Method for analyzing the match between the weavability and fiber length of recycled carbon fiber yarns

Abstract

The application provides a method for analyzing the match between the weavability and fiber length of regenerated carbon fiber yarns, comprising: calculating the length unevenness of the target batch of regenerated carbon fibers and the proportion of the bifurcated section in the full length of the fibers according to the depth of the broom-shaped bifurcation and the length of the single fiber, to obtain an end bifurcation length characteristic set; obtaining historical broom-shaped bifurcation depth records in each speed interval, screening the speed interval with a broom-shaped bifurcation depth not exceeding the upper limit of the target depth from the historical broom-shaped bifurcation depth records, and determining the target beater speed; adjusting the hybrid opening production line according to the target beater speed, obtaining the stretch-breaking data of the slivers obtained after the adjusted carding, and determining the critical draft multiple according to the stretch-breaking data of the slivers.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for analyzing the weavability and fiber length matching of recycled carbon fiber yarn. Background Technology

[0002] With the widespread application of carbon fiber composites in aerospace, wind turbine blades, and automotive lightweighting, the output of waste composite materials is increasing year by year. Recycled carbon fiber, obtained through pyrolysis, solvent decomposition, or mechanical stripping, has become an important source of reducing the overall energy consumption and cost of the carbon fiber industry. However, compared to virgin carbon fiber, which has uniform length specifications and flat end face morphology, recycled carbon fiber naturally exhibits a variety of original lengths due to diverse recycling processes, different original composite material structures, and random stress during stripping. A single batch may contain both long fibers (tens of millimeters) and short fibers (only a few millimeters). To ensure that recycled carbon fiber can smoothly enter the spinning and weaving process, the fibers need to be dispersed and mixed during the mixing and opening stage. The beater speed is a key process parameter in the opening stage. When the beater speed increases, the fiber ends are subjected to stronger impact and rubbing, gradually evolving from a relatively flat break into a multi-filament forked structure similar to a broom. This forked morphology increases the lateral friction and mechanical interlocking between fibers in the carding web, enhancing the overall cohesion of the yarn and thus improving its tensile strength to a certain extent. However, this originally beneficial morphological change of end bifurcation, under the reality of uneven length distribution in recycled carbon fibers, results in a situation where the absolute length of the end bifurcation is roughly similar at a certain beater speed, but the proportion of it on fibers of different lengths varies significantly. For longer fibers, the bifurcation segment only accounts for a small portion of its total length, and the fiber body still retains a relatively long solid segment, which can be stably held and smoothly transported by the front and rear rollers during drafting. But for shorter fibers, the bifurcation segment of the same length may occupy a considerable proportion of its total length, and the remaining solid portion is insufficient to cross the gripping distance between the front and rear rollers in the drafting zone, causing these short fibers to be unable to be effectively clamped by the rollers during drafting. After entering the drafting zone, the ungripped short fibers will deviate from the normal fiber movement trajectory, forming an uncontrolled slippage and floating state within the yarn. These free fibers not only cannot bear the drafting tension themselves, but also disrupt the continuous arrangement structure between surrounding fibers, causing the collapse of local cohesion. When the number of free fibers accumulates to a certain level, the tensile strength of the yarn will fluctuate significantly or even break unexpectedly, making it difficult to operate stably in subsequent weaving processes. Currently, most process control methods for fiber drafting stability follow the approach of setting globally fixed parameters for virgin carbon fibers or conventional chemical fibers, applying uniform beater speed and drafting conditions to the entire batch of fibers, ignoring the differentiated physical effects of the microscopic morphological changes of recycled carbon fibers on fiber groups of different lengths.In the prior art, a method and equipment for carding recycled carbon fiber short fibers into slivers, disclosed in CN118957828A, employs an opener, a feeder, and a carding machine to perform multi-machine graded opening and carding of recycled carbon fiber short fibers into slivers. This disclosed technology primarily reduces carding damage and improves fiber transfer capability through the shape of the carding cloth teeth, spacing, and airflow transfer structure. However, it still focuses on overall damage control and transfer efficiency in the carding process, failing to identify the matching relationship between the end-branching ratio of the recycled carbon fiber after opening and the fiber's own length. It also cannot determine whether short fibers will lose roller grip in the subsequent drafting zone due to insufficient solid segments. This extensive process control method creates a direct conflict between the strategy of increasing beater speed to enhance cohesion and the need to avoid short fibers losing grip due to excessive branching, making it difficult to simultaneously address the risks of overall cohesion and local drafting failure. Therefore, given the reality of varying fiber lengths, accurately identifying short fiber groups that are about to lose their grip in the drafting zone due to their excessively high proportion of bifurcation segments in their length has become a key issue that must be addressed in achieving stable matching of the weavability of recycled carbon fiber yarns. Summary of the Invention

[0003] This invention provides a method for analyzing the weavability and fiber length matching of recycled carbon fiber yarn, the method comprising: Acquire end morphology images and full-length images of single fibers of the target batch of recycled carbon fibers from the mixed opening production line, identify the end broom-like bifurcation contour and the fracture ends of single fibers, and extract the broom-like bifurcation depth and the cut length of single fibers. Based on the broom-like bifurcation depth and the cut-off length of a single fiber, the degree of length unevenness and the proportion of the bifurcation segment to the total fiber length are calculated to obtain the end bifurcation cut-off feature set. Identify single fibers in the end-bidding truncation feature set whose bifurcation segment accounts for more than a preset upper limit of the fiber length and cannot be held by the rollers in the stretching zone, and obtain a floating fiber group. The transverse cohesive friction force of the floating fiber group in the carding web is obtained, and the fiber strength of the floating fiber group is determined by comparing the measured value of the transverse cohesive friction force with the measured value of the length non-uniformity with a preset strength reference range. By screening the planktonic fiber population whose fiber strength is lower than the preset lower limit, a weak-strength planktonic fiber population is obtained. The upper limit of the allowable depth of the broom-shaped branching is dynamically adjusted according to the degree of length unevenness to obtain the target depth upper limit. Obtain historical broom-shaped fork depth records for each speed range, filter out speed ranges where the broom-shaped fork depth does not exceed the upper limit of the target depth from the historical broom-shaped fork depth records, and determine the target beater speed; Adjust the mixing and opening production line according to the target beater speed, and determine the critical draw ratio based on the tensile breaking data of the strips obtained after the adjustment and combing.

[0004] Preferably, the step of acquiring end morphology images and full-length images of a single fiber of the target batch of recycled carbon fiber from the mixed opening production line, identifying the broom-like bifurcation contour and the fracture ends of a single fiber, and extracting the broom-like bifurcation depth and the cut length of a single fiber includes: A linear industrial camera, in conjunction with backlighting, acquires a magnified image of the end region of a single fiber. An area-array industrial camera acquires a full-length image of the single fiber covering both ends of the fiber in a flat manner along the fiber spreading direction. A one-to-one correspondence between the two types of images is established according to the fiber number, resulting in an indexed set of fiber dual-view images.

[0005] Preferably, obtaining the indexed fiber dual-view set includes: The end morphology images in the indexed fiber dual-view set are segmented using the Otsu thresholding method to separate the fiber body and the broom-shaped bifurcation region. The binary image is refined to obtain the end skeleton. The bifurcation start point is marked at the bifurcation point of the skeleton. The connected region from the bifurcation start point to the farthest thin branch end point of the tip is determined as the broom-shaped bifurcation profile. Search for the locations of abrupt grayscale changes at both ends along the fiber principal axis of the full-length image of a single fiber as the locations of the break points at both ends; The axial projection distance between the bifurcation start point and the farthest thin branch end point is measured as the broom-shaped bifurcation depth, and the pixel span between the two ends of the break is converted into the single fiber cut length according to the camera calibration scale.

