Torsion measuring device for optical fiber production

Abstract

The invention discloses a torsion measuring device for optical fiber production, which belongs to the technical field of optical fiber production detection and comprises a laser measuring unit, a single laser transmitting module and a double laser differential transmitting module, the high-speed imaging unit carries a CMOS (Complementary Metal-Oxide-Semiconductor Transistor) image sensor with a frame rate not lower than 10kHz and a narrow-band filter and is used for capturing a laser speckle image on the surface of the optical fiber; a state machine logic circuit and a programmable timer are arranged in the self-adaptive control unit, and intelligent switching of measurement modes is achieved; the torsion analysis unit integrates a DSP processor and a storage medium and executes torsion position calibration and weight calculation; the abnormity diagnosis unit is connected with the cloud database and the local cache and is used for realizing historical data matching and cause probability sorting; and the alarm execution unit comprises an industrial-grade audible and visual alarm and a mechanical position marking device. According to the invention, the high-dynamic torsion scene can be effectively coped with while the measurement precision is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of optical fiber production detection, in particular to a torsion measuring device for optical fiber production. Background Art

[0002] Optical fiber structures are extremely precise, and the core and cladding geometry, in particular, significantly impacts performance. If fiber twist occurs during the production process, it can cause geometric distortion, such as an elliptical core or uneven cladding. This can affect optical signal transmission, increasing loss or reducing bandwidth. Furthermore, twisting can induce residual stress, potentially causing fiber degradation or breakage over long-term use. Therefore, monitoring twisting is critical to quality assurance.

[0003] Traditional torsion measurement devices suffer from issues such as insufficient measurement accuracy, delayed mode switching, and weak anomaly diagnosis capabilities, making them inadequate for the real-time monitoring requirements of high-speed optical fiber production lines. Existing single-laser measurement modes are unable to handle highly dynamic torsion scenarios, while fixed dual-laser modes suffer from excessive power consumption. Therefore, those skilled in the art have proposed a torsion measurement device for optical fiber production to address the issues raised in the background art. Summary of the Invention

[0004] The object of the present invention is to provide a torsion measurement device for optical fiber production, which can effectively cope with high-dynamic torsion scenarios while ensuring measurement accuracy, so as to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A torsion measurement device for optical fiber production, comprising:

[0007] The laser measurement unit is equipped with a dynamically switchable single laser emission module and a dual laser differential emission module. The single laser mode uses a semiconductor laser with a wavelength of 632nm to 1550nm, and the dual laser mode uses a paired laser with a wavelength difference of less than 20nm.

[0008] High-speed imaging unit, equipped with a CMOS image sensor with a frame rate of no less than 10kHz and a narrow-band filter, used to capture the laser speckle image on the optical fiber surface;

[0009] Adaptive control unit with built-in state machine logic circuit and programmable timer to achieve intelligent switching of measurement modes;

[0010] Torsion analysis unit, integrating DSP processor and storage medium, performs torsion position calibration and weight calculation;

[0011] The abnormality diagnosis unit connects the cloud database and local cache to achieve historical data matching and cause probability sorting;

[0012] Alarm execution unit, including industrial-grade sound and light alarms and mechanical position marking devices;

[0013] The working process of the torsion measuring device includes:

[0014] During the fiber startup phase, single laser low power measurement mode is enabled;

[0015] When the torsion angle is detected to exceed the preset threshold, it switches to the dual laser high-precision mode;

[0016] After running in high-precision mode for an adjustable period of time, it returns to single-laser mode if the torsional event disappears.

[0017] As a further solution of the present invention: the operating parameters of the single laser emission module meet the following requirements: the spot diameter does not exceed 1 / 3 of the optical fiber diameter, and the power density is positively correlated with the optical fiber transmittance; the acquisition frame rate of the high-speed imaging unit is set to more than twice the ratio of the maximum linear velocity of the optical fiber to the minimum spatial resolution; the projection angle of the two laser beams of the dual-laser differential emission module is dynamically adjusted according to the center spacing of the spot and the distance from the optical fiber to the projection surface.

[0018] As a further solution of the present invention: the calculation of the torsion angle θ in the dual laser differential measurement mode adopts the formula:

[0019]

[0020] Among them, I A and I B Respectively represents the reflected light intensity value of the two laser beams (unit: mW / cm 2 ), k1 is the refractive index correction coefficient of the optical fiber material (range: 1.05-1.25).

[0021] As a further solution of the present invention: the switching logic of the adaptive control unit includes:

[0022] Mode upgrade condition: when the torsion angle increment in 5-10 consecutive image frames exceeds the threshold range of 0.5°-2.0°;

[0023] Mode degradation condition: within a preset time window (5-60 seconds), all newly detected torsion angle increments are less than 20% of the current threshold;

[0024] The mode switching process is completed within 100ms, and a data buffer is used to maintain measurement continuity during the switching.

[0025] As a further solution of the present invention, the torsion anomaly weight value W generated by the torsion analysis unit adopts the formula:

[0026]

[0027] Where wi is the position weight coefficient (1.8-2.2 for the fixed end, 1.5-1.8 for the rewinding area, and 1.0 for the straight line segment); θ i is the angle value of the i-th torsion event (unit: degree); β is the angle amplification factor (range: 1.2-1.8); γ is the time attenuation factor (range: -0.05 to -0.02); Δti is the time difference between the event occurrence time and the current time (unit: second).

[0028] As a further solution of the present invention: the assignment rule of the position weight coefficient wi is related to the stress distribution characteristics of the optical fiber: the upper limit coefficient (2.0-2.2) is used in the optical fiber fixed end area because its torsion can easily lead to breakage; the median coefficient (1.5-1.8) is used in the winding and unwinding area because it has periodic tension changes; the reference coefficient (1.0) is used in the straight transmission section; all coefficient values meet the overall normalization constraint.

[0029] As a further solution of the present invention: the alarm execution unit adopts a dynamic threshold trigger mechanism: the basic threshold is proportional to the logarithm of the optical fiber linear velocity, and the actual threshold adds linear compensation of the standard deviation of the historical weight value on the basic threshold; the alarm level is divided into three levels, and when the weight value W>2.5, a red alarm is triggered and the torsion position is synchronously marked.

[0030] As a further solution of the present invention: the workflow of the abnormality diagnosis unit includes:

[0031] (1.1) Construct the current torsion feature vector, including the weight value W, the torsion density per unit length, the maximum torsion angle, and the angle change rate;

[0032] (1.2) Retrieve similar events from the historical database, and calculate the similarity by integrating the cosine similarity of the feature vector and the time decay factor;

[0033] (1.3) Output the probability ranking of the cause, and the probability value is determined by the weighted sum of the similarity and the historical recurrence rate of the cause.

[0034] As a further aspect of the present invention, a real-time processing architecture is also included, wherein:

[0035] The FPGA chip is responsible for performing image differential operations and angle calculations, with processing delays not exceeding 50% of the frame interval;

[0036] The multi-core ARM processor runs an adaptive threshold algorithm, and its background model is updated using an exponentially weighted moving average method;

[0037] The data bus adopts the PCIe3.0 protocol to ensure that the image data transmission rate is higher than 2Gbps.

[0038] As a further solution of the present invention: the implementation of the adaptive threshold algorithm includes:

[0039] (2.1) Initialize the background model to the grayscale value of the first frame image;

[0040] (2.2) Dynamically update the threshold value, which is the local pixel mean plus 3 times the standard deviation;

[0041] (2.3) The twist determination condition is: the difference between the current frame pixel value and the foreground and background models exceeds the threshold for more than 5 frames continuously;

[0042] (2.4) The learning rate is adaptively adjusted according to the fiber movement speed, and the learning rate is increased to 0.3-0.5 at high speed.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. This application achieves low power consumption and high precision through dynamic mode switching. During the startup phase, single laser mode is used, which reduces power consumption by 60% compared to dual laser mode. When the torsion threshold exceeds the threshold, it switches to dual laser differential mode. Combined with the k1 correction factor, it achieves ±0.1° measurement accuracy. Dynamic adjustment of the projection angle ensures consistency across different scenarios.

[0045] 2. The adaptive control unit completes mode switching within 100ms, and the data buffer maintains continuity; the state machine logic is combined with the 5-60 second window threshold to intelligently control mode upgrades and downgrades.

