transition period is theoretically continuous, avoiding the tension step caused by the sudden transfer of tension control and ensuring a smooth transition of material tension after the roll change is completed.

[0090] In addition, a damping coefficient is added to the main reel speed control loop in step 3. This allows the main reel servo system to apply braking torque more quickly when the cutter loses load after cutting the material, limiting the speed overshoot to a smaller range and further reducing the tension impact amplitude at the moment of reel change.

[0091] The combined effect of the above-mentioned technical means ensures the continuity of material tension control during high-speed roll changing, reduces the risk of material wrinkling, material breakage, or splicing failure caused by tension fluctuations, and improves the reliability of roll changing operations.

[0092] On the other hand, the present invention also proposes a high-speed roll changing control system based on tension prediction, comprising:

[0093] The parameter initialization unit is used to read the target tension setting value, main production line speed and material characteristic parameters from the host computer or local storage medium, and to perform calibration processing on the tension detection module and the distance measurement module.

[0094] The data acquisition and roll change prediction unit is used to collect material tension values ​​and main roll diameter in real time, and calculate the roll change trigger time based on the roll diameter change rate.

[0095] The acceleration curve and feedforward compensation generation unit is used to generate a nonlinear acceleration curve of the pre-roll spool based on the electronic cam algorithm, and to calculate the tension feedforward compensation based on the rotational inertia and angular acceleration of the pre-roll spool.

[0096] The integrated control command output unit is used to superimpose the speed error feedback term, tension error feedback term, and feedforward compensation term, and then continuously output them to the servo driver corresponding to the pre-reel at control cycle intervals.

[0097] The speed synchronization determination and roll changing execution unit is used to determine whether the speed error condition and phase alignment condition between the pre-roll and the main production line are met simultaneously, and when they are met, the cutter cylinder and pressure roller cylinder are triggered in sequence to complete the roll changing action.

[0098] The tension control gradual transfer unit is used to synchronously adjust the main reel torque output and the pre-reel tension closed-loop control gain coefficient according to the coordinated gradual transfer strategy, so as to complete the smooth transfer of tension control.

[0099] By way of example, the system proposed in this invention may also include various features and combinations thereof in the method embodiments, which will not be elaborated here.

[0100] Example 2

[0101] This invention discloses a multi-axis high-speed roll changing control method based on tension prediction, comprising the following steps:

[0102] Step 2-1: Obtain the running data of each preparatory reel and generate a rotational inertia prediction sequence.

[0103] Similar to steps 1 and 2 in Embodiment 1, the central control module completes the initialization control parameter reading, sensor calibration, and the current roll diameter of the main roll. Current angular velocity of the main scroll and volume change trigger time Real-time calculation.

[0104] In multi-axis roll changing scenarios, the system numbered as follows: For each pre-loaded reel, the central control module reads the current reel diameter measurement value from the corresponding ranging module. (Unit: m), and the current angular velocity is calculated from the corresponding encoder feedback value. (Unit: rad / s). Wherein, This refers to the total number of prepared scrolls.

[0105] For each pre-roll The central control module is the most recent Using 100 sampling points as samples, perform linear regression fitting on the roll diameter values ​​at adjacent sampling times to calculate the predicted slope of the roll diameter growth rate. (Unit: m / s). The input for linear regression is the most recent... The sequence of convolutional diameter values ​​corresponding to each sampling time point is output as a linear slope that minimizes the sum of squared residuals. ,in The regression window length (unit: number of windows) is determined during the system debugging phase based on the noise level measured by the roll diameter.

[0106] Predicting slope based on roll diameter growth rate The central control module controls the pre-roll reel. Extrapolate the roll diameter to calculate the future roll number. One control cycle ( Predicted volume size at time ) (Unit: m):

[0107]

[0108] in, Preparing the scroll The current roll diameter measurement (unit: m). To control the cycle duration (unit: s, i.e. 1ms). This represents the total number of predicted future control cycles.

[0109] Furthermore, for each prepared scroll In the future Prediction sequence of rotational inertia within each control cycle (Unit: kg·m²), calculate the moment of inertia according to the formula in step 3 of Example 1:

[0110]

[0111] in, Preparing the scroll The current total mass (unit: kg).

[0112] It should be noted that the predicted slope of the volume diameter growth rate obtained from the above linear regression fitting... This refers to the slope parameter obtained by least-squares estimation of the linear trend of roll diameter change over time within the regression window. Since the roll diameter growth rate during the roll change acceleration phase is affected by both the material linear velocity and the roll rotation speed, the predicted slope of the roll diameter growth rate... Based on the latest The sampling points are updated on a rolling basis to adapt to the dynamic characteristics of the roll diameter growth rate changing with the acceleration process.

[0113] Step 2-2: Calculate the transmission coefficient and reflection coefficient of each guide roller and generate the inter-axis multipath propagation characteristic matrix.

[0114] The central control module reads the material transport path parameters stored in the control register, including those for each guide roller. Position coordinates (Unit: m) Angle of wrap of material at the guide roller (Unit: rad) and the coefficient of friction of the guide roller surface (Dimensionless). Among them, This represents the total number of guide rollers along the material transport path.

