A method for monitoring carbon fiber carbonization quality based on electrical resistance method

Abstract

The present invention relates to a carbon fiber carbonization quality monitoring method based on the resistance method, comprising the following steps: connecting a bridge circuit measuring instrument to a pair of measuring shafts; winding a carbon fiber tow of a continuous production line on a pair of measuring shafts; starting the bridge circuit measuring instrument, automatically adjusting the value of an adjustable resistor through the bridge circuit measuring instrument, and maintaining the balance of the bridge; the bridge circuit measuring instrument continuously measures the resistance value of each carbon fiber tow between the measuring shafts; the resistance value of the carbon fiber tow fluctuates, and features related to the resistance value fluctuation are extracted by constructing a deep learning framework, and waveforms of predicted resistance and measured resistance are obtained from the known qualified sample resistance value and the carbon fiber tow resistance value measured in production; by comparing the waveforms of predicted resistance and measured resistance, it is determined whether the carbon fiber tow has quality problems. The present invention can efficiently and accurately monitor the resistance value in the carbon fiber production process, thereby performing online quality monitoring on the carbon fiber.

Description

Technical Field

[0001] The invention relates to a carbon fiber carbonization quality monitoring method, in particular to a carbon fiber carbonization quality monitoring method based on an electrical resistance method, and belongs to the technical field of quality detection in a carbon fiber production process. Background Art

[0002] As an important component of high-performance composite materials, carbon fiber is widely used in aerospace, automotive industry, sports equipment, building materials and other fields due to its light weight, high strength, high modulus and corrosion resistance.

[0003] The preparation process of carbon fiber usually includes the pyrolysis and carbonization stages of the original fiber. During the pyrolysis process, the original fiber is exposed to a high temperature environment to remove hydrogen and oxygen elements and form a structure with a high carbon content. The carbonization stage is a treatment carried out at a higher temperature to further increase the carbon content and crystallinity of the carbon fiber. Appropriate pyrolysis temperature and carbonization temperature can promote the removal of residual hydrogen and oxygen in the original fiber, while increasing the carbon content and crystallinity of the fiber. From the original fiber to carbonization, the appearance and performance of the carbon fiber undergo significant changes, forming a carbon structure with a high degree of purity and crystallinity, and possessing excellent mechanical and chemical properties.

[0004] However, insufficient carbonization of carbon fiber means that the residual hydrogen and oxygen in the raw silk in the corresponding production process exceeds the minimum standard of that link. These problems may reduce its durability and performance. Therefore, in order to improve the quality of carbon fiber and ensure the manufacture of high-quality composite materials, strict quality monitoring of the carbon fiber production process is required. However, due to the limitations of production equipment and the complexity of the operating environment, this test needs to be carried out after the carbon fiber production is completed, and it is impossible to perform online quality monitoring of the carbon fiber.

[0005] Therefore, in order to solve the above technical problems, it is necessary to provide an innovative carbon fiber carbonization quality monitoring method based on the resistance method to overcome the above defects in the prior art. Summary of the invention

[0006] The purpose of the present invention is to provide a carbon fiber carbonization quality monitoring method based on the resistance method, which can efficiently and accurately obtain the resistance value of the carbon fiber, and compare the waveform laws of the training model and the measured model obtained by deep learning between the resistance value of the carbon fiber of the produced qualified sample and the measured carbon fiber resistance value to monitor the quality of the carbon fiber.

[0007] To achieve the above object, the technical solution adopted by the present invention is: a method for monitoring the carbonization quality of carbon fiber based on the electrical resistance method, characterized in that it comprises the following steps:

[0008] 1) