A high-speed winding device and method

Abstract

This invention relates to the field of winding equipment control technology, and discloses a high-speed winding device and method, including a main drive and wire laying mechanism, an active tension swing arm mechanism, a contour vision acquisition component, and an electrical control terminal. The electrical control terminal is equipped with multiple control modules. During the winding process, the dynamic perturbation anti-overlap control module, when determining that the transient winding ratio falls into the resonance range, controls the main drive and wire laying mechanism to perform position perturbation actions to prevent wire stacking. Simultaneously, the reverse-phase feedforward tension compensation module synchronously controls the active tension swing arm mechanism to perform displacement compensation with the phase opposite to the perturbation action, offsetting tension fluctuations caused by sudden trajectory changes. The dual-scale state-space observation module combines stress feedback to update mechanical parameters online to correct the control benchmark. The spatiotemporal joint vision filtering module performs wire laying calibration based on image data. This invention avoids local bulges or collapse defects, maintains tension stability at high speeds, and improves the quality of wire winding.

Description

Technical Field

[0001] This invention relates to the field of winding equipment control technology, specifically to a high-speed winding device and method. Background Technology

[0002] High-speed winding equipment is widely used in the winding and forming process of linear materials such as metal wire, optical fiber, and textile yarn. During the high-speed winding process, the appearance and flatness of the coiled tube are important indicators for measuring the winding quality, which mainly depends on the trajectory control of the winding mechanism and the tension stability of the system.

[0003] In actual production, when the equipment operates at a constant linear speed and the transient winding ratio of the spindle speed and the reciprocating frequency of the wire falls within a specific resonance range, continuous wire tends to overlap at the same position on the bobbin surface. This overlap can cause local bulges on the winding surface, leading to edge collapse or difficulty in unwinding. Existing winding control systems typically attempt to disrupt the resonance condition by using simple speed fluctuations or position fine-tuning to prevent overlap. However, transient changes in the wire trajectory can cause abrupt changes in the length of the spatial wire path. Since existing tension lever mechanisms usually rely on delayed error feedback adjustment and cannot intervene in advance, such abrupt trajectory changes will inevitably cause severe fluctuations in wire tension, easily leading to wire damage or even breakage.

[0004] Furthermore, the winding process is a typical time-varying process. As the coil diameter increases, the wire length and stress state change, and the equivalent stiffness and damping of the system also change accordingly. Traditional control systems mostly rely on fixed mechanical model parameters set during the equipment commissioning phase, lacking online adaptive update capabilities, which leads to a gradual decrease in the accuracy of tension compensation as the coil diameter increases. At the same time, although some existing equipment is equipped with visual inspection components, they are mostly used only for offline defective product rejection after winding is completed, and are not deeply integrated into the closed-loop calibration of the wire placement position during operation, making it difficult to actively prevent coil morphology distortion under high-speed operation. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a high-speed winding device and method, which solves the problems of tension fluctuations easily caused by existing high-speed winding devices when performing anti-overlap trajectory fine-tuning, and the reduced tension compensation accuracy and poor roll forming quality caused by the control system's inability to adapt to the time-varying mechanical characteristics of winding due to the use of a fixed parameter model.

[0006] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides a high-speed winding device, comprising: The system comprises a main drive and wire winding mechanism, an active tension swing arm mechanism, a contour vision acquisition component, and an electrical control terminal. The main drive and wire winding mechanism winds the wire and controls its trajectory. The active tension swing arm mechanism measures and adjusts the wire tension. The contour vision acquisition component acquires image data of the outer contour of the coiled bobbin. The electrical control terminal contains a control system, which includes: a parameter identification and initialization module for identifying the wire's mechanical parameters; a spatiotemporal joint vision filtering module for processing image data to generate a position deviation compensation vector and perform wire winding position calibration; a dynamic perturbation anti-overlap control module for controlling the main drive and wire winding mechanism to perform position perturbation actions when wire winding overlaps; a reverse-phase feedforward tension compensation module for controlling the active tension swing arm mechanism to perform displacement compensation with the phase opposite to the position perturbation action; and a dual-scale state-space observation module for online prediction of system state, updating mechanical parameters, and outputting the feedforward compensation reference correction torque.

[0007] By adopting the above technical solution, the system can perform anti-overlapping position perturbation while introducing feedforward displacement compensation with opposite phase to offset tension fluctuations caused by position changes. At the same time, it can perform flatness calibration of the outer contour of the roll by combining visual images and adjust the mechanical parameters of the wire online through the state observer to adjust the control benchmark, thereby realizing the joint control of the wire laying space position and tension, improving the tension stability and roll forming quality of the winding process.

[0008] Preferably, the active tension swing arm mechanism is internally configured with a voice coil motor, an active tension swing arm, and a closed-loop tension sensor. The parameter identification and initialization module outputs a sweep frequency current command to the voice coil motor, receives tensile stress data collected in real time by the closed-loop tension sensor, and uses the reciprocating displacement generated by the active tension swing arm as the input excitation parameter. Based on the discretized viscoelastic dynamic equation, the equivalent elastic modulus and damping coefficient of the wire are calculated using the least squares method as the mechanical parameters.

[0009] By adopting the above technical solution, specific excitation is applied in the initial stage of operation using the inherent actuators and sensors of the winding equipment, and the mechanical parameters of the current wire are obtained online, reducing the initial model error caused by differences in wire material or batch.

[0010] Furthermore, the main drive and wiring mechanism is internally equipped with a spindle motor, a wiring servo motor, and a photoelectric encoder. The spatiotemporal joint visual filtering module performs an exponentially weighted moving average operation on the image data at the same mechanical phase angle within a continuous rotation cycle to extract the true slope change rate. When it is determined that the true slope change rate exceeds a preset safety threshold, the position deviation compensation vector is generated and output to the wiring servo motor for closed-loop correction of the winding shape. In the process of performing the exponentially weighted moving average operation, a negatively correlated linear mapping function between the smoothing factor and the rotational speed of the spindle motor is established to adaptively adjust the smoothing factor.

[0011] By adopting the above technical solution, visual data comparison is performed under a specific mechanical phase angle to eliminate periodic interference caused by spindle rotation asymmetry. Furthermore, by introducing an adaptive smoothing factor related to rotational speed, the filtering weight is increased in the high-speed stage and the response speed is accelerated in the low-speed stage, thereby obtaining an effective contour slope change as a calibration benchmark for wiring.

[0012] Preferably, the main drive and wiring mechanism is internally equipped with a spindle motor, a wiring servo motor, and a photoelectric encoder. The dynamic micro-disturbance anti-overlap control module obtains the rotational speed data output by the photoelectric encoder and the reciprocating frequency of the wiring servo motor to calculate the transient winding ratio. When it is determined that the transient winding ratio falls into the set resonance range, a digital triangular wave micro-disturbance signal with an amplitude constrained within one to three percent of the amplitude of the basic position control command of the main drive and wiring mechanism is generated. The digital triangular wave micro-disturbance signal is superimposed on the basic position control command of the main drive and wiring mechanism to drive the wiring servo motor to perform a lateral reciprocating motion with slight spatial jitter.

[0013] By adopting the above technical solution, the risk of wire roll overlap can be judged in advance by using the transient winding ratio, and the wire trajectory can be deflected by superimposing a digital triangular wave signal with a limited amplitude range, thereby reducing the accumulation of wire in a fixed spatial position.

[0014] Furthermore, the reverse-phase feedforward tension compensation module extracts the instantaneous amplitude of the digital triangular wave perturbation signal, and calculates the desired offset tension by combining the set geometric coupling coefficient with the equivalent elastic modulus in the mechanical parameters. The reverse-phase feedforward tension compensation module converts the desired offset tension into a feedforward thrust current command and outputs it to the voice coil motor. The motion phase of the feedforward thrust current command is opposite to the phase of the digital triangular wave perturbation signal.

[0015] By adopting the above technical solution, the anti-overlapping micro-disturbance signal is taken as a known disturbance source. The tension variable caused by it is calculated by combining the elastic modulus of the wire and the spatial geometric relationship of the wire. The tension swing arm mechanism is driven to generate a reverse displacement for feedforward compensation, thereby suppressing the tension fluctuation caused by the wire winding micro-disturbance.

[0016] Preferably, when the reverse-phase feedforward tension compensation module generates the feedforward thrust current command to be output to the active tension swing arm mechanism, it adds an advanced phase compensation amount based on the bus communication cycle to the feedforward thrust current command, and has a built-in feedforward limiting protection mechanism. It uses a preset smoothing saturation function to perform transition processing on the part of the expected displacement that exceeds 80% of the safety threshold of the maximum effective stroke of the voice coil motor.

[0017] By adopting the above technical solution, the communication delay in the electrical control system and the inertial lag of the mechanical actuator are offset by the advance phase compensation amount, ensuring the time consistency of the compensation action; at the same time, the displacement compensation range is constrained by the smooth saturation function to prevent the voice coil motor from exceeding the effective mechanical stroke and causing collision.

[0018] Furthermore, the dual-scale state-space observation module acquires the real-time stress feedback output by the closed-loop tension sensor, separates the low-frequency slowly changing component through a first-order low-pass filter with a cutoff frequency set to five Hz, and obtains the fast-changing high-frequency component by subtracting the low-frequency slowly changing component from the original real-time stress feedback. Based on the fast-changing high-frequency component and the mechanical parameters, a discrete state-space equation containing a discrete state transition matrix is ​​constructed to output a corrected estimated state vector.

