Online detection method, equipment and system for polyethylene fiber yarn

Abstract

The invention relates to an online detection method, equipment and system for polyethylene fiber yarns, and belongs to the technical field of intelligent manufacturing of high-performance fibers. The equipment comprises a yarn color detection module, a surface defect detection module, an optical coherence tomography (OCT) module and an ultrasonic diameter detection module, and can perform multi-dimensional detection on appearance and internal structure defects of yarns in the production process. By constructing a process parameter optimization model based on defect indexes and diameter errors, the system can dynamically adjust key process parameters such as the melting temperature, the high shear rate and the cooling rate according to real-time detection data, and closed-loop control is achieved. Process parameter updating adopts joint optimization of a sensitivity matrix and a weight matrix, and a parameter constraint mechanism is arranged to prevent border crossing and system oscillation. The whole system can improve the quality consistency and manufacturing stability of polyethylene fiber yarns, and is suitable for continuous and intelligent production scenes.

Description

Technical Field

[0001] This invention belongs to the field of online yarn detection technology, specifically a method, equipment, and system for online detection of polyethylene fiber yarn. Background Technology

[0002] With the widespread application of high-performance fiber materials in aerospace, protective equipment, and high-end manufacturing, higher requirements are being placed on the quality stability and consistency of fiber products. Polyethylene fiber (such as ultra-high molecular weight polyethylene fiber, UHMWPE), as an engineering fiber with excellent properties such as high strength, high modulus, and corrosion resistance, occupies an important position in industrial applications.

[0003] The production process of polyethylene fiber typically involves multiple steps, including melt spinning, stretching and setting, and cooling molding. In actual production, fluctuations in process parameters such as melt temperature, draw ratio, and cooling rate can easily lead to surface defects, bubbles, and uneven fiber thickness, thus affecting its mechanical properties and service life.

[0004] Traditional quality control methods rely heavily on manual sampling inspection or single online sensing technology, which cannot achieve full-process, full-size, and real-time defect detection of fibers, and lack effective process parameter feedback and adjustment mechanisms, making it difficult to meet the needs of intelligent manufacturing.

[0005] Therefore, there is an urgent need for an online detection system for polyethylene fibers that integrates multimodal detection technology and adaptive optimization algorithms. This system should be able to comprehensively detect the color, surface defects, internal structure, and diameter changes of the fibers, and dynamically optimize the production process based on the detection results. This would enable closed-loop control and refined management, thereby improving the consistency of fiber products and the stability of the manufacturing process. Summary of the Invention

[0006] This invention designs an online detection method, equipment, and system for polyethylene fiber yarn, which can effectively ensure the quality of polyethylene fiber yarn during the production and manufacturing process.

[0007] To achieve the above objectives, the present invention provides the following technical solution: An online detection device for polyethylene fiber yarn includes an upper housing, and a yarn color detection module is installed on the inlet of the upper housing; The upper housing is equipped with a first guide wheel and a second guide wheel. There is also a cavity in the upper housing with a first wire hole and a second wire hole. A surface defect detection module, including an industrial camera, an LED bar light source, and a coding machine, is installed inside. The partition separates the upper and lower boxes and has a third wire hole. A support plate is installed in the lower housing, on which the third and fourth guide wheels, the first and second reversing wheels, and the tensioning wheel are installed. The ultrasonic diameter detection module and the optical coherence tomography (OCT) module are also installed on it. An online detection method for polyethylene fiber yarns, based on the aforementioned equipment, includes the following steps: Adjust the wavelength of the LED strip light source to determine the yarn color; It bypasses the second guide wheel and passes through the first wire hole; The surface defect detection module performs the detection. If it passes the test, continue operation; If the result is unsatisfactory, the defect data will be fed back to the data processing module. The data processing module matches the corresponding process parameters in the database based on the defect type. according to The model calculates the defect index, where, For the first One process parameter, This is the optimal value for the parameter. , , The fitting coefficients are determined through regression analysis of experimental data. This is the random error term; Then optimize the data model based on process parameters.

