A prediction method for recycled carbon fiber pathways combined with neural networks

Abstract

The present application provides a prediction method for a recycled carbon fiber pathway combined with a neural network, comprising: using a continuous imaging method to monitor the entire cracking process in real time, obtaining a surface morphology image sequence, analyzing the surface morphology image sequence through an image processing method, and judging the changing trend of the carbon fiber surface groove depth and roughness micromorphology parameters; establishing a nonlinear mapping relationship between the carbon fiber surface structure and the recycled quality based on the carbon fiber surface groove depth and roughness micromorphology parameters, and outputting a corresponding recycled quality prediction value when a numerical value representing the carbon fiber surface structure parameter is input; predicting the quality and performance of the recycled carbon fiber for recycled carbon fiber samples obtained under different cracking process conditions, analyzing and comparing the prediction results, and optimizing the cracking process parameters.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a prediction method for recycling carbon fiber pathways combined with a neural network. Background Art

[0002] In the carbon fiber recycling process, the pyrolysis stage is one of the key steps. During this stage, the surface micromorphology of the carbon fiber changes significantly, such as the change in the depth of the surface grooves resulting in obvious differences in roughness. These changes in surface structure have an important impact on the performance of the final recycled carbon fiber, especially its mechanical properties and application potential. However, there is an urgent need to conduct in-depth research and establish a mapping relationship between surface structure and recycling quality to reveal the intrinsic mechanism of how structural changes affect the performance of carbon fiber. Specifically, due to the complex and diverse changes in the surface micromorphology of carbon fibers during the recycling process and their obvious anisotropic characteristics, traditional characterization methods such as scanning electron microscopy (SEM) and atomic force microscopy (AFM) are difficult to fully characterize these fine structures. At the same time, the mechanism by which external factors such as temperature and pressure affect the evolution of surface structure during the pyrolysis process is still unclear, resulting in insufficient accuracy in the correlation model between surface structure and recycling quality. In addition, existing quality assessment methods mainly rely on macroscopic mechanical property tests, ignoring the influence of surface structure and lacking quantitative correlation between surface micromorphology and mechanical properties. Therefore, it is urgent to develop new surface structure characterization technologies, combine them with advanced neural network algorithms, establish an accurate mapping model of surface micromorphology and recycling quality, clarify the evolution law of carbon fiber surface structure and its influence mechanism on mechanical properties, and provide theoretical guidance and technical support for high-quality recycled carbon fiber. Summary of the Invention

[0003] The present invention provides a method for predicting the recycling of carbon fiber pathways in combination with a neural network, which mainly includes:

[0004] Obtain recycled carbon fiber samples, observe the carbon fiber surface micromorphology using a scanning electron microscope to obtain surface morphology image data, scan the carbon fiber surface to obtain a three-dimensional topographic map, digitize the obtained surface morphology image, measure the carbon fiber surface groove depth using a line scan analysis method, and obtain the carbon fiber surface micromorphology characterization parameter values, including groove depth and roughness;

[0005] The entire cracking process is monitored in real time using a continuous imaging method to obtain a surface morphology image sequence, which is analyzed using an image processing method to determine the changing trends of the carbon fiber surface groove depth and roughness micromorphology parameters;

[0006] A nonlinear mapping relationship between carbon fiber surface structure and recycling quality is established based on the microscopic morphological parameters of carbon fiber surface groove depth and roughness. When the numerical value representing the carbon fiber surface structure parameter is input, the corresponding recycling quality prediction value is output;

[0007] X-ray tomography is used to obtain the internal three-dimensional structural data of carbon fiber. The annular structure image of the carbon fiber cross section is obtained through image reconstruction. The integrity and uniformity of the annular structure are quantitatively analyzed based on the image to determine the quality evaluation index of the annular structure.

[0008] The carbon fiber surface micromorphology characterization parameters and the ring structure quality evaluation index are integrated to evaluate the comprehensive performance of recycled carbon fiber. The effects of surface structure changes and ring structure characteristics on the mechanical properties and electrical conductivity of recycled carbon fiber are analyzed, and the correlation law of carbon fiber structure and performance is obtained.

[0009] For the recycled carbon fiber samples obtained under different pyrolysis process conditions, the quality and performance of the recycled carbon fiber are predicted, the prediction results are analyzed and compared, and the pyrolysis process parameters are optimized.

[0010] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0011] The present invention discloses a prediction method for recycling carbon fiber pathways combined with a neural network. The method obtains carbon fiber surface micromorphology data through scanning electron microscopy and nanometer-level resolution scanning, and uses line scanning analysis to measure surface groove depth and roughness. At the same time, X-ray tomography technology is used to obtain internal three-dimensional structural data of carbon fiber to analyze the integrity and uniformity of the annular structure. The present invention integrates surface micromorphology parameters with annular structure quality indicators to construct a comprehensive performance evaluation model to analyze the impact of structural changes on mechanical properties and conductive properties. In addition, a nonlinear mapping model of surface structure and recycling quality is established through a neural network algorithm to achieve recycling quality prediction. Finally, the present invention uses these models to predict the quality and performance of recycled carbon fibers under different cracking process conditions, optimize the cracking process parameters, and thus improve the quality and performance of the recycled carbon fibers. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 The present invention is a flowchart of a method for predicting a recycling carbon fiber pathway combined with a neural network. DETAILED DESCRIPTION

[0013] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0014] like Figure 1 In this embodiment, a prediction method for recycling carbon fiber pathways using a neural network may specifically include:

[0015] Step S101: obtain a recycled carbon fiber sample, observe the surface micromorphology of the carbon fiber using a scanning electron microscope to obtain surface morphology image data, scan the carbon fiber surface to obtain a three-dimensional topographic map, digitize the obtained surface morphology image, measure the groove depth of the carbon fiber surface by line scanning analysis, and obtain the numerical value of the carbon fiber surface micromorphology characterization parameter, including groove depth and roughness.