[0019] By adopting the above technical solution, the tension feedback signal is decoupled into low-frequency components that characterize the change in roll diameter and the drift of the tension reference, and high-frequency components that characterize equipment vibration and high-speed winding disturbance. Based on the high-frequency components, a state-space equation is constructed to improve the accuracy of the observer's prediction of the transient dynamic changes of the system.

[0020] Preferably, the dual-scale state-space observation module calculates the macroscopic tension residual predicted by the actual measurement vector and the discrete state-space equation, with the goal of minimizing the macroscopic tension residual, and uses the recursive least squares algorithm to update the equivalent elastic modulus and damping coefficient in the discrete state transition matrix in real time. The corrected estimated state vector is then combined with the system feedforward gain matrix for linear weighted transformation to output the feedforward compensation benchmark.

[0021] By adopting the above technical solution, the mechanical parameter variables inside the state space model are corrected in real time based on the deviation between the model prediction and the actual measurement, so that the system can adapt to the changes in plastic tension and damping characteristics of the wire during long-term winding, and improve the feedforward compensation accuracy of continuous operation.

[0022] Furthermore, the dual-scale state-space observation module calculates the observation information covariance. When it determines that the fault tolerance threshold has been exceeded, it triggers the anomaly prediction and hold mechanism, forcibly setting the state observer gain matrix to zero and performing open-loop recursion. When it is detected that the tensile stress data is continuously lower than the preset wire breakage protection threshold and approaches zero within a set time window, and the rotational speed data fed back by the photoelectric encoder configured in the main drive and wiring mechanism is higher than the set operating speed threshold, it determines that a wire breakage fault has occurred and immediately triggers the highest priority hardware interrupt to send an emergency braking command.

[0023] By adopting the above technical solution, parameter updates are stopped when the feedback signal exceeds the fault tolerance threshold to maintain system stability. At the same time, multi-condition logic judgment is performed by combining tension drop and high-speed operation status to improve the accuracy of wire breakage fault identification and response speed.

[0024] Secondly, the present invention also provides a high-speed winding method, comprising the following steps: S1. When the static threading of the wire is completed and the main spindle motor is stationary, the parameter identification and initialization module outputs a sweep frequency current command to the voice coil motor of the active tension swing arm mechanism to drive the active tension swing arm to reciprocate, collect tensile stress data and actual displacement data, and extract mechanical parameters. S2. After the winding equipment enters the running state, the main drive and the main shaft motor of the winding mechanism drive the winding tube to rotate and speed up according to the set process acceleration curve. During this period, the spatiotemporal joint vision filtering module triggers the monochrome line array industrial camera inside the contour vision acquisition component to perform synchronous exposure, performs exponential weighted moving average calculation on the image data and adaptively lowers the smoothing factor. When the safety threshold is exceeded, the output position deviation compensation vector fine-tunes the reversing pole position of the winding guide. S3. During the constant linear speed operation phase, the dynamic micro-disturbance anti-overlap control module calculates the transient winding ratio based on the rotation speed and reciprocating frequency. When the transient winding ratio falls into the resonance range, a digital triangular wave micro-disturbance signal is generated and superimposed on the basic position control command of the wiring servo motor to drive the wiring guide to superimpose a small high-frequency jitter on the wiring path. S4. The reverse phase feedforward tension compensation module synchronously extracts the instantaneous offset amplitude and current phase data of the digital triangular wave micro-disturbance signal and calls the mechanical parameters. It calculates the expected offset tension through the spatial linear geometric equation and outputs a feedforward thrust current command with advanced phase compensation to the active tension swing arm mechanism to actively counteract wire tension fluctuations. S5. The dual-scale state-space observation module separates the collected real-time stress feedback, the actual torque current of the main spindle motor, and the displacement and velocity data of the active tension swing arm mechanism into low-frequency slowly changing components and fast-changing high-frequency components through a low-pass filter. Combining the constructed discrete state-space equations and the recursive least squares algorithm, the mechanical parameters are updated in real time, and a feedforward compensation benchmark is output to correct the output torque of the voice coil motor of the active tension swing arm mechanism in real time.

[0025] By adopting the above technical solutions, the control timing logic of the equipment in different operating cycles such as static initialization, acceleration, and constant speed is standardized. Parameter identification, roll-to-roll visual correction, dynamic micro-disturbance anti-overlap, reverse phase tension compensation, and state space parameter updates are combined into the same control process to achieve stable control of the equipment.

[0026] This invention provides a high-speed winding device and method. It has the following beneficial effects: 1. This invention improves the winding quality of wire by working in concert with a dynamic micro-perturbation anti-overlap control module and a spatiotemporal joint visual filtering module. The system uses a contour visual acquisition component to monitor the winding shape in real time, and when the transient winding ratio falls into the resonance range, it superimposes a digital triangular wave micro-perturbation signal into the basic position command of the winding mechanism. This action can actively change the local layout trajectory of the wire, avoid local bulges or collapses caused by overlapping of windings during the winding process, and ensure the flatness of the spool.

[0027] 2. This invention uses a reverse-phase feedforward tension compensation module to solve the tension fluctuation problem caused by anti-overlapping micro-disturbance action. When the cable laying mechanism performs micro-disturbance, the module will synchronously extract the amplitude of the micro-disturbance signal and calculate the expected offset tension in combination with the mechanical parameters of the cable. Then, it controls the active tension swing arm mechanism to output a feedforward displacement that is opposite to the phase of the micro-disturbance action. This feedforward offset mechanism can offset the transient stretching of the cable caused by the sudden change in the cable laying trajectory in advance, and maintain the overall stability of the tension under high-speed operation.

[0028] 3. This invention utilizes a dual-scale state-space observation module to achieve online adaptive updating of control parameters, thereby improving the long-term accuracy of the tension control system. Since the mechanical properties of the wire change with the winding diameter and process conditions during the winding process, this module separates the rapidly changing high-frequency components from the real-time stress feedback and combines them with a recursive least squares algorithm to correct the equivalent elastic modulus and damping coefficient in the discrete state-space equation in real time. This dynamic update mechanism enables the feedforward compensation benchmark output by the system to continuously conform to the actual working conditions, avoiding control deviations caused by fixed parameter models. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the device structure of the present invention; Figure 2This is a schematic diagram of the internal system architecture of the device of the present invention; Figure 3 This is a schematic diagram of the architecture of the parameter identification and initialization module of the present invention; Figure 4 This is a schematic diagram of the spatiotemporal joint visual filtering module of the present invention; Figure 5 This is a schematic diagram of the architecture of the dynamic perturbation anti-overlap control module of the present invention; Figure 6 This is a schematic diagram of the architecture of the reverse-phase feedforward tension compensation module of the present invention; Figure 7 This is a schematic diagram of the architecture of the dual-scale state-space observation module of the present invention; Figure 8 This is a schematic flowchart of the high-speed winding method of the present invention; Figure 9 This is a comparison chart of the spatiotemporal joint visual filtering effects of the present invention; Figure 10 This is a schematic diagram of the basic position and superimposed perturbation command of the main drive and cable laying mechanism of the present invention; Figure 11 This is a schematic diagram of the feedforward displacement command for the active tension swing arm mechanism of the present invention; Figure 12 This is a comparison diagram of the macroscopic system tension control effect of the present invention.

[0030] Among them, 1. Main drive and wiring mechanism; 2. Active tension swing arm mechanism; 3. Contour vision acquisition component; 4. Electrical control terminal; 401. Parameter identification and initialization module; 402. Spatiotemporal joint vision filtering module; 403. Dynamic micro-disturbance anti-overlap control module; 404. Inverse phase feedforward tension compensation module; 405. Dual-scale state space observation module. Detailed Implementation

[0031] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] See attached document Figure 1 , Figure 1 This is a schematic diagram of the device structure according to an embodiment of the present invention. The present invention provides a high-speed winding device, including: a main drive and wire laying mechanism 1, an active tension swing arm mechanism 2, a contour vision acquisition component 3, and an electrical control terminal 4.

[0033] The main drive and wire laying mechanism 1 is used to drive the wire winding and shape and control the spatial arrangement trajectory of the wire. The main drive and wire laying mechanism 1 is equipped with a spindle motor, a wire laying servo motor and a photoelectric encoder. The spindle motor is used to drive the winding bobbin to rotate. The photoelectric encoder is located at the tail of the spindle motor. The photoelectric encoder outputs physical phase angle pulse signals and speed data. The wire laying servo motor is used to drive the wire laying guide to perform lateral reciprocating motion on the wire laying path.

[0034] The active tension swing arm mechanism 2 is located on the wiring path below the main drive and wiring mechanism 1. The active tension swing arm mechanism 2 is equipped with a voice coil motor, an active tension swing arm, and a closed-loop tension sensor. The voice coil motor is connected to the active tension swing arm and is used to drive the active tension swing arm to generate axial displacement. The closed-loop tension sensor is installed at the wire passing node above the active tension swing arm and is fixed on an independent vibration damping mounting base of the frame. The measuring end of the closed-loop tension sensor contacts the wire through a guide wheel with a micro-damping bearing to physically isolate the mechanical vibration noise transmitted to the sensor by the high-frequency reciprocating motion of the active tension swing arm, thereby ensuring the authenticity and stability of the tensile stress data of the wire measured and output by the closed-loop tension sensor.

[0035] The contour vision acquisition component 3 is fixed on the frame panel behind the main drive and wiring mechanism 1. The contour vision acquisition component 3 is used to install vision perception hardware. The contour vision acquisition component 3 is equipped with a monochrome line array industrial camera. The lens of the monochrome line array industrial camera is aimed at the outer contour of the rolled bobbin on the main drive and wiring mechanism 1. The monochrome line array industrial camera is electrically connected to the photoelectric encoder. The monochrome line array industrial camera receives the physical phase angle pulse signal output by the photoelectric encoder to perform synchronous exposure.

[0036] The electrical control terminal 4 is located on the back of the high-speed winding equipment. The electrical control terminal 4 is used to integrate the control panel and the underlying computing unit. The electrical control terminal 4 integrates a central processing unit and a field-programmable gate array.

[0037] See attached document Figure 2 , Figure 2 This is a schematic diagram of the internal system architecture of a device according to an embodiment of the present invention. Based on the main drive and winding mechanism 1, the active tension swing arm mechanism 2, the contour vision acquisition component 3, and the electrical control terminal 4, the present invention provides a control system inside a high-speed winding device, including: a parameter identification and initialization module 401, a spatiotemporal joint vision filtering module 402, a dynamic micro-disturbance anti-overlap control module 403, a reverse phase feedforward tension compensation module 404, and a dual-scale state space observation module 405.

[0038] The parameter identification and initialization module 401 is fixed inside the electrical control terminal 4. The parameter identification and initialization module 401 is used to extract the mechanical parameters of the wire. The parameter identification and initialization module 401 is communicatively connected to the active tension swing arm mechanism 2. The parameter identification and initialization module 401 outputs a sweep frequency current command to the voice coil motor of the active tension swing arm mechanism 2.

[0039] The parameter identification and initialization module 401 synchronously receives the tensile stress data output by the closed-loop tension sensor. The parameter identification and initialization module 401 constructs a dynamic equation based on the viscoelastic model of the wire. The parameter identification and initialization module 401 uses the least squares method to calculate the equivalent elastic modulus and damping coefficient of the wire. The parameter identification and initialization module 401 stores the equivalent elastic modulus and damping coefficient in the register of the electrical control terminal 4.

[0040] The spatiotemporal joint vision filtering module 402 is fixed inside the electrical control terminal 4. The spatiotemporal joint vision filtering module 402 is used to process the roll contour data and perform position calibration. The spatiotemporal joint vision filtering module 402 is communicatively connected to the contour vision acquisition component 3 and receives the image data output by the contour vision acquisition component 3.

[0041] The spatiotemporal joint visual filtering module 402 combines the physical phase angle pulse signal to perform synchronous processing logic on the image data. The spatiotemporal joint visual filtering module 402 performs exponential weighted moving average calculation on the image data within a continuous rotation cycle. The spatiotemporal joint visual filtering module 402 filters out interference signals. The spatiotemporal joint visual filtering module 402 generates a position deviation compensation vector and outputs the position deviation compensation vector to the main drive and the wiring servo motor of the wiring mechanism 1.

[0042] The dynamic disturbance anti-overlap control module 403 is fixed inside the electrical control terminal 4. The dynamic disturbance anti-overlap control module 403 is used to change the spatial arrangement trajectory of the wire. The dynamic disturbance anti-overlap control module 403 is communicatively connected to the main drive and wire laying mechanism 1. The dynamic disturbance anti-overlap control module 403 obtains the speed data output by the photoelectric encoder and the reciprocating frequency of the wire laying servo motor.

[0043] The dynamic disturbance anti-overlap control module 403 calculates the transient winding ratio based on the rotational speed data and reciprocating frequency. When the dynamic disturbance anti-overlap control module 403 determines that the transient winding ratio falls into the set resonance range, the dynamic disturbance anti-overlap control module 403 generates a digital triangular wave disturbance signal. The dynamic disturbance anti-overlap control module 403 superimposes the digital triangular wave disturbance signal onto the position control command of the main drive and the wiring mechanism 1.

[0044] The reverse-phase feedforward tension compensation module 404 is fixed inside the electrical control terminal 4. The reverse-phase feedforward tension compensation module 404 is used to physically offset the additional tension fluctuations caused by the change in the cable trajectory. The reverse-phase feedforward tension compensation module 404 and the dynamic micro-disturbance anti-overlap control module 403 operate synchronously based on the same clock cycle. The reverse-phase feedforward tension compensation module 404 extracts the characteristic data of the digital triangular wave micro-disturbance signal.

[0045] The reverse-phase feedforward tension compensation module 404, combined with the equivalent elastic modulus calculated by the parameter identification and initialization module 401, calculates the expected counteracting tension that the voice coil motor needs to output. The reverse-phase feedforward tension compensation module 404 outputs a feedforward displacement command to the voice coil motor of the active tension rocker arm mechanism 2. The phase of the feedforward displacement command is opposite to the phase of the digital triangular wave perturbation signal.

[0046] The dual-scale state space observation module 405 is fixed inside the electrical control terminal 4. The dual-scale state space observation module 405 is used to maintain the steady-state output of the system tension. The dual-scale state space observation module 405 is communicatively connected to the parameter identification and initialization module 401 and the active tension swing arm mechanism 2, respectively.

[0047] The dual-scale state-space observation module 405 receives tensile stress data output by the closed-loop tension sensor and equivalent elastic modulus and damping coefficient extracted by the parameter identification and initialization module 401. The dual-scale state-space observation module 405 calculates the macroscopic tension residual. The dual-scale state-space observation module 405 updates the equivalent elastic modulus and damping coefficient using a recursive least squares algorithm. The dual-scale state-space observation module 405 outputs a feedforward compensation reference to the main control loop of the electrical control terminal 4 to correct the output torque of the voice coil motor of the active tension rocker arm mechanism 2.

[0048] See attached document Figure 3 , Figure 3 This is a schematic diagram of the architecture of a parameter identification and initialization module according to an embodiment of the present invention. In this embodiment, the parameter identification and initialization module 401 is fixed inside the electrical control terminal 4. As a preferred embodiment, the parameter identification and initialization module 401 is used to automatically extract the mechanical characteristic data of the wire when the winding device is in a static state after the wire is threaded and the spindle is not rotating.

[0049] In order to obtain an accurate underlying physical reference, the parameter identification and initialization module 401 outputs a sweep frequency current command to the voice coil motor of the active tension swing arm mechanism 2. The voice coil motor of the active tension swing arm mechanism 2 receives the sweep frequency current command and generates an axial electromagnetic thrust, thereby driving the active tension swing arm to generate a high-frequency reciprocating motion. The active tension swing arm forces the wire on the wire path to undergo axial tensile deformation and contraction deformation.

[0050] Considering the viscoelastic nature of wires, the elastic modulus and damping characteristics usually exhibit significant frequency response differences at different stretching rates. Therefore, the sweep current command is preferably configured as a discrete sine wave signal sequence containing multiple preset frequencies. The range of preset frequencies is usually set to cover the low-frequency to mid-frequency range of the wire's mechanical resonance frequency. This multi-band continuous physical excitation can fully excite and expose the complete physical response characteristics of the wire under static and dynamic alternating rates.

[0051] Under the continuous action of the above physical excitation, the closed-loop tension sensor on the active tension swing arm mechanism 2 measures the tensile stress data generated by the wire during axial deformation in real time. The parameter identification and initialization module 401 synchronously receives the tensile stress data output by the closed-loop tension sensor. Then, the parameter identification and initialization module 401 uses the reciprocating displacement generated by the active tension swing arm as the input excitation parameter and the tensile stress data output by the closed-loop tension sensor as the output response parameter.

[0052] Therefore, the parameter identification and initialization module 401 constructs a discretized viscoelastic dynamic equation describing the deformation law of the wire based on the input excitation parameters and output response parameters. The systematic identification and construction process of this discretized viscoelastic dynamic model does not rely on a large sample library for black-box mapping, but is based on the classic Kelvin-Woyt physical model architecture. The parameter identification and initialization module 401 equates the microscopic wire to a parallel physical combination of a macroscopic ideal spring and an ideal damper, thereby establishing the first-order differential deterministic physical relationship between the material tension response and displacement change.

[0053] The mathematical expression for the discretized viscoelastic dynamics equation is: ; In the formula, For closed-loop tension sensors, the discrete time step Tensile stress data collected below; The equivalent elastic modulus to be identified; The absolute deformation of the wire caused by the displacement of the active tension swing arm; The basic free length of the wire between the front and rear nodes of the active tension swing arm mechanism 2; The damping coefficient to be identified; For absolute deformations in continuous time The first derivative; This represents the discrete sequence number of time sampling in the dynamic model.

[0054] For the discretized viscoelastic dynamics model constructed above, the parameter identification and initialization module 401 performs data fitting calculations on the discretized viscoelastic dynamics equations using the least squares algorithm. Specifically, the parameter identification and initialization module 401 establishes a cost function with the objective of minimizing the squared error between the tensile stress data and the predicted tension data output by the discretized viscoelastic dynamics equations. The mathematical expression of the cost function is as follows: ; In the formula, The cost function is the least squares iterative optimization method. This represents the total number of data points within the sampling window. For the first The actual tensile stress data output by the closed-loop tension sensor at each sampling point; In order to identify the parameter matrix Below, the discretized viscoelastic dynamics equations are in the first... The predicted tension data output from each sampling point This is the set of parameters to be identified, including the equivalent elastic modulus and damping coefficient. This is the index of the traversal sequence of the sampled data points within the window.

[0055] Subsequently, the parameter identification and initialization module 401 calculates the partial derivatives of the variables to be identified in the cost function and sets the partial derivatives to zero, thereby solving for the equivalent elastic modulus and damping coefficient. The specific derivation process of the least squares algorithm for solving the partial derivatives and performing data matrix operations is implemented based on the existing standard numerical analysis methods to realize the calculation logic of the least squares algorithm for solving the partial derivatives and performing data matrix operations.