[0008] In the formula, For process parameters The adjustment amount, For learning rate, For the first One defect indicator, For defects For parameters Sensitivity (partial derivative) This represents the actual detected defect value. The target defect value; If multiple defects exist, then according to... Update process parameters, where Adjustment vector for process parameters , This is the adjustment value for the melting temperature. This is an adjustment value for high shear rates. For cooling rate, Defect deviation vector , Bubble density, Target value for bubble density For surface roughness, For the target roughness, This is a weighted diagonal matrix; Sensitivity matrix ; The new process parameters are determined using the following formula:

[0009] For parameter adjustment amount; These are the latest process parameter values; These are the current process parameter values; , These are the minimum and maximum values ​​of the process parameters; If the process parameter values ​​are unreasonable, a secondary update is performed using the parameter constraint model, which is as follows:

[0010]

[0011] It is an adaptive adjustment factor that controls the adjustment intensity and prevents oscillation; It refers to the unconstrained adjustment of process parameters; It is an adjustment coefficient, used to control the adjustment intensity; When the error is small It allows for larger adjustments while ensuring rapid convergence; When the error is large Reduce the adjustment range to avoid instability; The high-polyethylene fiber yarn passes through the inlet hole on the partition and successively around the tensioning wheel, the first reversing wheel, and the second reversing wheel, and then passes through the optical coherence tomography (OCT) module; If the test is passed, operation will continue; If the result is not satisfactory, a low-resolution scan based on polyethylene fibers will be performed to locate the suspected defective area. Dense sampling is performed in the defect area, and multi-angle scanning is carried out simultaneously. The fiber is rotated and scanned repeatedly at 30° and 60° to enhance the accuracy of the three-dimensional reconstruction of the defect. The obtained data is fed back to the data processing module, which then processes it. The acquired slice data is subjected to Fourier transform to generate a 2D image; The generated 2D image is denoised using mean filtering, and then enhanced using histogram equalization. The obtained images are matched with defect data in the database to determine the type of internal defects; Match relevant process parameters according to the type of defect; Defect indexes are calculated using process parameters: ; In the formula, For the first One process parameter, This is the optimal value for the parameter. , , The fitting coefficients are determined through regression analysis of experimental data. This is the random error term; Then optimize the model based on process parameters.

[0012] In the formula, For process parameters The adjustment amount, For learning rate, For the first One defect indicator, For defects For parameters Sensitivity (partial derivative) This represents the actual detected defect value. The target defect value; If multiple defects exist, then according to... Update process parameters, where Adjustment vector for process parameters , This is the adjustment value for the melting temperature. This is an adjustment value for high shear rates. For cooling rate, Defect deviation vector , Bubble density, Target value for bubble density For surface roughness, For the target roughness, This is a weighted diagonal matrix; Sensitivity matrix ; The new process parameters are determined using the following formula:

[0013] For parameter adjustment amount; These are the latest process parameter values; These are the current process parameter values; , These are the minimum and maximum values ​​of the process parameters; If the process parameter values ​​are unreasonable, a secondary update is performed using the parameter constraint model, which is as follows:

[0014]

[0015] It is an adaptive adjustment factor that controls the adjustment intensity and prevents oscillation; It refers to the unconstrained adjustment of process parameters; It is an adjustment coefficient, used to control the adjustment intensity; When the error is small It allows for larger adjustments while ensuring rapid convergence; When the error is large Reduce the adjustment range to avoid instability; Control the inkjet printer to mark the location of defects; The high-polyethylene fiber yarn passes around the third guide wheel and the fourth guide wheel; Using an ultrasonic diameter detection module; The ultrasonic probe emits a high-frequency ultrasonic beam, which passes through polyethylene fiber and reaches the receiving end. Ultrasonic waves are partially reflected at the fiber-coupled medium interface, and partially penetrate to the other side. The receiving probe on the other side captures the transmitted signal, while the transmitting probe receives the reflected signal. According to the formula: , The speed of propagation of ultrasound. and This represents the propagation time of the ultrasonic wave to both interfaces. For error compensation; If diameter error rate If the set qualified threshold is exceeded, The error rate formula is as follows, which is fed back to the data processing module. , This is the optimal diameter.

[0016] Then, the parameters are optimized based on the process parameter optimization model, using the following formula:

[0017] Process parameter adjustment values Learning rate Sensitivity of diameter to process parameters : Measuring diameter Optimal diameter The new process parameters are determined using the following formula:

[0018] For parameter adjustment amount; These are the latest process parameter values; These are the current process parameter values; , These are the minimum and maximum values ​​of the process parameters; If the process parameter values ​​are unreasonable, a secondary update is performed using the parameter constraint model, which is as follows:

[0019]

[0020] It is an adaptive adjustment factor that controls the adjustment intensity and prevents oscillation; It refers to the unconstrained adjustment of process parameters; It is an adjustment coefficient, used to control the adjustment intensity; When the error is small It allows for larger adjustments while ensuring rapid convergence; When the error is large Reduce the adjustment range to avoid instability; Control the inkjet printer to mark the location of defects; Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall structure of the online detection device of the present invention. Figure 2 This is a schematic diagram of the online detection method of the present invention. Detailed Implementation

[0022] This invention relates to an online detection method, equipment, and system for polyethylene fiber yarns, which is particularly suitable for intelligent quality control in the production process of high-performance fibers.