[0006] Preferably, the step of calculating the degree of length unevenness and the proportion of the forked segment to the total fiber length based on the broom-like bifurcation depth and the cut-off length of the single fiber to obtain the end bifurcation cut-off feature set includes: The lengths of all individual fibers cut off are compiled into a length series according to the fiber number. The ratio of the standard deviation to the arithmetic mean of the length series is used as the length dispersion coefficient to determine the degree of length non-uniformity. The proportion of the bifurcation segment to the total length of the fiber is obtained by dividing the broom-shaped bifurcation depth under the same fiber number by the cut length of a single fiber. The degree of length unevenness and the proportion of the bifurcation segment to the total fiber length are summarized into the same data table according to the fiber number to obtain the end bifurcation truncation feature set.

[0007] Preferably, the identification of single fibers in the end-bifuzzing feature set where the proportion of the bifurcation segment to the total fiber length exceeds a preset upper limit and thus cannot be held by the rollers in the drafting zone, to obtain a floating fiber population, includes: The proportion of the bifurcation segment to the total fiber length under each fiber number in the end bifurcation truncation feature set is compared with the preset upper limit of proportion for each fiber. Fiber numbers that exceed the preset upper limit of proportion and their corresponding single fiber truncation lengths are written into the planar fiber data table in ascending order of fiber number. The planar fiber data table is then identified as the planar fiber population.

[0008] Preferably, the step of obtaining the lateral cohesion friction of the planar fiber community in the carding web, and determining the fiber strength of the planar fiber community by comparing the measured value of the lateral cohesion friction with the measured value of the length non-uniformity to a preset strength reference range, includes: For each fiber numbered individual fiber in the floating fiber group, it is pulled along a direction perpendicular to the main fiber axis until it slides relative to the adjacent fiber. The peak value of the reaction force is collected by a micro-force sensor as the lateral cohesive friction force. The measured values ​​of the lateral cohesive friction force and the length non-uniformity are combined into binary data points, which are then projected into a two-dimensional lookup table that divides several strength level intervals based on the lateral cohesive friction force and the length non-uniformity. The corresponding strength value is determined according to the position where the data falls. The arithmetic mean of all representative strength values ​​of the planar fiber population is taken as the fiber strength.

[0009] Preferably, the process of screening planktonic fiber groups with fiber strength lower than a preset lower limit to obtain a weak-strength planktonic fiber group, and dynamically adjusting the allowable upper limit of the broom-shaped branching depth based on the degree of length unevenness to obtain the target upper limit, includes: The cut lengths of all individual fibers in the weak-strength planktonic fiber population are compiled into a weak population length series, and the ratio of the standard deviation to the arithmetic mean is used as the weak population length dispersion coefficient. The corresponding reduction ratio is obtained by looking up the corresponding reduction ratio in the pre-established table of the relationship between the degree of length unevenness and the reduction ratio. The target depth upper limit is obtained by multiplying the current upper limit of the broom-shaped branching depth by 1 and the difference between the result and the reduction ratio.

[0010] Preferably, the step of acquiring historical broom-shaped fork depth records for each speed range, and filtering out speed ranges from the historical broom-shaped fork depth records where the broom-shaped fork depth does not exceed the upper limit of the target depth to determine the target beater speed includes: The historical broom-shaped fork depth values ​​are collected from the historical process database according to the preset beater speed range, and the historical broom-shaped fork depth records are obtained. The beater speed ranges with a broom-shaped fork depth not exceeding the upper limit of the target depth are selected as candidate speed ranges, and the highest speed value among the candidate speed ranges is determined as the target beater speed.

[0011] Preferably, the step of adjusting the mixed opening production line according to the target beater speed and determining the critical draw ratio based on the tensile breaking data of the sliver obtained after the adjustment and carding process includes: The target beater speed is written back to the beater motor control unit, and the strips at the output end of the combing process are cut according to the preset initial length of the sample as tensile test samples. Tensile force is applied to each tensile specimen at a constant rate along the specimen axis, and the tensile force value and specimen elongation at the moment of fracture are collected and written into the tensile fracture data of the strip. The sum of the sample elongation and the initial length of the sample is divided by the initial length of the sample to obtain the draw ratio at break. The arithmetic mean is taken and the preset safety margin is lowered to obtain the critical draw ratio.

[0012] Preferably, the step of summarizing the degree of length unevenness and the proportion of the bifurcation segment to the total fiber length in the same data table according to fiber number to obtain the end bifurcation truncation feature set includes: Using the fiber number as the primary key, the degree of length unevenness is appended to each row of records as a common field, and the proportion of the forked segment to the total fiber length and the cut length of a single fiber are stored as fiber-level fields aligned.

[0013] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a method for analyzing the weavability and fiber length matching of recycled carbon fiber yarn, addressing the problems of fiber holding failure and uneven strength caused by end-broom-like bifurcations during the processing of recycled carbon fiber in a mixed opening production line. This invention acquires end morphology and full-length images of the target batch of fibers, identifies the broom-like bifurcation contour and fracture characteristics, extracts the bifurcation depth and cut length, calculates the degree of length unevenness and the proportion of bifurcated segments, and identifies floating fiber groups that cannot be effectively held by the rollers in the drafting zone due to excessively large bifurcated segment proportions. By measuring the lateral cohesion friction of the floating fibers in the carding web and combining it with the degree of length unevenness to assess fiber strength, weak floating fiber groups are screened out, and the allowable upper limit of the broom-like bifurcation depth is dynamically adjusted accordingly. Based on historical speed records, speed ranges that meet the depth upper limit requirements are selected, the target beater speed is determined, and production line parameters are adjusted. The critical draft ratio is determined through sliver tensile fracture data, achieving precise matching between fiber length characteristics and processing technology, effectively improving the weavability of recycled carbon fiber yarn. Attached Figure Description

[0014] Figure 1 This is a flowchart of a method for analyzing the weavability and fiber length matching of recycled carbon fiber yarn according to the present invention.

[0015] Figure 2 This is a schematic diagram of a method for analyzing the weavability and fiber length matching of recycled carbon fiber yarn according to the present invention.

[0016] Figure 3 This is another schematic diagram of the method for analyzing the weavability and fiber length matching of recycled carbon fiber yarn according to the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0018] like Figures 1-3 This embodiment of the method for analyzing the weavability and fiber length matching of recycled carbon fiber yarn may specifically include: Step S101: Obtain the end morphology image and the full length image of a single fiber of the target batch of recycled carbon fiber from the mixed opening production line, identify the end broom-shaped fork contour and the fracture ends of the single fiber, and extract the broom-shaped fork depth and the cut length of the single fiber.

[0019] For the target batch of recycled carbon fiber laid on the conveyor belt after the mixed opening production line, a linear array industrial camera with backlighting is used to capture individual fibers one by one at the end of the conveyor belt. The linear array industrial camera acquires the end morphology image of the fiber tip area in a local magnification manner, while an area array industrial camera acquires the full-length image of the single fiber covering both ends of the fiber in a flat manner along the fiber laying direction. The grayscale normalization and background removal of the end morphology image and the full-length image of the single fiber are performed, and a one-to-one correspondence relationship is established between the end image and the full-length image according to the fiber number to obtain an indexed fiber dual-view set. For the end morphology images in the indexed fiber dual-view set, the Otsu thresholding method is used to segment the fiber body and the broom-shaped bifurcation region from the background. The segmented binary images are then thinned to obtain the end skeleton. The skeleton is scanned from the root to the tip, and the bifurcation starting point is marked at the point where the skeleton splits from a single trunk into multiple fine branches. The connected region between the bifurcation starting point and the farthest tip end is determined as the broom-shaped bifurcation contour. For the full-length image of a single fiber in the indexed fiber dual-view set, the gray-scale abrupt change positions at both ends are searched along the fiber main axis, and these gray-scale abrupt change positions at both ends are determined as the break positions at both ends of the single fiber. Based on the bifurcation starting point and the farthest thin branch endpoint of the broom-shaped bifurcation profile, the axial projection distance between the two points is measured along the fiber main axis. This axial projection distance is recorded as the broom-shaped bifurcation depth under the fiber number. Based on the break positions at both ends of the single fiber, the pixel span between the break positions is calculated along the fiber main axis. Combined with the scale of the actual length corresponding to the unit pixel obtained in advance during the camera calibration stage, the pixel span is converted into a length value. This length value is recorded as the cut length of the single fiber under the same fiber number, resulting in the end break extraction results with the broom-shaped bifurcation depth and the cut length of the single fiber labeled one by one.

