Hollow core polarization maintaining optical fiber cutting machine tension control system and control method

Abstract

The application provides a kind of hollow core polarization maintaining optical fiber cutting machine tension control system and control method, belong to optical fiber cutting control technical field, system uses four layers architecture of perception layer, processing layer, control layer, execution layer: perception layer is measured by double strain gauge symmetrical compensation device and temperature sensor synchronous acquisition mechanics and environmental signal;Processing layer adopts Wheatstone bridge, instrument amplifier, low-pass filter, 16-bit ADC and temperature compensation unit to realize signal multistage conditioning;Control layer deploys feedforward adaptive PID compound control system, combined with optical fiber diameter-tension parameter library feedforward compensation and dynamic PID adjustment, and integrates step loading control and fault early warning mechanism;Execution layer adopts torque mode motor and outputs linear tension through lead screw mechanism.

Description

Technical Field

[0001] This invention relates to the field of tension control technology for hollow polarization-maintaining fiber cleavers, specifically to a tension control system and control method for hollow polarization-maintaining fiber cleavers. Background Technology

[0002] Hollow-core polarization-maintaining fiber, a special type of optical fiber, features an air-hole core structure. It possesses excellent polarization-maintaining characteristics, low nonlinearity, and potential low-loss transmission capabilities, making it valuable for applications in fiber optic sensing, high-power laser transmission, quantum communication, and precision interferometry. During the fabrication of fiber optic devices, especially during end-face dicing, precise and stable axial tension must be applied to ensure the perpendicularity and flatness of the cut surface. This prevents fiber damage, breakage, or polarization state degradation caused by stress concentration or sudden changes in tension, thereby ensuring the efficiency of subsequent welding and coupling, and maintaining device performance.

[0003] Traditional tension control in hollow-core polarization-maintaining fiber cleavers typically employs simple open-loop or PID control methods, relying on force sensors or pressure feedback devices for adjustment. However, existing technologies generally have the following limitations:

[0004] 1. Conventional force sensors are susceptible to changes in ambient temperature, mechanical vibration, and electromagnetic interference, leading to signal drift and increased noise, making it difficult to achieve accurate and stable measurements in the μN to N range. Especially during long-term continuous operation or changes in operating conditions, zero-point drift and sensitivity temperature drift can significantly reduce control accuracy.

[0005] 2. Existing PID control parameters are usually fixed, making it difficult to adapt to optical fibers of different diameters and material properties, as well as dynamic changes in load during the cutting process. During high-speed cutting or significant adjustments to the tensile setpoint, the system is prone to overshoot, oscillation, or slow response, failing to achieve fast and stable tensile tracking.

[0006] 3. The control system often fails to fully utilize known fiber optic specifications for feedforward compensation, resulting in a complete reliance on feedback regulation and limited response speed. Furthermore, it lacks real-time, reliable early warning and rapid protection measures for abnormal conditions such as impending fiber breakage, easily leading to abnormal fiber breakage and production interruptions.

[0007] 4. The transmission backlash in the actuator can cause dead zones and nonlinearities in the tension control, affecting accuracy. In addition, step loading of the target tension can easily generate instantaneous impact forces, causing implicit damage or direct breakage to the brittle hollow polarization-maintaining optical fiber.

[0008] Therefore, there is an urgent need to develop a high-precision, high-response-speed, highly adaptive, and well-protected tension control system and method for hollow-core polarization-maintaining fiber cleavers to solve the above-mentioned technical bottlenecks and meet the stringent requirements of high-end fiber optic device manufacturing for the reliability and consistency of the cleaver process. Summary of the Invention

[0009] To address the above problems, this invention proposes a tension control system for a hollow-core polarization-maintaining fiber cleaver, comprising:

[0010] The sensing layer is used to detect tensile force signals and ambient temperature signals.

[0011] The processing layer is used to perform multi-level conditioning on the signals detected by the sensing layer and output digital signals.

[0012] The control layer is used to generate control commands based on the digital signals output by the processing layer;

[0013] An execution layer is used to apply a linear tension to the optical fiber according to the control command;

[0014] The sensing layer includes a dual strain gauge symmetrical compensation measurement device and a temperature sensor. The dual strain gauge symmetrical compensation measurement device is located in the stress concentration area of ​​the elastic deformation part of the fiber clamping mechanism. It adopts a layout of attaching one strain gauge to each of the upper and lower surfaces. The two strain gauges are connected to the differential input terminal of the Wheatstone bridge circuit.

[0015] The processing layer, along the signal flow direction, includes a Wheatstone bridge circuit, an instrumentation amplifier, a filter, an ADC chip, and a temperature compensation unit in sequence.

[0016] The control layer is equipped with a feedforward adaptive PID composite control system, which includes a feedforward compensation algorithm that outputs a reference tension value according to a preset fiber specification, and an adaptive PID adjustment module that dynamically adjusts PID parameters based on tension deviation.

[0017] The execution layer includes a servo motor operating in torque mode, which converts rotational motion into linear tension through a lead screw mechanism.

[0018] Furthermore, the elastic deformation section is a cantilever beam structure with its central axis coinciding with the optical fiber clamping axis. Two strain gauges are respectively connected to the opposite arms of the Wheatstone bridge to output differential doubling signals and suppress bending interference under axial tensile force.

[0019] Furthermore, the Wheatstone bridge circuit adopts a full-bridge connection, with two strain gauges serving as two resistors on opposite sides of the bridge arm, and the other two bridge arms being fixed resistors.

