Method for accurately testing and controlling tension in rotating state

Abstract

The invention discloses a method for accurately testing and controlling tension in a rotating state, and belongs to the technical field of tension control. The method comprises the following steps: acquiring a tension value, a pulley inclination angle and a vibration acceleration of each strand of wire rod in a plurality of strands of wire rods; the virtual instrument architecture platform executes real-time operation; the virtual instrument architecture platform writes a frequency instruction into the frequency converter and sends a target position pulse sequence to the servo motor at the same time; the frequency converter adjusts the rotating speed of the winch driving motor according to the frequency instruction. The servo motor controls the linear speed of the traction roller according to the target position pulse sequence. The control precision is optimized through multi-mode sensing data fusion and a dynamic compensation mechanism, a dynamic weight distribution PID control strategy is combined, the weight coefficients of proportion, integral and differential items are adjusted in a self-adaptive mode according to the tension change rate, a feed-forward compensation mechanism is introduced, collaborative optimization control over the frequency converter and the servo motor is achieved, and the control precision is improved. And the tension control precision and the abnormal response efficiency under the complex working condition are obviously improved.

Description

Technical Field

[0001] The present invention relates to a wire tension testing and control method, in particular to a method for accurately testing and controlling tension in a rotating state, belonging to the technical field of tension control. Background Art

[0002] Real-time tension control in multi-strand wire processing equipment is a critical process in wire manufacturing, and its stability directly impacts the quality consistency of the stranded wire. Traditional control solutions currently face several challenges in practical application. Indirect control methods based on magnetic powder brakes infer tension from resistance torque, but are susceptible to environmental factors under complex operating conditions, leaving room for improvement in control accuracy. Direct detection solutions, while acquiring tension data through contact sensors, are limited by the mechanical structure of the capstan. The sensor's mounting position and force direction may deviate slightly, and vibration interference under high-speed rotation can also affect signal quality. Existing systems often employ a decentralized control architecture, leaving room for improvement in the efficiency of coordinated regulation between sensor signals and actuators. Periodic sensor parameter calibration is required during operation, making the process relatively cumbersome. Furthermore, real-time compensation mechanisms for pulley inclination changes and mechanical vibration interference are still incomplete, requiring further improvement in safety response efficiency under unexpected operating conditions. These factors pose a challenge to the efficient and stable operation of multi-strand wire processing equipment, necessitating a control solution that integrates multi-source data fusion processing and dynamic compensation capabilities. Summary of the Invention

[0003] Purpose of the invention: In view of the above problems, the purpose of the present invention is to provide a method for accurately testing and controlling tension in a rotating state.

[0004] Technical solution: A method for accurately testing and controlling tension in a rotating state of the present invention comprises the following steps:

[0005] Synchronously collect the tension value, pulley inclination angle and vibration acceleration of each strand of wire in multiple strands to obtain composite sensing data;

[0006] Upload the composite sensor data to the virtual instrument architecture platform via wireless transmission;

[0007] The virtual instrument architecture platform performs real-time calculations;

[0008] Based on the calculation results, the virtual instrument architecture platform writes frequency instructions to the inverter and sends the target position pulse sequence to the servo motor;

[0009] The frequency converter adjusts the speed of the capstan drive motor according to the frequency command, and the servo motor controls the linear speed of the traction roller according to the target position pulse sequence, so that the actual tension of the single strand of tested wire converges to the set range;

[0010] The virtual instrument architecture platform continuously monitors the tension status of the tested wire and triggers the protection mechanism if the trigger conditions are met.

[0011] Furthermore, the steps of performing real-time calculations on the virtual instrument architecture platform include:

[0012] Tension F of a single wire based on the pulley inclination raw To compensate, the formula is:

[0013] F true =F raw -Mg sinθ-Mω 2 r p ,

[0014] Where, F true represents the tension value after compensation, M represents the weight of the pulley, θ represents the pulley inclination, ω represents the winch speed, r p Indicates the distance from the center of the pulley to the center of the main spindle of the wire processing equipment.

[0015] Furthermore, the steps of executing real-time calculations on the virtual instrument architecture platform further include:

[0016] The vibration suppression coefficient is calculated based on the vibration acceleration. The formula is:

[0017]

[0018] Where a x 、a y 、a z is the triaxial vibration acceleration, a max is the vibration threshold;

[0019] Combined with the tension compensation value, the tension value of the single-strand wire based on vibration suppression is calculated using the formula:

[0020] F output =F true ·K v .

[0021] Furthermore, the steps of executing real-time calculations on the virtual instrument architecture platform further include:

[0022] Based on the tension value F after vibration suppression output The predicted value F of the capstan-traction roller dynamic model model , calculate the final tension estimate F through the Kalman filter algorithm final ,include:

[0023] Step 31, define the state variable x k , expressed as:

[0024]

[0025] in, is the rate of change of tension;

[0026] Step 32, define the observation variable z k , expressed as:

[0027]

[0028] Among them, the predicted value F model It is calculated from the capstan speed ω and the traction roller linear speed v, and the formula is:

[0029]

[0030] Where r is the winch radius, K 系统 is the system stiffness coefficient;

[0031] Step 33, performing Kalman filter operation, including a prediction phase and an update phase;

[0032] In the prediction stage, the tensor value at the next moment is predicted using the formula:

[0033]

[0034] Where A is the state transfer matrix, Δt is the control period;

[0035] In the update phase, the Kalman gain and tensor value are updated, and the formulas are:

[0036]

[0037] Where H is the observation matrix, The observation matrix H maps the state variables to the observation space; represents the k-time prior state covariance matrix, T represents transpose, Q represents the process noise covariance matrix in Kalman filtering, and R represents the noise covariance matrix, which is expressed as:

[0038]

[0039] Among them, σ F is the standard deviation of the tension sensor measurement error, σ model is the standard deviation of the model prediction error; E v is vibration energy, E th is the vibration energy threshold, when the vibration energy E v >E th When , increase the model prediction noise weight;

[0040] Step 34, since the two observations Foutput and F model All reflect the tension value. The matrix is ​​designed to take only the tension part of the state variable, that is, the first column of the matrix, and ignore the rate of change, that is, set the second column to zero, to obtain the final tension value.

[0041] Furthermore, based on the calculation result, the virtual instrument architecture platform writes a frequency instruction to the frequency converter and simultaneously sends a target position pulse sequence to the servo motor, including the following steps:

[0042] Step 41: Set the final tension value Input to the PID controller and compare it with the single wire tension threshold F set Perform real-time comparison and generate real-time tension deviation signal Transmitting real-time deviation signals to the virtual instrument architecture platform;

[0043] Step 42 , the total control amount of the virtual instrument architecture platform includes the compensation amount generated in the feedforward channel and the coordination adjustment amount generated in the feedback channel;

[0044] Among them, in the feedforward channel, the tension change rate predicted by Kalman filter is Generate compensation, the formula is:

[0045]

[0046] Where K ff is the feedforward coefficient, obtained by system inertia calibration;

[0047] In the feedback channel, the coordinated adjustment amount Δn is calculated based on the real-time tension deviation signal e(t) through the proportional-integral-differential control algorithm. The formula is:

[0048]

[0049] Where Δn is the coordinated adjustment of the capstan speed and the traction roller speed, K p is the proportional coefficient, Represents the proportional term; K i is the integration coefficient, represents the integral term; K d is the differential coefficient, represents the differential term;

[0050] Among them, when Exceeding the threshold When , the proportional coefficient, integral coefficient and differential coefficient are dynamically updated, and the update formulas are:

[0051]

[0052] Where K′p , K′ i and K′ d Represent the updated proportional coefficient, integral coefficient and differential coefficient respectively, F max is the maximum tension change rate threshold;

[0053] The collaborative adjustment amount Δn is calculated based on the updated coefficients. The formula is:

[0054]

[0055] The total control quantity is:

[0056]

[0057] Step 43: Allocate the total control quantity u(t) to the control instructions of the inverter and the servo motor according to the frequency domain characteristics, wherein the instructions received by the inverter are expressed as:

[0058] f inv =α(t)·u(t),

[0059] In the formula, α(t) represents the weight coefficient,

[0060] The command received by the servo motor is expressed as:

[0061] N pulse =(1-α(t))·u(t).

[0062] Furthermore, the steps of triggering the protection mechanism include:

[0063] When the final tension estimate output by the Kalman filter is Exceeding the preset safety range [F set -ΔF th1 ,F set +ΔF th1 ], the first level alarm is triggered, that is, the sound and light alarm is triggered and the abnormal log is recorded; Among them, ΔF th1 is the first-level tension deviation threshold, representing the allowable tension offset, ΔF th1 =k·F set , k is the process accuracy coefficient;

[0064] If the residual is detected at the same time It is judged that the sensor is faulty, and the first level alarm is skipped and the second level alarm is entered directly. If the abnormality lasts longer than 5 seconds, the dynamic speed reduction strategy is implemented;

[0065] like Less than the emergency stop tension threshold F e , and the capstan speed ω and the traction roller linear speed v satisfy η th is the speed mismatch rate threshold, triggering the third-level alarm, that is, cutting off the power supply and locking the equipment; where v nominal To calibrate the linear speed of the traction roller.

[0066] Furthermore, the dynamic speed reduction strategy includes:

[0067] Adjust the speed of the winch drive motor and reduce the speed to β(t) times the reference value according to the gradient, where β(t) = 50% + 30% e -0.5t ;

[0068] Adjust the linear speed of the traction roller servo motor, and adjust the linear speed synchronously to v′=v·β(t), maintaining the pay-off or take-up speed difference Δv=|ωr-v| constant;

[0069] Adjust PID parameters and proportional coefficient K' s =0.5K′ p , the integration time is doubled T′ i =2T i .

[0070] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0071] 1. The present invention uses multiple sensors to collect tension, pulley inclination and vibration data in real time. After fusion processing on the virtual instrument architecture platform, it dynamically compensates for measurement errors caused by gravity deviation and mechanical vibration, significantly improving the tension control accuracy and stability.

[0072] 2. This invention optimizes control accuracy through multimodal sensor data fusion and dynamic compensation mechanisms. Based on a multi-source data fusion algorithm based on Kalman filtering, the real-time measurement data from tension sensors, tilt sensors, and vibration sensors are dynamically integrated with the predicted values ​​of the capstan-traction roller dynamics model to generate an optimal tension estimate with enhanced anti-interference capabilities.

[0073] 3. This invention combines a dynamic weight distribution PID control strategy to adaptively adjust the weight coefficients of the proportional, integral, and differential terms according to the rate of change of tension, and introduces a feedforward compensation mechanism to achieve coordinated optimization control of the inverter and servo motor. The inverter responds to high-frequency disturbances to quickly adjust the capstan speed, and the servo motor accurately controls the linear speed of the traction roller through a pulse sequence to eliminate steady-state errors.

[0074] 4. The present invention triggers hierarchical safety protection through a multi-source verification mechanism based on Kalman filter output. It detects sensor failures through residual analysis and switches to model predictive control mode. Dynamic gradient deceleration maintains a constant payout / takeup speed difference. Furthermore, an emergency shutdown mechanism with dual confirmation of multiple signals, including tension estimation, capstan speed, and traction line speed, significantly improves tension control accuracy and abnormal response efficiency under complex working conditions. Wireless transmission and automatic calibration reduce maintenance complexity, making the system suitable for high-precision wire processing.

[0075] 5. The present invention automatically retrieves parameters and completes closed-loop control, greatly simplifying the operating process and reducing dependence on professional skills. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 is a flow chart of the present invention;

[0077] Figure 2 This is a control framework diagram of the present invention;

[0078] Figure 3 It is a flowchart of data processing and control algorithm;

[0079] Figure 4 This is the pulley-wire mechanical analysis diagram;

[0080] Figure 5 This is a flowchart of the Kalman filter dynamic fusion process;

[0081] Figure 6 Generate and distribute flow charts for control instructions;

[0082] Figure 7 This is the flow chart of the dynamic speed reduction strategy. DETAILED DESCRIPTION

[0083] In order to make the purpose, technical solutions and advantages of this application more clear, this application is further described in detail below with reference to the accompanying drawings and embodiments.

[0084] Combine Figures 1 to 3 The present embodiment provides a method for accurately testing and controlling tension in a rotating state, the method comprising the following steps:

[0085] Step 1: Synchronously collect the tension value, pulley inclination angle, and vibration acceleration of each of the multiple wires to obtain composite sensing data.

[0086] In one example, a multi-strand wire processing device can be pre-installed with six cantilever beam single-pulley tension sensors, six angle sensors, and one vibration sensor. The tension sensors are mounted on a rotating capstan support via a fixed bracket to measure the real-time tension of each strand of the multi-strand wire. The angle sensors are embedded in the tension sensor mounting bracket to measure the pulley inclination angle θ, compensating for measurement errors caused by gravity. The vibration sensor is fixed to the capstan bearing to monitor high-frequency vibration interference and generate a suppression signal. The sensor measurements are combined into composite sensor data consisting of tension value, pulley inclination angle, and vibration acceleration.

[0087] Step 2: Upload the composite sensor data to the virtual instrument architecture platform via wireless transmission.

[0088] A LoRa wireless transmitter is integrated on the sensor side to encapsulate the composite sensor data into a data packet format: [ID][tension][angle][vibration]. A LoRa gateway receiver is integrated on the virtual instrument architecture platform side to upload the tension value, pulley angle, and vibration data of each wire to the virtual instrument architecture platform via wireless transmission. The virtual instrument architecture platform can use LabVIEW.

[0089] Step 3: The virtual instrument architecture platform performs real-time calculations.

[0090] The virtual instrument architecture control platform is used to realize functions such as data fusion processing, error compensation, dynamic control and alarm.

[0091] Furthermore, the steps of performing real-time calculations on the virtual instrument architecture platform include:

[0092] Combine Figure 4 As shown, the tension F of the single wire is affected by the pulley inclination angle. raw To compensate, the formula is:

[0093] F true =F raw -Mg sinθ-Mω 2 r p ,

[0094] Where, F true represents the tension value after compensation, M represents the weight of the pulley, θ represents the pulley inclination, ω represents the winch speed, r p Indicates the distance from the center of the pulley to the center of the main spindle of the wire processing equipment.

[0095] Furthermore, the steps of executing real-time calculations on the virtual instrument architecture platform further include:

[0096] The vibration suppression coefficient is calculated based on the vibration acceleration. The formula is:

[0097]

[0098] Where a x 、a y 、a z is the triaxial vibration acceleration, a max is the vibration threshold;

[0099] Combined with the tension compensation value, the tension value of the single-strand wire based on vibration suppression is calculated using the formula:

[0100] F output =F true ·K v .

[0101] Combine Figure 5 Furthermore, the step of executing real-time calculations on the virtual instrument architecture platform also includes:

[0102] Based on the tension value F after vibration suppression output The predicted value F of the capstan-traction roller dynamic model model , calculate the final tension estimate F through the Kalman filter algorithm final ,include:

[0103] Step 31, define the state variable x k , expressed as:

[0104]

[0105] in, is the rate of change of tension;

[0106] Step 32, define the observation variable z k , expressed as:

[0107]

[0108] Among them, the predicted value F model It is calculated from the capstan speed ω and the traction roller linear speed v, and the formula is:

[0109]

[0110] Where r is the winch radius; K 系统 is the system stiffness coefficient, which is the equivalent comprehensive stiffness coefficient of the wire tension transmission path. Its unit is N·s / m. It is obtained through step-speed test calibration and includes mechanical transmission stiffness, wire elasticity, and dynamic coupling effects.

[0111] Step 33, performing Kalman filter operation, including a prediction phase and an update phase;

[0112] In the prediction stage, the tensor value at the next moment is predicted using the formula:

[0113]

[0114] Where A is the state transfer matrix, Δt is the control period;

[0115] In the update phase, the Kalman gain and tensor value are updated, and the formulas are:

[0116]

[0117] Where H is the observation matrix, The observation matrix H maps the state variables to the observation space; represents the k-time prior state covariance matrix, T represents transpose, Q represents the process noise covariance matrix in Kalman filtering, and R represents the noise covariance matrix, which is expressed as:

[0118]

[0119] Among them, σ F is the standard deviation of the tension sensor measurement error, σ model is the standard deviation of the prediction error of the capstan-traction roller dynamic model; E v is vibration energy, E th is the vibration energy threshold, when the vibration energy E v >E th When , increase the model prediction noise weight;

[0120] Step 34, since the two observations F output and F model All reflect the tension value. The matrix is ​​designed to take only the tension part of the state variable, that is, the first column of the matrix, and ignore the rate of change, that is, set the second column to zero, to obtain the final tension value.