[0023] The specific implementation method is as follows: An online testing device for polyethylene fiber yarn includes an upper chamber and a lower chamber.

[0024] A yarn color detection module is installed at the inlet of the upper housing. This module uses a color sensor to detect whether the yarn color conforms to a preset range. Based on the detection results, the emission wavelength of the LED strip light source is automatically adjusted to achieve optimal illumination and improve the quality of subsequent image acquisition.

[0025] The upper housing contains a first guide wheel and a second guide wheel to guide the yarn into the internal cavity. Inside the cavity is a surface defect detection module, including an industrial camera, an LED bar light source, and a coding machine. This module detects defects by performing high-speed imaging of the yarn surface and using image processing algorithms.

[0026] If the test results are satisfactory, the yarn continues to run; if they are unsatisfactory, the defect image and location are recorded and sent to the data processing module, and the inkjet printer marks the corresponding location.

[0027] The data processing module retrieves relevant parameters from the process parameter database based on the defect type and calculates them using the following defect index model: according to The model calculates the defect index, where, For the first One process parameter, This is the optimal value for the parameter. , , The fitting coefficients are determined through regression analysis of experimental data. This is the random error term; Then optimize the data model based on process parameters.

[0028] In the formula, For process parameters The adjustment amount, For learning rate, For the first One defect indicator, For defects For parameters Sensitivity (partial derivative) This represents the actual detected defect value. The target defect value; If multiple defects exist, then according to... Update process parameters, where Adjustment vector for process parameters , This is the adjustment value for the melting temperature. This is an adjustment value for high shear rates. For cooling rate, Defect deviation vector , Bubble density, Target value for bubble density For surface roughness, For the target roughness, This is a weighted diagonal matrix; Sensitivity matrix ; The new process parameters are determined using the following formula: , For parameter adjustment amount, These are the latest process parameter values. These are the current process parameter values. , These are the minimum and maximum values ​​of the process parameters. If the process parameter values ​​are unreasonable, a secondary update is performed using the parameter constraint model, as follows.

[0029] , It is an adaptive adjustment factor that controls the adjustment intensity and prevents oscillations. It refers to the unconstrained adjustment of process parameters. It is an adjustment coefficient, used to control the adjustment intensity; This system also includes an optical coherence tomography (OCT) module. Once the surface defects are deemed acceptable, the yarn continues through the tensioning and reversing rollers before entering the OCT module for internal structure scanning. If the detection fails, a low-resolution scan is first used to initially locate the defect area, followed by repeated 30° and 60° multi-angle scans to improve the accuracy of the 3D reconstruction. The resulting sliced ​​images are converted into 2D frequency domain images using Fourier transform, and mean filtering and histogram equalization are employed to enhance image clarity. The final image is then matched with typical defect templates in the database to determine the type of internal defect, and process parameters are further optimized accordingly.

[0030] The yarn then passes through the third and fourth guide rollers and enters the ultrasonic diameter detection module. This module acquires fiber diameter information by transmitting and receiving signals through an ultrasonic probe.

[0031] If diameter error rate If the set qualified threshold is exceeded, The error rate formula is as follows, which is fed back to the data processing module. , This is the optimal diameter.

[0032] Then, the parameters are optimized based on the process parameter optimization model, using the following formula: , Process parameter adjustment values, Learning rate The sensitivity of diameter to process parameters. : Measure the diameter Optimal diameter. The formula for determining whether process parameters exceed limits is as follows: , For parameter adjustment amount, These are the latest process parameter values. These are the current process parameter values. , These represent the minimum and maximum values ​​of the process parameters. If a process parameter exceeds these limits, a secondary update is performed using the parameter constraint model, which is as follows: ,

[0033] It is an adaptive adjustment factor that controls the adjustment intensity and prevents oscillations. It refers to the unconstrained adjustment of process parameters. It is an adjustment coefficient used to control the adjustment intensity if it exceeds the range.

[0034] The entire system enables online detection and closed-loop process optimization of polyethylene fibers, improving product consistency and production efficiency, and is suitable for industrial-grade intelligent manufacturing platforms.

Claims

1. An online testing device for polyethylene fibers, characterized by: (1) Upper box, (2) First guide wheel, (12) Second guide wheel, fixed on the upper box, the image detection module includes: (14) Industrial camera fixed in the closed space of the upper box, with LED strip light source installed inside to provide suitable light source conditions, (15) Inkjet printer. (11) Partition can separate the upper box and (4) lower box, the lower box is equipped with support plate (15), on which (5) tension wheel and (3) first reversing wheel, (8) second reversing wheel, (10) third guide wheel, (9) Optical coherence tomography (OCT) module, (7) fourth guide wheel, (6) ultrasonic diameter detection module, (17) industrial camera; also includes data processing module (16) for receiving defect data from fiber surface defect detection module, optical coherence tomography (OCT) module and ultrasonic diameter detection module, calculating the defect generation rate every once in a period of time, and adjusting process parameters to reduce defect generation if the defect rate exceeds the threshold.