[0046] 3. The torsion analysis unit quantifies anomalies through a weighted formula, integrating position coefficient, angle factor and time attenuation. W>2.5 triggers a level 3 alarm. The dynamic threshold is related to the logarithm of the linear velocity and compensates for the historical standard deviation.

[0047] 4. The abnormality diagnosis unit constructs feature vectors to match historical data, and the probability of cause sorting reduces fault location to minutes. The FPGA and ARM real-time processing architecture ensures 2Gbps data transmission with a delay of less than 50% of the frame interval. The device has normalized weight coefficients for all process steps, dual hardware alarms, dual backup storage, and an MTBF of ≥8,000 hours, ensuring strong engineering adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a structural block diagram of a torsion measurement device for optical fiber production.

[0049] In the figure: 101, laser measurement unit; 102, high-speed imaging unit; 103, adaptive control unit; 104, torsion analysis unit; 105, abnormality diagnosis unit; 106, alarm execution unit; 107, real-time processing architecture. DETAILED DESCRIPTION

[0050] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0051] As mentioned in the background of this application, the inventors discovered that traditional torsion measurement devices suffer from issues such as insufficient measurement accuracy, delayed mode switching, and weak anomaly diagnosis capabilities, making them incapable of meeting the real-time monitoring requirements of high-speed fiber optic production lines. Existing single-laser measurement modes are unable to handle highly dynamic torsion scenarios, while fixed dual-laser modes suffer from excessive power consumption.

[0052] In order to solve the above-mentioned defects, the present application discloses a torsion measurement device for optical fiber production, which can effectively cope with high-dynamic torsion scenarios while ensuring measurement accuracy.

[0053] The following will describe in detail how the solution of this application solves the above technical problems with reference to the accompanying drawings.

[0054] See also Figure 1In an embodiment of the present invention, a torsion measurement device for optical fiber production includes: a laser measurement unit 101, which is equipped with a single laser emission module and a dual laser differential emission module that can be dynamically switched, wherein the single laser mode uses a semiconductor laser with a wavelength of 632nm to 1550nm, and the dual laser mode uses a paired laser with a wavelength difference of less than 20nm; a high-speed imaging unit 102, equipped with a CMOS image sensor with a frame rate of not less than 10kHz and a narrow-band filter, for capturing laser speckle images on the surface of the optical fiber; an adaptive control unit 103, with a built-in state machine logic circuit and a programmable timer, to implement the measurement mode. The torsion measurement device features intelligent switching; the torsion analysis unit 104, integrating a DSP processor and storage media, performs torsion position calibration and weight calculation; the anomaly diagnosis unit 105 connects to a cloud database and local cache to match historical data and rank cause probability; and the alarm execution unit 106 includes an industrial-grade audible and visual alarm and a mechanical position marker. The torsion measurement device's workflow includes: During the fiber startup phase, single-laser low-power measurement mode is activated; when the torsion angle exceeds a preset threshold, the device switches to dual-laser high-precision mode; after operating in high-precision mode for an adjustable period of time, the device returns to single-laser mode if the torsion event disappears. This device integrates six major units: laser measurement, high-speed imaging, and adaptive control, to create a fiber torsion measurement system that combines low power consumption with high precision. The laser measurement unit dynamically switches between single and dual-laser modes to adapt to different torsion scenarios; the high-speed imaging unit captures laser speckle patterns, providing the data foundation for torsion analysis; the adaptive control unit implements intelligent mode switching to balance energy consumption and accuracy; the torsion analysis unit performs angle calculation and anomaly weight assessment; and the anomaly diagnosis and alarm unit provides fault location and early warning. The workflow solves the problems of high power consumption and slow response of traditional devices through closed-loop control of "single laser start-dual laser switching-mode fallback".

[0055] In this embodiment, the operating parameters of the single laser emission module meet the following requirements: the spot diameter does not exceed 1 / 3 of the fiber diameter, and the power density is positively correlated with the fiber transmittance; the acquisition frame rate of the high-speed imaging unit 102 is set to be at least twice the ratio of the fiber's maximum linear velocity to the minimum spatial resolution; and the projection angle of the two laser beams of the dual-laser differential emission module is dynamically adjusted based on the center spacing of the spot and the distance from the fiber to the projection surface. By limiting the matching relationship between the spot diameter of the single laser module (≤ 1 / 3 of the fiber diameter) and the power density, it is ensured that the laser projection does not cover the entire cross-section of the fiber and that energy utilization is maximized; the frame rate of the high-speed imaging unit is set to "linear velocity / resolution × 2" to meet the Nyquist sampling theorem and avoid motion blur; the dual laser module dynamically adjusts the angle based on the spot spacing and projection distance to ensure the clarity of the interference fringes and improve the accuracy of torsion angle calculation.

[0056] In this embodiment, the torsion angle θ is calculated in the dual laser differential measurement mode using the formula:

[0057]

[0058] Among them, I A and I B Respectively represents the reflected light intensity value of the two laser beams (unit: mW / cm 2 ), k1 is the refractive index correction coefficient of the optical fiber material (the range is 1.05-1.25). A -I B ) and light intensity and (I A +I B ) and combined with the fiber material refractive index correction factor k1 to eliminate the interference of material properties on light propagation, achieving quantitative measurement of small fiber twist angles. This formula can achieve an accuracy of ±0.1° in polarization-maintaining fiber testing.

[0059] In this embodiment, the switching logic of the adaptive control unit 103 includes the following: mode upgrade condition: when the torsion angle increment in 5-10 consecutive frames exceeds the threshold range of 0.5°-2.0°; mode downgrade condition: within a preset time window (5-60 seconds), all newly detected torsion angle increments are less than 20% of the current threshold; the mode switching process is completed within 100ms, and a data buffer is used to maintain measurement continuity during the switching. The adaptive control unit triggers mode upgrade based on the "continuous frame angle increment threshold" (such as when the increment exceeds 0.5°-2.0° in 5-10 frames) to capture sudden torsion events; mode downgrade is achieved based on "incremental decay within the time window" (such as when the increment is less than 20% of the threshold within 5-60 seconds) to avoid frequent switching; the 100ms switching time and data buffer design ensure that the measurement data is seamless.

[0060] In this embodiment, the torsional anomaly weight value W generated by the torsional analysis unit 104 adopts the formula:

[0061]

[0062] Where wi is the position weight coefficient (1.8-2.2 for the fixed end, 1.5-1.8 for the rewinding area, and 1.0 for the straight line segment); θ i is the angle value of the i-th torsional event (unit: degrees); β is the angle amplification factor (range: 1.2-1.8); γ is the time decay factor (range: -0.05 to -0.02); and Δti is the time difference between the event occurrence and the current time (unit: seconds). The formula for calculating the torsional anomaly weight W combines the position weight (with higher weights for fixed ends), the angle value, and the time decay factor to convert the spatial location, angle, and duration of the torsional event into a quantitative risk value. This allows the system to prioritize anomaly severity and prioritize high-risk torsional events.

[0063] In this embodiment, the position weight coefficient wi is assigned according to the stress distribution characteristics of the optical fiber: an upper limit coefficient (2.0-2.2) is used in the fixed end region of the optical fiber because torsion can easily lead to breakage; a median coefficient (1.5-1.8) is used in the reeling and unreeling region because of periodic tension variations; and a reference coefficient (1.0) is used in the straight transmission segment. All coefficient values satisfy the overall normalization constraint. The position weight coefficients are assigned according to the stress distribution characteristics of the optical fiber: an upper limit coefficient (2.0-2.2) is used at the fixed end due to its susceptibility to breakage, a median coefficient (1.5-1.8) is used in the reeling and unreeling region due to tension fluctuations, and a reference value of 1.0 is used in the straight transmission segment. The normalization constraint ensures that the weight values are horizontally comparable, enabling the precise location of torsional faults.

[0064] In this embodiment, the alarm execution unit 106 employs a dynamic threshold triggering mechanism: a base threshold is proportional to the logarithm of the fiber's linear velocity, and the actual threshold is a linear compensation of the standard deviation of the historical weight value added to the base threshold. Alarm levels are divided into three levels: when the weight value W > 2.5, a red alarm is triggered and the torsion position is simultaneously marked. The alarm execution unit employs a dynamic threshold mechanism based on the logarithm of the linear velocity × the standard deviation of the historical weight, adapting to the real-time monitoring requirements of high-speed production lines. The three-level alarm system (e.g., W > 2.5 triggers a red alarm) combined with mechanical position marking enables hierarchical fault response and physical location, shortening maintenance time.