[0121] In Kalman filtering, weight adjustment is achieved by modifying the observation noise covariance matrix R. When E v >E th When E v =2E th , the model noise variance is expanded to 3 times of the original; if E v =3E th , the model noise variance is expanded to 5 times of the original. According to the Kalman gain calculation formula It can be seen that the model noise term As it increases, the lower right corner element of the noise covariance matrix R increases, and the inverse matrix The weight of the corresponding model is reduced, and the Kalman gain K k The gain value of the model channel is reduced. The sensor weight is Inversely proportional to Inversely proportional. When the vibration increases, As becomes larger, the model weight decreases, and the system becomes more dependent on sensor data. Then from the state equation: After weight adjustment, the system judges under strong vibration: sensor data F output Protected by vibration suppression algorithm (more reliable); model predicts F model Due to mechanical structure vibration, it is easy to be distorted (unreliable). Final output More adoption of F output Quantity.

[0122] Step 4: Based on the calculation results, the virtual instrument architecture platform writes the frequency instruction to the inverter and sends the target position pulse sequence to the servo motor.

[0123] Combine Figure 6 ,Furthermore, based on the calculation result, the virtual instrument architecture platform writes a frequency instruction to the frequency converter, and simultaneously sends a target position pulse sequence to the servo motor, the steps including:

[0124] Step 41: Set the final tension value Input to the PID controller and compare it with the single wire tension threshold F set Perform real-time comparison and generate real-time tension deviation signal Transmitting real-time deviation signals to the virtual instrument architecture platform;

[0125] Step 42 , the total control amount of the virtual instrument architecture platform includes the compensation amount generated in the feedforward channel and the coordination adjustment amount generated in the feedback channel;

[0126] Among them, in the feedforward channel, the tension change rate predicted by Kalman filter is Generate compensation, the formula is:

[0127]

[0128] Where K ff is the feedforward coefficient, obtained by system inertia calibration;

[0129] In the feedback channel, the coordinated adjustment amount Δn is calculated based on the real-time tension deviation signal e(t) through the proportional-integral-differential control algorithm. The formula is:

[0130]

[0131] Where Δn is the coordinated adjustment of the capstan speed and the traction roller speed, Kp is the proportional coefficient, Represents the proportional term; K i is the integration coefficient, represents the integral term; K d is the differential coefficient, represents the differential term;

[0132] Among them, when Exceeding the threshold When , the proportional coefficient, integral coefficient and differential coefficient are dynamically updated, and the update formulas are:

[0133]

[0134] Where K′ p , K′ i and K′ d Represent the updated proportional coefficient, integral coefficient and differential coefficient respectively, F max is the maximum tension change rate threshold;

[0135] The collaborative adjustment amount Δn is calculated based on the updated coefficients. The formula is:

[0136]

[0137] The total control quantity is:

[0138]

[0139] Step 43: Allocate the total control quantity u(t) to the control instructions of the inverter and the servo motor according to the frequency domain characteristics, wherein the instructions received by the inverter are expressed as:

[0140] f inv =α(t)·u(t),

[0141] f inv That is, the virtual instrument architecture platform writes frequency instructions to the frequency converter;

[0142] In the formula, α(t) represents the weight coefficient,

[0143] The changing rule of weight coefficient α(t) is:

[0144] When the tension deviation is large, the weight is adjusted to 1, and the inverter responds with full authority;

[0145] When the tension deviation is very small, the weight is adjusted to 0, and the servo is precisely adjusted;

[0146] When the tension deviation is in the middle state, the weight is adjusted to 0.5, and the inverter and servo motor are controlled in a coordinated manner;

[0147] The command received by the servo motor is expressed as:

[0148] N pulse =(1-α(t))·u(t).

[0149] N pulse That is the target position pulse sequence sent by the virtual instrument architecture platform to the servo motor.

[0150] In step 5, the frequency converter adjusts the speed of the capstan drive motor according to the frequency command, and the servo motor controls the linear speed of the traction roller according to the target position pulse sequence, so that the actual tension of the single strand of tested wire converges to the set range.

[0151] The virtual instrument architecture platform writes frequency commands to the inverter via the Modbus TCP protocol and simultaneously sends the target position pulse train to the servo motor via the EtherCAT bus. Based on the frequency commands, the inverter adjusts the capstan drive motor speed to a range of 0-50 Hz with a response time of less than 0.1 seconds. The servo motor controls the linear speed of the traction roller to a range of 0-100 kpps with a positioning accuracy of ±0.01 mm. The coordinated adjustment of the capstan speed and traction speed ensures that the actual pulling force converges within the set range.

[0152] Step 6: The virtual instrument architecture platform continuously monitors the tension state of the tested wire. If the trigger condition is met, the protection mechanism is triggered.

[0153] The protection mechanism is a hierarchical alarm system consisting of a first-level audible and visual warning, a second-level automatic speed reduction, and a third-level emergency stop mechanism. The audible and visual warnings are used to warn when tension exceeds the limit. The emergency stop function cuts off power and locks the machine when tension returns to zero.

[0154] Furthermore, the steps of triggering the protection mechanism include:

[0155] When the final tension estimate output by the Kalman filter is Exceeding the preset safety range [F set -ΔF th1 ,F set +ΔF th1 ], the first level alarm is triggered, that is, the sound and light alarm is triggered and the abnormal log is recorded; Among them, ΔF th1 is the first-level tension deviation threshold, representing the allowable tension offset, ΔF th1 =k·F set , where k is the process accuracy coefficient;

[0156] If the residual is detected at the same time It is judged that the sensor is faulty, and the first level alarm is skipped and the second level alarm is entered directly. If the abnormality lasts longer than 5 seconds, the dynamic speed reduction strategy is executed; at this time, it automatically switches to the predicted value Fmodel As PID input;

[0157] like Less than the emergency stop tension threshold F e , and the capstan speed ω and the traction roller linear speed v satisfy η th is the speed mismatch rate threshold, triggering the third-level alarm, that is, cutting off the power supply and locking the equipment; where v nominal To calibrate the linear speed of the traction roller.

[0158] In one example, the emergency stop tension threshold F e For example, set it to 1N.

[0159] Combine Figure 7 ,Furthermore, the dynamic speed reduction strategy includes:

[0160] Adjust the speed of the winch drive motor and reduce the speed to β(t) times the reference value according to the gradient, where β(t) = 50% + 30% e -0.5t , to avoid a sudden drop causing a secondary shock;

[0161] Adjust the linear speed of the traction roller servo motor, and adjust the linear speed synchronously to v′=v·β(t), maintaining the pay-off or take-up speed difference Δv=|ωr-v| constant;

[0162] Switch to conservative control mode during deceleration, adjust PID parameters, and adjust the proportional coefficient K' s =0.5K′ p , the integration time is doubled T′ i =2T i .

[0163] Through the dynamic optimization design of multi-source data fusion using Kalman filtering and PID control, closed-loop optimization from signal processing to execution control is achieved. The Kalman filter, through state expansion and noise adaptation, can provide high-precision tension estimation even under strong interference. The PID controller significantly improves response speed and stability through weight distribution, feedforward compensation, and coordination with the actuator. This solution is not only suitable for high-precision wire processing but can also be extended to tension control scenarios in other rotating machinery, demonstrating its broad technical versatility and industrial value.

[0164] The method described in this embodiment uses a tension sensor to measure the instantaneous tension value of a single strand of wire at the capstan, an angle sensor synchronously detects the pulley inclination angle θ to compensate for measurement deviations caused by the gravity component, and a vibration sensor captures high-frequency mechanical vibration interference from the capstan bearing; the sensor data is encapsulated into a standardized protocol frame via LoRa wireless communication and transmitted to the virtual instrument architecture control platform via a predetermined frequency band; the virtual instrument architecture platform performs multi-dimensional fusion processing on the received data—first, an angle compensation algorithm is used to eliminate the static error introduced by the inclination angle, then a Butterworth low-pass filter is called to suppress vibration noise above 50 Hz, and finally, a PID closed-loop operation is performed based on the preset target tension value to generate a collaborative control instruction; this instruction dynamically adjusts the inverter output frequency to change the capstan speed via the Modbus TCP protocol, and simultaneously sends a pulse sequence to the servo motor via the bus to accurately control the linear speed of the traction roller, so that the system tension quickly converges to the set range. During operation, the virtual instrument architecture platform can display the tension fluctuation curve and vibration spectrum in real time. When it detects that the tension exceeds the limit or returns to zero, it immediately activates the graded response mechanism - first triggering the sound and light alarm to alert the operator, and automatically reducing the speed if the abnormality persists. Finally, in an emergency, the independent DO circuit will cut off the power supply of the equipment, ensuring high precision and high reliability of the multi-strand wire processing process.

[0165] The method described in this paper optimizes measurement deviations caused by environmental interference and structural limitations in traditional tension control methods by integrating dynamic compensation mechanisms for tension, angle, and vibration data. A multi-threaded processing architecture based on a virtual instrument framework enables efficient conversion of sensor data into control instructions. Through coordinated adjustment of the frequency converter and servo motor, the method improves the consistent control of multi-strand tension. Furthermore, the system integrates hierarchical safety protection and maintenance modes, simplifying the manual operation process and reducing reliance on professional experience, providing a viable technical path for high-precision and high-stability operation of multi-strand wire processing equipment.

Claims

1. A method for accurately testing and controlling tension in a rotating state, characterized in that: The following steps are involved: Synchronously collect the tension value, pulley inclination angle and vibration acceleration of each strand of wire in multiple strands to obtain composite sensing data; Upload the composite sensor data to the virtual instrument architecture platform via wireless transmission; The virtual instrument architecture platform performs real-time calculations; Based on the calculation results, the virtual instrument architecture platform writes frequency instructions to the inverter and sends the target position pulse sequence to the servo motor; The frequency converter adjusts the speed of the capstan drive motor according to the frequency command, and the servo motor controls the linear speed of the traction roller according to the target position pulse sequence, so that the actual tension of the single strand of tested wire converges to the set range; The virtual instrument architecture platform continuously monitors the tension status of the tested wire and triggers the protection mechanism if the trigger conditions are met.

2. The method for accurately testing and controlling tension in a rotating state according to claim 1, characterized in that: The steps for the virtual instrument architecture platform to perform real-time operations include: Tension F of a single wire based on the pulley inclination raw To compensate, the formula is: F true =F raw -Mg sin θ-Mω 2 r p , Where, F true represents the tension value after compensation, M represents the weight of the pulley, θ represents the pulley inclination, ω represents the winch speed, r p Indicates the distance from the center of the pulley to the center of the main spindle of the wire processing equipment.

3. The method for accurately testing and controlling tension in a rotating state according to claim 2, characterized in that: The steps of executing real-time calculations on the virtual instrument architecture platform also include: The vibration suppression coefficient is calculated based on the vibration acceleration. The formula is: Where a x 、a y 、a z is the triaxial vibration acceleration, a max is the vibration threshold; Combined with the tension compensation value, the tension value of the single-strand wire based on vibration suppression is calculated using the formula: F output =F true ·K v 。 4. The method for accurately testing and controlling tension in a rotating state according to claim 3, characterized in that: The steps of executing real-time calculations on the virtual instrument architecture platform also include: Based on the tension value F after vibration suppression output The predicted value F of the capstan-traction roller dynamic model model , calculate the final tension estimate F through the Kalman filter algorithm final ,include: Step 31, define the state variable x k , expressed as: in, is the rate of change of tension; Step 32, define the observation variable z k , expressed as: Among them, the predicted value F model It is calculated from the capstan speed ω and the traction roller linear speed v, and the formula is: Where r is the winch radius, K 系统 is the system stiffness coefficient; Step 33, performing Kalman filter operation, including a prediction phase and an update phase; In the prediction stage, the tensor value at the next moment is predicted using the formula: Where A is the state transfer matrix, Δt is the control period; In the update phase, the Kalman gain and tensor value are updated, and the formulas are: Where H is the observation matrix, The observation matrix H maps the state variables to the observation space; represents the k-time prior state covariance matrix, T represents transpose, Q represents the process noise covariance matrix in Kalman filtering, and R represents the noise covariance matrix, which is expressed as: Among them, σ F is the standard deviation of the tension sensor measurement error, σ model is the standard deviation of the model prediction error; E v is vibration energy, E th is the vibration energy threshold, when the vibration energy E v >E th When , increase the model prediction noise weight; Step 34, since the two observations F output and F model All reflect the tension value. The matrix is ​​designed to take only the tension part of the state variable, that is, the first column of the matrix, and ignore the rate of change, that is, set the second column to zero, to obtain the final tension value.

5. The method for accurately testing and controlling tension in a rotating state according to claim 4, characterized in that: Based on the calculation results, the virtual instrument architecture platform writes a frequency instruction to the inverter and sends a target position pulse sequence to the servo motor. The steps include: Step 41: Set the final tension value Input to the PID controller and compare it with the single wire tension threshold F set Perform real-time comparison and generate real-time tension deviation signal Transmitting real-time deviation signals to the virtual instrument architecture platform; Step 42 , the total control amount of the virtual instrument architecture platform includes the compensation amount generated in the feedforward channel and the coordination adjustment amount generated in the feedback channel; Among them, in the feedforward channel, the tension change rate predicted by Kalman filter is Generate compensation, the formula is: Where K ff is the feedforward coefficient, obtained by system inertia calibration; In the feedback channel, the coordinated adjustment amount Δn is calculated based on the real-time tension deviation signal e(t) through the proportional-integral-differential control algorithm. The formula is: Where Δn is the coordinated adjustment of the capstan speed and the traction roller speed, K p is the proportional coefficient, Represents the proportional term; K i is the integration coefficient, represents the integral term; K d is the differential coefficient, represents the differential term; Among them, when Exceeding the threshold When , the proportional coefficient, integral coefficient and differential coefficient are dynamically updated, and the update formulas are: Where K′ p , K′ i and K′ d Represent the updated proportional coefficient, integral coefficient and differential coefficient respectively, F max is the maximum tension change rate threshold; The collaborative adjustment amount Δn is calculated based on the updated coefficients. The formula is: The total control quantity is: Step 43: Allocate the total control quantity u(t) to the control instructions of the inverter and the servo motor according to the frequency domain characteristics, wherein the instructions received by the inverter are expressed as: f inv =α(t)·u(t), In the formula, α(t) represents the weight coefficient, The command received by the servo motor is expressed as: N pulse =(1-α(t))·u(t)。 6. The method for accurately testing and controlling tension in a rotating state according to claim 5, characterized in that: The steps to trigger the protection mechanism include: When the final tension estimate output by the Kalman filter is Exceeding the preset safety range [F set -ΔF th1 ,F set +ΔF th1 ], the first level alarm is triggered, that is, the sound and light alarm is triggered and the abnormal log is recorded; among them, ΔF th1 is the first-level tension deviation threshold, representing the allowable tension offset, ΔF th1 =k·F set , k is the process accuracy coefficient; If the residual is detected at the same time It is judged that the sensor is faulty, and the first level alarm is skipped and the second level alarm is entered directly. If the abnormality lasts longer than 5 seconds, the dynamic speed reduction strategy is implemented; like Less than the emergency stop tension threshold F e , and the capstan speed ω and the traction roller linear speed v satisfy η th is the speed mismatch rate threshold, triggering the third-level alarm, that is, cutting off the power supply and locking the equipment; where v nominal To calibrate the linear speed of the traction roller.

7. The method for accurately testing and controlling tension in a rotating state according to claim 6, characterized in that: Dynamic speed reduction strategies include: Adjust the speed of the winch drive motor and reduce the speed to β(t) times the reference value according to the gradient, where β(t) = 50% + 30% e -0.5t ; Adjust the linear speed of the traction roller servo motor, and adjust the linear speed synchronously to v′=v·β(t), maintaining the pay-off or take-up speed difference Δv=|ωr-v| constant; Adjust PID parameters and proportional coefficient K' s =0.5K′ p , the integration time is doubled T′ i =2T i .

Images