[0056] Meanwhile, in order to avoid matrix irreversibility caused by severe interference from sensor signals or accidental loosening of wires, the parameter identification and initialization module 401 performs variance consistency verification on the collected tensile stress data before the least squares iterative calculation, and automatically calls the general default parameter matrix pre-stored in the electrical control terminal 4 when it is determined that the calculation does not converge, so as to ensure the integrity of the underlying control logic and the safety of equipment operation.

[0057] The final parameter identification and initialization module 401 stores the equivalent elastic modulus and damping coefficient in the register inside the electrical control terminal 4. When the winding equipment starts up and enters the running state, the electrical control terminal 4 calls the equivalent elastic modulus and damping coefficient in the register, so that the equivalent elastic modulus and damping coefficient can be used as the reference parameters for the tension control algorithm to perform torque calculation.

[0058] See attached document Figure 4 , Figure 4This is a schematic diagram of the architecture of a spatiotemporal joint visual filtering module according to an embodiment of the present invention. In this embodiment, the spatiotemporal joint visual filtering module 402 is fixed inside the electrical control terminal 4. As a preferred embodiment, the spatiotemporal joint visual filtering module 402 is used to process the roll contour data and perform dynamic calibration of the wiring position in an environment where the winding equipment is running at high speed and there is random flying fluff interference in the surroundings.

[0059] To ensure spatial consistency in visual acquisition, the spatiotemporal joint visual filtering module 402 binds the physical phase angle pulse signal output by the main drive and the wiring mechanism 1 with the exposure action of the monochrome line array industrial camera at the underlying hardware level. By reading the physical phase angle pulse signal output by the photoelectric encoder, the spatiotemporal joint visual filtering module 402 defines the current mechanical phase angle. The monochrome line array industrial camera performs synchronous exposure and outputs one-dimensional grayscale image data at each specific mechanical phase angle.

[0060] The image processing logic within the spatiotemporal joint visual filtering module 402 extracts the pixel coordinates of the edge of the roll end face. Specifically, the spatiotemporal joint visual filtering module 402 uses a one-dimensional difference operator to identify grayscale step change points, extracts the physical coordinates of the end face edge under a single mechanical phase angle, and combines the edge coordinates extracted under each mechanical phase angle in a continuous rotation cycle according to the time sequence, thereby converting them into the original slope change rate matrix characterizing the flatness of the roll end face.

[0061] Considering that the floating fluff in the textile field will obscure the field of view of the monochrome line array industrial camera at random spatial locations and random time points, the spatiotemporal joint visual filtering module 402 performs an exponentially weighted moving average calculation on the original slope change rate matrix of continuous rotation cycles under the same mechanical phase angle. The spatiotemporal joint visual filtering module 402 smooths the contour data of the fixed phase angle in space on the time axis and filters out the high-frequency interference signal caused by the instantaneous passing of the fluff.

[0062] The mathematical expression for the exponentially weighted moving average is: ; In the formula, Current rotation period At a specific mechanical phase angle The rate of change of the true slope after down-filtering; Current rotation period At a specific mechanical phase angle The original slope change rate matrix collected below; It is a smoothing factor; For the previous rotation cycle At the same specific mechanical phase angle The rate of change of the true slope after down-filtering; A specific mechanical phase angle bound to the synchronous exposure of a monochrome line scan industrial camera; This is the discrete counting sequence number of the spindle rotation cycle of the equipment.

[0063] In order to balance the smoothness of the filter with the system’s response speed to real wiring anomalies, the smoothing factor is usually set between 0.1 and 0.3. The spatiotemporal joint vision filtering module 402 adaptively adjusts the smoothing factor according to the real-time speed of the spindle motor. The specific adjustment mechanism is that the spatiotemporal joint vision filtering module 402 establishes a negative correlation linear mapping function between the smoothing factor and the spindle motor speed.

[0064] When the spindle motor speed increases, the spatiotemporal joint vision filtering module 402 appropriately reduces the smoothing factor according to the negative correlation linear mapping function to enhance the spatiotemporal joint vision filtering module 402's dependence on historical stable data. At the same time, in order to avoid the control dead zone caused by the continuous obstruction of the lens of the monochrome linear array industrial camera by large clumps of flying catkins, the spatiotemporal joint vision filtering module 402 evaluates the connectivity and brightness characteristics of the original slope change rate matrix before calculation. When the image data of several consecutive cycles loses effective edge features, the spatiotemporal joint vision filtering module 402 triggers a visual anomaly alarm mechanism and temporarily stops the subsequent compensation calculation logic.

[0065] After obtaining the true contour of the roll, the spatiotemporal joint vision filtering module 402 determines whether the true slope change rate after filtering exceeds the preset safety threshold. The safety threshold here is determined by the maximum allowable end face runout in the process parameters. Once the limit is exceeded, it indicates that there is a real bulge or collapse defect on the end face of the roll. The spatiotemporal joint vision filtering module 402 generates a position deviation compensation vector according to the direction and amplitude of the limit exceedance, and converts the position deviation compensation vector into a drive pulse output to the main drive and the wiring servo motor of the wiring mechanism 1. The wiring servo motor of the main drive and the wiring mechanism 1 dynamically fine-tunes the reversing pole position of the wiring guide according to the position deviation compensation vector and realizes the closed-loop correction of the roll shape.

[0066] See attached document Figure 5 , Figure 5 This is a schematic diagram of the architecture of a dynamic perturbation anti-overlap control module according to an embodiment of the present invention. In this embodiment, the dynamic perturbation anti-overlap control module 403 is fixed inside the electrical control terminal 4. As a preferred embodiment, the dynamic perturbation anti-overlap control module 403 is used to monitor the wire laying status during the winding process and dynamically change the spatial arrangement trajectory of the wire to avoid the occurrence of wire overlap defects.

[0067] In order to accurately identify potential overlap risks, the dynamic perturbation anti-overlap control module 403 is connected to the main drive and the wiring mechanism 1 to obtain the speed data output by the photoelectric encoder and the reciprocating frequency of the wiring servo motor in real time. Based on the continuity of the underlying hardware sensor data, the dynamic perturbation anti-overlap control module 403 calculates the transient winding ratio to characterize the overlap of the wiring trajectory based on the speed data and the reciprocating frequency.

[0068] The mathematical formula for calculating the transient winding ratio is: ; In the formula, For continuous time Rotational speed data output by the photoelectric encoder; For continuous time The reciprocating frequency of the ribbon cable servo motor; It is a continuous-time variable.

[0069] After obtaining the transient winding ratio, the dynamic perturbation anti-overlap control module 403 determines whether the transient winding ratio falls into the set resonance range. The resonance range is the dangerous area of ​​overlapping wires, which is determined by extending a fixed tolerance band from the center point of the integer or simple fractional winding ratio to both ends. In order to make the system's judgment conditions adaptable to engineering, the specific range of the fixed tolerance band is quantitatively determined by combining the outer diameter of the wire and the lateral movement speed of the current wire guide.

[0070] When the dynamic perturbation anti-overlap control module 403 determines that the transient winding ratio falls into the resonance range, it indicates that the current wiring trajectory of the wire has a serious tendency to overlap and stack. At this time, the dynamic perturbation anti-overlap control module 403 triggers the underlying hardware interrupt of the electrical control terminal 4 and generates a digital triangular wave perturbation signal in real time in the field programmable logic gate array inside the electrical control terminal 4. Under normal circumstances, the digital triangular wave perturbation signal adopts a high-frequency, low-amplitude triangular waveform to ensure that it produces a continuous and stepless offset effect on the wiring position.

[0071] To prevent the introduction of extreme mechanical resonance, the amplitude of the digital triangular wave perturbation signal is strictly constrained to within one to three percent of the amplitude of the basic position control command of the main drive and winding mechanism 1. That is, the perturbation range is limited to between 1.5 mm and 4.5 mm to ensure that it meets the precision winding process requirements for the flatness of the sewing or textile thread edges, and to avoid excessive perturbation causing edge collapse or mesh defects at both ends of the roll. In addition, the fundamental frequency of the digital triangular wave perturbation signal is set to less than one-third of the mechanical resonance frequency of the main drive and winding mechanism 1. For the specific configuration of the waveform signal generated by the field programmable gate array, the signal generation logic and the configuration of the underlying registers are implemented according to the existing standard hardware description language.

[0072] After generating the digital triangular wave perturbation signal, the dynamic perturbation anti-overlap control module 403 adds the digital triangular wave perturbation signal as an additional position offset to the basic position control command of the main drive and the winding mechanism 1. The winding servo motor of the main drive and the winding mechanism 1 receives the basic position control command after the digital triangular wave perturbation signal is added and executes a lateral reciprocating motion with slight spatial jitter. The winding trajectory misalignment in physical space disrupts the winding pattern that was originally going to overlap. At the same time, in order to prevent the long-term output of the digital triangular wave perturbation signal from damaging the overall flatness of the roll end face and causing the algorithm dead zone, the dynamic perturbation anti-overlap control module 403 continuously monitors the change of the transient winding ratio. Once the transient winding ratio naturally leaves the resonance range as the roll diameter increases.