[0016] The image is locally enhanced according to the grayscale value distribution characteristics of the carbon fiber surface microscopic image. After the edge of the carbon fiber surface groove contour is extracted, the image noise is filtered out using wavelet transform. The carbon fiber surface groove area is marked through morphological operation, and the contour data after the marked grayscale value distribution is obtained according to the grayscale value distribution in the marked area; a carbon fiber surface depth map is established based on the contour data after the marked grayscale value distribution, the groove area is segmented, and the local depth change is calculated by the grayscale value difference of adjacent pixel points to obtain the groove depth value; for the carbon fiber surface depth map, the texture features are extracted through the grayscale co-occurrence matrix to obtain the surface roughness data.

[0017] Exemplarily, an image acquisition parameter library is established based on the grayscale distribution characteristics of a carbon fiber surface microscopic image. A subregion segmentation method is used to perform local image enhancement. A contrast adaptive adjustment algorithm is used to highlight surface texture features. Edges are extracted from the carbon fiber surface groove contours to obtain first contour data. For the first contour data, a wavelet transform is used to filter image noise. Morphological operations are used to mark the carbon fiber surface groove regions. The spacing between adjacent contour lines is calculated based on the grayscale distribution within the marked regions to obtain second contour data. A carbon fiber surface depth map is established based on the second contour data. A region growing algorithm is used to segment the groove regions. Local depth variations are calculated based on the grayscale differences between adjacent pixels to obtain groove depth values. For the depth map, a grayscale co-occurrence matrix is ​​used to extract surface texture features. Surface roughness data is obtained by calculating statistical parameters such as angular second moment, correlation, and entropy. A three-dimensional point cloud model is constructed based on the groove depth values ​​and surface roughness data. The point cloud data is smoothed using a Gaussian kernel function, and the three-dimensional surface equation of the carbon fiber surface is fitted using the least squares method. Based on the three-dimensional surface equation, a surface feature extraction algorithm is used to calculate the principal curvature and Gaussian curvature. The curvature distribution characteristics are used to obtain the carbon fiber surface concavity parameters, resulting in the quantitative results of the carbon fiber surface micromorphology. In carbon fiber surface micromorphology analysis, the grayscale value distribution characteristics are an important data basis for characterizing the carbon fiber surface structure. Statistical analysis of the grayscale histogram reveals that the grayscale value distribution in the groove area of ​​the carbon fiber surface exhibits a bimodal characteristic, where the brighter areas correspond to the raised portions of the fiber surface, with grayscale values ​​ranging from 180 to 220, and the darker areas correspond to the grooves, with grayscale values ​​ranging from 90 to 130. Based on this characteristic, an image acquisition parameter library can be established. During the local image enhancement process, the subregion size is set to 16×16 pixels, and adaptive contrast adjustment is performed within each subregion. When the difference between the maximum and minimum grayscale values ​​within the region is less than 50, histogram equalization is used to enhance the regional contrast, making the groove edge features more prominent. The enhanced image was subjected to wavelet transform for noise filtering, using the db4 wavelet basis function. Soft thresholding was applied to the high-frequency coefficients, with the threshold set at three times the standard deviation. During the depth map construction phase, the region-growing algorithm's growth criterion was based on the grayscale value differences between adjacent pixels. Seed points were selected in the region with the lowest grayscale value. When the grayscale value difference between adjacent pixels was less than 20, they were included in the same region. This method accurately delineated the boundaries of the groove region. For marked groove regions, the vertical distance between adjacent contour lines was calculated to obtain the depth value. For example, the spacing between adjacent contour lines in a particular groove region was 25 pixels, corresponding to an actual depth of 2.5 microns.Surface texture feature extraction utilizes a grayscale co-occurrence matrix method, with pixel displacement set to 1 and directional angles set to 0, 45, 90, and 135 degrees. The calculated second-order angular moment reflects the image's uniformity, the entropy characterizes its complexity, and the correlation describes the correlation of local grayscale values. Together, these parameters constitute a quantitative indicator of surface roughness. In 3D surface reconstruction, point cloud data is smoothed using a Gaussian kernel function with a standard deviation of 3 and a window size of 5×5. This parameter setting effectively removes measurement noise while preserving surface detail. The principal curvature calculated from surface feature extraction reflects the degree of surface curvature in different directions, while the Gaussian curvature comprehensively characterizes the surface's local geometric properties. These curvature parameters accurately describe the microscopic morphology of the carbon fiber surface. The quantified micromorphology results show a root mean square (RMS) surface roughness of 0.8 microns, a maximum groove depth of 3.2 microns, and an average groove spacing of 12 microns.

[0018] Step S102: Using a continuous imaging method to monitor the entire cracking process in real time, obtaining a surface morphology image sequence, and analyzing the surface morphology image sequence using an image processing method to determine the change trend of the carbon fiber surface groove depth and roughness micromorphology parameters.