[0065] In this embodiment, the abnormality diagnosis unit 105's workflow includes: (1.1) constructing a current torsion feature vector, including the weight value W, the torsion density per unit length, the maximum torsion angle, and the rate of change of the angle; (1.2) searching a historical database for similar events, calculating a similarity that integrates the cosine similarity of the feature vector and the time decay factor; and (1.3) outputting a probability ranking of the causes, where the probability value is determined by the weighted sum of the similarity and the historical recurrence rate of the cause. By constructing a feature vector containing parameters such as the weight value and torsion density, and combining cosine similarity with the time decay factor to search the historical database, the abnormality diagnosis unit outputs a probability ranking of the causes (e.g., a 71% probability of abnormal winding tension), thus upgrading fault location from manual troubleshooting to data-driven intelligent diagnosis.

[0066] This embodiment also includes a real-time processing architecture 107, in which: an FPGA chip is responsible for performing image difference operations and angle calculations, with processing delays not exceeding 50% of the frame interval; a multi-core ARM processor runs an adaptive threshold algorithm, which uses an exponentially weighted moving average method to update the background model; and a data bus uses the PCIe 3.0 protocol to ensure image data transmission rates exceeding 2 Gbps. The real-time processing architecture utilizes a heterogeneous computing model of FPGA and ARM: the FPGA is responsible for image difference and angle calculations (with delays less than 50% of the frame interval), meeting high-speed processing requirements; the ARM runs an adaptive threshold algorithm (exponentially weighted background model updates), improving interference resistance in complex scenes; and the PCIe 3.0 bus ensures data transmission rates exceeding 2 Gbps, preventing image frame dropouts.

[0067] In this embodiment, the implementation of the adaptive threshold algorithm includes the following steps: (2.1) initializing the background model to the grayscale value of the first frame; (2.2) dynamically updating the threshold to the local pixel mean plus three standard deviations; (2.3) determining a torsion: the difference between the current frame pixel value and the previous background model exceeds the threshold for more than five consecutive frames; and (2.4) adaptively adjusting the learning rate based on the fiber's motion speed, increasing the learning rate to 0.3-0.5 at high speeds. The adaptive threshold algorithm dynamically updates the detection threshold using the "local mean + three standard deviations" method, combining this with the five-frame threshold-exceeding determination rule to reduce false alarms. The learning rate adaptively adjusts with fiber speed (0.3-0.5 at high speeds), enabling the background model to quickly adapt to changes in motion and improving the real-time detection capability of torsion events.

[0068] In order to further illustrate the present invention, a torsion measuring device for optical fiber production provided by the present invention is described in detail below with reference to embodiments.

[0069] Example 1: Single laser low power measurement mode application

[0070] On a certain G.652 optical fiber drawing production line, the device automatically enters single laser measurement mode after startup. At this point, the single laser emission module uses a semiconductor laser with a wavelength of 650nm, adjusts the spot diameter to 1 / 4 of the fiber diameter (approximately 5μm), and sets the power density to 12mW / cm based on the fiber transmittance (0.92). 2 The frame rate of the CMOS sensor of the high-speed imaging unit 102 is set to 15 kHz (fiber linear velocity 10 m / s, minimum spatial resolution 5 μm, ratio 2000, 2 times or more is 4000, and 15 kHz is actually used), and the center wavelength of the narrowband filter matches the laser wavelength.

[0071] After the initial tension of the fiber stabilizes during startup, the adaptive control unit 103 continuously monitors the laser speckle pattern output by the high-speed imaging unit. If the torsion angle increment is less than 0.5° for 10 consecutive image frames, the device maintains single-laser mode. This mode reduces system power consumption by 60% compared to dual-laser mode, making it suitable for routine monitoring of straight fiber transmission segments.

[0072] Example 2: Dual laser high-precision mode and abnormality diagnosis

[0073] During the winding and unwinding process of a specialized polarization-maintaining optical fiber, when the fiber torsion angle exceeds a preset threshold (1.5°), the adaptive control unit 103 triggers a mode switch: the dual-laser differential transmission module activates, selecting paired lasers with wavelengths of 1310nm and 1320nm. The projection angle is dynamically adjusted to 0.34° based on the center-to-center spacing of the beam spots (0.3mm) and the distance from the fiber to the projection surface (50mm). The high-speed imaging unit 102 increases its frame rate to 20kHz to ensure capture of high-frequency torsion signals.