[0115] For each guide roller The central control module calculates the transmission coefficient based on an elastic wave transmission model with frictional contact. (Dimensionless) and reflection coefficient (Dimensionless):

[0116]

[0117]

[0118] in, This is an exponential decay factor based on the cumulative amplification effect of frictional force within the guide roller wrap angle range. This represents the natural constant, reflecting the attenuation effect of the guide roller on the elastic waves of material tension. When When it approaches zero, the transmittance coefficient When the value approaches 1, it indicates that the elastic wave is almost completely transmitted; when When the transmission coefficient increases, A value approaching zero indicates that the elastic wave is significantly attenuated at the guide roller.

[0119] The propagation speed of elastic waves in different segments of the material (Unit: m / s) Calculated by the following formula:

[0120]

[0121] in, It is the elastic modulus of the material (unit: Pa, i.e., kg / (m·s²)). Both and represent the material's mass per unit area (unit: kg / m²), and are parameters read and written into the control register in step 1 of Example 1. In the above formula, The dimensions are After taking the square root, the velocity dimension is obtained in m / s, which is consistent with the dimension.

[0122] For the pre-roll To the preparatory scroll ( The propagation path between the segments is determined by the central control module based on the length of each material segment. (Unit: m, i.e., the length of the material segment between two adjacent guide rollers or between the guide roller and the roll) and propagation speed Calculate the propagation delay time of the direct path. (Unit: s):

[0123]

[0124] in, Indicates from scroll To the scroll The collection of all material segments through which the sensor's position passes.

[0125] For each order of reflection path, the central control module calculates the first... Delay time of the first reflection path (Unit: s) and amplitude attenuation coefficient (dimensionless), where , The maximum reflection order to be considered (determined during system commissioning based on material elastic properties and the number of guide rollers). The delay time of the first reflection path is the time it takes for the elastic wave to travel through the reflection path. The cumulative propagation time from the secondary guide roller to the target sensor is represented by the amplitude attenuation coefficient, which is the product of all transmission and reflection coefficients along that path. Specifically, for the path passing through the guide roller set... Transmission and guide roller assembly After reflection, it reaches the scroll. The sensor's first For a path of order one, the amplitude attenuation coefficient is:

[0126]

[0127] in, For the first The collection of guide rollers that undergo transmission in the step path. For the first The set of guide rollers that undergo reflection in the first-order path. and Guide rollers The transmission coefficient and reflection coefficient.

[0128] For direct routes ( Elastic waves along Each material segment in the middle is transmitted sequentially without being reflected by any guide rollers, and its amplitude attenuation coefficient is:

[0129]

[0130] That is, the product of the transmission coefficients of all guide rollers on the direct path. It is an empty set.

[0131] The central control module organizes the delay and attenuation parameters of each of the above-mentioned paths into an inter-axis multipath propagation characteristic matrix. The matrix dimension is , of which element For a containing A sequence of delay and decay parameter pairs:

[0132]

[0133] diagonal elements Represents scroll The direct path characteristics of the self-compensation amount to the sensor on this axis, and the off-diagonal elements. ( ) indicates a scroll The compensation effect propagates to the reel through the material transport path. Multipath characteristics at the sensor.

[0134] In this embodiment of the application, in order to improve the calculation accuracy of the transmission coefficient and the reflection coefficient, the wrap angle of the guide roller is adjusted. In terms of acquisition methods, the central control module can also be based on the position coordinates of each guide roller on the material conveying path. The geometric relationship between the material segments and adjacent guide rollers is used to calculate the actual wrap angle of the material at each guide roller in real time, rather than directly using the static wrap angle value stored during the commissioning phase. When significant changes in material tension or roll diameter cause a shift in the material path shape, the real-time calculated wrap angle value can more accurately reflect the elastic wave propagation characteristics under the current operating conditions.

[0135] Steps 2-3: Calculate the independent feedforward compensation sequence for each axis and generate the time-domain signal matrix of cross-axis coupling influence.

[0136] For each pre-roll The central control module is based on the rotational inertia prediction sequence. and their respective angular acceleration programming curves (Unit: rad / s², generated by electronic cam algorithm, and is the same as the current angular acceleration of the preparatory scroll in step 3 of Example 1) (with consistent definitions), calculate the future of each axis separately. Independent feedforward compensation sequence for each control cycle (Unit: N·m):

[0137]

[0138] For each pair of reels The central control module will roll the reel Independent feedforward compensation sequence Inter-axis multipath propagation feature matrix elements Perform temporal convolution and calculate the convolution axis. The compensation effect reaches the scroll Actual arrival waveform at the sensor location (Unit: N·m):

[0139]

[0140] in, To perform floor operations, the continuous propagation delay time is discretized into the number of control cycles.

[0141] The central control module organizes the arriving waveforms of all reel combinations into a time-domain signal matrix representing the cross-axis coupling effects. , dimension , of which The time-domain sequences corresponding to the rows and columns are , indicating a scroll The compensation effect on the scroll The arrival waveform at the sensor over time.