[0019] According to the carbon fiber surface morphology image sequence, the continuous frame images are spatially aligned using the image registration algorithm, and the displacement vector field between adjacent frames is calculated by the optical flow method to obtain the first sequence of image data; for the first sequence of image data, the image sequence background is extracted by temporal median filtering, and the change area is extracted by the background difference method to obtain the second sequence of image data; for the second sequence of image data, the surface groove contour is extracted by the morphological operator, and the three-dimensional coordinates of the contour feature points are calculated by the least squares method to obtain the groove depth time series data; for the second sequence of image data, the surface texture features are extracted by the gray level co-occurrence matrix, and the local statistical parameters are calculated by the sliding window to obtain the surface roughness time series data.

[0020] Exemplarily, a time-series image library is established based on a sequence of carbon fiber surface topography images, and continuous frame images are spatially aligned using an image registration algorithm. The displacement vector field between adjacent frames is calculated using the optical flow method to obtain a first sequence of image data. For the first sequence of image data, a temporal median filter is used to extract the image sequence background, and the changing area in each frame of the image is extracted using the background difference method. The changing area is marked using an adaptive threshold segmentation algorithm to obtain a second sequence of image data. Based on the second sequence of image data, a morphological operator is used to extract the surface groove contour, and the three-dimensional coordinates of the contour feature points are calculated using the least squares method to obtain the groove depth time-series data. For the groove depth time-series data, a Kalman filter is used for data smoothing, and the depth change rate is calculated using the recursive least squares method to obtain a depth trend curve. Based on the second sequence of image data, a gray-level co-occurrence matrix is ​​used to extract surface texture features, and local statistical parameters are calculated using a sliding window to obtain surface roughness time-series data. Based on the depth trend curve and roughness time series data, a support vector regression method was used to construct a morphology parameter prediction function. Local curvature analysis was used to determine the inflection point of parameter changes, and the evolution characteristics of the carbon fiber surface micromorphology parameters were obtained. In the real-time monitoring of carbon fiber surface morphology, image registration is a key step in ensuring data continuity. Precise alignment is achieved by calculating the displacement vector field between adjacent frames. When the image acquisition frequency is set to 10 frames per second, the maximum displacement between adjacent frames typically does not exceed 5 pixels. In this case, the optical flow method can accurately capture the subtle deformation process. The time series image background extraction method uses a temporal median filtering method. For each pixel position, the median of the grayscale values ​​of the corresponding position in all frames within the time window is taken as the background value. When the window size is set to 15 frames, it has good robustness, effectively suppressing random noise and accurately reflecting slow background changes. The change area obtained by background subtraction is segmented using an adaptive threshold. The threshold is set to twice the standard deviation of the local area to achieve accurate detection of weak changes. To extract groove contour features, a morphological operator uses a 3×3 structuring element for opening and closing operations, eliminating small noise while maintaining contour integrity. The extracted contour points are discretely distributed in space, and a continuous depth curve is obtained through least squares fitting. The time series of depth data shows a clear phased pattern. In the initial stage of cracking, the groove depth increases rapidly, increasing by an average of 0.5 microns per minute. As the cracking process progresses, the growth rate gradually decreases. Surface texture feature analysis uses the gray-level co-occurrence matrix method, with a displacement of 1 pixel and directions including 0, 45, 90, and 135 degrees. The calculated statistical parameters such as the angular second moment and entropy value are plotted over time to quantitatively characterize the evolution of surface roughness. Experimental data show that during the cracking process, the surface roughness increases first and then stabilizes. The root mean square roughness is 0.3 microns in the initial stage, increases to 0.8 microns after 15 minutes, and remains essentially stable after 30 minutes.The morphology parameters were predicted using support vector regression, with a nonlinear mapping relationship established through a radial basis kernel function. A cross-validation method was used for kernel parameter optimization, with the root mean square error on the validation set kept within 0.05 microns. Local curvature analysis revealed that the curves for the change in groove depth and surface roughness exhibited a clear inflection point at approximately 40% of the cracking rate. This is when the carbon fiber surface microstructure began to undergo a significant transformation, with the depth growth rate decreasing to less than 0.1 microns per minute and the roughness growth also tending to flatten.

[0021] Step S103: establishing a nonlinear mapping relationship between the carbon fiber surface structure and the recycling quality based on the microscopic morphology parameters of the carbon fiber surface groove depth and roughness. When the numerical value representing the carbon fiber surface structure parameter is input, the corresponding recycling quality prediction value is output.

[0022] A feature data set is constructed based on the numerical values ​​of the groove depth and roughness of the carbon fiber surface. The median filling method is used to process the missing values ​​in the feature data set. The outliers are eliminated by the triple standard deviation principle to obtain a data set after eliminating the outliers. The numerical values ​​are normalized by the standardization method, and the key dimensions in the data set after eliminating the outliers are extracted by the principal component analysis method to obtain a data set with extracted key dimensions. Based on the data set of the key dimensions, a nonlinear mapping structure is constructed using a residual neural network. The output of the hidden layer neurons is nonlinearly transformed by the rectified linear unit function to obtain a mapping function. For the mapping function, a fixed-length sliding window is used to segment the input data, and the predicted values ​​of each segment are obtained by forward calculation.

[0023] For example, a feature dataset is constructed based on the surface groove depth and roughness values ​​of the carbon fiber. Missing values ​​are addressed using the median filling method, and outliers are eliminated using the triple standard deviation principle, resulting in a preprocessed first dataset. For the first dataset, the values ​​are normalized using a standardization method, and key dimensions of the feature data are extracted using principal component analysis, resulting in a reduced-dimensional second dataset. For the second dataset, a nonlinear mapping structure is constructed using a five-layer residual neural network. The outputs of the hidden layer neurons are nonlinearly transformed using a rectified linear unit function to obtain a first mapping function. For the first mapping function, training and validation data are divided using a five-fold cross-validation method. The prediction deviation is calculated using a mean squared error loss function, and the network parameters are updated using a backpropagation algorithm to obtain a second mapping function. For the second mapping function, the input data is segmented using a fixed-length sliding window. Predicted values ​​for each segment are obtained through forward computation. The prediction results for each segment are fused using a weighted average method to obtain a carbon fiber recycling quality value. For the recycled quality value, the reliability range of the prediction result is calculated using a confidence interval method. Anomalous prediction results are identified using an outlier detection algorithm to obtain a carbon fiber recycling quality evaluation index. The carbon fiber surface feature dataset includes two core dimensions: groove depth and roughness. In the data preprocessing stage, missing values ​​are handled by filling the median. For example, if a set of groove depth data sequences is 2.5, 3.1, missing values, 2.8, and 3.2 microns, the median of 2.95 microns is used to fill the missing value positions to maintain the continuity and rationality of the data. The three-standard deviation criterion is used to eliminate outliers. When the measured value deviates from the mean by more than three standard deviations, it is identified as an outlier and eliminated. The minimum-maximum normalization method is used for data standardization to map the eigenvalues ​​of different dimensions to the range of 0 to 1, making the groove depth and roughness data comparable. The results of principal component analysis show that the cumulative variance contribution rate of the first two principal components is more than 95%. The first principal component mainly reflects the groove depth characteristics with a weight coefficient of 0.82, and the second principal component mainly reflects the surface roughness characteristics with a weight coefficient of 0.76. The residual neural network adopts a five-layer structure. The input layer corresponds to the feature dimension after dimensionality reduction. The number of neurons in the three middle hidden layers is 64, 32, and 16 respectively. The output layer corresponds to the recycling quality prediction value. The skip connection structure in the network effectively alleviates the gradient vanishing problem in the deep network training process. The introduction of the rectified linear unit function enhances the model's ability to express nonlinear features. During the cross-validation process, the dataset is evenly divided into 5 parts, 4 of which are used as training sets each time, and the remaining 1 is used as a validation set. The generalization performance of the model is ensured through 5 rounds of rotation validation. The training process adopts a mini-batch method with a batch size of 32. The initial value of the learning rate is set to 0.001. When there is no significant decrease in the error of the validation set for 3 consecutive rounds, the learning rate is automatically reduced to 0.1 times the original value.During the online prediction phase, the input data is segmented using a sliding window of length 10, with an overlap of 50% between adjacent windows to ensure smooth and stable predictions. The segmented predictions are then combined using an exponentially weighted average, with a weight coefficient of 0.7 for newer data and a weight coefficient of 0.3 for older data. A 95% confidence level is used for the confidence intervals of the predictions. Predictions with a confidence interval width exceeding twice the mean are marked as low confidence. Outlier detection is based on a local anomaly factor algorithm, calculating the density ratio of each predicted point to its neighboring points. When the density ratio is significantly lower than the mean, the prediction is marked as an outlier. Experimental data show that the model achieves a root mean square error of 0.15 on the validation set, with an average confidence interval width of 0.3 and a low percentage of outlier predictions below 5%.

[0024] In step S104, an X-ray tomography method is used to obtain the internal three-dimensional structural data of the carbon fiber, and an annular structure image of the carbon fiber cross section is obtained by an image reconstruction method. The integrity and uniformity of the annular structure are quantitatively analyzed based on the image to determine the quality evaluation index of the annular structure.

[0025] For carbon fiber tomography data, the scanning plane is spatially aligned using a marker point matching method, and the original data is denoised using a Gaussian filter to obtain a denoised image sequence; based on the denoised image sequence, the scattering attenuation data is compensated using an iterative back-projection algorithm, and edge features are extracted to obtain an image sequence after edge feature extraction; based on the image sequence after edge feature extraction, the cross-sectional profile is extracted, and the annular structure boundary is obtained through a morphological closing operation to obtain an image sequence after the annular structure boundary is obtained; for the image sequence after the annular structure boundary is obtained, the circularity calculation method is used to quantify the integrity of the annular structure, the annular structure density distribution is calculated using a gray-level co-occurrence matrix, and the internal pores are identified based on regional connectivity analysis to obtain structural defect data; based on the structural defect data, a feature extraction structure is constructed to obtain an annular structure quality evaluation index.

[0026] For example, spatial registration is performed on carbon fiber tomography data, alignment correction is performed on the scanned slices using a landmark matching method, and noise reduction is performed on the raw data using Gaussian filtering to obtain a first image sequence. For the first image sequence, scatter attenuation data is compensated using an iterative backprojection algorithm, edge features are extracted using a four-layer wavelet decomposition, and edge saliency is calculated based on gradient amplitude to obtain a second image sequence. For the second image sequence, cross-sectional contours are extracted using a region growing method, edge gaps are filled using a morphological closing operation, and the annular structure boundary is determined using a contour tracing algorithm to obtain a third image sequence. For the third image sequence, the annular structure integrity is quantified using a circularity calculation method, and the structural center offset is calculated using a radial symmetry assessment algorithm to obtain structural characteristic parameters. Based on these structural characteristic parameters, the annular structure density distribution is calculated using a gray-level co-occurrence matrix, structural uniformity is assessed using an entropy method, and internal pores are identified based on regional connectivity analysis to obtain structural defect data. Based on these structural defect data and structural characteristic parameters, a six-layer feature extraction structure is constructed using a deep convolutional neural network. Residual connections are used to optimize the feature transfer process to obtain an annular structure quality evaluation index. During the spatial registration of X-ray tomography images, the selection of markers significantly impacts the accuracy of the results. Four high-contrast markers are typically placed on the exterior of the carbon fiber, with the relative positional deviation between markers kept within 0.1 mm. After registration, the spatial correspondence error between adjacent slices is less than 0.05 mm. A Gaussian filter with a 5×5 kernel and a standard deviation of 1.2 effectively suppresses random noise while preserving edge features. For scatter attenuation compensation, an iterative backprojection algorithm employs 10 rounds of optimization, with a projection angle interval of 1 degree within each round. The signal-to-noise ratio (SNR) of the compensated image is improved by approximately 40%. Wavelet decomposition employs a two-dimensional discrete wavelet transform with a four-level decomposition using the db4 wavelet basis function, resulting in 16 subband images. The high-frequency subbands represent edge features, and the edge significance threshold is set at 2.5 times the local region standard deviation. For ring structure extraction, a region growing method uses the highest grayscale value as the seed point, with a growth threshold set at 85% of the seed value. Morphological closing employs a circular structuring element with a radius of 3 pixels to fill in small gaps in the edge contour. In the sequence of boundary points obtained by contour tracing, the spacing between adjacent points is maintained between 1 and 2 pixels, ensuring the continuity of the boundary description. Structural integrity evaluation uses a circularity parameter to calculate the deviation of the actual contour from the ideal circle. A circularity value between 0.95 and 1 indicates good structural integrity. For radial symmetry assessment, the contour is divided into 36 sectors, and the standard deviation of the distance from the center of gravity to the center point of each sector is calculated. A standard deviation less than 0.2 mm is considered to indicate good structural symmetry.The density distribution of the ring structure was calculated using a gray-level co-occurrence matrix method, with a displacement of 1 pixel and directional angles of 0, 45, 90, and 135 degrees. The calculated entropy value reflects the uniformity of the structure, with an entropy value less than 3.5 indicating good structural uniformity. Internal pores were identified using an 8-neighborhood connectivity analysis. Connected regions with an area less than 5 square pixels were labeled as micropores, and regions with an area greater than 20 square pixels were labeled as significant defects. The deep convolutional neural network consists of six convolutional layers with a 3×3 kernel size and the number of channels, 32, 64, 128, 256, 512, and 512, respectively. A residual connection was added between every two layers to avoid loss of deep feature information. The last two layers use a fully connected structure to output a quality score for the carbon fiber ring structure, ranging from 0 to 100, with scores above 90 indicating excellent structural quality, 80 to 90 indicating good quality, 70 to 80 indicating acceptable quality, and scores below 70 indicating significant defects.

[0027] In step S105, the carbon fiber surface micromorphology characterization parameters and the annular structure quality evaluation index are integrated to evaluate the comprehensive performance of the recycled carbon fiber, analyze the influence of the surface structure changes and the annular structure characteristics on the mechanical properties and conductive properties of the recycled carbon fiber, and obtain the correlation law of the carbon fiber structure performance.

[0028] After normalization processing based on the carbon fiber surface micromorphology characterization parameters and the annular structure quality indicators, outlier data points are eliminated; the principal component analysis method is used to reduce the dimension of the characteristic data, the number of principal components is determined by the variance contribution rate, and the coupling coefficient between parameters is calculated based on correlation analysis to obtain a characteristic matrix; based on the characteristic matrix, an adaptive neural network is used to construct a performance prediction structure, the network parameters are optimized by the back propagation algorithm, and the prediction accuracy is calculated based on the mean square error function; the mechanical property prediction function is constructed by the multivariate regression method, the key influencing factors are screened by the significance test, and the contribution of each parameter is calculated based on variance analysis to obtain the mechanical property mapping relationship; the conductivity performance prediction function is established by the sensitivity analysis method, and the parameter action order is determined by partial correlation analysis to obtain the conductivity performance mapping relationship; based on the mechanical property mapping relationship and the conductivity performance mapping relationship, the carbon fiber structure performance correlation law is obtained.

[0029] For example, an original data set is established based on the carbon fiber surface micromorphology characterization parameters and the annular structure quality index. The numerical values ​​are normalized using a standardization method, and outlier data points are removed through outlier detection to obtain a first characteristic matrix. For the first characteristic matrix, the characteristic data are reduced in dimension using principal component analysis, the number of principal components is determined by variance contribution rate, and the coupling coefficient between parameters is calculated based on correlation analysis to obtain a second characteristic matrix. Based on the second characteristic matrix, a five-layer adaptive neural network is used to construct a performance prediction structure, the network parameters are optimized using a backpropagation algorithm, and the prediction accuracy is calculated based on the mean square error function to obtain a third characteristic matrix. For the third characteristic matrix, a multivariate regression method is used to construct a mechanical property prediction function, the key influencing factors are screened through significance testing, and the contribution of each parameter is calculated based on variance analysis to obtain a mechanical property mapping relationship. Based on the mechanical property mapping relationship, a sensitivity analysis method is used to establish a conductive property prediction function, the parameter action order is determined through partial correlation analysis, and the prediction accuracy is verified based on the orthogonal test method to obtain a conductive property mapping relationship. A weighted fusion algorithm was used to construct a comprehensive evaluation function for the mapping relationship between mechanical properties and electrical conductivity. The weight coefficients were determined through cross-validation, resulting in a correlation between carbon fiber structure and performance. The original dataset, consisting of carbon fiber surface micromorphology and annular structure quality indicators, encompassed multiple dimensions: groove depth ranged from 0.5 to 5 microns, surface roughness from 0.2 to 2 microns, annular structure roundness from 0.85 to 1, and structural uniformity from 0.6 to 0.95. Data normalization employed a minimum-maximum normalization method, and outlier detection was based on the triple standard deviation criterion, removing data points that deviated significantly from the mean. Principal component analysis revealed that the cumulative variance contribution of the first three principal components reached 92%. The first principal component primarily reflects surface morphology, explaining 48% of the data variance; the second principal component reflects annular structure characteristics, explaining 32% of the variance; and the third principal component, associated with structural defects, explains 12% of the variance. Analysis of the coupling coefficient between features found that the surface groove depth and the annular structure uniformity showed a significant negative correlation, with a correlation coefficient of -0.75. The adaptive neural network adopts a five-layer structure. The input layer corresponds to the feature dimension after dimensionality reduction. The number of neurons in the three hidden layers is 64, 32, and 16 respectively. The output layer corresponds to the predicted values ​​of mechanical properties and conductive properties. The network training adopts the stochastic gradient descent method with a batch size of 32. The initial learning rate is 0.001. When the error of the validation set does not decrease significantly after 5 consecutive rounds, the learning rate is automatically reduced to 0.1 times the original value. In the prediction of mechanical properties, the significance test adopts a confidence level of 0.05. The results of variance analysis show that the surface groove depth has the highest contribution, reaching 45%, followed by the annular structure roundness, with a contribution of 30%, and the structural uniformity index with a contribution of 15%. The relative error of the prediction model on the validation set is controlled within 8%.In terms of conductive performance prediction, sensitivity analysis shows that the integrity of the annular structure is the most critical factor, with a sensitivity coefficient of 0.68, followed by surface defect density, with a sensitivity coefficient of 0.42. In the comprehensive evaluation function, the weight coefficients of mechanical properties and conductive properties were determined by five-fold cross validation, which were 0.6 and 0.4 respectively. The evaluation results show that when the surface groove depth is below 2.5 microns, the roundness of the annular structure is greater than 0.95, and the structural uniformity index is higher than 0.85, the comprehensive performance of carbon fiber is optimal. Correlation analysis found that the influence of surface micromorphology on mechanical properties is more significant, explaining 75% of the performance variation, while the annular structure characteristics have a greater impact on conductive properties, explaining 82% of the conductive performance changes.

[0030] Step S106 , predicting the quality and performance of the recycled carbon fiber samples obtained under different pyrolysis process conditions, analyzing and comparing the prediction results, and optimizing the pyrolysis process parameters.

[0031] The surface structure data of recycled carbon fibers under different pyrolysis process conditions were obtained, and the correlation coefficients between process parameters and structural characteristics were calculated through correlation analysis. A nonlinear mapping function was used to calculate the recovery mass prediction value of each group of samples, and the accuracy of the recovery mass prediction value was evaluated through cross-validation. A comprehensive evaluation function was used to construct a multi-objective optimization equation, and the weight coefficients of performance indicators were determined through the entropy weight method. The process parameters were grouped, and the optimal solution of the performance indicators was found through the gradient descent method to obtain the process parameter optimization scheme.

[0032] Exemplarily, a feature database is established based on surface structural data of recycled carbon fibers under different pyrolysis process conditions. Correlation analysis is used to calculate the correlation coefficient between process parameters and structural features. Regularization is used to eliminate dimensional effects, resulting in a first data set. For the first data set, a nonlinear mapping function is used to calculate the recovery mass prediction value for each sample group. A validation sample set is generated using the Monte Carlo method. The prediction accuracy is evaluated based on cross-validation, resulting in a second data set. Based on the second data set, a multi-objective optimization equation is constructed using a comprehensive evaluation function. The weight coefficients of each performance indicator are determined using the entropy weight method. The prediction results are calibrated using a Bayesian optimizer, resulting in a third data set. For the third data set, a clustering algorithm is used to group sample features. The significance of the influence of process parameters is calculated using variance analysis. A parameter-response relationship is constructed using a decision tree algorithm, resulting in a fourth data set. Based on the fourth data set, a grid search method is used to traverse the pyrolysis process parameters. The parameter variation boundaries are determined using a constraint propagation algorithm. The optimal solution for the performance indicator is found using a gradient descent method, resulting in a fifth data set. For the fifth data set, an adaptive optimization algorithm was used to locally tune the process parameters. Sensitivity analysis was used to determine the direction of parameter adjustment. The reliability of the optimization results was assessed based on confidence intervals to obtain an optimized solution for the pyrolysis process parameters. The pyrolysis process conditions affected the surface structure of carbon fibers in multiple ways. Correlation analysis revealed a significant positive correlation between pyrolysis temperature and surface groove depth, with a correlation coefficient of 0.82. When the temperature increased from 350 to 450 degrees Celsius, the groove depth increased from 0.8 microns to 2.5 microns. The correlation coefficient between pyrolysis time and surface roughness was 0.65, with surface roughness increasing by 0.3 microns for every 30-minute extension of pyrolysis time. The Monte Carlo method was used to evaluate the prediction accuracy of 1,000 validation samples. During cross-validation, the root mean square error of the recovered mass prediction was 0.12, and the correlation coefficient between the predicted and measured values ​​reached 0.89. 95% of the prediction errors in the validation samples were within ±15%, demonstrating the good generalization performance of the prediction model. In the comprehensive performance evaluation, the entropy weight method was used to determine the weight of each performance indicator, where the weight of tensile strength was 0.35, the weight of elastic modulus was 0.25, the weight of electrical conductivity was 0.25, and the weight of surface structural integrity was 0.15. The Bayesian optimization process established the parameter response surface through Gaussian process regression, and the expected improvement criterion was selected as the acquisition function. 10 sampling points were selected for evaluation in each round of optimization. Cluster analysis results showed that the samples can be divided into three characteristic groups. The first group is characterized by high strength and low conductivity, the second group is characterized by medium strength and high conductivity, and the third group is characterized by balanced performance. Variance analysis showed that the pyrolysis temperature had the most significant effect on tensile strength, with an F statistic of 28.6, and the pyrolysis time had the greatest effect on surface structural integrity, with an F statistic of 22.4.During parameter optimization, the grid search was set to a temperature step of 10 degrees Celsius and a time step of 5 minutes. Constraints included an upper temperature limit of 500 degrees Celsius and an upper time limit of 120 minutes. Gradient descent employed an adaptive learning rate with an initial value of 0.01. When no significant improvement was observed after three consecutive rounds of optimization, the learning rate was reduced to 0.5 times the original value. Local optimization employed an adaptive algorithm, dynamically adjusting the parameter search range based on the degree of improvement in performance indicators. The temperature adjustment accuracy was ±5 degrees Celsius, and the time adjustment accuracy was ±2 minutes. Sensitivity analysis showed that near the optimal parameter point, a 1-degree Celsius temperature change resulted in a 0.5% change in strength, and a 1-minute time change resulted in a 0.3% change in structural integrity. A 95% confidence interval analysis of the optimization results indicated that, under the selected optimal process parameters, the expected tensile strength was 85% ± 3% of the original strength, and the expected electrical conductivity was 90% ± 2% of the original conductivity.

[0033] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for predicting the recycling of carbon fiber pathways using a neural network, characterized in that: The method comprises: Obtain recycled carbon fiber samples, observe the carbon fiber surface micromorphology using a scanning electron microscope to obtain surface morphology image data, scan the carbon fiber surface to obtain a three-dimensional topographic map, digitize the obtained surface morphology image, measure the carbon fiber surface groove depth using a line scan analysis method, and obtain the carbon fiber surface micromorphology characterization parameter values, including groove depth and roughness; The entire cracking process is monitored in real time using a continuous imaging method to obtain a surface morphology image sequence, which is analyzed using an image processing method to determine the changing trends of the carbon fiber surface groove depth and roughness micromorphology parameters; A nonlinear mapping relationship between carbon fiber surface structure and recycling quality is established based on the microscopic morphological parameters of carbon fiber surface groove depth and roughness. When the numerical value representing the carbon fiber surface structure parameter is input, the corresponding recycling quality prediction value is output; X-ray tomography is used to obtain the internal three-dimensional structural data of carbon fiber. The annular structure image of the carbon fiber cross section is obtained through image reconstruction. The integrity and uniformity of the annular structure are quantitatively analyzed based on the image to determine the quality evaluation index of the annular structure. The carbon fiber surface micromorphology characterization parameters and the ring structure quality evaluation index are integrated to evaluate the comprehensive performance of recycled carbon fiber. The effects of surface structure changes and ring structure characteristics on the mechanical properties and electrical conductivity of recycled carbon fiber are analyzed, and the correlation law of carbon fiber structure and performance is obtained. For the recycled carbon fiber samples obtained under different pyrolysis process conditions, the quality and performance of the recycled carbon fiber are predicted, the prediction results are analyzed and compared, and the pyrolysis process parameters are optimized.

2. The method according to claim 1, characterized in that The method comprises obtaining a recycled carbon fiber sample, observing the surface micromorphology of the carbon fiber using a scanning electron microscope to obtain surface topography image data, scanning the carbon fiber surface to obtain a three-dimensional topographic map, digitizing the obtained surface topography image, measuring the groove depth of the carbon fiber surface by a line scanning analysis method, and obtaining the carbon fiber surface micromorphology characterization parameter value, including the groove depth and roughness, including: The image is locally enhanced based on the grayscale distribution characteristics of the carbon fiber surface microscopic image. After edge extraction of the carbon fiber surface groove contour, the image noise is filtered out using wavelet transform. The carbon fiber surface groove area is marked through morphological operation, and the contour data after the marked grayscale value distribution is obtained based on the grayscale value distribution in the marked area. Establishing a carbon fiber surface depth map based on the contour data after the grayscale value distribution of the markers, segmenting the groove area, and calculating the local depth change by the grayscale value difference of adjacent pixels to obtain the groove depth value; For the carbon fiber surface depth map, the surface roughness data is obtained by extracting texture features through the gray level co-occurrence matrix.

3. The method according to claim 1, characterized in that The method adopts a continuous imaging method to monitor the entire cracking process in real time, obtains a surface topography image sequence, and analyzes the surface topography image sequence by an image processing method to determine the change trend of the carbon fiber surface groove depth and roughness micro-topography parameters, including: According to the carbon fiber surface topography image sequence, the continuous frame images are spatially aligned using the image registration algorithm, and the displacement vector field between adjacent frames is calculated using the optical flow method to obtain the first sequence of image data; For the first sequence of image data, a temporal median filter is used to extract the image sequence background, and a background difference method is used to extract the changed area to obtain a second sequence of image data; For the second sequence of image data, a morphological operator is used to extract the surface groove contour, and the three-dimensional coordinates of the contour feature points are calculated by the least squares method to obtain the groove depth time series data; For the second sequence of image data, the gray level co-occurrence matrix is ​​used to extract surface texture features, and the local statistical parameters are calculated through a sliding window to obtain surface roughness time series data.

4. The method according to claim 1, wherein The nonlinear mapping relationship between the carbon fiber surface structure and the recycling quality is established based on the microscopic morphology parameters of the carbon fiber surface groove depth and roughness. When the numerical value representing the carbon fiber surface structure parameter is input, the corresponding recycling quality prediction value is output, including: A feature data set was constructed based on the groove depth and roughness values ​​of the carbon fiber surface. The missing values ​​in the feature data set were processed using the median filling method. Outliers were eliminated using the triple standard deviation principle to obtain a data set after outlier elimination. Normalizing the values ​​using a standardization method, extracting key dimensions from the data set after removing outliers using a principal component analysis method, and obtaining a data set with extracted key dimensions; According to the data set of the key dimension, a nonlinear mapping structure is constructed using a residual neural network, and a nonlinear transformation is performed on the output of the hidden layer neurons through a rectified linear unit function to obtain a mapping function; For the mapping function, a fixed-length sliding window is used to segment the input data, and the prediction value of each segment is obtained through forward calculation.

5. The method according to claim 1, characterized in that The method uses an X-ray tomography method to obtain the internal three-dimensional structure data of the carbon fiber, obtains the annular structure image of the carbon fiber cross section through an image reconstruction method, and quantitatively analyzes the integrity and uniformity of the annular structure based on the image to determine the quality evaluation index of the annular structure, including: For carbon fiber tomography data, the scanning planes are spatially registered using the marker point matching method, and the original data are denoised using Gaussian filtering to obtain a denoised image sequence. According to the image sequence after the noise reduction processing, after compensating the scattering attenuation data using an iterative back-projection algorithm, edge features are extracted to obtain an image sequence after edge feature extraction; Extracting the cross-sectional contour according to the image sequence after edge feature extraction, obtaining the annular structure boundary through morphological closing operation, and obtaining the image sequence after the annular structure boundary is obtained; For the image sequence after obtaining the annular structure boundary, the circularity calculation method is used to quantify the integrity of the annular structure, the density distribution of the annular structure is calculated by gray-level co-occurrence matrix, and the internal pores are identified based on regional connectivity analysis to obtain structural defect data; According to the structural defect data, a feature extraction structure is constructed to obtain a ring structure quality evaluation index.

6. The method according to claim 1, characterized in that The carbon fiber surface micromorphology characterization parameters and the ring structure quality evaluation index are integrated to evaluate the comprehensive performance of the recycled carbon fiber, analyze the effects of surface structure changes and ring structure characteristics on the mechanical properties and conductive properties of the recycled carbon fiber, and obtain the correlation law of carbon fiber structure performance, including: After normalization based on the carbon fiber surface micromorphology characterization parameters and the ring structure quality index, outlier data points were removed; The principal component analysis method is used to reduce the dimension of the characteristic data, the number of principal components is determined by the variance contribution rate, and the coupling coefficient between parameters is calculated based on the correlation analysis to obtain the characteristic matrix; Based on the characteristic matrix, an adaptive neural network is used to construct a performance prediction structure, network parameters are optimized through a back propagation algorithm, and prediction accuracy is calculated based on a mean square error function; The mechanical properties prediction function was constructed using the multivariate regression method. The key influencing factors were screened through significance test. The contribution of each parameter was calculated based on variance analysis to obtain the mechanical properties mapping relationship. The conductivity prediction function was established by using sensitivity analysis method, and the order of parameter action was determined by partial correlation analysis to obtain the conductivity mapping relationship. Based on the mechanical property mapping relationship and the conductive property mapping relationship, a correlation law of carbon fiber structure properties is obtained.

7. The method according to claim 1, characterized in that The method predicts the quality and performance of the recycled carbon fiber samples obtained under different pyrolysis process conditions, analyzes and compares the prediction results, and optimizes the pyrolysis process parameters, including: Obtain surface structure data of recycled carbon fibers under different pyrolysis process conditions, and calculate the correlation coefficient between process parameters and structural characteristics through correlation analysis; A nonlinear mapping function is used to calculate the recovery mass prediction value of each group of samples, and the accuracy of the recovery mass prediction value is evaluated by cross-validation; A comprehensive evaluation function is used to construct a multi-objective optimization equation, and the weight coefficients of performance indicators are determined by the entropy weight method; The process parameters are grouped, and the optimal solution of the performance indicators is found through the gradient descent method to obtain the process parameter optimization plan.

Images