[0073] The dynamic perturbation anti-overlap control module 403 controls the field programmable gate array to smoothly attenuate the amplitude of the digital triangular wave perturbation signal until it stops outputting, so that the main drive and the wiring mechanism 1's wiring servo motor can resume operation under the smooth basic position control command. In addition, if the dynamic perturbation anti-overlap control module 403 detects that the transient winding ratio stays in the resonance range for a longer than the preset safe time threshold, the dynamic perturbation anti-overlap control module 403 will activate the timeout forced exit mechanism. That is, the dynamic perturbation anti-overlap control module 403 sends a short acceleration fine-tuning command to the main spindle motor of the main drive and the wiring mechanism 1 to actively destroy the current resonance condition, thereby completely avoiding the underlying control logic from falling into the local execution dead zone.

[0074] See attached document Figure 6 , Figure 6 This is a schematic diagram of the architecture of a reverse-phase feedforward tension compensation module according to an embodiment of the present invention. In this embodiment, the reverse-phase feedforward tension compensation module 404 is fixed inside the electrical control terminal 4. As a preferred embodiment, the reverse-phase feedforward tension compensation module 404 is used to use the physical parameters extracted by the front-end module to offset the instantaneous tension change caused by the cable disturbance through the reverse phase action.

[0075] Since the anti-overlapping disturbances performed by the cable servo motor inevitably change the spatial path length of the cable and cause transient tension fluctuations, based on the basic physical requirement of maintaining the dynamic tension balance of the system, in order to actively offset transient tension fluctuations at the physical level, the inverse-phase feedforward tension compensation module 404 and the dynamic disturbance anti-overlap control module 403 share the system clock cycle. The inverse-phase feedforward tension compensation module 404 extracts the feature data of the digital triangular wave disturbance signal generated by the dynamic disturbance anti-overlap control module 403 in real time. The feature data specifically covers the instantaneous amplitude and phase state of the digital triangular wave disturbance signal.

[0076] After acquiring the feature data, the reverse-phase feedforward tension compensation module 404 calls the equivalent elastic modulus pre-stored in the parameter identification and initialization module 401. The reverse-phase feedforward tension compensation module 404 calculates the expected offset tension of the voice coil motor of the active tension swing arm mechanism 2 to counteract micro-disturbance fluctuations in combination with the basic wire-passing geometric model. Regarding the construction logic of the basic wire-passing geometric model, this geometric model is completely independent of the sample data drive, but is strictly established by the electrical control terminal 4 according to the absolute coordinates of the physical space when it is factory-fixed. Specifically, the reverse-phase feedforward tension compensation module 404 uses the center wire hole of the wire guide as the geometric moving point of the three-dimensional coordinate system, and the fixed hinge rotation axis of the active tension swing arm as the geometric origin. It uses the Euclidean space trigonometric law to construct the spatial distance vector and accurately establishes the rigid geometric mapping analytical solution between the change in the macroscopic tension path length of the wire and the microscopic compensation displacement of the swing arm.

[0077] The mathematical formula for calculating the expected counteracting force is: ; In the formula, For continuous time The calculated voice coil motor of the active tension rocker mechanism is expected to counteract the tension. The geometric coupling coefficient is the ratio of the cable position to the tensile deformation. The geometric coupling coefficient integrates the free length of the cable base and the equivalent cross-sectional area parameters. It is used to directly convert the perturbation displacement of the cable position into a dimensionless strain coefficient that characterizes the equivalent stress factor. Extract the solidified equivalent elastic modulus for parameter identification and initialization module 401; For continuous time Instantaneous amplitude of digital triangular wave perturbation signal; It is a continuous-time variable.

[0078] After calculating the desired offset tension, the reverse-phase feedforward tension compensation module 404, in conjunction with the electromagnetic thrust coefficient of the voice coil motor of the active tension swing arm, converts the desired offset tension into a corresponding feedforward thrust current command. In order to overcome the time delay caused by the data transmission of the underlying hardware bus, the reverse-phase feedforward tension compensation module 404 introduces an advance phase compensation amount based on the bus communication cycle into the feedforward thrust current command. The reverse-phase feedforward tension compensation module 404 outputs a feedforward thrust current command with an advance phase compensation amount to the voice coil motor of the active tension swing arm mechanism 2.

[0079] Due to the introduction of the negative sign in the formula, the motion phase of the feedforward thrust current command with the advance phase compensation is actually opposite to the phase of the digital triangular wave perturbation signal at the controlled execution end. Then, the voice coil motor of the active tension swing arm mechanism 2 drives the active tension swing arm to perform a compensation action opposite to the cable perturbation. The reverse phase feedforward tension compensation module 404 directly offsets and weakens the instantaneous stress change generated inside the wire to the greatest extent by relying on the spatial reverse superposition at the physical and mechanical level.

[0080] For the low-level closed-loop drive logic of the voice coil motor of the active tension lever mechanism 2 to perform displacement control, the low-level closed-loop drive logic is implemented based on the existing servo motor vector control method to perform servo closed-loop drive control. In addition, in order to avoid the extreme abnormality of wire slack and desquamation caused by excessive feedforward compensation action amplitude, the reverse phase feedforward tension compensation module 404 has a built-in feedforward limiting protection mechanism based on the maximum effective stroke of the voice coil motor of the active tension lever mechanism 2.

[0081] When the reverse-phase feedforward tension compensation module 404 determines that the expected displacement corresponding to the calculated feedforward thrust current command exceeds 80% of the safety threshold of the maximum effective stroke of the voice coil motor of the active tension rocker mechanism 2, the reverse-phase feedforward tension compensation module 404 does not perform a rigid direct hardware cutoff, but uses a preset smooth saturation function to perform a flexible transition process on the part exceeding the safety threshold, thereby ensuring that the winding equipment can maintain the basic stability of the wire tension under any large disturbance conditions without structural loss of control or inducing secondary mechanical impact.

[0082] See attached document Figure 7 , Figure 7 This is a schematic diagram of the architecture of a dual-scale state-space observation module according to an embodiment of the present invention. In this embodiment, the dual-scale state-space observation module 405 is fixed inside the electrical control terminal 4. As a preferred embodiment, the dual-scale state-space observation module 405 is used to estimate the internal physical state that cannot be directly measured by hardware sensors in real time during the operation of the winding equipment and to provide a more timely feedback reference for subsequent closed-loop control.

[0083] In order to accurately capture the dynamic behavior of the winding system with significant frequency differences, the dual-scale state space observation module 405 synchronously receives the real-time stress feedback from the closed-loop tension sensor, the actual torque current of the spindle motor, and the displacement and velocity data fed back in real time by the active tension swing arm mechanism 2. The dual-scale state space observation module 405 performs time-scale separation logic based on low-pass filtering on the real-time stress feedback and the actual torque current.

[0084] Specifically, the dual-scale state-space observation module 405 inputs the real-time stress feedback and the actual torque current into a first-order low-pass filter with a cutoff frequency set to five Hz. The dual-scale state-space observation module 405 separates the low-frequency slowly changing components that characterize the slow changes in winding radius and moment of inertia through the first-order low-pass filter. At the same time, the dual-scale state-space observation module 405 uses the original real-time stress feedback to subtract the low-frequency slowly changing components to obtain the fast-changing high-frequency components that characterize the instantaneous stress fluctuations of the wire. The dual-scale state-space observation module 405 avoids numerical aliasing of multi-scale variables in the same observation matrix by forcibly stripping the physical frequency band.

[0085] For the separated fast-changing high-frequency components, the dual-scale state-space observation module 405 constructs a discrete state-space equation with the internal tension of the wire and the first-order rate of change of the internal tension of the wire as the core. The dual-scale state-space observation module 405 performs iterative prediction and measurement correction to output the optimal estimated state vector. Regarding the construction steps of this discrete state-space model, the dual-scale state-space observation module 405 is directly derived from the differential theory of Newton's second law of motion.

[0086] The specific construction mechanism is that the dual-scale state-space observation module 405 discretizes the system dynamics differential equations in the time domain with zero order preservation, generating a discrete state transition matrix and a discrete input control matrix with actual physical connotations. The core elements inside the matrix are calibrated with underlying physical parameters by extracting the equivalent elastic modulus and damping coefficients from the direct mapping parameter identification and initialization module 401, thereby completing the white-box construction of the state-space white-box model.

[0087] The mathematical expression for the discrete state-space equation is: ; In the formula, For the next discrete time step The estimated state vector for prediction; It is a discrete state transition matrix; The current discrete time step The next updated estimated state vector; For discrete input control matrix; The current discrete time step The control input vector below; Here is the gain matrix of the state observer; The current discrete time step The actual measurement vector output by the lower closed-loop tension sensor; For measuring the observation matrix; The discrete control step size sequence number for performing computational iterations for the state-space observer.

[0088] To ensure the convergence speed and estimation accuracy of the discrete state space equations, the specific value of the state observer gain matrix is ​​not arbitrarily set. Instead, it is pre-calculated by the dual-scale state space observation module 405 calling the parameter identification and initialization module 401 to extract the equivalent elastic modulus and damping coefficient, and using the pole placement method to place the closed-loop poles of the system in the stable region within the unit circle. The specific derivation process of solving the state observer gain matrix using the pole placement method is as follows.

[0089] Based on the state feedback design method in existing modern control theory, the calculation logic of the gain matrix of the state observer is implemented by the pole placement method. At the same time, in view of the risk that the closed-loop tension sensor in the industrial field may be subjected to sudden mechanical impact or signal disconnection, which may cause the output error jump value, the dual-scale state space observation module 405 calculates the observation information covariance between the actual measurement vector and the model prediction vector before performing the measurement correction step.