[0074] Torsion analysis unit 104 uses the formula Calculate the torsion angle, where k1 is 1.15 (optical fiber material refractive index correction coefficient), I A =28mW / cm 2 , I B =22mW / cm 2 , θ = 0.138°. Meanwhile, the position weight coefficient is set to 1.7 (for the reeling area), the angle magnification factor is 1.5, and the time decay factor is -0.03. If the angle of a torsional event is 3.2° and it occurs 10 seconds ago, the weight W ≈ 6.04.

[0075] When W > 2.5, the alarm execution unit 106 triggers a red alarm, and the mechanical position marking device sprays a fluorescent marker on the corresponding position on the optical fiber. The abnormality diagnosis unit 105 constructs a feature vector [6.04, 0.8 pieces / m, 3.2°, 0.5° / s]. Searching the historical database, it finds an 85% similarity with a torsion event caused by winding tension fluctuations three months prior. Combined with the historical recurrence rate of this cause (32%), the output probability ranking of the causes is: abnormal winding tension (71% probability) > guide wheel wear (21%) > laser drift (8%).

[0076] In the real-time processing architecture 107, the delay of the FPGA chip to complete the image differential operation and angle calculation is 40μs (80% of the frame interval of 50μs), and the multi-core ARM processor uses the exponentially weighted moving average method (learning rate 0.4) to update the background model. The PCIe3.0 data bus transmission rate is maintained at 2.5Gbps to ensure high-speed data transmission without frame loss.

[0077] This invention achieves low power consumption and high precision through dynamic mode switching. During the startup phase, it uses a single laser mode, which reduces power consumption by 60% compared to the dual laser mode. When the torsion exceeds the threshold, it switches to the dual laser differential mode. This, combined with the k1 correction factor, achieves ±0.1° measurement accuracy. Dynamic adjustment of the projection angle ensures consistency across different scenarios. The adaptive control unit completes mode switching within 100ms, maintaining data buffer continuity. State machine logic, combined with a 5-60 second window threshold, intelligently controls mode upgrades and downgrades. Furthermore, the torsion analysis unit quantifies anomalies using a weighted formula, integrating position coefficients, angle factors, and time decay. A level 3 alarm is triggered when W > 2.5. The dynamic threshold is related to the logarithm of the linear velocity and compensates for historical standard deviations. The anomaly diagnosis unit constructs a feature vector to match historical data, and the probability of cause ranking reduces fault location to the minute level. The FPGA and ARM real-time processing architecture ensures 2Gbps data transmission with a latency less than 50% of the frame interval. The device features normalized weight coefficients for all process steps, dual hardware alarms, and dual backup storage, with an MTBF of ≥ 8,000 hours, ensuring strong engineering adaptability.

[0078] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

[0079] The above is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solutions and inventive concepts of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A torsion measuring device for optical fiber production, characterized in that: include: The laser measurement unit (101) is configured with a single laser emission module and a dual laser differential emission module that can be dynamically switched, wherein the single laser mode uses a semiconductor laser with a wavelength of 632nm to 1550nm, and the dual laser mode uses a pair of lasers with a wavelength difference of less than 20nm; A high-speed imaging unit (102) equipped with a CMOS image sensor with a frame rate of not less than 10 kHz and a narrow-band filter for capturing a laser speckle image on the surface of the optical fiber; An adaptive control unit (103) having a built-in state machine logic circuit and a programmable timer to realize intelligent switching of measurement modes; a torsion analysis unit (104), integrating a DSP processor and a storage medium, and performing torsion position calibration and weight calculation; An abnormality diagnosis unit (105), connected to the cloud database and local cache, realizes historical data matching and cause probability sorting; an alarm execution unit (106), comprising an industrial-grade sound and light alarm and a mechanical position marking device; The working process of the torsion measuring device includes: During the fiber startup phase, single laser low power measurement mode is enabled; When the torsion angle is detected to exceed the preset threshold, it switches to the dual laser high-precision mode; After running in high-precision mode for an adjustable period of time, it returns to single-laser mode if the torsional event disappears.