2. The online detection device for polyethylene fibers according to claim 1, characterized in that, Includes the following steps: The dimmable module uses an industrial camera and a data processor to adjust the wavelength of the strip LED light source based on the color of the polyethylene fiber; the defect detection module detects the image; if it is unqualified, the information is fed back to the data processor, the data processing module performs statistical analysis on the data, and controls the inkjet printer to mark the defective parts; If the result is satisfactory, the process continues. The optical coherence tomography (OCT) module is used to determine internal structural defects in the fiber. If the result is satisfactory, the process continues. If the result is unsatisfactory, the defect location is marked using an inkjet printer. The ultrasonic diameter detection module is used to determine if the diameter meets the standard. If the result is satisfactory, the process continues. If the result is unsatisfactory, the non-compliant location is marked.

3. The online detection method for polyethylene fibers according to claim 2, characterized in that, The dimmable module adjusts the light source, including the following steps: The industrial camera (17) sends the yarn color data to the data processing module (16), which then adjusts the wavelength of the strip LED light source (13) in the dimmable module to create the most suitable lighting conditions for the defect detection module.

4. The online detection method for polyethylene fibers according to claim 2, characterized in that, The defect detection module determines whether the acquired image is acceptable based on image data, including the following steps: Acquire images of polyethylene fiber yarn before defect detection is performed by the online inspection equipment. The image is fed into the defect detection module, which uses a convolutional neural network to detect defects in the image. If the test is passed, operation can continue; If the defect is not met, the defect signal is transmitted to the data processing module, which performs data statistics and controls the inkjet printer (15) to mark the defect.

5. The online detection method for polyethylene fibers according to claim 2, characterized in that, Optical coherence tomography (OCT) modules are used to detect internal defects in fibers, including the following steps: Low-resolution scanning based on polyethylene fiber is used to locate suspected defect areas. Dense sampling is performed in the defect area, and multi-angle scanning is carried out simultaneously. The fiber is rotated and scanned repeatedly at 30° and 60° to enhance the accuracy of the three-dimensional reconstruction of the defect. The obtained data is fed back to the data processing module, which then processes it. The acquired slice data is subjected to Fourier transform to generate a 2D image; The generated 2D image is denoised using mean filtering, and then enhanced using histogram equalization. If the image detection is successful, continue the process; If the result is not satisfactory, the data will be fed back to the data processing module, and the inkjet printer (15) will be controlled to mark the defect.

6. The online detection method for polyethylene fibers according to claim 2, characterized in that, The ultrasonic diameter detection module detects the diameter of polyethylene fibers, including the following steps: An ultrasonic probe emits a high-frequency ultrasonic beam, which passes through a polyethylene fiber and reaches the receiving end. At the fiber-coupling medium interface, the ultrasonic waves are partially reflected, and partially penetrate to the other side. The receiving probe on the other side captures the transmitted signal, while the transmitting probe receives the reflected signal. According to the formula: , The speed at which ultrasound propagates. and This represents the propagation time of the ultrasonic wave to both interfaces. For error compensation; if the diameter error rate If the set qualified threshold is exceeded, The error rate formula is as follows, which is fed back to the data processing module. , This is the optimal diameter.

7. The online detection method for polyethylene fibers according to any one of claims 4-6, characterized in that, The data processing module processes data, including the following steps: If there are defects on the fiber surface, in the fiber diameter error rate, or inside the fiber, the defect information will be fed back to the data processing module. The data processing module performs statistical analysis based on the average value of defect indicators over a period of time, and optimizes the process parameters through a parameter optimization model. Based on the process parameters and relevant fitting coefficients, the defect index can be obtained, as shown below: In the formula, For the first One process parameter, This is the optimal value for the parameter. , , The fitting coefficients are determined through regression analysis of experimental data. This is the random error term; The process parameter optimization model is shown below: In the formula, For process parameters The adjustment amount, For learning rate, For the first One defect indicator, For defects For parameters Sensitivity (partial derivative) This represents the actual detected defect value. The target defect value; when or hour; To ensure stable system operation, constraints will be applied to the process parameters. The constraint model is as follows: ; ; It is an adaptive adjustment factor that controls the adjustment intensity and prevents oscillation; It refers to the unconstrained adjustment of process parameters; It is an adjustment coefficient used to control the adjustment level; when the error is small, It allows for larger adjustments to ensure rapid convergence; however, when the error is large... Reduce the adjustment range to avoid instability.

Images