[0090] When the dual-scale state-space observation module 405 determines that the covariance of the observed information exceeds the preset fault tolerance threshold, the fault tolerance threshold is determined by multiplying the statistical value of the Gaussian white noise variance of the closed-loop tension sensor in the static no-load state by a factor of three, i.e., by the 3σ criterion. The dual-scale state-space observation module 405 triggers the anomaly prediction and hold mechanism, that is, the dual-scale state-space observation module 405 forces the state observer gain matrix to be set to zero, so that the discrete state-space equation completely depends on the discrete state transition matrix and the discrete input control matrix for pure mathematical open-loop recursion. The dual-scale state-space observation module 405 thus completely blocks the state estimation divergence dead zone caused by the anomaly of the closed-loop tension sensor.

[0091] In response to the yarn breakage or head breakage faults that are prone to occur during high-speed textile winding, when the dual-scale state space observation module 405 detects that the tensile stress data output by the closed-loop tension sensor is continuously lower than the preset yarn breakage protection threshold and approaches zero within the set time window, and at this time the speed data fed back by the photoelectric encoder is higher than the set operating speed threshold, the dual-scale state space observation module 405 outputs a yarn breakage abnormality flag bit to the main control circuit of the electrical control terminal 4.

[0092] The main control circuit of the electrical control terminal 4 determines that a wire breakage fault has occurred and immediately triggers the highest priority hardware interruption. The main control circuit not only forcibly blocks the recursive calculation of state estimation, but also immediately sends an emergency braking command to the spindle motor and the servo motor of the main drive and the thread laying mechanism 1, so that the equipment stops quickly, thereby preventing the free disconnected sewing thread or textile thread from getting into the spindle or causing a large-area flying yarn entanglement accident.

[0093] Meanwhile, to ensure continuous matching between the physical model and the actual state of the wire, the dual-scale state-space observation module 405 calculates the macroscopic tension residuals predicted by the actual measurement vector and the discrete state-space equations. With the goal of minimizing the macroscopic tension residuals, it uses a recursive least squares algorithm to update the equivalent elastic modulus and damping coefficients in the discrete state transition matrix online in real time. The discrete recursive formulas for updating the macroscopic tension residuals and approximating the parameter matrix are as follows: ; In the formula, For the first The Kalman gain matrix is ​​obtained by discrete iterative calculation. For the first The error covariance matrix of the step; For the first The step consists of a data input vector composed of the current displacement and velocity; This is a forgetting factor used to reinforce the tracking weights of the current state; Input vector for data The transpose of the matrix; For the first The updated dynamic parameter matrix includes the equivalent elastic modulus and damping coefficient. For the first The step includes a dynamic parameter matrix containing the equivalent elastic modulus and damping coefficient; For the first The macroscopic tension residuals obtained from step calculation; For the first The calculated error covariance matrix is ​​updated step by step; It is the identity matrix; This is the discrete control step sequence number for synchronous execution of the recursive least squares algorithm.

[0094] Finally, the dual-scale state space observation module 405 transforms the corrected estimated state vector into a usable feedforward compensation benchmark. During the specific transformation, the dual-scale state space observation module 405 combines the system feedforward gain matrix to perform a linear weighted transformation on the estimated state vector and transmits the feedforward compensation benchmark to the main control loop of the electrical control terminal 4. The main control loop corrects the output torque of the voice coil motor of the active tension rocker mechanism 2 in real time based on the feedforward compensation benchmark. The dual-scale state space observation module 405 realizes the early perception and active suppression of unmeasurable interference inside the winding equipment by outputting the feedforward compensation benchmark.

[0095] See attached document Figure 8 , Figure 8This is a flowchart illustrating a high-speed winding method according to an embodiment of the present invention. Based on the aforementioned high-speed winding equipment and its internal control system, the present invention provides a high-speed winding method that integrates the static logic of each control module into dynamic execution steps in the actual production sequence. In this embodiment, the method includes the following steps: S1. When the winding equipment completes the static threading of the wire and the main drive and the main spindle motor of the wire laying mechanism 1 are in a stationary state, the parameter identification and initialization module 401 is activated to execute the initialization program. In order to obtain the physical characteristics of the current batch of wire, the parameter identification and initialization module 401 outputs a constant DC command to the voice coil motor of the active tension swing arm mechanism 2 to drive the active tension swing arm to displace in order to eliminate the initial slack state of the sewing thread or textile thread and establish a basic pre-tension environment. Under the premise of maintaining the basic pre-tension, the parameter identification and initialization module 401 outputs a broadband sweep frequency current command to the voice coil motor.

[0096] As a preferred approach, the frequency range of the broadband sweep frequency command is typically set within twice the typical mechanical resonant frequency of the device. The active tension swing arm executes an axial high-frequency reciprocating motion containing multiple frequency bands according to the broadband sweep frequency current command, thereby forcing the wire to produce continuous dynamic deformation. The closed-loop tension sensor in the active tension swing arm mechanism 2 collects the tensile stress data of the wire caused by dynamic deformation in real time and transmits it back along with the actual displacement data of the voice coil motor.

[0097] The parameter identification and initialization module 401 is based on the classical first-order Kelvin-Woyt viscoelastic dynamics formula, namely: ; In the formula, The measured tensile stress; The damping coefficient to be identified; The velocity data for wire deformation caused by the displacement of the active tension swing arm; The equivalent elastic modulus to be identified; This represents the absolute deformation data caused by the displacement of the active tension pendulum.

[0098] The parameter identification and initialization module 401 uses the least squares method to perform numerical fitting, calculates the equivalent elastic modulus and damping coefficient of the current wire, and then writes the obtained equivalent elastic modulus and damping coefficient into the register inside the electrical control terminal 4 as the physical reference parameters for the closed-loop control of the tension of the current batch of wire.

[0099] S2. After the winding equipment enters the operating state, the main drive and the main spindle motor of the winding mechanism 1 drive the winding tube to rotate and increase speed according to the set process acceleration curve. During this period, the photoelectric encoder synchronously outputs the physical phase angle pulse signal that represents the mechanical rotation angle of the main spindle. The spatiotemporal joint vision filtering module 402 receives the physical phase angle pulse signal and triggers the monochrome line array industrial camera inside the contour vision acquisition component 3 to perform synchronous exposure when the mechanical phase angle reaches the preset angle point. The monochrome line array industrial camera outputs one-dimensional image data containing the edge pixels of the winding end face.

[0100] To extract true shape features, the spatiotemporal joint visual filtering module 402 uses a one-dimensional difference operator to identify the coordinates of edge feature points in the one-dimensional image data, combines the continuous edge feature point sequences collected within a single period into an original slope change rate matrix, and performs an exponentially weighted moving average operation on the slope data within continuous rotation periods. The operation logic satisfies the formula: ; In the formula, This is the smoothed estimate; It is a smoothing factor; This is the measurement value for the current period; This is a smoothed estimate from the previous period.

[0101] During the calculation process, the spatiotemporal joint vision filtering module 402 adaptively lowers the smoothing factor based on the current speed of the spindle motor. Generally, the higher the speed, the smaller the smoothing factor value, so as to filter out the short-term occlusion interference caused by flying fluff in the workshop and obtain the smooth true slope change rate. When the spatiotemporal joint vision filtering module 402 determines that the true slope change rate exceeds the set end face jump safety threshold, it generates the corresponding position deviation compensation vector.

[0102] The safety threshold is generally preset based on the wire diameter and the maximum end face protrusion allowed by the process. The spatiotemporal joint vision filtering module 402 converts the position deviation compensation vector into a wire routing commutation correction pulse and sends the wire routing commutation correction pulse to the wire routing servo motor of the main drive and wire routing mechanism 1. The wire routing servo motor fine-tunes the commutation pole position of the wire routing guide according to the wire routing commutation correction pulse to correct the end face shape.

[0103] S3. When the spindle motor is running at a constant linear speed, the dynamic disturbance anti-overlap control module 403 reads the spindle motor speed data fed back by the photoelectric encoder and the reciprocating frequency of the lateral reciprocating motion of the wire servo motor in real time. The dynamic disturbance anti-overlap control module 403 calculates the ratio of the spindle motor speed data to the reciprocating frequency to obtain the transient winding ratio. The dynamic disturbance anti-overlap control module 403 compares the transient winding ratio with the integer or fractional resonance range set in the system. This resonance range is usually defined as a tolerance percentage set by the integer or specific fractional fluctuation. When it is determined that the transient winding ratio falls into the resonance range, it indicates that there is a risk of overlapping wires in the wire arrangement trajectory when the transient winding ratio falls into the resonance range.

[0104] To address this operating condition, the dynamic perturbation anti-overlap control module 403 triggers the field-programmable gate array (FPGA) within the electrical control terminal 4. The waveform generation logic within the FPGA instantly generates a high-frequency, low-amplitude digital triangular wave perturbation signal. The dynamic perturbation anti-overlap control module 403 superimposes the digital triangular wave perturbation signal onto the basic position control command of the cable servo motor, driving the cable guide to superimpose tiny high-frequency jitters on the cable routing path. This creates a resonance condition for the overlapping cables by disrupting the cable trajectory misalignment in physical space.

[0105] S4. While the cable guide is performing a perturbation action, the length of the cable on the spatial geometric path will change instantaneously, thereby causing a sudden change in tension. The reverse phase feedforward tension compensation module 404, based on the same clock beat as the dynamic perturbation anti-overlap control module 403, synchronously extracts the instantaneous offset amplitude and current phase data of the digital triangular wave perturbation signal.