2. The torsion measuring device for optical fiber production according to claim 1, characterized in that: The operating parameters of the single laser emission module meet the following requirements: the spot diameter does not exceed 1 / 3 of the optical fiber diameter, and the power density is positively correlated with the optical fiber transmittance; the acquisition frame rate of the high-speed imaging unit (102) is set to be more than twice the ratio of the maximum linear velocity of the optical fiber to the minimum spatial resolution; and the projection angle of the two laser beams of the dual laser differential emission module is dynamically adjusted according to the center spacing of the spot and the distance from the optical fiber to the projection surface.

3. The torsion measuring device for optical fiber production according to claim 2, characterized in that: The calculation formula of the torsion angle θ in the dual laser differential measurement mode is: Among them, I A and I B They represent the reflected light intensity values of the two laser beams respectively, and k1 is the refractive index correction coefficient of the optical fiber material.

4. The torsion measuring device for optical fiber production according to claim 3, characterized in that: The switching logic of the adaptive control unit (103) includes: Mode upgrade condition: when the torsion angle increment in 5-10 consecutive image frames exceeds the threshold range of 0.5°-2.0°; Mode degradation condition: within the preset time window, all newly detected torsion angle increments are less than 20% of the current threshold; The mode switching process is completed within 100ms, and a data buffer is used to maintain measurement continuity during the switching.

5. The torsion measuring device for optical fiber production according to claim 4, characterized in that: The torsion anomaly weight value W generated by the torsion analysis unit (104) adopts the formula: Among them, wi is the position weight coefficient; θ i is the angle value of the i-th torsion event; β is the angle amplification factor; γ is the time attenuation factor; Δti is the time difference between the event occurrence time and the current time.

6. The torsion measuring device for optical fiber production according to claim 5, characterized in that: The assignment rule of the position weight coefficient wi is related to the stress distribution characteristics of the optical fiber: an upper limit coefficient of 2.0-2.2 is used in the optical fiber fixed end area because its torsion can easily lead to breakage; a median coefficient of 1.5-1.8 is used in the winding and unwinding area because it has periodic tension changes; a reference coefficient of 1.0 is used in the straight transmission section; all coefficient values meet the overall normalization constraint.

7. The torsion measuring device for optical fiber production according to claim 6, characterized in that: The alarm execution unit (106) adopts a dynamic threshold trigger mechanism: the basic threshold is proportional to the logarithm of the optical fiber linear velocity, and the actual threshold is a linear compensation of the standard deviation of the historical weight value added to the basic threshold; the alarm level is divided into three levels, and when the weight value W>2.5, a red alarm is triggered and the torsion position is marked synchronously.

8. The torsion measuring device for optical fiber production according to claim 7, characterized in that: The workflow of the abnormality diagnosis unit (105) includes: (1.1) Construct the current torsion feature vector, including the weight value W, the torsion density per unit length, the maximum torsion angle, and the angle change rate; (1.2) Retrieve similar events from the historical database, and calculate the similarity by integrating the cosine similarity of the feature vector and the time decay factor; (1.3) Output the probability ranking of the cause, and the probability value is determined by the weighted sum of the similarity and the historical recurrence rate of the cause.

9. The torsion measuring device for optical fiber production according to claim 8, characterized in that: Also included is a real-time processing architecture (107), wherein: The FPGA chip is responsible for performing image differential operations and angle calculations, with processing delays not exceeding 50% of the frame interval; The multi-core ARM processor runs an adaptive threshold algorithm, and its background model is updated using an exponentially weighted moving average method; The data bus adopts the PCIe3.0 protocol to ensure that the image data transmission rate is higher than 2Gbps.

10. The torsion measuring device for optical fiber production according to claim 9, characterized in that: The implementation of the adaptive threshold algorithm includes: (2.1) Initialize the background model to the grayscale value of the first frame image; (2.2) Dynamically update the threshold value, which is the local pixel mean plus 3 times the standard deviation; (2.3) The twist determination condition is: the difference between the current frame pixel value and the foreground and background models exceeds the threshold for more than 5 frames continuously; (2.4) The learning rate is adaptively adjusted according to the fiber movement speed, and the learning rate is increased to 0.3-0.5 at high speed.

Images