[0020] After the target batch of recycled carbon fiber is processed in the mixing and opening production line, it is evenly spread onto the surface of a matte black conveyor belt by an end-spreading roller, ensuring that individual fibers remain non-overlapping to guarantee that images of each fiber can be captured subsequently. The conveyor belt passes through the image acquisition station at a constant linear speed. Above this station, linear array industrial cameras and area array industrial cameras are arranged side-by-side in the fiber travel direction. Together with a backlight panel arranged below, they form a transmissive illumination structure, so that the fiber body and the broom-like branching fine branches appear as high-contrast dark patterns on a light-colored background, making it easy to clearly identify the complete outline of each fiber and its end branching characteristics.

[0021] Specifically, the linear scan industrial camera is equipped with a high-magnification telecentric lens, which only magnifies and images the distal ends of a single fiber. The position encoder of the conveyor belt triggers the linear scan camera to begin scanning line by line, with each line achieving a resolution of over 100 pixels per millimeter. These lines are then stitched together to form a magnified end-fiber image, in which the individual fine branches and the main trunk of the broom-like forks can be distinguished. Simultaneously, the area scan industrial camera is equipped with a low-distortion lens with a large field of view, covering the entire projection of a single fiber in a flat manner along the fiber's spreading direction. At the same moment the conveyor belt position encoder triggers the linear scan camera to start scanning, the area scan camera is simultaneously triggered to perform a single exposure, acquiring a full-length image covering both ends of the fiber. The end-fiber image and the full-length image describe different scale information of the same fiber at the same conveyor position. The time required for the linear scan camera to complete the scanning and stitching, and the exposure time of the area scan camera, are both controlled within 0.5 seconds. The end-fiber morphology image and the full-length image of a single fiber are processed using histogram equalization for grayscale normalization, and then differentially removed using a constant brightness area of ​​the backlight as a background template. The system records the time each fiber enters the acquisition station using a position encoder and assigns a unique fiber number, establishing a one-to-one correspondence between the end-fiber morphology image and the full-length image of the same fiber through this fiber number, resulting in an indexed set of dual-view images of the fibers. The position encoder uses a photoelectric sensor to detect the timestamp of the fiber passing through the acquisition station's starting point, and calculates the fiber position based on the transmission speed, ensuring that each fiber receives a unique identifier. This allows two images corresponding to the same fiber to be associated through the fiber number, resulting in an indexed set of dual-view images of the fibers.

[0022] It should be noted that the end-shape images in the indexed fiber dual-view set are segmented using the Otsu thresholding method. In the end-shape images, the fiber body and broom-shaped branching tendons are darker in grayscale, while the background is brighter, exhibiting a bimodal grayscale distribution. The Otsu thresholding method determines the optimal segmentation threshold by maximizing the inter-class variance, classifying the fiber body and broom-shaped branching region as the foreground and the backlit region as the background, resulting in a binary image that retains only the fiber and branching tendons.

[0023] Preferably, the binary image undergoes small-area connected component removal to filter out isolated noise points formed by residual fluff particles in the background. During removal, areas with fewer than 50 connected component pixels are set as background. The foreground edges of the noise-removed binary image are peeled off layer by layer. Each iteration executes two sub-iterations, deleting edge points that meet the conditions based on the number of foreground pixels in the 8-neighborhood of a pixel, the number of 0-to-1 transitions of adjacent pixels, and the product condition of specified neighboring pixels. This process is repeated until no pixels can be deleted, and the foreground region is reduced to a centerline of a single pixel width. The resulting centerline is the end skeleton. The end skeleton runs linearly along the fiber axis and branches at the tip. Pixel by pixel, the skeleton is scanned from root to tip, and the number of pixels belonging to the skeleton in the 8-neighborhood of each skeleton pixel is counted, denoted as N. If N is greater than 2, the pixel is determined to be a branch node. The first branch node to appear during the scan from root to tip is the bifurcation starting point described later. In step S101, when searching for the gray-level abrupt change positions at both ends along the fiber's main axis, a pixel-by-pixel gray-level profile extraction method is used. Gray-level profile curves perpendicular to the main axis are extracted at 0.1 mm intervals along the fiber's main axis. These profile curves exhibit a continuous low gray-level plateau with a gray-level value below 50 in the fiber-covered area and a high gray-level plateau with a gray-level value above 200 in the backlit area outside the fiber. A steep transition segment with a gray-level gradient greater than 15 gray levels per pixel exists between these two plateaus. By detecting locations in the profile curve where the gray-level gradient exceeds 15 gray levels per pixel, the gray-level abrupt change positions at both ends of the fiber can be identified, thereby determining the range of several dispersed branches within the broom-like bifurcation contour and the transition region connecting the roots of these branches.

[0024] For example, the grayscale profile curve is subjected to first-order difference calculation along the fiber main axis to obtain the grayscale gradient value of each pixel position. Pixels with an absolute difference value exceeding the transition threshold T are identified as grayscale abrupt change positions. The two outermost grayscale abrupt change positions at both ends of the fiber main axis are the break positions at both ends of a single fiber. The transition threshold T is adaptively determined according to the statistical characteristics of the current grayscale profile curve and is taken as follows: T=μ+k×σ, where μ is the mean of the absolute values ​​of the first-order differences of all adjacent pixels in the profile, σ is its standard deviation, and k is the sensitivity coefficient, with a value range of 2.5 to 3.5. For samples with normal fiber-background contrast in 8-bit grayscale images, k is usually taken as 3.0, and the corresponding T value generally falls between 15 and 40 grayscale levels. When the sample contrast is low or the background noise is large, the k value can be appropriately adjusted down or up within this range to avoid missing weak breaks or misjudging noise as breaks.

[0025] Understandably, the furthest branch endpoint at the tip is defined as the largest projected coordinate of each branch endpoint in the broom-shaped bifurcation profile along the fiber main axis. Based on the marked bifurcation start point and the furthest branch endpoint in the broom-shaped bifurcation profile, the coordinates of the two points are projected onto the fiber main axis. The absolute value of the difference between the coordinates of the two projected points along the main axis is taken as the broom-shaped bifurcation depth Ld, and written into a data table with the fiber number as the index, denoted as Ld=|x2-x1|, where x1 and x2 are the projected coordinates of the bifurcation start point and the furthest branch endpoint in the main axis, respectively. Finally, based on the break positions at both ends of the single fiber, the number of pixels between the two breaks is calculated along the fiber main axis to obtain the pixel span Lp.

[0026] For example, during the system deployment phase, a calibration target plate with known scales is used to capture images within the field of view of the area array industrial camera to complete camera calibration. The scale corresponding to the actual length per unit pixel is obtained by dividing the physical length of the scale by the number of corresponding pixels. Based on this scale, the pixel span is converted into a length value, which is used as the cut length of a single fiber. This length value, along with the broom-like forking depth value and fiber number, is stored in an end fracture feature data table. The data table contains three fields: fiber number, forking depth, and cut length. Finally, the end fracture extraction results with broom-like forking depth and single fiber cut length marked for each fiber are obtained.

[0027] Step S102: Based on the broom-shaped bifurcation depth and the cut length of a single fiber, calculate the length non-uniformity of the target batch of recycled carbon fibers and the proportion of the bifurcation segment to the total fiber length to obtain the end bifurcation cut-off feature set.