[0106] The reverse-phase feedforward tension compensation module 404 calls the pre-stored equivalent elastic modulus in the register and substitutes the pre-stored equivalent elastic modulus into the spatial line geometry equation to calculate the additional path change caused by the cable perturbation inside the cable and the corresponding expected offset tension. The reverse-phase feedforward tension compensation module 404, combined with the electromagnetic thrust coefficient of the voice coil motor of the active tension swing arm, converts the expected offset tension into a feedforward thrust current command. Considering the lag in the hardware communication of the control system, the reverse-phase feedforward tension compensation module 404 extracts the periodic delay time of the communication bus and adds an advance phase compensation amount to the feedforward thrust current command.

[0107] The reverse-phase feedforward tension compensation module 404 sends a feedforward thrust current command with an advanced phase compensation to the voice coil motor of the active tension swing arm mechanism 2. Before sending the feedforward thrust current command, the reverse-phase feedforward tension compensation module 404 uses a smooth saturation function, such as a piecewise linear limiting function, to limit and intercept the command amplitude to prevent the active tension swing arm from exceeding its maximum effective stroke. The voice coil motor receives the command and drives the active tension swing arm to generate a mechanical displacement opposite to the direction of the cable perturbation, actively counteracting the cable tension fluctuation through reverse geometric spatial movement.

[0108] S5. Throughout the entire winding cycle, the dual-scale state space observation module 405 continuously collects real-time stress feedback from the closed-loop tension sensor, actual torque current of the spindle motor, and displacement and velocity data from the active tension swing arm mechanism 2. The dual-scale state space observation module 405 calls a low-pass filter to separate the signal into a low-frequency component representing the increase in winding radius and a fast-changing high-frequency component representing instantaneous force fluctuations. The dual-scale state space observation module 405 inputs the fast-changing high-frequency component into the discrete state space equation to predict the current internal system state.

[0109] Meanwhile, the dual-scale state-space observation module 405 calculates the macroscopic tension residual based on actual stress feedback, and uses the recursive least squares algorithm to correct the equivalent elastic modulus and damping coefficient matrix in the discrete state-space equation in real time, so as to adapt to the physical property drift caused by tension accumulation during the winding process of the wire. The dual-scale state-space observation module 405 transforms the estimated state into a feedforward compensation reference through the feedforward gain matrix, and superimposes the feedforward compensation reference into the torque output reference of the main control circuit of the electrical control terminal 4.

[0110] The dual-scale state space observation module 405 continuously monitors the data output status of the closed-loop tension sensor. When the tensile stress data falls below the lower limit of the wire breakage protection within the set time window and the spindle speed fed back by the photoelectric encoder is still higher than the stop speed, the electrical control terminal 4 determines that a wire breakage fault has occurred. The main control circuit blocking status of the electrical control terminal 4 is updated and calculated, and an electrical braking command is sent to the spindle motor and the cable servo motor. If the wire breakage emergency stop condition is not triggered, when the meter counter in the electrical control terminal 4 reaches the full roll target length, the electrical control terminal 4 stops outputting the position deviation compensation vector and digital triangular wave perturbation signal.

[0111] The main control circuit reduces the operating speed of the spindle motor and the cable servo motor in coordination according to the deceleration curve. The voice coil motor of the active tension swing arm mechanism 2 synchronously reduces the foundation preload torque until the equipment stops rotating smoothly. The configuration of the underlying main frequency clock of the electrical control terminal and the register mapping of the servo motor hardware interface are implemented according to the existing chip hardware manual. The clock synchronization of the underlying main frequency clock of the electrical control terminal and the communication handshake mechanism of the servo motor hardware interface are implemented.

[0112] Specific application examples In this application example, polyester sewing thread with a specification of 40S / 2 is selected as the winding material, the rated speed of the main spindle motor of the winding equipment is set to 4000 rpm, and the target running tension is set to 35.00 centinewtons.

[0113] After the equipment is powered on and the initial threading is completed, the spindle motor remains stationary. The parameter identification and initialization module 401 in the electrical control terminal 4 outputs a sweep current command with a frequency range of 10 Hz to 50 Hz to the voice coil motor of the active tension swing arm mechanism 2. The active tension swing arm generates a slight reciprocating displacement. The closed-loop tension sensor synchronously collects tensile stress data at the corresponding frequency. The parameter identification and initialization module 401 uses the least squares method to fit the discrete viscoelastic dynamic equation, extracts the equivalent elastic modulus and damping coefficient of the polyester sewing thread of the test batch, and stores the data in the register.

[0114] The main drive and wiring mechanism 1 starts to operate. The monochrome linear array industrial camera in the contour vision acquisition component 3 reads the physical phase angle pulse signal of the photoelectric encoder and performs synchronous exposure under a fixed mechanical phase angle. In response to the image noise caused by free-floating fluff in the production environment, the spatiotemporal joint vision filtering module 402 performs an exponential weighted moving average operation on the extracted original slope change rate matrix to extract the real roll end face runout data and generate a position deviation compensation vector output to the wiring servo motor to adjust the commutation point.

[0115] As the wire continues to wind and the coil diameter increases, the dynamic micro-disturbance anti-overlap control module 403 calculates the transient winding ratio in real time. When the transient winding ratio value enters the resonance range calibrated near the integer fraction, the dynamic micro-disturbance anti-overlap control module 403 superimposes a digital triangular wave micro-disturbance signal into the basic position control command, changing the lateral reciprocating motion trajectory of the wire servo motor. Within the synchronous clock cycle, the reverse phase feedforward tension compensation module 404 extracts the amplitude of the digital triangular wave micro-disturbance signal, calls the equivalent elastic modulus to calculate the expected offset tension, and outputs a reverse phase feedforward displacement command to the voice coil motor of the active tension swing arm mechanism 2. The active tension swing arm performs a reverse compensation action to offset the tension path change caused by the wire disturbance.

[0116] During the winding process, the dual-scale state space observation module 405 uses a first-order low-pass filter with a 5 Hz cutoff frequency to divide the stress data output by the closed-loop tension sensor into low-frequency slowly changing components and fast-changing high-frequency components. The dual-scale state space observation module 405 performs state observation based on the fast-changing high-frequency components, synchronously calculates the macroscopic tension residual predicted by the actual measurement vector and the discrete state space equation, uses the recursive least squares algorithm to update the equivalent elastic modulus and damping coefficient in the state transition matrix in real time, and inputs the corrected feedforward compensation reference into the main control loop to correct the torque until the current roll reaches the preset number of winding layers and the winding is completed.

[0117] This test was conducted on the same mechanical hardware platform. Test group A was set as a basic solution using conventional proportional-integral-derivative control and single vision detection, while test group B was set as a system using the control system of this invention. Three different batches of polyester sewing thread were selected for each group to conduct full-spindle operation tests, and the sensor monitoring data during the operation were recorded.

[0118] Table 1: Comparison Test Results of Operating Parameters of Winding Equipment According to Table 1 and Figure 9 , Figure 10 , Figure 11 and Figure 12 Based on the data and the technical solution of this invention, test group A and test group B exhibit specific differences in multiple operating parameters. The dynamic tension deviation statistics of test group A are distributed between 7.91 centine Newtons (cN) and 8.65 cN, while the dynamic tension deviation statistics of test group B remain between 1.44 cN and 1.68 cN. According to the difference analysis of the dynamic tension deviation statistics, the dual-scale state space observation module 405 extracts the system state through frequency separation logic and uses the recursive least squares algorithm to update the equivalent elastic modulus and damping coefficient in the viscoelastic dynamic equation in real time, so that the control model parameters can adapt to the changes in the mechanical properties of polyester sewing thread during the winding process. The main control circuit in the electrical control terminal 4 corrects the torque according to the feedforward compensation reference to suppress the basic tension fluctuation under high-speed operation.

[0119] During the operation phase of the anti-overlapping disturbance program triggered by the electrical control terminal 4, test group A experienced a tension peak of up to 51.33 centinenewtons, deviating from the set basic tension value of 35.00 centinenewtons. Test group B's tension peak during the operation phase of the anti-overlapping disturbance program was constrained to within 36.88 centinenewtons. Since the sudden change in the cable trajectory introduced by the anti-overlapping operation will generate additional spatial path length stretching, this invention adopts a reverse-phase feedforward tension compensation module 404. The reverse-phase feedforward tension compensation module 404 constructs a spatial distance vector mapping based on the basic cable passing geometry model, and directly calculates the expected offset tension at the physical space level using the equivalent elastic modulus extracted in the early stage. The reverse-phase feedforward tension compensation module 404 drives the active tension swing arm mechanism 2 to output a reverse compensation displacement with an advanced phase, thereby offsetting the transient stress change caused by the disturbance operation at the physical and mechanical level.

[0120] Regarding the forming quality and interference response characteristics of the roll end face, test group A recorded 11 to 16 visual intervention misjudgments per full-axis cycle, accompanied by end face runout of 1.94 mm to 2.38 mm. Test group B had zero visual intervention misjudgments, and the end face runout data decreased to 0.38 mm to 0.52 mm. The comparison results of the number of visual intervention misjudgments and the end face runout data correspond to the underlying binding mechanism of the spatiotemporal joint visual filtering module 402. The spatiotemporal joint visual filtering module 402 uses the physical phase angle pulse signal generated by the photoelectric encoder to control the exposure of the linear array industrial camera, locks the mechanical phase angle profile of the roll fixed, and uses exponential weighted moving average calculation to process the edge pixel coordinates of continuous rotation cycles.

[0121] The exponentially weighted moving average operation attenuates the high-frequency step components generated in the original slope change rate matrix by unstructured noise such as random flying fluff. This limits the amplitude of the single pulse interference signal to below the set threshold for triggering visual anomaly intervention, thereby ensuring that the number of visual intervention misjudgments in test group B is 0. This guarantees that the spatiotemporal joint visual filtering module 402 generates a position deviation compensation vector based on the real end face shape, avoiding unexpected displacement caused by the servo motor receiving incorrect correction commands.

[0122] Based on the operation of the above system, the roll forming unwinding rate of test group B was controlled below 0.23%, which confirms that stable winding tension and orderly winding trajectory directly ensure the adhesion stability of flexible textile yarns at the edge of the roll.

[0123] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A high-speed winding device, characterized in that, include: The main drive and wiring mechanism (1), the active tension swing arm mechanism (2), the contour vision acquisition component (3), and the electrical control terminal (4); The main drive and wire winding mechanism (1) is used to drive the wire to wind and shape and control the spatial arrangement trajectory of the wire; The active tension swing arm mechanism (2) is located on the wiring path below the main drive and wiring mechanism (1). It is equipped with a voice coil motor, an active tension swing arm and a closed-loop tension sensor. The voice coil motor is used to drive the active tension swing arm to generate axial displacement, and the closed-loop tension sensor is used to measure the tensile stress data of the wire. The contour vision acquisition component (3) is fixed on the frame panel behind the main drive and wiring mechanism (1). The contour vision acquisition component (3) is equipped with a monochrome line array industrial camera. The lens of the monochrome line array industrial camera is aimed at the outer contour of the coiled tube on the main drive and wiring mechanism (1) to acquire image data of the outer contour of the coiled tube. The electrical control terminal (4) is located on the back of the high-speed winding equipment and is communicatively connected to the main drive and wire laying mechanism (1), the active tension swing arm mechanism (2), and the contour vision acquisition component (3). The electrical control terminal (4) is equipped with a control system, which is configured as follows: The mechanical parameters of the wire are extracted through the parameter identification and initialization module (401); The image data is processed by the spatiotemporal joint visual filtering module (402) to generate a position deviation compensation vector and perform cable position calibration. When the dynamic perturbation anti-overlap control module (403) determines that the wiring overlaps, it generates a digital triangular wave perturbation signal and superimposes it into the position control command of the main drive and wiring mechanism (1); The active tension swing arm mechanism (2) is controlled by the inverse phase feedforward tension compensation module (404) to perform displacement compensation that is opposite in phase to the digital triangular wave perturbation signal; The system state is predicted online by the dual-scale state space observation module (405), the mechanical parameters are updated, and the output torque of the voice coil motor of the active tension rocker mechanism (2) is corrected by the feedforward compensation benchmark.

2. The high-speed winding device according to claim 1, characterized in that, The active tension swing arm mechanism (2) is internally configured with the voice coil motor, the active tension swing arm and the closed-loop tension sensor. The parameter identification and initialization module (401) outputs a sweep frequency current command to the voice coil motor, receives the tensile stress data collected in real time by the closed-loop tension sensor, and uses the reciprocating displacement generated by the active tension swing arm as the input excitation parameter. Based on the discretized viscoelastic dynamic equation, the equivalent elastic modulus and damping coefficient of the wire are calculated by the least squares method as the mechanical parameters.

3. The high-speed winding device according to claim 1, characterized in that, The main drive and wiring mechanism (1) is equipped with a spindle motor, a wiring servo motor and a photoelectric encoder. The spatiotemporal joint visual filtering module (402) performs an exponential weighted moving average operation on the image data under the same mechanical phase angle within a continuous rotation cycle to extract the true slope change rate. When it is determined that the true slope change rate exceeds the preset safety threshold, the position deviation compensation vector is generated and output to the wiring servo motor for closed-loop correction of the winding shape. When performing the exponential weighted moving average operation, a negative correlation linear mapping function between the smoothing factor and the rotation speed of the spindle motor is established to adaptively adjust the smoothing factor.

4. The high-speed winding device according to claim 1, characterized in that, The main drive and wiring mechanism (1) is equipped with a spindle motor, a wiring servo motor and a photoelectric encoder. The dynamic micro-disturbance anti-overlap control module (403) obtains the speed data output by the photoelectric encoder and the reciprocating frequency of the wiring servo motor to calculate the transient winding ratio. When it is determined that the transient winding ratio falls into the set resonance range, a digital triangular wave micro-disturbance signal with an amplitude constrained to within one percent to three percent of the amplitude of the basic position control command of the main drive and wiring mechanism (1) is generated. The digital triangular wave micro-disturbance signal is superimposed on the basic position control command of the main drive and wiring mechanism (1) to drive the wiring servo motor to perform a lateral reciprocating motion with slight spatial jitter.

5. The high-speed winding device according to claim 4, characterized in that, The reverse-phase feedforward tension compensation module (404) extracts the instantaneous amplitude of the digital triangular wave perturbation signal, and calculates the expected offset tension by combining the set geometric coupling coefficient with the equivalent elastic modulus in the mechanical parameters. The reverse-phase feedforward tension compensation module (404) converts the expected offset tension into a feedforward thrust current command and outputs it to the voice coil motor. The motion phase of the feedforward thrust current command is opposite to the phase of the digital triangular wave perturbation signal.

6. The high-speed winding device according to claim 5, characterized in that, When generating the feedforward thrust current command to the active tension swing arm mechanism (2), the reverse-phase feedforward tension compensation module (404) adds an advanced phase compensation amount based on the bus communication cycle to the feedforward thrust current command, and has a built-in feedforward limiting protection mechanism. It uses a preset smooth saturation function to perform transition processing on the part of the expected displacement that exceeds 80% of the safety threshold of the maximum effective stroke of the voice coil motor.

7. The high-speed winding device according to claim 2, characterized in that, The dual-scale state-space observation module (405) acquires the real-time stress feedback output by the closed-loop tension sensor, separates the low-frequency slowly changing component by a first-order low-pass filter with a cutoff frequency set to five Hz, and obtains the fast-changing high-frequency component by subtracting the low-frequency slowly changing component from the original real-time stress feedback. Based on the fast-changing high-frequency component and the mechanical parameters, a discrete state-space equation containing a discrete state transition matrix is ​​constructed to output a corrected estimated state vector.

8. The high-speed winding device according to claim 7, characterized in that, The dual-scale state space observation module (405) calculates the macroscopic tension residual between the actual measurement vector and the macroscopic tension residual predicted by the discrete state space equation. With the goal of minimizing the macroscopic tension residual, it uses the recursive least squares algorithm to update the equivalent elastic modulus and damping coefficient in the discrete state transition matrix in real time. It then combines the corrected estimated state vector with the system feedforward gain matrix for linear weighted transformation and outputs the feedforward compensation benchmark.

9. The high-speed winding device according to claim 7, characterized in that, The dual-scale state space observation module (405) calculates the observation information covariance. When it is determined that the fault tolerance threshold is exceeded, the abnormal prediction and maintenance mechanism is triggered, and the state observer gain matrix is ​​forced to be set to zero to perform open-loop recursion. When it is detected that the tensile stress data is continuously lower than the preset wire breakage protection threshold and approaches zero within the set time window, and the speed data fed back by the photoelectric encoder configured in the main drive and wiring mechanism (1) is higher than the set operating speed threshold, it is determined that a wire breakage fault has occurred and then the highest priority hardware interrupt is triggered to send an emergency braking command.

10. A high-speed winding method, characterized in that, Using the high-speed winding apparatus according to any one of claims 1 to 9 includes the following steps: S1. When the static threading of the wire is completed and the main spindle motor is in a stationary state, the parameter identification and initialization module (401) outputs a sweep frequency current command to the voice coil motor of the active tension swing arm mechanism (2) to drive the active tension swing arm to reciprocate, collect tensile stress data and actual displacement data, and extract mechanical parameters. S2. After the winding equipment enters the running state, the main drive and the main shaft motor of the winding mechanism (1) drive the winding tube to rotate and speed up according to the set process acceleration curve. During this period, the spatiotemporal joint vision filtering module (402) triggers the monochrome line array industrial camera inside the contour vision acquisition component (3) to perform synchronous exposure, perform exponential weighted moving average calculation on the image data and adaptively lower the smoothing factor. When the safety threshold is exceeded, the output position deviation compensation vector fine-tunes the reversing pole position of the winding guide. S3. During the constant linear speed operation phase, the dynamic micro-disturbance anti-overlap control module (403) calculates the transient winding ratio based on the rotation speed and reciprocating frequency. When the transient winding ratio falls into the resonance range, it generates a digital triangular wave micro-disturbance signal and superimposes it onto the basic position control command of the wiring servo motor to drive the wiring guide to superimpose a small high-frequency jitter on the wiring path. S4, the reverse phase feedforward tension compensation module (404) synchronously extracts the instantaneous offset amplitude and current phase data of the digital triangular wave micro-disturbance signal and calls the mechanical parameters, calculates the expected offset tension through the spatial linear geometric equation, and outputs a feedforward thrust current command with advanced phase compensation to the active tension swing arm mechanism (2) to actively counteract the tension fluctuation of the wire. S5, the dual-scale state space observation module (405) separates the collected real-time stress feedback, the actual torque current of the main shaft motor and the displacement and velocity data of the active tension swing arm mechanism (2) into low-frequency slowly changing components and fast-changing high-frequency components through a low-pass filter. Combined with the constructed discrete state space equation and the recursive least squares algorithm, the mechanical parameters are updated in real time, and the feedforward compensation benchmark is output to correct the output torque of the voice coil motor of the active tension swing arm mechanism (2) in real time.

Images