[0142] It should be noted that in the above temporal convolution operation, This refers to the scroll The independent feedforward compensation sequence according to the first The propagation delay of the first reflection path is shifted in the time domain and then multiplied by the corresponding amplitude attenuation coefficient. When the shifted time index is less than 1, the corresponding independent feedforward compensation sequence... A value of zero indicates that the scroll is at that moment. Acceleration has not yet begun, and the compensation amount is zero.

[0143] Steps 2-4: Identify the peak values ​​of coupling effects and generate a collaborative peak-shifting compensation scheme.

[0144] For each pre-roll The central control module affects the time-domain signal matrix from all other reels due to cross-axis coupling. ( The arrival waveform is summed in the time domain to calculate the roll. The total coupling effect on the signal at the sensor location (Unit: N·m):

[0145]

[0146] The central control module affects the overall coupling signal. Search along the timeline to identify the peak moments of coupling effects at each reel. and peak amplitude :

[0147]

[0148] By each scroll The peak value of the coupling effect does not exceed the allowable fluctuation range of the target tension. (Unit: N·m, determined during system commissioning based on material tension specifications) This is the constraint condition, representing the offset at which the central control module applies compensation to each axis. (Unit: control cycles) and amplitude scaling ratio (dimensionless, A joint solution is performed. The solution process uses independent feedforward compensation sequences for each axis. Inter-axis multipath propagation feature matrix and allowable fluctuation range As input, the optimal time offset for each axis. and optimal amplitude scaling ratio For output, the constraints are expressed as follows:

[0149]

[0150] The above constraint satisfaction problem is solved using a step-by-step search method: the central control module first searches according to the peak times of each axis. The axes are sorted in order of priority, and the application time offset of the sorted axes is fixed in sequence. and amplitude scaling ratio The value represents the offset of the current axis at the moment the constraint condition is met. Within the search range, the algorithm enumerates each control cycle with a step size of 1, selecting the optimal time offset that maximizes the constraint margin. Offset at the optimal time Once determined, adjust the amplitude scaling ratio. In the interval The search uses a preset step size to select the scaling factor that maximizes the amplitude while satisfying the constraints. The maximum value is used as the optimal amplitude scaling ratio. This is done to preserve the compensation effect of each axis as much as possible.

[0151] The optimal time offset for each axis is obtained by solving the problem. and optimal amplitude scaling ratio Subsequently, the central control module generates a coordinated peak-shifting compensation sequence for each axis. (Unit: N·m):

[0152]

[0153] In this embodiment of the application, in order to reduce the online computational load of the above-mentioned joint solution, in addition to steps 2-4, a time offset can also be applied to each axis. The search range is pre-pruned. The central control module performs pre-pruning based on the direct path delay time between each axis. Estimate the shortest propagation time required for the compensation effect of each axis to reach other axis sensors, and search only within the application time offset range that can stagger the peak of the coupling effect, thereby reducing the search space of the optimization variables to a physically meaningful subset and reducing the number of iterations.

[0154] Steps 2-5: Apply propagation inverse compensation and generate a time-calibrated roll change control command sequence.

[0155] The central control module provides coordinated peak-shaving compensation sequences. Apply propagation inverse compensation to eliminate the compensation effect from the reel. The impact of the propagation time required for the servo driver output to propagate to the shared material transport segment on the control timing.

[0156] For each pre-roll The central control module is based on a scroll. Direct path length to key locations in the shared material transport section (Unit: m), calculate the propagation lead of the compensation effect. (Unit: control cycles)

[0157]

[0158] Time-calibrated roll change control command sequence (Unit: N·m) Advance by the coordinated peak-shifting compensation sequence The output of each control cycle is:

[0159]

[0160] That is, during the planning period When the compensation effect reaches the shared segment, the central control module in The moment is towards the scroll The servo driver issues corresponding compensation commands to align the actual compensation effect at the time it reaches the shared segment with the tension prediction planning time.

[0161] Steps 2-6: Output integrated control commands to each axis servo driver.

[0162] The central control module operates on a 1ms control cycle for each pre-loaded reel. The sequence of time-calibrated roll change control commands Replace the tension feedforward compensation amount in step 4 of Example 1 Output integrated control quantity to the corresponding servo driver (Unit: N·m):

[0163]

[0164] Among them, the speed error proportional gain coefficient Tension error gain coefficient Target tension set value The definition is consistent with step 4 of Example 1; Preparing the scroll The current linear velocity (unit: m / s) is calculated from the corresponding servo drive encoder feedback value; For scrolls The current actual tension value (unit: N) at the corresponding sensor location. The dimension matching method for each term in the above formula is consistent with step 4 of Example 1, and the velocity error proportional gain coefficient... and tension error gain coefficient During the system identification phase, calibration is used to unify the dimensions of all output quantities to N·m.

[0165] For each pre-rolled reel that has passed steps 2-6, speed synchronization determination and rewinding can be further performed. The determination of speed synchronization conditions, the execution of rewinding, and the coordinated and gradual transfer of tension control after rewinding are the same as steps 5 and 6 in Example 1.

[0166] In this embodiment of the application, in order to continuously maintain the consistency between the propagation characteristic parameters and the actual operating conditions during operation, the following steps are also included.

[0167] Steps 2-7: Update the propagation characteristic parameters and the predicted slope of the roll diameter growth rate online.

[0168] During the rewinding acceleration phase of each pre-reel, the central control module continuously acquires the actual tension signals from each sensor location. (Unit: N).

[0169] For each pair of reels The central control module controls the reel. Time-calibrated roll change control command sequence With scroll Actual tension signal at the location Perform cross-correlation analysis and calculate the cross-correlation function. :

[0170]

[0171] in, Time delay (unit: control cycles). To control the cycle number, the cross-correlation function is used. peak position As a scroll To the scroll The actual propagation delay estimate is used to correlate the corresponding peak amplitude with the time-calibrated roll change control command sequence. The ratio of the root mean square is used as an estimate of the actual amplitude decay.

[0172] It should be noted that the above cross-correlation function In the middle, the sequence of time-calibrated roll change control commands The unit is N·m, and the actual tension signal The unit of is N, and the unit of their product is N²·m. The cross-correlation function itself is a dimensionless quantity. This will be used later to extract the peak position. When using only the cross-correlation function The time index position of the peak value is not directly used for physical quantity calculations; when extracting the estimated value of the actual amplitude decay, the cross-correlation function is used. Peak amplitude and time-calibrated roll change control command sequence The ratio of root mean squares is used as a normalization process to eliminate the influence of dimensions, and the resulting attenuation estimate is a dimensionless quantity, which is compared with the direct path amplitude attenuation coefficient. With consistent dimensions, it can be directly used for subsequent comparisons and updates.

[0173] The central control module combines the actual propagation delay and attenuation characteristics extracted from the above cross-correlation analysis with the inter-axis multipath propagation characteristic matrix. The predicted values ​​for direct routes are compared, and the prediction error is calculated. (Unit: s) and (Dimensionless). Among them, The estimated actual propagation delay and the propagation delay time of the direct path. difference, The actual amplitude attenuation estimate and the direct path amplitude attenuation coefficient are given. difference.

[0174] Using a recursive identification algorithm, the central control module propagates the feature matrix across multiple paths between axes. speed of propagation in and transmission coefficient Perform online updates:

[0175]

[0176]

[0177] in, The propagation speed before the update (unit: m / s). The updated propagation speed (unit: m / s). The learning step size (dimensionless, determined during the system debugging phase based on the rate of parameter change) is used to recursively identify the propagation speed. For scrolls To the scroll The length of the direct path (in meters). To update the front guide roller The transmission coefficient (dimensionless). For the replacement of the guide rollers The transmission coefficient (dimensionless). The recursive identification learning step size for the transmission coefficient (dimensionless, determined during system debugging based on the parameter change rate). Reflectance coefficient. From the updated transmittance according to Synchronized updates.

[0178] It should be noted that the above transmission coefficient update formula middle, From the scroll To the scroll The overall amplitude attenuation error of the direct path, while the transmission coefficient This represents the transmission coefficient of a single guide roller along the path. When the path is direct... When multiple guide rollers are involved, a single cross-correlation analysis can only extract the overall amplitude attenuation error of the path and cannot directly distinguish the contribution of each guide roller. During actual updates, the central control module performs path... All guide rollers on Apply the same correction amount to all Through the cross constraints of multiple paths, the transmission coefficient of each guide roller gradually converges to the actual transmission coefficient of each guide roller in multiple roll change cycles.

[0179] Simultaneously, the central control module controls each reel Actual tension deviation at the location Arrange the scrolls according to their axes. Coupling weights affected by tension (Dimensionless) Linear regression model with inverse assignment to each axis, correcting the predicted slope of volume diameter growth rate. Coupling weights It is derived from the normalization of the peak amplitude of the waveforms arriving at each axis in the inter-axis multipath propagation characteristic matrix, specifically:

[0180]

[0181] in, The scrolls calculated in steps 2-3 The compensation effect on the scroll The arriving waveform at the sensor, For removing the scroll All roll numbers except those listed above. The slope correction formula is:

[0182]

[0183] in, Prepare scrolls before updating The predicted slope of the roll diameter growth rate (unit: m / s). Preparing scrolls for the update The predicted slope of the roll diameter growth rate (unit: m / s). The learning step size for slope correction (unit: m / (s·N), determined during the system debugging phase).

[0184] The online update results of the propagation velocity, transmission coefficient, and reflection coefficient are synchronously written into the control register for use in step 2-2 when regenerating the inter-axis multipath propagation characteristic matrix in the next roll change cycle, thereby realizing the joint adaptive calibration of tension prediction parameters and propagation characteristic parameters.

[0185] It should be noted that the actual propagation delay extracted by the above cross-correlation analysis refers to the delay from the roll. Issue a compensation command to the scroll to achieve the compensation effect. Estimated time difference between the observable tension response at the sensor location. Friction coefficient due to guide roller wear. The elastic modulus changes with the duration of use and batch variations of the material. and mass per unit area If a shift occurs, the above recursive identification method can continuously track these parameter drifts without stopping the machine, maintaining the consistency between the inter-axis multipath propagation feature matrix and the actual working conditions.

[0186] Based on the single-axis roll-changing feedforward compensation in Embodiment 1, this invention addresses the problem of mutual coupling of tension feedforward compensation amounts of each axis through a shared material transmission segment in multi-axis simultaneous acceleration scenarios. It introduces an inter-axis multipath propagation feature matrix based on an elastic wave transmission model, extending the traditional static coupling coefficient into a time-domain convolution operator that can describe the delay and amplitude attenuation of each order of reflection path.

[0187] In steps 2-3, by performing temporal convolution between the independent feedforward compensation sequence of each axis and the inter-axis multipath propagation feature matrix, the cross-axis coupling influence temporal signal matrix can reflect the multipath propagation process of the compensation effect of each axis in the shared material segment, enabling the system to predict the actual arrival waveform of the compensation effect of each axis at other axis sensors during the planning stage.

[0188] In steps 2-4, by adjusting the application time and amplitude scaling ratio of the compensation amount of each axis, it is ensured that the peak value of the coupling effect at each axis is still within the allowable range after multiple reflections are superimposed, thus avoiding the oscillation phenomenon caused by the compensation peak value re-overlapping in the time domain at the downstream axis due to propagation delay.

[0189] In steps 2-5, the propagation inverse compensation eliminates the timing deviation caused by the compensation effect propagating from the driver output to the shared material segment by issuing the compensation command in advance according to the corresponding propagation time, thereby aligning the actual control time with the predicted planning time and further improving the timing accuracy of multi-axis collaborative compensation.

[0190] In steps 2-7, the actual propagation delay and attenuation features extracted by cross-correlation analysis are used to continuously update the transmission coefficient, reflection coefficient, and propagation speed parameters in the inter-axis multipath propagation feature matrix in a recursive manner. At the same time, the tension deviation is reverse-allocated according to the coupling weight to correct the predicted slope of the roll diameter growth rate of each axis. This enables the system to maintain the consistency between the propagation feature parameters and the actual working conditions under changes in working conditions such as guide roller wear and material batch changes, thereby maintaining the tension prediction accuracy and roll change control quality.

[0191] The combined effect of the above-mentioned technical means effectively controls the coupling effect of tension feedforward compensation of each axis in multi-axis simultaneous roll changing scenarios, reduces the risk of tension oscillation in the shared material section caused by the superposition of multi-axis compensation, and improves the reliability of roll changing operation of high-speed multi-axis winding system.

[0192] Application example:

[0193] A film manufacturing company has configured a dual-axis synchronous winding production line with a main production line speed of... The target tension setpoint is 18 m / s. The value is 120 N, the material is biaxially oriented polypropylene film (BOPP), and the elastic modulus is... for Pa, mass per unit area of ​​material The value is 0.018 kg / m². The production line is equipped with two pre-loaded reels (reel 1 and reel 2), which are connected by a shared material transport section. Three guide rollers (guide roller A, guide roller B, and guide roller C) are distributed along the transport path. During a certain reel change cycle, the reel change trigger time windows of the two reels overlap, and the system needs to perform multi-axis coordinated reel change control.

[0194] Step 2-1: Obtain the running data of each preparatory reel and generate a rotational inertia prediction sequence.

[0195] After the central control module completes initialization, it reads the current roll diameter measurement value from the ranging modules corresponding to roll 1 and roll 2 respectively, and calculates the current angular velocity from the encoder feedback value. (Using the closest...) Using a set of sampling points as the regression window, linear regression was performed on the time-varying trend of the two axes' roll diameter to obtain the predicted slope of the roll diameter growth rate for each axis. .based on For the future Extrapolate the roll diameter of each control cycle, and then generate a rotational inertia prediction sequence for each axis according to the rotational inertia formula. .

[0196] Table 1 Current operating status and predicted moment of inertia parameters of each pre-roll reel

[0197]

[0198] The two-axis rotational inertia prediction sequence increases slowly with the increase of roll diameter, providing a basis for the calculation of the subsequent independent feedforward compensation sequence for each axis.

[0199] Step 2-2: Calculate the transmission coefficient and reflection coefficient of each guide roller and generate the inter-shaft multipath propagation characteristic matrix.

[0200] The central control module reads the process parameters of the three guide rollers stored in the control register and calculates the transmission coefficient of each guide roller based on the elastic wave transmission model. and reflection coefficient .

[0201] The propagation speed of elastic waves in a material is determined by Calculation, where Pa, kg / m², propagation speed It is approximately 10,000 m / s.

[0202] Table 2. Elastic wave propagation characteristic parameters of each guide roller

[0203]

[0204] Based on the length and propagation speed of each material segment, the central control module calculates the propagation delay time of the direct path from scroll 1 to scroll 2. The direct path passes through guide rollers A and B, with a total material section length of 3.20 m and a propagation delay. The delay is approximately 0.00032 s, corresponding to approximately 0 control cycles (less than one 1ms cycle, rounded to 0). Considering the first-order reflection path (arriving after reflection via guide roller C), the delay increases to approximately 0.00058 s, corresponding to one control cycle, with an amplitude attenuation coefficient... It is the product of the transmission coefficients of guide rollers A and B and the reflection coefficient of guide roller C, i.e. .

[0205] Inter-axis multipath propagation feature matrix off-diagonal elements and It contains two sets of delay and attenuation parameter pairs for the direct path and the first-order reflection path, respectively, for use in subsequent temporal convolution.

[0206] Steps 2-3: Calculate the independent feedforward compensation sequence for each axis and generate the time-domain signal matrix of cross-axis coupling influence.

[0207] The central control module outputs angular acceleration planning curves for each axis based on the electronic cam algorithm. With the rotational inertia prediction sequence Calculate the independent feedforward compensation sequence for each axis in the future control cycle. Then, the compensation sequence of each axis is convolved with the elements of the inter-axis multipath propagation feature matrix in the time domain to generate the time domain signal matrix of cross-axis coupling influence. .

[0208] Table 3 Independent feedforward compensation and cross-axis arrival waveforms for each axis under typical control cycles.

[0209]

[0210] in, This indicates the compensation effect of reel 1 via the direct path (attenuation coefficient). The waveform that arrives at sensor 2 after being convolved in the temporal domain with the first-order reflection path; The waveform represents the arrival waveform of the compensation effect of reel 2 propagating to the sensor on reel 1. The first-order reflection path attenuation coefficient is small (approximately 0.039), contributing little to the amplitude of the arrival waveform, but forming a tailing effect near the peak moment.

[0211] Steps 2-4: Identify the peak values ​​of coupling effects and generate a collaborative peak-shifting compensation scheme.

[0212] The central control module performs time-domain summation on the arriving waveforms from other axes at each axis to calculate the total coupling effect signal. And identify the peak time of coupling effect at each axis. and peak amplitude .

[0213] Table 4. Peak value identification results and collaborative peak-shifting compensation parameters for each axis coupling effect

[0214]

[0215] The peak values ​​of the two-axis coupling effects both exceeded the allowable fluctuation limit. In N·m, the central control module determines the optimal time offset and amplitude scaling ratio through stepwise search. The compensation application time of reel 1 is delayed by 8 control cycles, and the amplitude scaling is reduced to 0.92 times; the compensation application time of reel 2 is delayed by 6 control cycles, and the amplitude scaling is reduced to 0.88 times. After the compensation peaks of the two axes are staggered on the time axis, the coupled superimposed amplitudes at each axis converge to within the allowable range. The coordinated peak-shifting compensation sequence for each axis... It will be generated immediately.

[0216] Steps 2-5: Apply propagation inverse compensation and generate a time-calibrated roll change control command sequence.

[0217] The central control module calculates the propagation lead of the compensation effect for each axis based on the direct path length from each axis to key locations in the shared material transport section. The coordinated peak-shifting compensation sequence is output in advance according to the corresponding number of cycles to generate a time-calibrated roll-changing control command sequence. .

[0218] Table 5. Inverse compensation parameters and time calibration results for each axis of propagation

[0219]

[0220] In this scenario, since the path length from each axis to the shared segment is 1.60 m, the corresponding propagation time is approximately 0.00016 s, which is less than one control cycle, resulting in a propagation lead. Rounded to 0. The sequence of roll change control commands after time calibration is consistent with the cooperative peak-shifting compensation sequence. The calibration step will produce a significant timing advance effect under conditions of long path or low propagation speed.

[0221] Steps 2-6: Output integrated control commands to each axis servo driver.

[0222] The central control module uses a 1ms control cycle to send the time-calibrated roll change control command sequence. Replace the single-axis feedforward compensation value and output the comprehensive control value to the servo drives corresponding to reel 1 and reel 2 respectively. .

[0223] Table 6. Comprehensive control output of each axis under typical control cycles

[0224]

[0225] As the speed of the pre-roll axis gradually approaches the speed of the main production line, the speed error term and tension error term continue to decrease. The main contribution of the comprehensive control quantity comes from the time-calibrated roll change control command sequence. The compensation peaks of the two axes are staggered on the time axis, avoiding the synchronous superposition impact of tension in the shared material section.

[0226] Steps 2-7: Update the propagation characteristic parameters and the predicted slope of the volume diameter growth rate online.

[0227] During the roll change acceleration process, the central control module continuously collects the actual tension signals from the sensors on each axis, and... and Perform cross-correlation analysis to extract the estimated values ​​of actual propagation delay and amplitude attenuation. Compare these values ​​with the predicted values ​​in the inter-axis multipath propagation feature matrix, calculate the prediction error, and update the propagation speed using a recursive identification algorithm. and the transmission coefficient of each guide roller .

[0228] Table 7. Results of cross-correlation analysis and updates of propagation characteristic parameters

[0229]

[0230] The actual amplitude attenuation estimate is slightly lower than the direct path prediction, indicating that the friction coefficient of guide roller B has increased under the current operating conditions (corresponding to the guide roller wear trend). A recursive identification algorithm is used to learn the step size. The transmission coefficient of guide roller B is adjusted downwards, and the updated parameters are written to the control register for use in regenerating the inter-axis multipath propagation characteristic matrix during the next roll change cycle. Roll diameter growth rate prediction slope. Synchronization by Coupling Weights Reverse correction is used to maintain the accuracy of the predicted sequence of rotational inertia for each axis.

[0231] Throughout the implementation process, data starts from the initial measurements of the roll diameters of each axis and process parameters. Step 2-1 generates a predicted sequence of rotational inertia. Step 2-2 establishes an inter-axis propagation characteristic matrix describing the multipath propagation characteristics of elastic waves. Step 2-3 performs temporal convolution of the independent compensation sequence for each axis with the elements of the propagation characteristic matrix to obtain the actual arrival waveform of the coupling effect at each axis. Step 2-4 identifies coupling peaks and solves for the optimal peak-shifting parameters. Step 2-5 eliminates propagation timing deviations. Step 2-6 superimposes the compensation values ​​after collaborative planning and timing calibration onto the integrated control output. The online cross-correlation analysis in Step 2-7 feeds back the actual tension response signal to the recursive update of the propagation characteristic matrix and the predicted slope of the roll diameter growth rate, forming a complete data flow closed loop from physical measurement to predictive planning and then to closed-loop correction, ensuring the continuous effectiveness of multi-axis collaborative roll-changing control under varying operating conditions.

[0232] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. A high-speed roll change control method based on tension prediction, characterized in that, include: Real-time acquisition of the material's current actual tension value and the main roll diameter; calculation of roll change trigger time based on the roll diameter change rate. When the current roll diameter of the main roll decreases to the roll diameter triggering threshold, the acceleration curve of the pre-roll roll is generated based on the electronic cam algorithm, and the tension feedforward compensation is calculated based on the rotational inertia and angular acceleration of the pre-roll roll. Using the control cycle as the interval, the speed error feedback term, tension error feedback term, and the tension feedforward compensation amount are superimposed and output to the servo driver corresponding to the pre-reel; When the speed error condition and phase alignment condition between the pre-roll and the main production line are met simultaneously, the cutter cylinder and the pressure roller cylinder are triggered in sequence to perform the roll changing action. After the roll changing operation is completed, the main roll torque output is linearly reduced synchronously and the pre-roll tension closed-loop control gain coefficient is linearly increased during the transition time, thus completing the smooth transfer of tension control.

2. The high-speed roll changing control method based on tension prediction according to claim 1, characterized in that, The calculation of the roll change trigger time based on the roll diameter change rate includes: Perform differential processing on two consecutive roll diameter samples to obtain the roll diameter change rate. : ， in, For the first The roll diameter value at the next sampling time. For the first The roll diameter value at the next sampling time. The time interval between two samplings; Calculate the roll change trigger time based on the roll diameter change rate. : ， in, The current scroll diameter of the main scroll. The preset volume size is the volume diameter that triggers the volume change.

3. The high-speed roll changing control method based on tension prediction according to claim 1, characterized in that, The calculation of tension feedforward compensation based on the rotational inertia and angular acceleration of the pre-rolled reel includes: Calculate the moment of inertia of the pre-roll reel in each control cycle. : ， in, To prepare for the current total mass of the scroll, Prepare the current roll diameter for the roll; Preparatory reel rotation angle command output by the electronic cam algorithm Perform a second-order difference to obtain the current angular acceleration of the pre-roll reel. : ， in, , , The first , , The angle command value for each control cycle. To control cycle duration; The tension feedforward compensation amount .

4. The high-speed roll changing control method based on tension prediction according to claim 3, characterized in that, The step of superimposing the speed error feedback term, the tension error feedback term, and the tension feedforward compensation amount and outputting them to the servo driver corresponding to the pre-reel includes: Output integrated control quantity to the servo driver corresponding to the pre-roll reel : ， in, This is the speed error proportional gain coefficient. Main production line speed, To prepare the current linear velocity of the reel, This is the tension error gain coefficient. Set the target tension value. This represents the current actual tension value of the material. This is the tension feedforward compensation amount; Wherein, the current linear velocity of the pre-roll reel Angular velocity fed back by the servo drive encoder Current roll diameter of the pre-roll reel according to The current roll diameter of the prepared roll is calculated and recalculated based on the distance measuring module update in each control cycle.

5. The high-speed roll changing control method based on tension prediction according to claim 1, characterized in that, The velocity error condition and phase alignment condition are simultaneously satisfied, including: The speed error condition is the current linear velocity of the pre-roll reel. With the main production line speed Between ; The phase alignment condition is the angular phase difference between the pre-reel and the master reel. satisfy ,in This is the preset phase alignment tolerance; Once both of the above conditions are met and the duration exceeds the set stable confirmation time, the judgment speed synchronization is completed.

6. The high-speed roll changing control method based on tension prediction according to claim 1, characterized in that, The method of synchronously and linearly decreasing the main reel torque output and linearly increasing the pre-reel tension closed-loop control gain coefficient during the transition time includes: During the transition time Internal, main reel torque output command and the gain coefficient of the pre-roll tension closed-loop control Adjust as follows: ， ， in, To switch the initial torque output value of the main reel at the start, To prepare the target gain coefficient for the closed-loop control of the preheated reel tension after the switching is completed. This represents the elapsed time since the switch began. ; During the transition period, the current actual tension value of the material is continuously monitored. If it exceeds the allowable deviation range, the gradient slope is automatically adjusted to maintain tension stability.

7. The high-speed roll changing control method based on tension prediction according to claim 1, characterized in that, Before generating the acceleration curve for the pre-reel, the process also includes: pre-adding a damping coefficient to the speed control loop of the main reel during the pre-reel acceleration phase. The damping coefficient Used to apply braking torque to the servo drive to suppress speed overshoot when the main reel experiences a sudden increase in speed due to cut-off and loss of load. Before starting the acceleration of the pre-roll reel, the process also includes: controlling the pre-roll reel to make axial micro-movements, and adjusting the axial position of the pre-roll reel in a closed-loop manner by reading the feedback value of the shaft end displacement sensor, so that the axial deviation between the edge of the pre-roll reel end face and the edge of the main reel end face converges to the preset tolerance range.

8. The high-speed roll changing control method based on tension prediction according to claim 1, characterized in that, In a multi-axis rewinding scenario involving multiple pre-roll reels, the calculated tension feedforward compensation also includes: For each pre-roll Based on recent The slope of the predicted volume diameter growth rate is calculated by linear regression fitting of each sampling point. Extrapolate the predicted volume size for each future control cycle And calculate the corresponding rotational inertia prediction sequence. and independent feedforward compensation sequence ; Based on the wrap angle of each guide roller along the material transport path and coefficient of friction Calculate the transmission coefficient of each guide roller. and reflection coefficient And based on the elastic wave propagation speed of the material Calculate the propagation delay time and amplitude attenuation coefficient for each path order to generate the inter-axis multipath propagation feature matrix. ; The independent feedforward compensation sequence of each axis is convolved with the inter-axis multipath propagation feature matrix in the time domain to obtain the cross-axis coupling effect time domain signal matrix; the peak time and peak amplitude of the total coupling effect signal at each axis are identified; with the constraint that the superimposed peak of the coupling effect at each axis does not exceed the allowable fluctuation range of the target tension, the application time offset and amplitude scaling ratio of the compensation amount of each axis are jointly solved to generate a collaborative peak-shifting compensation sequence. The collaborative peak-shifting compensation sequence is subjected to propagation inverse compensation processing. The compensation command is issued in advance according to the corresponding propagation time, and a time-calibrated roll-changing control command sequence is generated to replace the tension feedforward compensation output to each axis servo driver.

9. The high-speed roll changing control method based on tension prediction according to claim 8, characterized in that, Also includes: During the roll-changing acceleration period of each pre-reel combination, cross-correlation analysis is performed on the time-calibrated roll-changing control command sequence of each roll combination and the actual tension signal at the corresponding sensor position to extract the estimated value of actual propagation delay and the estimated value of actual amplitude attenuation. The actual propagation delay estimate and the actual amplitude attenuation estimate are compared with the predicted values ​​of the direct path in the inter-axis multipath propagation feature matrix to calculate the prediction error; The propagation velocity and transmission coefficient in the inter-axis multipath propagation feature matrix are updated online using a recursive identification algorithm. ， ， in, and These represent the propagation speed before and after the update. To recursively identify the learning step size for propagation speed, For propagation delay prediction error, For scrolls To the scroll The length of the direct path, and These are the guide rollers before and after the update. The transmission coefficient, The recursive identification learning step size for the transmission coefficient. This refers to the amplitude attenuation prediction error; Simultaneously, the actual tension deviations at each axis are back-allocated to the linear regression model of each axis according to the coupling weights, and the predicted slope of the roll diameter growth rate is corrected.

10. A high-speed roll change control system based on tension prediction, used to execute the high-speed roll change control method based on tension prediction as described in any one of claims 1 to 9, characterized in that, include: The parameter initialization unit is used to read the target tension setpoint, main production line speed and material property parameters, and to perform calibration processing on the tension detection module and the distance measurement module. The data acquisition and roll change prediction unit is used to collect the current actual tension value of the material and the current roll diameter of the main roll in real time, and calculate the roll change trigger time based on the roll diameter change rate. The acceleration curve and feedforward compensation generation unit is used to generate the acceleration curve of the pre-roll spool based on the electronic cam algorithm, and to calculate the tension feedforward compensation based on the rotational inertia and angular acceleration of the pre-roll spool. The integrated control command output unit is used to continuously output the speed error feedback item, tension error feedback item and the tension feedforward compensation amount to the servo driver corresponding to the pre-reel at control cycle intervals. The speed synchronization determination and roll changing execution unit is used to determine whether the speed error condition and phase alignment condition between the pre-roll and the main production line are met simultaneously, and when they are met, the cutter cylinder and pressure roller cylinder are triggered in sequence to perform the roll changing action. The tension control gradual transfer unit is used to synchronously and linearly decrease the main reel torque output and linearly increase the pre-reel tension closed-loop control gain coefficient during the transition time, so as to complete the smooth transfer of tension control.

Images