[0028] For the cut length of a single fiber in the target batch of recycled carbon fiber, indexed by fiber number, the cut lengths of all fibers are compiled into a length series in order of fiber number. The arithmetic mean and standard deviation of this length series are calculated, and the ratio of the standard deviation to the arithmetic mean is used as the length dispersion coefficient. This length dispersion coefficient is used to determine the degree of length non-uniformity of the target batch of recycled carbon fiber. For the broom-like bifurcation depth and cut length of a single fiber under the same fiber number in the target batch of recycled carbon fiber, the broom-like bifurcation depth is divided by the cut length of the single fiber to obtain the proportion of the bifurcation segment to the total fiber length. This process is repeated for all fiber numbers in the target batch of recycled carbon fiber, storing the proportion of the bifurcation segment to the total fiber length for each fiber indexed by its fiber number, resulting in a record of the bifurcation segment proportion for each fiber. The non-uniformity of the length of the target batch of recycled carbon fiber, the percentage of bifurcation segments marked on each fiber, and the cut length and broom-like bifurcation depth of each fiber number are summarized into the same data table according to the fiber number. The fiber number is used as the primary key, and the non-uniformity of the length is added as a common field to each row of records. The percentage of the bifurcation segment to the total fiber length and the cut length of the single fiber are stored as fiber-level fields to obtain the end bifurcation cut-off feature set.

[0029] In one implementation, after the target batch of recycled carbon fibers passes through the image acquisition station of the mixing and opening production line, each fiber is assigned a unique fiber number in a data table, and the broom-like bifurcation depth and single fiber cut-off length of that fiber are recorded under that fiber number. Subsequent calculations are all performed around this data table.

[0030] Specifically, the cut lengths of individual fibers in the target batch of recycled carbon fiber, indexed by fiber number, are first collected. Following the ascending order of fiber number, the values ​​of the individual fiber cut length field in the data table are read row by row, and all read values ​​are written into a one-dimensional storage structure in the order of reading, forming a length sequence.

[0031] It should be noted that the recycling process for recycled carbon fiber is diverse, and the stress during the stripping process is random, resulting in a natural variation in the cut length of individual fibers within the same batch. Therefore, the fluctuation within the length series is significantly greater than that of the virgin carbon fiber length series. Generally, the standard deviation of the virgin carbon fiber length series is usually within 0.3 mm, while the standard deviation of the recycled carbon fiber length series can reach 1.5 mm to 5 mm. To describe this fluctuation, a two-step statistical calculation is performed on the length series: First, the arithmetic mean u of the length series is calculated as the center length level of the fibers in that batch; second, the standard deviation s of the length series is calculated, and the coefficient of variation cv = s / u is calculated to reflect the relative intensity of the fluctuation. For example, if the length series of a certain batch of recycled carbon fiber has u = 6.2 mm and s = 2.4 mm, then cv ≈ 0.39, which is significantly higher than the virgin carbon fiber cv, which is generally no more than 0.05. Therefore, this batch is determined to be a highly discrete batch. In one embodiment, the arithmetic mean is obtained by summing all elements in the length series and then dividing by the number of elements. The square of the difference between each element in the length sequence and the arithmetic mean is calculated. The sum of all squared differences is then divided by the number of elements, and the square root is taken to obtain the standard deviation. The arithmetic mean is represented by μ, and the standard deviation by σ. Based on this, the length dispersion coefficient is obtained by dividing the standard deviation by the arithmetic mean. The length dispersion coefficient is calculated using the formula CV = σ / μ, where CV is the length dispersion coefficient, σ is the standard deviation, and μ is the arithmetic mean. The length dispersion coefficient is a dimensionless value that characterizes the relative dispersion of the target batch of recycled carbon fibers in the length dimension. The length dispersion coefficient is used as the degree of length non-uniformity of the target batch of recycled carbon fibers.

[0032] It is understood that the aforementioned degree of length non-uniformity describes a batch-level characteristic of the entire batch, and applies to any fiber within that batch. Based on the aforementioned proportional calculation, the broom-like bifurcation depth and single-fiber cut-off length are extracted for each fiber in the target batch of recycled carbon fibers according to fiber number, and a division operation is performed according to the number to obtain the bifurcation segment ratio of each fiber, denoted as r. i Where i is the fiber number; when the cut length of a fiber is less than 0.5 mm, it is marked as an abnormal sample and removed to avoid the denominator being too small and causing r to be incorrect. i Distortion. Ultimately, all will be valid. i Summarize them in numerical order to form a bifurcation percentage sequence for this batch, which can be used for subsequent joint analysis with the degree of length unevenness.

[0033] For example, for a fiber with fiber number k, the broom-shaped fork depth under that fiber number is read from the data table, and then the single fiber cut length under that fiber number is read. The broom-shaped fork depth is used as the dividend, and the single fiber cut length is used as the divisor. The resulting quotient is the proportion of the fork segment of the fiber to the total length of the fiber.

[0034] Preferably, all fiber numbers of the target batch of recycled carbon fibers are traversed in ascending order of fiber number. The above proportional calculation is repeated for each fiber number, and the obtained proportion of the bifurcation segment to the total fiber length is written back to the "Bifurcation Segment Proportion" field of the row containing the fiber number in the data table, resulting in a bifurcation segment proportion record labeled for each fiber. Further, the length unevenness of the target batch of recycled carbon fibers, the bifurcation segment proportion record labeled for each fiber, and the cut-off length and broom-like bifurcation depth of a single fiber under each fiber number are summarized into the same data table according to fiber number.

[0035] Specifically, the data table uses fiber number as the primary key, with each row corresponding to one fiber. The fiber-level fields in each row include the single fiber cut-off length, broom-like bifurcation depth, and the proportion of the bifurcation segment to the total fiber length. An additional common field is added to the data table, with the length unevenness as its value appended to each row. This allows any row to read both the fiber-level characteristics of the fiber itself and the shared length unevenness across the batch.

[0036] It is understandable that the data table organized in the above manner is the end bifurcation truncation feature set. This set simultaneously carries two types of information: batch-level length unevenness and fiber-level bifurcation ratio, which can support the joint discrimination of individual fibers and the whole batch.

[0037] Step S103: Evaluate the proportion of the bifurcation segment to the total fiber length in the end bifurcation truncation feature set, identify single fibers whose proportion exceeds the preset upper limit and cannot be effectively held by the rollers in the stretching zone, and obtain a floating fiber group.

[0038] For each fiber number in the end-bifing cut-off feature set, the proportion of the bifurcation segment to the total fiber length is compared with a preset upper limit for each fiber. If the proportion exceeds the preset upper limit, the single fiber corresponding to that fiber number is determined to be a floating fiber that cannot be clamped by the rollers in the drafting zone. This comparison and determination is performed on all fiber numbers in the end-bifing cut-off feature set. All fiber numbers determined to be floating fibers and their corresponding cut-off lengths are arranged in ascending order by fiber number and written into a floating fiber data table. The floating fiber data table contains two columns: a fiber number field and a cut-off length field. The established floating fiber data table is defined as a floating fiber group.

[0039] In one embodiment, the set of end-branching cut-off features is stored in a data table using fiber number as the primary key. Each row of the data table contains the cut-off length of a single fiber under that fiber number, the depth of the broom-like bifurcation, and the proportion of the bifurcation segment to the total fiber length. The preset upper limit of the proportion is derived based on the gripping distance between the front and rear rollers and the shortest grippable length of the solid segment of a single fiber in the subsequent drawing process of the mixed opening production line. Specifically, let the gripping distance between the front and rear rollers be L_grip, and the shortest grippable length of the solid segment of a single fiber be L_solid = L_grip. The allowance is determined to be 10%–15% of the grip distance; then, the average fiber length L of the batch is used as an approximate estimate of the total length of a single fiber, requiring the total solid segment length L of each fiber to be... If Dfork is not less than Lreal, and under the condition that Laverage is uniformly approximated by Ltotal, then Rupper limit = 1. Lactual / Laverage. Taking a grip distance Lgrip = 32 mm, a margin of 12.5%, a minimum grip length Lactual = 28 mm, and an average fiber length Laverage = 38 mm as an example, substituting these values, we get Rupper limit ≈ 1. 28 / 38≈0.263, 0.26 is used in engineering. This upper limit is predetermined during the production line process setting stage based on the roller specifications and fiber raw material characteristics and is used as a fixed process parameter. In this embodiment, the end bifurcation cut-off feature set is read line by line in ascending order of fiber number. For each fiber number, the proportion of the corresponding bifurcation segment to the total fiber length is taken out, and the proportion of the bifurcation segment to the total fiber length is compared with the preset upper limit. Wherein, the solid segment length Lsolid = Ltotal - Dfork, where Lsolid is the solid segment length, Ltotal is the cut-off length of a single fiber, and Dfork is the broom-like bifurcation depth. If the proportion exceeds the preset upper limit, the bifurcation segment of the fiber occupies too large a share of its total length, and the solid segment length is insufficient to cross the gripping distance between the front and rear rollers. The single fiber corresponding to this fiber number is determined to be a floating fiber that cannot be clamped by the rollers in the drafting zone. If the proportion does not exceed the preset upper limit, the next fiber number is determined. Furthermore, the comparison and determination are repeated for all fiber numbers in the aforementioned end-bidding truncation feature set. The fiber numbers of all identified planar fibers and their corresponding single-fiber truncation lengths are written into the same record sequentially according to fiber number. After completing the identification and end-point truncation of individual planar fibers, the fiber number and its actual truncation length within the observation interval are compiled into the same summary table, forming a planar fiber record based on the sampling area.

[0040] Understandably, this record represents a group of planktonic fibers, which retains the fiber number and cut length of each individual fiber, enabling further identification of the length distribution within the group.

[0041] Step S104: Obtain the lateral cohesion friction force of the floating fiber group in the carding net, and determine the fiber strength of the floating fiber group by comparing the measured value of the lateral cohesion friction force with the measured value of the length non-uniformity with a preset strength reference range.

[0042] For each individual fiber recorded under its fiber number in the floating fiber group, a carding mesh is laid out on a flat support surface in the area where the floating fiber group is located. Micro-force sensor probes are arranged vertically along the fiber main axis, and the micro-force sensor probes are pressed against the overlapping positions between the floating fiber and its adjacent fibers. The floating fiber is pulled at a constant rate in a direction perpendicular to the fiber main axis until it slides relative to the adjacent fiber. The micro-force sensor probes collect the reaction force signal during the pulling process, and the peak value of the reaction force signal is determined as the lateral cohesive friction force of the floating fiber in the carding mesh. The pulling and collecting process is repeated for all fiber numbers in the planktonic fiber population to obtain a transverse cohesion friction force sequence indexed by fiber number. Each value in the transverse cohesion friction force sequence is combined with the measured value of the length non-uniformity to form a binary data point. This binary data point is projected into a two-dimensional lookup table formed by a preset strength reference range. The preset strength reference range is divided into several strength level intervals with transverse cohesion friction force as the horizontal axis and length non-uniformity as the vertical axis. Each strength level interval corresponds to a representative strength value. The strength level interval corresponding to the binary data point is determined according to the position where the binary data point falls. For each fiber number in the planktonic fiber group that falls into the strength level range, the representative strength value corresponding to the strength level range is assigned to the single fiber under that fiber number, resulting in a representative strength value labeled for each fiber. The representative strength values ​​of all fiber numbers in the planktonic fiber group are aggregated into a strength series according to the fiber number. The arithmetic mean of the strength series is calculated, and the resulting mean is used as the fiber strength of the planktonic fiber group. This mean is then appended to the record of the planktonic fiber group as a group-level field, resulting in a planktonic fiber group with a fiber strength identifier.

[0043] In one implementation, after identifying the floating fibers through image processing, the system automatically assigns a unique number to each floating fiber and measures its single fiber cut length. The fiber number and its corresponding cut length value are recorded one by one and stored in a data table using the fiber number as an index. Each row contains a set of correspondences between fiber numbers and cut lengths. Before determining the fiber strength, the carding web of the section containing the floating fibers is first stripped off by the carding machine doffer. The doffer is a cylindrical component in the carding machine responsible for stripping the fibers from the cylinder surface and transferring them to subsequent processes. The stripped carding web is laid flat on the support table, ensuring that the fiber spindle direction is roughly aligned with the long axis of the table, facilitating subsequent lateral traction operations.

[0044] Specifically, a two-dimensionally movable probe drive mechanism is installed above the carding mesh in the area where the floating fiber group is located. A micro-force sensor probe with a range in the millinewtons is fixed to the end of the probe drive mechanism. The probe tip of the micro-force sensor probe is blunted to avoid scratching the fiber surface. According to the fiber number of the first floating fiber in the data table, and using the two-dimensional coordinates of the fiber's center point recorded in the data table, the probe drive mechanism is moved horizontally to above that coordinate position. The micro-force sensor probe is then lowered onto the mesh surface in a direction perpendicular to the carding mesh surface, pressing it against the overlap between the floating fiber and its adjacent fibers. The initial coordinates of the probe at this point are recorded as the reference point for subsequent displacement calculations. Further, along a direction perpendicular to the fiber's main axis, the probe drive mechanism drives the micro-force sensor probe to move horizontally at a constant rate of 0.5 to 2 mm per second, pulling the floating fiber. During the pulling process, the overlap point between the floating fiber and its adjacent fibers gradually receives force, and the micro-force sensor probe continuously outputs a reaction force signal at a sampling rate of 100 to 500 Hz. When the tensile displacement continues to increase to a certain point, relative sliding occurs between the floating fibers and adjacent fibers, and the reaction force signal shows a clear inflection point.

[0045] In one embodiment, the reaction force signal is first subjected to a sliding average noise reduction with a window length of 10 to 50 milliseconds to obtain a smooth signal curve. The smooth signal curve shows an overall upward trend before the inflection point, reaches a global maximum value at the inflection point, and then drops sharply. The global maximum value is determined as the peak value of the reaction force signal of the floating fiber.

[0046] Specifically, the peak value of the reaction force signal is determined as the lateral cohesion friction force of the floating fiber in the carding web, and the "lateral cohesion friction force" field in the data table is written back using the fiber number of the floating fiber as an index.

[0047] It should be noted that the above-described process of probe pressing, constant-speed pulling, signal acquisition, and peak identification is repeated for all fiber numbers in the planktonic fiber group. Each fiber is processed sequentially from smallest to largest according to its fiber number, resulting in a transverse cohesion friction force sequence indexed by its fiber number. The number of elements in this transverse cohesion friction force sequence is the same as the number of planktonic fibers in the planktonic fiber group. Further, each value in the transverse cohesion friction force sequence is combined with a measurement of the length non-uniformity of the target batch of recycled carbon fibers to form a binary data point, denoted as (F, U), where F is the transverse cohesion friction force of that planktonic fiber, and U is the length non-uniformity of that batch.

[0048] It is understood that the preset strength benchmark range is a two-dimensional lookup table obtained by reverse calibration of yarn breakage test results of virgin carbon fiber and multiple batches of recycled carbon fiber before the production line is put into operation. The two-dimensional lookup table divides the entire two-dimensional plane into several rectangular areas with the lateral cohesive friction force as the horizontal axis and the degree of length non-uniformity as the vertical axis. Each rectangular area is a strength level interval. A representative strength value is pre-filled into the strength level interval. The magnitude of the representative strength value reflects the tensile breaking strength that the fiber falling into the interval can withstand during drafting.

[0049] Preferably, the representative strength values ​​are arranged in a monotonically decreasing pattern, with the lateral cohesive friction force increasing from small to large and the degree of length unevenness decreasing from large to small.

[0050] For example, each binary data point is sequentially mapped to the two-dimensional lookup table. Based on the rectangular area the data point falls into in the two-dimensional lookup table, the strength level interval corresponding to that rectangular area is taken as the strength level interval under that fiber number. In this embodiment, for each fiber number in the planktonic fiber group, the representative strength value corresponding to the strength level interval is assigned to a single fiber under that fiber number, resulting in a representative strength value labeled for each fiber. This value is then written back to the representative strength value field in the data table according to the fiber number. Subsequently, the representative strength values ​​of each fiber are collected in order of fiber number to form a strength sequence. The fiber strength of the planktonic fiber group is characterized by calculating the arithmetic mean of all elements in this strength sequence. Specifically, the sum of all elements in the strength sequence is divided by the total number of elements, and the resulting arithmetic mean is the fiber strength of the planktonic fiber group. This arithmetic mean is appended to the record of the planktonic fiber group as a group-level common field, resulting in a planktonic fiber group with a fiber strength identifier. The fiber strength of this floating fiber community characterizes the overall holding and bearing capacity of the floating fiber community at the carding web level under the condition of excessive bifurcation ratio.

[0051] Step S105: Evaluate the fiber strength of the planktonic fiber group, screen planktonic fiber groups with fiber strength lower than the preset strength lower limit to obtain a weak-strength planktonic fiber group, identify the degree of length unevenness of the weak-strength planktonic fiber group, and dynamically adjust the allowable upper limit of the broom-shaped branching depth to obtain the target depth upper limit.

[0052] For the planktonic fiber group with attached fiber strength identifier, the fiber strength of the planktonic fiber group is read from the group-level common field of the planktonic fiber group. The fiber strength is compared with a preset lower limit. If the fiber strength is lower than the preset lower limit, the planktonic fiber group is marked as a weak-strength planktonic fiber group, and all single fiber cut-off length data stored in the planktonic fiber group record with fiber number as the index are retained. For the single fiber cut-off lengths stored in the weak-strength planktonic fiber group record with fiber number as the index, all the single fiber cut-off lengths are aggregated into a weak group length sequence in order of fiber number. The arithmetic mean and standard deviation of the weak group length sequence are calculated. The ratio of the standard deviation to the arithmetic mean is used as the weak group length dispersion coefficient. The weak group length dispersion coefficient is determined as the length non-uniformity of the weak-strength planktonic fiber group. For the length unevenness of the weak-strength planktonic fiber community, the corresponding reduction ratio is found according to the pre-established correspondence table between the length unevenness and the reduction ratio. The correspondence table records the depth reduction ratio values ​​corresponding to different length unevenness intervals. The difference between the current allowable upper limit of the broom-shaped bifurcation depth and the reduction ratio is multiplied by 1, and the calculated value is determined as the target depth upper limit.

[0053] In one embodiment, the planktonic fiber group with attached fiber strength identifier has been stored as a record set with fiber number as the primary key. The group-level common field in the set records the fiber strength of the planktonic fiber group, and each record retains fiber-level fields such as fiber number, single fiber cut-off length, and broom-like forking depth of the planktonic fiber.

[0054] Specifically, the preset lower strength limit is derived by working backward from the minimum tensile strength requirement of the yarn in the subsequent weaving process of the mixed opening production line. It is a scalar value and is pre-written into the production line process parameter table. The fiber strength of the floating fiber group is read from the group-level common field of the floating fiber group and compared with the preset lower strength limit. If the fiber strength is lower than the preset lower strength limit, the entire floating fiber group record is marked as a weak-strength floating fiber group, and all the single fiber cut-off length data stored in the floating fiber group record with the fiber number as the index are retained as is in the weak-strength floating fiber group record.

[0055] Understandably, a weak-strength planktonic fiber population means that the combination of the lateral cohesion friction of each planktonic fiber in the carding web and the batch length unevenness results in a lower strength level. For example, if the average lateral cohesion friction of each planktonic fiber in the population is less than 0.15 cN and the batch length unevenness (CV) exceeds 25%, it is classified as a weak-strength population. In this case, relying solely on batch-level length unevenness is insufficient to reflect the length distribution characteristics within the population. It is necessary to perform length statistics separately for this weak-strength planktonic fiber population, extracting histograms of fiber length distribution within the population at 2mm intervals, and calculating the mean, standard deviation, and coefficient of variation of the length within the population as a basis for subsequent process adjustments. In one embodiment, the cut lengths of individual fibers stored in the weak-strength planktonic fiber population record, indexed by fiber number, are read row by row in ascending order of fiber number. The read values ​​are then sequentially written into an independent one-dimensional storage structure, forming a weak population length sequence. The number of elements in the weak population length sequence is the same as the number of fibers in the weak-strength planktonic fiber population, excluding other fibers from the batch. Further, the arithmetic mean of the weak population is obtained by summing all elements in the weak population length sequence and dividing by the number of elements. Then, the square of the difference between each element in the weak population length sequence and the arithmetic mean of the weak population is calculated. The square root of all these squared differences is then obtained by summing them, dividing by the number of elements, and finally taking the square root. The formula for calculating this standard deviation is as follows: Where σ is the standard deviation of the weak population, and Li is the cut-off length of the i-th fiber. Let Σ be the arithmetic mean of the weak population, n be the number of elements in the sequence of weak population lengths, and Σ represent the summation over all elements. Dividing the standard deviation of the weak population by its arithmetic mean yields the dimensionless value of the weak population length dispersion coefficient, which is used to determine the degree of length inhomogeneity in the weak-strength planktonic fiber population.

[0056] It should be noted that the pre-established correspondence table is a two-column lookup table obtained by reverse calibration based on batch draw fracture test results under several different length non-uniformity levels before the production line goes into operation. The first column of the correspondence table lists several value ranges of length non-uniformity from smallest to largest, and the second column records the reduction ratio that should be adopted for each length non-uniformity value range. The reduction ratio is a dimensionless value between 0 and 1. In the correspondence table, the larger the value range of the length non-uniformity, the larger the reduction ratio recorded in the second column, and the two have a monotonically increasing relationship overall.

[0057] For example, the numerical value of the length unevenness of the weak-strength planktonic fiber group is compared row by row with the value range in the first column of the correspondence table. The interval is determined by the left-closed and right-open principle. That is, when the length unevenness value is greater than or equal to the lower limit of a certain interval and less than the upper limit of the interval, it is determined to fall into the interval. At this time, the reduction ratio recorded in the second column of the row is taken as the reduction ratio for this adjustment. Finally, the allowable upper limit H0 of the current broom-shaped branching depth is taken from the production line process parameter table. The allowable upper limit H0 of the current broom-shaped branching depth is multiplied by 1-R to obtain the reduced allowable upper limit Ht of the broom-shaped branching depth, where R is the reduction ratio. The reduced allowable upper limit Ht of the broom-shaped branching depth is determined as the target depth upper limit and passed to step S106. The target depth upper limit Ht changes dynamically with the degree of length unevenness of the weak-strength planktonic fiber group, and is used to constrain the bifurcation control range of the beater process in step S106. When the fiber end bifurcation depth, i.e. the broom-shaped unfolding length of the fiber end, is detected to exceed the target depth upper limit Ht, the beater speed is reduced or the beater rack angle is adjusted, thereby tightening the end bifurcation allowance allowed by the beater process in real time according to the degree of short fiber aggregation inside the current weak-strength planktonic fiber group.

[0058] Step S106: Obtain historical broom-shaped fork depth records for each speed range, filter out speed ranges where the broom-shaped fork depth does not exceed the upper limit of the target depth from the historical broom-shaped fork depth records, and determine the target beater speed.

[0059] For several preset beater speed ranges, the broom-shaped fork depth values ​​recorded during previous trial runs in each beater speed range are retrieved from the historical process database. These values ​​are then categorized and compiled according to the beater speed range, forming historical broom-shaped fork depth records for each speed range. For each historical broom-shaped fork depth record, the broom-shaped fork depth in each beater speed range is compared with the target depth upper limit. Beater speed ranges whose broom-shaped fork depth does not exceed the target depth upper limit are selected as candidate speed ranges. The highest speed value among the candidate speed ranges is determined as the target beater speed.