[0020] Furthermore, the temperature compensation unit uses a piecewise polynomial model to correct the temperature drift error in real time:

[0021] ΔT total = a1·ΔT + a2·ΔT 2 + a3·ΔT 3 + a4·ΔT·F measured ;

[0022] Where: ΔT = T - T0, T0 is the reference temperature; a1 is the temperature drift of the strain gauge resistance; a2 and a3 characterize the nonlinear temperature characteristics of the elastic modulus of the adhesive; and a4 is the thermal expansion coupling coefficient of the elastomer.

[0023] Furthermore, the execution layer also includes an encoder installed at the tail of the servo motor. The encoder feeds back the position information of the servo motor to the control layer in real time, which is used to realize real-time compensation and error elimination of transmission backlash at the software level.

[0024] Furthermore, the control layer also includes:

[0025] The stepped loading control module is used to decompose the target tension into 3-5 steps for gradual loading based on the encoder's position information;

[0026] A fault warning mechanism is used to trigger an emergency stop of the servo motor when the resistance change rate of the strain gauge exceeds 5% / s.

[0027] This invention also proposes a tension control method for a hollow polarization-maintaining fiber cleaver, which, based on the aforementioned tension control system for a hollow polarization-maintaining fiber cleaver, includes the following steps:

[0028] Step 1: Symmetrically attach the first strain gauge and the second strain gauge to the upper and lower surfaces of the elastic deformation part of the fiber clamping mechanism. Connect the two strain gauges to the Wheatstone bridge circuit with three wires. The elastic deformation part is a cantilever beam structure and its central axis coincides with the fiber clamping axis. The two strain gauges are respectively connected to the opposite side of the Wheatstone bridge. When the clamping mechanism is subjected to tensile force and produces a slight deformation, the bridge outputs a differential doubled voltage signal that is proportional to the tensile force.

[0029] Step 2: The ambient temperature signal is collected synchronously by the temperature sensor built into the clamping mechanism, the temperature drift compensation is calculated according to the piecewise polynomial model, and the compensation is added to the subsequent tensile force calculation in real time.

[0030] Step 3: The output signal of the Wheatstone bridge circuit is sequentially sent to the instrumentation amplifier for fixed gain amplification, filtered out mechanical vibration noise by a low-pass filter, and then converted into a digital signal by an ADC chip.

[0031] Step 4: The microprocessor converts the digital signal into a real-time tensile force value based on the calibration model;

[0032] Step 5: Based on the preset fiber diameter d, query the built-in diameter-tension parameter library to obtain the corresponding feedforward reference tension value F. ref , will F ref Fine-tuning the offset F with the setpoint bias The sum of these values ​​constitutes the total setpoint F of the PID controller. set ;

[0033] Step 6: Perform PID calculation once every fixed time period to calculate the tension deviation e. Based on the absolute value and rate of change of the deviation e, dynamically adjust the PID parameters and calculate the output torque command.

[0034] Step 7: When the target tension changes, the microprocessor decomposes the target tension into N steps, and the delay Δt of each step increases step by step. The torque command is sent to the motor driver through the CAN bus, which drives the servo motor to apply linear tension to the fiber optic clamping mechanism through the lead screw mechanism.

[0035] Step 8: During the cutting process, steps 3-7 are continuously executed in a loop. The encoder provides real-time feedback on the motor position, and the microprocessor corrects the torque command based on the rate of position change.

[0036] Step 9: Synchronously monitor the strain gauge resistance change rate dR / dt. When dR / dt > preset value, it is determined that the optical fiber is about to break. The microprocessor immediately completes the emergency stop of the servo motor by triggering the motor driver enable terminal.

[0037] Furthermore, in step 7, a stepped loading trigger condition is used; during the initial loading, the system enters the cutting preparation state, and then the tension is increased from zero to the target working tension.

[0038] Furthermore, during the initial startup, the motor is preheated at 10% of its rated torque to stabilize the substrate temperature of the strain gauge. When the ambient temperature is lower than the preset temperature, temperature data is continuously collected after startup to establish a temperature-drift curve and calculate the temperature drift compensation amount.

[0039] Furthermore, the method also includes periodically performing self-calibration: applying standard forces of 10N, 20N, and 30N sequentially, recording the corresponding digital signals, and updating the coefficients of the calibration model using the least squares method.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] By employing a dual-strain gauge symmetrical compensation measurement device in conjunction with a temperature compensation unit, and through a Wheatstone bridge full-bridge connection and instrumentation amplifier conditioning, the temperature drift error is suppressed to within 0.1%FS, which is an order of magnitude better than the traditional single-strain gauge solution.

[0042] The feedforward adaptive PID composite control system deployed in the control layer outputs the reference torque command in advance through the diameter-tension parameter library, which reduces the system response time from 150ms of traditional PID to less than 60ms, an improvement of 60%. It can still achieve fast and stable tracking of tension under high-speed cutting conditions, avoiding fiber damage caused by response lag.

[0043] The fuzzy PID control module dynamically adjusts parameters based on tensile force deviation and its rate of change, achieving compatible control of optical fibers with diameters ranging from 200 to 1000 μm. The parameter library contains reference tensile force values ​​for five different fiber specifications. Combined with stepped loading control, the system can automatically adapt to the mechanical properties of fibers with varying stiffness.

[0044] By monitoring the strain gauge resistance change rate dR / dt in real time, an emergency stop of the motor can be completed within 20ms, a response speed far faster than the 100ms-level protection of traditional pressure sensors. This mechanism reduces the fiber breakage rate to <0.5%, a reduction of more than 90% compared to existing control methods. Simultaneously, the stepped loading control limits the rate of tensile force increase to within 100N / s, avoiding the instantaneous impact caused by step loading and significantly reducing the risk of latent microcracks. Attached Figure Description

[0045] Figure 1 This is a structural diagram of the tension control system of the present invention;

[0046] Figure 2 This is a schematic diagram of the tension control method of the present invention;

[0047] Figure 3 This is a perspective view of the clamping mechanism in a specific embodiment;

[0048] Figure 4 This is a top view of the clamping mechanism in a specific embodiment. Detailed Implementation

[0049] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0050] like Figure 1 As shown, the tension control system of the present invention includes a sensing layer, a processing layer, and an execution layer, which achieves precise control through mechanical, electrical, and software collaboration.

[0051] 1. Perception layer

[0052] The sensing layer features a dual-strain gauge symmetrical compensation measurement device. In the stress concentration area of ​​the elastic deformation part of the fiber optic clamping mechanism, a symmetrical layout is adopted, with one metal foil strain gauge attached to each of the upper and lower surfaces. The two strain gauges are connected to the differential input terminal of a Wheatstone bridge. In conjunction with the PT100 temperature sensor built into the clamping mechanism, mechanical signals and ambient temperature signals are acquired simultaneously.

[0053] The dual-strain gauge symmetrical compensation measurement device achieves selective measurement based on the difference in strain distribution between axial tension and bending deformation of a cantilever beam:

[0054] (1) Pure axial tensile force: When the elastic deformation part is only subjected to axial tensile force F, according to the plane section assumption in mechanics of materials, the normal strain value E generated on the upper and lower surfaces is top E bottom Equal in size but opposite in sign, i.e., E top = -E bottom Connect the upper surface strain gauge R1 and the lower surface strain gauge R3 to opposite arms of a Wheatstone bridge. The bridge output voltage U out for:

[0055] U out = K·U exc ·(E top -E bottom ) = 2K·U exc ·E top ;

[0056] Where K is the strain gauge sensitivity coefficient, U exc The excitation voltage is used. This connection method increases the sensitivity by 2 times. At the same time, temperature changes cause the resistance drift of the two strain gauges to be in the same direction, and the bridge circuit automatically cancels out the drift, achieving temperature self-compensation.

[0057] (2) Suppression of bending moment interference: When there is an additional bending moment M, the strain on the upper and lower surfaces has the same sign E. top =E bottom Bridge output U out ≈0, which suppresses bending interference in principle.

[0058] The present invention, through the centering design of the optical fiber clamping mechanism, ensures that the deviation between the optical fiber axis and the neutral layer of the elastic deformation part is <0.1mm, thus ensuring that the bending moment is <0.05N·m and the introduced tensile force measurement error is <0.3%.

[0059] (3) Compensation for non-ideal force error: For the composite deformation caused by residual eccentric force, the system monitors the attitude angle of the fiber optic clamping mechanism through an encoder. ,when Software compensation is triggered when the temperature exceeds 0.5°.

[0060] F corrected = Fmeasured / cos ;

[0061] This compensation reduces the non-ideal stress error to <0.5%FS.

[0062] In one embodiment, the elastic deformation part is a cantilever beam structure with a yield strength q. s ≥780MPa; the length of the elastic deformation section is L=30mm, the width is w=15mm, and the thickness is t=3mm. The stress concentration area of ​​the elastic deformation section is at the root of the cantilever beam structure; the center of the strain gauge is 5±0.2mm away from the fixed end of the cantilever beam.

[0063] The strain gauge is preferably made of constantan alloy, with a rated strain covering 150% of the maximum load. The 75N range corresponds to a working tensile force of 50N. The substrate is made of phenolic resin, and the surface has a surface finish of Ra < 1.6. After polishing, the surface is bonded with epoxy resin and coated with a silicone protective layer.

[0064] 2. Processing layer

[0065] The processing layer features a multi-precision collaborative signal conditioning module. The signal conditioning module, along the signal flow direction, includes:

[0066] Wheatstone bridge circuit: The strain gauges are connected using a full-bridge method. The resistance change at the level is converted into a mV-level differential voltage signal to achieve primary cancellation of temperature and humidity interference.

[0067] Instrumentation amplifier: Fixed gain 1000 times, signal-to-noise ratio improved to over 45dB;

[0068] Low-pass filter: cutoff frequency 10Hz, filters out interference from cutting table vibration;

[0069] 16-bit ADC chip: sampling frequency 250Hz, tensile strength resolution up to 0.01N;

[0070] Temperature compensation unit: The PT100 temperature sensor signal is input to the microprocessor to correct the temperature drift error in real time. After correction, the temperature error is <0.1%FS, which means that the measurement error caused by temperature change is less than 0.1% of the full scale value.

[0071] In a specific embodiment, the Wheatstone bridge uses a full-bridge configuration to achieve optimal sensitivity and temperature compensation. The specific configuration is as follows:

[0072] Bridge arm composition: All four arms are 120mm. Resistors. R1 is the strain gauge attached to the upper surface of the elastic deformation section, and R3 is the strain gauge on the lower surface, located on opposite bridge arms. R2 and R4 are two high-precision, low-temperature drift fixed surface-mount resistors on the signal conditioning board.

[0073] Connection method: R1 and R2 are connected in series, and R3 and R4 are connected in series, forming the two arms of the bridge. The differential voltage signal is drawn from the midpoint between R1 and R2 and the midpoint between R3 and R4 and sent to the instrumentation amplifier.

[0074] Excitation voltage: The excitation voltage of the bridge is provided by a precision voltage reference source in the signal conditioning module to ensure high stability and low noise. This reference source is powered by the system main power supply after being regulated by an LDO.

[0075] 3. Control Layer

[0076] The control layer has an adaptive PID adjustment module, an embedded microprocessor, and a feedforward adaptive PID composite control system, with the feedforward adaptive PID composite control system deployed on the embedded microprocessor.

[0077] The feedforward adaptive PID composite control system employs composite control logic:

[0078] Feedforward compensation algorithm: Built-in fiber diameter-tension parameter library, 200-1000 The system outputs a reference tension value in advance based on the preset fiber specifications, which improves the system response speed by 60% compared to pure PID.

[0079] Adaptive PID control module: Calculates the tension deviation ΔF = preset value - measured value in real time, and dynamically adjusts the proportional coefficient K. p With integration time T i K during high-speed cutting p Increase by 20-30%, T at low speed i Reduce by 40-50%, control cycle of 10ms, and suppress tension fluctuations to within ±1%;

[0080] Step-by-step loading control: Based on encoder position feedback, the target tension is decomposed into 3-5 steps for gradual loading, avoiding damage to the optical fiber by instantaneous impact force;

[0081] Fault warning mechanism: When the resistance change rate is detected to be greater than 5% / s, it is determined that the optical fiber is about to break and the motor is triggered to stop within 20ms.

[0082] 4. Execution layer

[0083] The execution layer features a servo motor with torque mode. A side-start micro-screw mechanism converts the rotational motion of the servo motor into linear tension. The conversion formula from the servo motor output torque Tm to the fiber tension F is as follows:

[0084]

[0085] in: The lead screw drive efficiency is preferably 0.92; P=5mm is the lead screw pitch, Ff =0.3N is the system friction force, obtained from actual measurement.

[0086] The motor operates in torque mode with an output torque accuracy of ±0.5%; an encoder is coaxially mounted at the tail of the motor to provide real-time position information to the software layer, enabling stroke control.

[0087] The fiber optic clamping mechanism drives the strain gauge to generate deformation signals through mechanical connection. The strain gauge and PT100 temperature sensor signals are connected to the input terminal of the signal conditioning module via shielded cable. The signal conditioning module outputs digital signals and connects to the microprocessor via SPI bus. The embedded microprocessor sends torque commands to the servo motor driver via CAN bus. The motor reacts to the fiber optic clamping mechanism via the lead screw mechanism. The encoder feedback signal is returned to the microprocessor via RS485 interface, forming a control link.

[0088] The motor output shaft is directly connected to a THK BTK1405 side-start lead screw via a compact coupling. The lead screw employs a fixed-support bearing housing structure at both ends to enhance rigidity. Instead of a mechanical backlash compensation mechanism, the system uses an encoder to provide real-time position feedback. The microprocessor dynamically corrects torque commands based on the rate of position change, achieving real-time compensation and error elimination of transmission backlash at the software level.

[0089] In a specific embodiment, system error analysis and verification can be performed to quantify the measurement error of the system under three non-ideal operating conditions:

[0090] Under eccentric force conditions, for example, when the fiber optic axis is offset by 0.2 mm, a bending moment M = 0.01 N·m is introduced, with a theoretical error of 0.42% and an actual measurement error of 0.35%. After software compensation, the error is reduced to 0.08%.

[0091] Under temperature shock conditions, such as when a temperature jump occurs from 25℃ to 40℃, the error is 1.2N without compensation, and <0.03N after polynomial compensation.

[0092] Under dynamic loading conditions, such as a slope of 50 N / s, stepped loading reduces the overshoot from 15% to 0.8%.

[0093] In summary, the confidence level of this system is >99.7% within the ±1% accuracy range, which meets the requirements of industrial cutting.

[0094] Based on the above control system, based on the perception, processing, and execution architecture, such as Figure 2 As shown, precise tension control can be achieved through the following steps:

[0095] Step 1: Mechanical signal perception and conversion.

[0096] On the upper and lower surfaces of the elastic deformation section of the fiber optic clamping mechanism, a first strain gauge and a second strain gauge are symmetrically attached. The leads of the two strain gauges are connected to a Wheatstone bridge circuit using a three-wire connection. When the clamping mechanism undergoes slight deformation under tension, the resistance difference between the two strain gauges changes, and the bridge outputs a millivolt-level voltage signal U proportional to the tension. b .

[0097] Step 2: Temperature synchronous acquisition and compensation.

[0098] The ambient temperature T is collected synchronously by the PT100 temperature sensor built into the clamping mechanism. The microprocessor calculates the temperature drift compensation amount according to the temperature compensation unit and adds the compensation amount to the subsequent tensile force calculation in real time to eliminate temperature interference.

[0099] The temperature compensation unit uses a piecewise polynomial model to correct the temperature drift error ΔT in real time. total :

[0100] ΔT total = a1·ΔT + a2·ΔT 2 + a3·ΔT 3 + a4·ΔT·F measured ;

[0101] in:

[0102] ΔT = T - T0, where T0 is the reference temperature; preferably, T0 = 25℃;

[0103] a1 represents the temperature drift of the strain gauge resistance, in N / ℃.

[0104] a2、 a3 characterizes the nonlinear temperature characteristics of the elastic modulus of the adhesive, in N / ℃. 2 N / ℃ 3 ;

[0105] a4 is the thermal expansion coupling coefficient of the elastic body, in units of 1 / ℃, used to correct the effect of thermal stress on the stiffness of the elastic deformation part; F measured The current measured tensile force value; a1-a4 were obtained through standard force loading calibration at at least 5 temperature points.

[0106] Step 3: Multi-level signal conditioning.

[0107] The bridge output signal is sequentially fed into an instrumentation amplifier for 1000x fixed gain amplification, then filtered by a 10Hz low-pass filter to remove mechanical vibration noise, and finally converted into a digital signal D by a 16-bit ADC chip at a 250Hz sampling frequency. s .

[0108] Step 4: Calculate the tension value.

[0109] The microprocessor is based on the calibration model F s =k×(D s -D0)+b, converts the digital signal into a real-time tension value F. s Where k is the sensitivity coefficient, preferably 0.02. D0 is the zero-point reference value, and b is the offset compensation amount.

[0110] Step 5: Preset feedforward reference value.

[0111] Based on the preset fiber diameter d, the built-in diameter-tensile parameter library is consulted to obtain the corresponding reference tensile value F. ref When d=250 Time F ref =20N, d=800 Time F ref =35N; F ref The initial output value is used for PID control to shorten the response time. The parameter library is a piecewise linear interpolation table, as shown in Table 1.

[0112] Table 1

[0113] Fiber diameter d <![CDATA[Reference tensile force value F ref > Elastic modulus correction factor ke 200 15.0 1.00 250 20.0 1.00 400 25.0 0.98 600 30.0 0.95 800 35.0 0.92 1000 40.0 0.90

[0114] The reference tension value obtained in step 5 is not directly used as a motor command, but rather as the setpoint bias for the PID controller. Therefore, the total target tension setpoint F of the system... set It consists of a feedforward reference value and an optional fine-tuning bias (initially 0), i.e., F set = F ref + F bias Among them, F bias For optional fine-tuning of the bias, the adjustment calculated by the PID controller is then based on the measured tension F. measured With this total setting value F set The deviation between them.

[0115] Immediately after powering on the system or switching fiber optic specifications, connect F ref Insert F set Since the feedforward value is based on empirical data and may contain errors, the system always runs a closed-loop PID controller. However, to achieve a smooth transition, the PID parameters (K...) are adjusted. p T i According to the real-time deviation ΔF = F set - F measured The size is dynamically adjusted with a blur effect:

[0116] when When the value is large, such as > 2N, the system is in the rapid approach phase. In this case, a higher proportional gain Kp is used, and the integral action is temporarily frozen.i The goal is to use the feedforward value to quickly approach the target while avoiding integral saturation.

[0117] when Reduce to a moderate range, such as 0.5N ≤ When ≤ 2N, the system enters the precise adjustment phase. The PID controller switches to the normal parameter group, and the integrator starts working to eliminate steady-state error. At this time, the control is dominated by the PID controller, and the feedforward value is used as compensation for the static operating point.

[0118] when When the value is very small, such as <0.5N, the system enters the steady-state maintenance phase. A lower K is then used. p and longer T i It focuses on stability and anti-interference.

[0119] In this way, feedforward control provides rapid coarse adjustment, while closed-loop PID is responsible for subsequent fine adjustment and stabilization. The two achieve a smooth transition and work together through deviation-based fuzzy parameter adjustment.

[0120] Step 6: Deviation calculation and fuzzy PID adjustment.

[0121] PID calculation is performed every 10ms: Calculate the tension deviation ΔF=F set +F temp -F s Based on the absolute value of the deviation Dynamically adjust PID parameters:

[0122] like >2N, using the high-response parameter set: K p =2.0, T i As it approaches infinity, the integral closes.

[0123] If 0.5N≤ ≤2N, using the balanced parameter set: K p =1.2, T i =50ms;

[0124] like <0.5N, using steady-state parameter set: K p =0.8, T i =100ms;

[0125] Calculate the output torque command T cmd =K p ×(ΔF+ (ΔF / T i )dt).

[0126] Step 7: Execute in a step-by-step manner.

[0127] When the target tension changes, the microprocessor will F set Decompose into N steps, N=3~5, with a delay for each step. =Incrementing in increments of 50ms; Torque command T is transmitted via CAN bus. cmd The signal is sent to the motor driver, which drives the servo motor to apply a linear tension to the clamping mechanism via a lead screw mechanism, thus avoiding instantaneous impact.

[0128] In the step-by-step loading execution in step 7, the triggering condition is one of the following two cases:

[0129] Initial loading: When the system completes alignment and enters the cutting preparation state, the initial loading is from zero tension or low holding tension to the target working tension.

[0130] Dynamic adjustment: During the cutting process, if the target tension setting needs to be adjusted due to process requirements, and the adjustment range exceeds the preset threshold, then...

[0131] When any of the above conditions are met, the microprocessor will assign a new target tensile force value F. target According to the preset number of steps N and the delay of each step It is decomposed into a series of progressively increasing sub-targets and gradually realized through closed-loop control, thereby avoiding instantaneous impact on the optical fiber.

[0132] Step 8: Dynamic monitoring and feedback correction.

[0133] During the cutting process, steps 3-7 are continuously executed in a loop, completing 250 closed-loop adjustments per second; the encoder provides real-time feedback on the motor position, and the microprocessor corrects the torque command based on the rate of position change, eliminating mechanical transmission backlash errors.

[0134] Step 9: Fault warning and emergency stop.

[0135] The strain gauge resistance change rate dR / dt is monitored synchronously. When dR / dt > 5% / s, it is determined that the optical fiber is about to break. The microprocessor immediately triggers the motor driver enable terminal through GPIO hardwire and completes the emergency stop of the motor within 20ms.

[0136] Step 10: Regular self-calibration

[0137] Perform standard weight loading calibration weekly: sequentially apply standard forces of 10N, 20N, and 30N, and record the corresponding digital signal D. s The least squares method is used to update the calibration model coefficients k and b to ensure long-term accuracy.

[0138] Example 1

[0139] 1. Perception layer

[0140] This embodiment uses a dual-strain gauge symmetrical compensation measurement device and a PT100 temperature sensor as the signal detection unit. The PT100 temperature sensor is model WZP-035 with an accuracy of ±0.1℃. In the stress concentration area of ​​the elastic deformation part of the fiber optic clamping mechanism, the elastic deformation part is made of 65Mn spring steel with a thickness of 3mm; a metal foil strain gauge is attached to each of the upper and lower surfaces. The metal foil strain gauge is model BF120-3AA with a resistance of 120Ω. The sensitivity coefficient is 2.0. The rated strain coverage range of the two strain gauges is 0-7500. The strain gauges are designed to withstand a maximum load of 75N. The substrate is phenolic resin, bonded to the surface-polished elastic deformation section using epoxy resin adhesive. A 0.5mm thick silicone protective layer is applied to the surface for moisture protection. The two strain gauge leads are connected to the differential input of a Wheatstone bridge using a three-wire system. A PT100 temperature sensor is embedded in the clamping mechanism 5mm from the strain gauges, simultaneously acquiring ambient temperature signals.

[0141] 2. Processing layer

[0142] The signal processing unit includes, in sequence along the signal flow direction:

[0143] Wheatstone bridge circuit: A full-bridge configuration is used. The two strain gauges mentioned above serve as resistors R1 and R3 on opposite bridge arms, respectively. The other two bridge arms, R2 and R4, are high-precision, low-temperature drift fixed resistors, model Vishay S102K, with a resistance of 120Ω. The temperature drift is ±5ppm / ℃. The excitation voltage is provided by a precision voltage reference source, model ADR4525, with an output voltage of 5V, which converts the strain gauge resistance change into a millivolt-level differential voltage signal.

[0144] Instrumentation amplifier: The ADI AD8221 chip is selected, with a fixed gain of 1000 times, an input impedance of 10GΩ, and a common-mode rejection ratio of 80dB, which amplifies the bridge output signal to the volt level;

[0145] Filter: A second-order active low-pass filter is used, with an OP27 operational amplifier and a cutoff frequency of 10Hz, which effectively filters out mechanical vibration noise.

[0146] ADC chip: TI ADS1115, 16-bit resolution, 250Hz sampling frequency, outputs digital signals via SPI bus;

[0147] Temperature compensation unit: It adopts an STM32F407 microprocessor with a main frequency of 168MHz, and corrects the temperature drift error in real time based on a piecewise polynomial model. The corrected system error is <0.1%FS.

[0148] 3. Control Layer

[0149] The control layer is based on the STM32F407 microprocessor and deploys a feedforward adaptive PID composite control system, which specifically includes:

[0150] Feedforward compensation algorithm: Built-in diameter-tension parameter library, storing reference tension values ​​corresponding to 250μm, 400μm, 600μm and 800μm optical fibers, which are 20N, 25N, 30N and 35N respectively, and outputting reference tension values ​​in advance according to preset optical fiber specifications;

[0151] Adaptive PID control module: Dynamically adjusts PID parameters with a control cycle of 10ms, when the tension deviation... When the coefficient is greater than 2N, the proportionality constant K p =2.0, integration time T i It tends to ∞; when 0.5N≤ When K ≤ 2N p =1.2, T i =50ms; when When K < 0.5N, p =0.8, T i =100ms;

[0152] Step-type loading control module: Based on encoder position feedback, the target tensile force is decomposed into 3-5 steps for gradual loading, with a delay of 50ms for each step;

[0153] Fault warning mechanism: When the strain gauge resistance change rate dR / dt>5% / s is detected, the microprocessor triggers the motor driver enable terminal within 20ms through the GPIO port to realize emergency stop protection.

[0154] 4. Execution Layer

[0155] This embodiment uses a servo motor operating in torque mode as the actuator. The servo motor is a Mitsubishi HG-KR43J with a rated torque of 4 N·m and a torque accuracy of ±0.5%. The motor output shaft is directly connected to a THK BTK1405 side-start lead screw via a rigid coupling of model SFC-050DA2, converting rotary motion into linear tension. A 17-bit incremental encoder with a resolution of 131072 p / r is coaxially mounted at the motor tail, providing real-time position information to the microprocessor for software compensation of transmission backlash. The fiber optic clamping mechanism is mounted on the lead screw nut, with a maximum stroke of 100 mm and a repeatability of ±0.02 mm.

[0156] System Connections: Strain gauge and temperature sensor signals are connected to a signal conditioning circuit board (80mm x 60mm) via shielded twisted-pair cables. The conditioned digital signals are then connected to the STM32F407 core board via an SPI interface. The core board communicates with the motor driver via a CAN bus to send torque commands. Encoder signals are returned to the core board via an RS485 interface, forming a closed-loop control link. The system operates at 24VDC, with a total power consumption of <50W.

[0157] Tests have shown that the system in this embodiment can achieve a stable tensile force of 20±0.15N when cutting 250μm hollow-core polarization-maintaining fiber, with a steady-state error of <0.75% and an overshoot of <1%, fully meeting the process requirements for the fabrication of high-end optical fiber devices.

[0158] Example 2

[0159] This embodiment uses the control system of Embodiment 1 to perform end-face cutting on a hollow polarization-maintaining fiber with a diameter of 400μm. The target tension is set to 25N, and the specific steps include:

[0160] Step 1: Mechanical Signal Sensing and Conversion

[0161] A 400μm hollow-core polarization-maintaining fiber is installed in the clamping mechanism, such as Figure 3 , Figure 4 As shown, the clamping force is controlled at 0.5N to avoid damaging the optical fiber. After the system is powered on, the clamping mechanism undergoes slight deformation under the initial tensile force, and the strain gauge resistance values ​​on the upper and lower surfaces are R1=120.048Ω respectively. R3 = 119.952 After being connected to a Wheatstone bridge, the output differential voltage Ub = 2.5mV, corresponding to an initial tensile force of 0.5N. The strain gauges use a three-wire connection to eliminate the influence of lead resistance.

[0162] Step 2: Synchronous Temperature Acquisition and Compensation

[0163] Start the PT100 temperature sensor, collect ambient temperature T=22.5℃ and reference temperature T0=25℃, and calculate the temperature drift compensation ΔT. total =0.0075N, and this compensation amount is added to the subsequent tensile force calculation in real time.

[0164] Step 3: Multi-level signal conditioning and conversion

[0165] The 2.5mV signal output from the bridge is fed into the AD8221 instrumentation amplifier. After amplification, the voltage is 2.5V. After filtering out the vibration noise of the cutting table by a 10Hz low-pass filter, the ADS1115 converts it into a digital signal Ds=16584 with a sampling rate of 250Hz and a zero reference value D0=16500. The total signal conditioning time is <2ms.

[0166] Step 4: Calculation of tension value

[0167] The microprocessor is based on the calibration model F s =k×(D s -D0)+b;k=0.02N / Ω;b=0.5N;Calculate the real-time tension F s =0.02×(16584-16500)+0.5=2.18N, and the actual tensile force F is obtained after considering the temperature compensation. actual =2.18-0.0075≈2.17N; the digital signal is D s .

[0168] Step 5: Presetting the feedforward baseline value and setting the target

[0169] Based on an optical fiber diameter d = 400 μm, the reference tensile force value F is obtained by consulting the diameter-tensile force parameter library. ref =25N. F ref As the setpoint bias of the PID controller, the total target tension F set =F ref +F bias =25+0=25N, where the fine-tuning bias F bias Initially 0.

[0170] Step 6: Step-by-step loading execution

[0171] The system enters the cutting preparation state, triggering stepped loading. The microprocessor decomposes the target tensile force of 25N into five steps: 5N, 10N, 15N, 20N, and 25N, each step held for 50ms. PID calculations are performed within each step.

[0172] First step, target 5N: =2.83N>2N, so K is used. p =2.0, T i As it approaches infinity, the output torque command T is given. cmd =5.66 N·m;

[0173] Steps 2-4: Gradual Approach Process Switch to K between 0.5 and 2N. p =1.2, T i =50ms;

[0174] Fifth step, target 25N: when Switch to K after <0.5N p =0.8, T i =100ms.

[0175] Torque commands are sent to the motor driver via the CAN bus. The motor applies linear tension to the optical fiber via a lead screw mechanism. The total loading time is approximately 250ms, and the tension rise rate is controlled within 100N / s.

[0176] Step 7: Dynamic monitoring and feedback correction

[0177] During the cutting process, the system continuously cycles through signal acquisition, conditioning, calculation, PID calculation, and torque output at 10ms intervals, completing 250 closed-loop adjustments per second. The encoder provides real-time feedback on the motor position; when an abnormal rate of position change is detected, such as a sudden change caused by gap, the microprocessor immediately corrects the torque command, compensating by ΔT. cmd =0.05× , The change rate of position feedback from the encoder is abruptly adjusted to ensure that the tension fluctuation is <±0.2N.

[0178] Step 8: Fault Warning and Emergency Stop Protection

[0179] The strain gauge resistance change rate dR / dt was monitored synchronously. When dR / dt reached 5.2% / s, the optical fiber was nearing the breakage threshold. The microprocessor triggered the motor driver enable pin within 18ms via GPIO hardwire, causing the motor to stop urgently and preventing fiber optic splintering and device damage. During this cutting process, dR / dt remained stable between 0.3-0.8% / s, and no emergency stop was triggered.

[0180] Step 9: Regular self-calibration

[0181] After cutting, perform self-calibration once a week: suspend 10N, 20N, and 30N standard weights sequentially on the clamping mechanism and record the corresponding digital signal D. s1 =17050、D s2 =17520、D s3 =18010, the calibration coefficient k=0.0198N / is updated using the least squares method. b=0.48N to ensure long-term system accuracy.

[0182] Using the method of this embodiment, 400 The perpendicularity of the cut end face of the hollow-core polarization-maintaining fiber reaches 89.5°±0.3°, and the end face roughness Ra<0.5. The fiber breakage rate is less than 0.5%, which improves the cutting qualification rate by 35% compared with the traditional open-loop control method and significantly improves product consistency.

[0183] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A tension control system for a hollow-core polarization-maintaining fiber cleaver, characterized in that, include: The sensing layer is used to detect tensile force signals and ambient temperature signals. The processing layer is used to perform multi-level conditioning on the signals detected by the sensing layer and output digital signals. The control layer is used to generate control commands based on the digital signals output by the processing layer; An execution layer is used to apply a linear tension to the optical fiber according to the control command; The sensing layer includes a dual strain gauge symmetrical compensation measurement device and a temperature sensor. The dual strain gauge symmetrical compensation measurement device is located in the stress concentration area of ​​the elastic deformation part of the fiber clamping mechanism. It adopts a layout of attaching one strain gauge to each of the upper and lower surfaces. The two strain gauges are connected to the differential input terminal of the Wheatstone bridge circuit. The processing layer, along the signal flow direction, includes a Wheatstone bridge circuit, an instrumentation amplifier, a filter, an ADC chip, and a temperature compensation unit in sequence. The control layer is equipped with a feedforward adaptive PID composite control system, which includes a feedforward compensation algorithm that outputs a reference tension value according to a preset fiber specification, and an adaptive PID adjustment module that dynamically adjusts PID parameters based on tension deviation. The execution layer includes a servo motor operating in torque mode, which converts rotational motion into linear tension through a lead screw mechanism.

2. The tension control system for the hollow-core polarization-maintaining fiber optic cleaver according to claim 1, characterized in that, The elastic deformation section is a cantilever beam structure with its central axis coinciding with the optical fiber clamping axis. Two strain gauges are respectively connected to the opposite arms of the Wheatstone bridge to output differential doubling signals and suppress bending interference under axial tensile force.

3. The tension control system for the hollow-core polarization-maintaining fiber optic cleaver according to claim 2, characterized in that, The Wheatstone bridge circuit adopts a full-bridge configuration, with two strain gauges serving as two resistors on opposite sides of the bridge arm, and the other two arms being fixed resistors.

4. The tension control system for the hollow-core polarization-maintaining fiber optic cleaver according to claim 1, characterized in that, The temperature compensation unit uses a piecewise polynomial model to correct temperature drift error in real time. ΔT total = a1·ΔT + a2·ΔT 2 + a3·ΔT 3 + a4·ΔT·F measured ; Where: ΔT = T - T0, T0 is the reference temperature; a1 is the dominant strain gauge resistance temperature drift; a2 and a3 characterize the nonlinear temperature properties of the adhesive's elastic modulus; a4 is the elastomer thermal expansion coupling coefficient; F measured This represents the current measured tensile force value.

5. The tension control system for the hollow-core polarization-maintaining fiber optic cleaver according to claim 1, characterized in that, The execution layer also includes an encoder installed at the tail of the servo motor. The encoder feeds back the position information of the servo motor to the control layer in real time, which is used to realize real-time compensation and error elimination of transmission backlash at the software level.

6. The tension control system for the hollow-core polarization-maintaining fiber optic cleaver according to claim 5, characterized in that, The control layer also includes: The stepped loading control module is used to decompose the target tension into 3-5 steps for gradual loading based on the encoder's position information; A fault warning mechanism is used to trigger an emergency stop of the servo motor when the resistance change rate of the strain gauge exceeds 5% / s.

7. A tension control method for a hollow polarization-maintaining fiber cleaver, based on the tension control system for a hollow polarization-maintaining fiber cleaver as described in any one of claims 1-6, characterized in that, Includes the following steps: Step 1: Symmetrically attach the first strain gauge and the second strain gauge to the upper and lower surfaces of the elastic deformation part of the fiber clamping mechanism. Connect the two strain gauges to the Wheatstone bridge circuit with three wires. The elastic deformation part is a cantilever beam structure and its central axis coincides with the fiber clamping axis. The two strain gauges are respectively connected to the opposite side of the Wheatstone bridge. When the clamping mechanism is subjected to tensile force and produces a slight deformation, the bridge outputs a differential doubled voltage signal that is proportional to the tensile force. Step 2: The ambient temperature signal is collected synchronously by the temperature sensor built into the clamping mechanism, the temperature drift compensation is calculated according to the piecewise polynomial model, and the compensation is added to the subsequent tensile force calculation in real time. Step 3: The output signal of the Wheatstone bridge circuit is sequentially sent to the instrumentation amplifier for fixed gain amplification, filtered out mechanical vibration noise by a low-pass filter, and then converted into a digital signal by an ADC chip. Step 4: The microprocessor converts the digital signal into a real-time tensile force value based on the calibration model; Step 5: Based on the preset fiber diameter d, query the built-in diameter-tension parameter library to obtain the corresponding feedforward reference tension value F. ref , will F ref Fine-tuning the offset F with the setpoint bias The sum of these values ​​constitutes the total setpoint F of the PID controller. set ; Step 6: Perform PID calculation once every fixed time period to calculate the tension deviation e. Based on the absolute value and rate of change of the deviation e, dynamically adjust the PID parameters and calculate the output torque command. Step 7: When the target tension changes, the microprocessor decomposes the target tension into N steps, and the delay Δt of each step increases step by step. The torque command is sent to the motor driver through the CAN bus, which drives the servo motor to apply linear tension to the fiber optic clamping mechanism through the lead screw mechanism. Step 8: During the cutting process, steps 3-7 are continuously executed in a loop. The encoder provides real-time feedback on the motor position, and the microprocessor corrects the torque command based on the rate of position change. Step 9: Synchronously monitor the strain gauge resistance change rate dR / dt. When dR / dt > preset value, it is determined that the optical fiber is about to break. The microprocessor immediately completes the emergency stop of the servo motor by triggering the motor driver enable terminal.

8. The tension control method for a hollow-core polarization-maintaining fiber cleaver according to claim 7, characterized in that, In step 7, the system is triggered by a stepped loading condition; during the initial loading, the system enters the cutting preparation state, and then the tension is increased from zero to the target working tension.

9. The tension control method for a hollow-core polarization-maintaining fiber cleaver according to claim 7, characterized in that, During the initial startup, the motor is preheated at 10% of its rated torque to stabilize the substrate temperature of the strain gauge. When the ambient temperature is lower than the preset temperature, temperature data is continuously collected after startup to establish a temperature-drift curve and calculate the temperature drift compensation.

10. The tension control method for a hollow-core polarization-maintaining fiber cleaver according to claim 7, characterized in that, The method also includes periodically performing self-calibration: applying standard forces of 10N, 20N, and 30N sequentially, recording the corresponding digital signals, and updating the coefficients of the calibration model using the least squares method.

Images