[0060] In one implementation, the aforementioned speed ranges of the beaters are pre-defined as several continuous segments with equal intervals, based on the speed range of the beater motor during the production line design phase. The width of each segment is 1 / 10 of the speed range. For example, if the motor speed range is 600 rpm to 1600 rpm, each segment is 100 rpm wide, resulting in 10 segments. Each segment is represented by a segment number and its corresponding upper and lower speed limits. The numbers increase sequentially from low to high speed, i.e., number 1 corresponds to the lowest speed segment, number 10 corresponds to the highest speed segment, and these are stored in the production line process parameter table. The end-shape image refers to a grayscale image obtained by close-up photography of the regenerated carbon fiber bundle tip after the beater action. It reflects the broom-like forking morphology of the fiber tip due to the breakage, and the depth of the broom-like forking is extracted from this image after morphological processing. The historical process database records, in chronological order, the end-shape images and corresponding broom-shaped fork depth extraction results synchronously collected during the trial operation of the production line at different beater speeds. Each historical record includes fields such as batch number, collection timestamp, equipment number, beater speed value, end-shape image storage path, and broom-shaped fork depth value to ensure data traceability and reusability.

[0061] Specifically, following the operator speed range numbering in ascending order, all historical records whose speed values ​​fall within the upper and lower limits of a given operator speed range are sequentially selected from the historical process database. Each operator speed range is non-overlapping to ensure that each historical record belongs to only one range. The broom-like bifurcation depth values ​​in the selected historical records are categorized and aggregated according to the operator speed range number. The arithmetic mean of multiple broom-like bifurcation depth values ​​within the same operator speed range is taken as the representative broom-like bifurcation depth for that range, and these values ​​are arranged in order of operator speed range number to form historical broom-like bifurcation depth records for each range. For each historical broom-like bifurcation depth record under each range, the representative broom-like bifurcation depth is extracted item by item in ascending order of operator speed range number and compared with the target depth upper limit. If the representative broom-like bifurcation depth of a certain operator speed range does not exceed the target depth upper limit, then that operator speed range is marked as a candidate speed range.

[0062] Understandably, there are usually more than one candidate speed range marked. Following the "capacity priority" principle, the range with the highest upper limit value among all candidate speed ranges is selected as the final range, and this upper limit value is taken as the target beater speed n. For example, if one candidate range is 800 to 1000 rpm and another is 950 to 1200 rpm, the range with the higher upper limit [950, 1200] is selected as the final range, and n = 1200 rpm. Finally, the target beater speed n is written back to the beater speed setpoint field in the production line process parameter table.

[0063] Step S107: Adjust the mixed opening production line according to the target beater speed, obtain the tensile breaking data of the strips obtained after the adjustment and combing into a web, and determine the critical draw ratio based on the tensile breaking data of the strips.

[0064] Before the production line starts operating, the target beater speed is written back to the beater motor control unit of the mixing and opening production line, so that the beater motor operates at the target beater speed. After the target batch of recycled carbon fiber enters the carding process from the mixing and opening process, it forms continuous slivers. At the output end of the carding process, several segments of the slivers are cut according to the preset initial length of the sample as tensile test samples. The tensile test samples are numbered in the cutting order and collected into a sliver sample set. For each tensile test sample in the sliver sample set, indexed by the sample number, both ends of the tensile test sample are fixed between the upper and lower clamps of the electronic tensile testing machine. The tensile testing machine continuously applies tensile force along the sample axis at a constant rate. At the same time, the force sensor and displacement sensor collect the tensile force value and the sample elongation during the tensile process, respectively. The tensile force value and the corresponding sample elongation at the moment of sample fracture are written into the tensile fracture data of the sliver according to the sample number. For the elongation of the specimen under each specimen number in the tensile fracture data of the strip, the sum of the specimen elongation and the initial length of the specimen is divided by the initial length of the specimen to obtain the draw ratio at fracture under that specimen number. The arithmetic mean of the draw ratios at fracture under all specimen numbers in the strip specimen set is taken, and then the arithmetic mean is reduced by a preset safety margin to obtain the reduced value. The reduced value is determined as the critical draw ratio at the target beater speed.

[0065] In one implementation, before the production line is started, the target beater speed is written back to the beater speed setting value field of the production line process parameter table. After the frequency converter reads the field, it drives the beater motor to run at the target beater speed.

[0066] Specifically, the target batch of recycled carbon fiber is fed into the mixing and opening production line by the feed roller of the mixing and opening process. After being loosened by a beater motor operating at the target beater speed, it enters the carding process. The carding machine's card cloth further separates the loose fibers, which are then stripped by a doffer, resulting in a continuous sliver with stable cross-sectional dimensions at the output end. A graduated cutting device is arranged at the output end of the carding process. The blade spacing of the cutting device is preset according to the initial length of the sample. The continuous sliver is cut segment by segment after entering the cutting device from the output end, and each segment of the sliver is a tensile test specimen. Each tensile test specimen is assigned a unique specimen number in chronological order, and all the tensile test specimens are collected according to their specimen numbers to form a sliver specimen set. The tensile test specimens in the sliver specimen set are clamped sequentially between the upper and lower clamps of an electronic tensile testing machine according to their numbers, and the clamping force is adjusted by a clamping force adjustment device to reach a preset threshold. The gripping force threshold is determined based on the sample linear density and the clamp jaw area, ranging from 50N to 150N. The lower limit is used when the linear density is less than 5 ktex, and the upper limit is used when it is greater than 10 ktex, to balance anti-slip properties and avoid sample damage at the clamp jaws; in this embodiment, 100N is used. Subsequently, the upper clamp moves vertically upward at a constant rate, continuously applying tensile force along the sample axis. Force and displacement sensors synchronously collect tensile force and elongation data at a sampling frequency of 100Hz until the sample breaks.

[0067] Understandably, when the tensile specimen breaks at a certain moment, the tensile force value abruptly changes from a continuously increasing trend to a sharp drop to near zero. This tensile force value at that moment is identified as the tensile force value at the instant of specimen breakage. The specimen elongation sequence output by the displacement sensor is read at the same timestamp, and the resulting value is the corresponding specimen elongation. The tensile force value at the instant of breakage and the specimen elongation are then written back into the strip tensile fracture data according to the specimen number.

[0068] Specifically, the tensile fracture data of the strips is a data table with the specimen number as the primary key. Each row contains the tensile force and elongation of the specimen at the moment of fracture. For each record in this data table, the elongation ratio D at fracture for that specimen number is calculated using the formula D=(L0+ΔL) / L0, where L0 is the initial length of the specimen and ΔL is the elongation of the specimen. This formula is equivalent to D=1+ΔL / L0. The resulting elongation ratio D at fracture represents the ratio of the total length of the specimen at fracture to its initial length. The elongation ratios D at fracture for all N specimen numbers are summed and divided by N to obtain the arithmetic mean of the elongation ratios at fracture. A preset safety margin ratio is read from the production line process parameter table. This ratio is preset within the range of 0.05 to 0.15 based on production experience, with a typical value of 0.1. The arithmetic mean is multiplied by 1 and subtracted from a preset safety margin ratio to obtain the adjusted value. This adjusted value is determined as the critical draft ratio at the target beater speed, and the critical draft ratio is written back to the critical draft ratio field corresponding to the beater speed in the production line process parameter table. Subsequent weaving processes read this critical draft ratio as the upper limit of the draft ratio and monitor the actual draft ratio during the weaving process in real time. When the actual draft ratio exceeds the critical value, a draft speed reduction or alarm is triggered to ensure operation within the safe range defined by the critical draft ratio.

[0069] The above description of the embodiments is only for the purpose of helping to understand the technical solutions and core ideas of this application; those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for analyzing the weavability and fiber length matching of recycled carbon fiber yarn, characterized in that, The method includes: Acquire end morphology images and full-length images of single fibers of the target batch of recycled carbon fibers from the mixed opening production line, identify the end broom-like bifurcation contour and the fracture ends of single fibers, and extract the broom-like bifurcation depth and the cut length of single fibers. Based on the broom-like bifurcation depth and the cut-off length of a single fiber, the degree of length unevenness and the proportion of the bifurcation segment to the total fiber length are calculated to obtain the end bifurcation cut-off feature set. Identify single fibers in the end-bidding truncation feature set whose bifurcation segment accounts for more than a preset upper limit of the fiber length and cannot be held by the rollers in the stretching zone, and obtain a floating fiber group. The transverse cohesive friction force of the floating fiber group in the carding web is obtained, and the fiber strength of the floating fiber group is determined by comparing the measured value of the transverse cohesive friction force with the measured value of the length non-uniformity with a preset strength reference range. By screening the planktonic fiber population whose fiber strength is lower than the preset lower limit, a weak-strength planktonic fiber population is obtained. The upper limit of the allowable depth of the broom-shaped branching is dynamically adjusted according to the degree of length unevenness to obtain the target depth upper limit. Obtain historical broom-shaped fork depth records for each speed range, filter out speed ranges where the broom-shaped fork depth does not exceed the upper limit of the target depth from the historical broom-shaped fork depth records, and determine the target beater speed; Adjust the mixing and opening production line according to the target beater speed, and determine the critical draw ratio based on the tensile breaking data of the strips obtained after the adjustment and combing.

2. The method according to claim 1, characterized in that, The process of acquiring end morphology images and full-length images of individual fibers of the target batch of recycled carbon fibers from the mixed opening production line, identifying the broom-like forking contours and the fracture ends of individual fibers, and extracting the broom-like forking depth and the cut length of individual fibers includes: A linear industrial camera, in conjunction with backlighting, acquires a magnified image of the end region of a single fiber. An area-array industrial camera acquires a full-length image of the single fiber covering both ends of the fiber in a flat manner along the fiber spreading direction. A one-to-one correspondence between the two types of images is established according to the fiber number, resulting in an indexed set of fiber dual-view images.

3. The method according to claim 2, characterized in that, The obtained set of indexed fiber dual views includes: The end morphology images in the indexed fiber dual-view set are segmented using the Otsu thresholding method to separate the fiber body and the broom-shaped bifurcation region. The binary image is refined to obtain the end skeleton. The bifurcation start point is marked at the bifurcation point of the skeleton. The connected region from the bifurcation start point to the farthest thin branch end point of the tip is determined as the broom-shaped bifurcation profile. Search for the locations of abrupt grayscale changes at both ends along the fiber principal axis of the full-length image of a single fiber as the locations of the break points at both ends; The axial projection distance between the bifurcation start point and the farthest thin branch end point is measured as the broom-shaped bifurcation depth, and the pixel span between the two ends of the break is converted into the single fiber cut length according to the camera calibration scale.

4. The method according to claim 1, characterized in that, The method involves calculating the degree of length unevenness and the proportion of the forked segment to the total fiber length based on the broom-like bifurcation depth and the cut-off length of the single fiber, thereby obtaining a set of end-bifurcation cut-off features, including: The lengths of all individual fibers cut off are compiled into a length series according to the fiber number. The ratio of the standard deviation to the arithmetic mean of the length series is used as the length dispersion coefficient to determine the degree of length non-uniformity. The proportion of the bifurcation segment to the total length of the fiber is obtained by dividing the broom-shaped bifurcation depth under the same fiber number by the cut length of a single fiber. The degree of length unevenness and the proportion of the bifurcation segment to the total fiber length are summarized into the same data table according to the fiber number to obtain the end bifurcation truncation feature set.

5. The method according to claim 1, characterized in that, The identification of single fibers in the end-branching feature set where the proportion of the bifurcation segment to the total fiber length exceeds a preset upper limit and thus cannot be held by the rollers in the drafting zone, yields a floating fiber population, including: The proportion of the bifurcation segment to the total fiber length under each fiber number in the end bifurcation truncation feature set is compared with the preset upper limit of proportion for each fiber. Fiber numbers that exceed the preset upper limit of proportion and their corresponding single fiber truncation lengths are written into the planar fiber data table in ascending order of fiber number. The planar fiber data table is then identified as the planar fiber population.

6. The method according to claim 1, characterized in that, The step of obtaining the lateral cohesive friction force of the planar fiber community in the carding web, and determining the fiber strength of the planar fiber community by comparing the measured value of the lateral cohesive friction force with the measured value of the length non-uniformity to a preset strength reference range, includes: For each fiber numbered individual fiber in the floating fiber group, it is pulled along a direction perpendicular to the main fiber axis until it slides relative to the adjacent fiber. The peak value of the reaction force is collected by a micro-force sensor as the lateral cohesive friction force. The measured values ​​of the lateral cohesive friction force and the length non-uniformity are combined into binary data points, which are then projected into a two-dimensional lookup table that divides several strength level intervals based on the lateral cohesive friction force and the length non-uniformity. The corresponding strength value is determined according to the position where the data falls. The arithmetic mean of all representative strength values ​​of the planar fiber population is taken as the fiber strength.

7. The method according to claim 1, characterized in that, The process of screening planktonic fiber populations with fiber strength below a preset lower limit yields a weak-strength planktonic fiber population. The upper limit of the allowable depth of the broom-like branching is dynamically adjusted based on the degree of length unevenness to obtain the target upper limit, including: The cut lengths of all individual fibers in the weak-strength planktonic fiber population are compiled into a weak population length series, and the ratio of the standard deviation to the arithmetic mean is used as the weak population length dispersion coefficient. The corresponding reduction ratio is obtained by looking up the corresponding reduction ratio in the pre-established table of the relationship between the degree of length unevenness and the reduction ratio. The target depth upper limit is obtained by multiplying the current upper limit of the broom-shaped branching depth by 1 and the difference between the result and the reduction ratio.

8. The method according to claim 1, characterized in that, The process of acquiring historical broom-shaped bifurcation depth records for each speed range, filtering out speed ranges where the broom-shaped bifurcation depth does not exceed the upper limit of the target depth from the historical broom-shaped bifurcation depth records, and determining the target hitter speed includes: The historical broom-shaped fork depth values ​​are collected from the historical process database according to the preset beater speed range, and the historical broom-shaped fork depth records are obtained. The beater speed ranges with a broom-shaped fork depth not exceeding the upper limit of the target depth are selected as candidate speed ranges, and the highest speed value among the candidate speed ranges is determined as the target beater speed.

9. The method according to claim 1, characterized in that, The process of adjusting the mixed opening production line according to the target beater speed, and determining the critical draw ratio based on the tensile breaking data of the sliver obtained after the adjustment and carding, includes: The target beater speed is written back to the beater motor control unit, and the strips at the output end of the combing process are cut according to the preset initial length of the sample as tensile test samples. Tensile force is applied to each tensile specimen at a constant rate along the specimen axis, and the tensile force value and specimen elongation at the moment of fracture are collected and written into the tensile fracture data of the strip. The sum of the sample elongation and the initial length of the sample is divided by the initial length of the sample to obtain the draw ratio at break. The arithmetic mean is taken and the preset safety margin is lowered to obtain the critical draw ratio.

10. The method according to claim 4, characterized in that, The method of summarizing the degree of length unevenness and the proportion of the bifurcation segment to the total fiber length in the same data table according to fiber number, yields the end bifurcation truncation feature set, including: Using the fiber number as the primary key, the degree of length unevenness is appended to each row of records as a common field, and the proportion of the forked segment to the total fiber length and the cut length of a single fiber are stored as fiber-level fields aligned.

Images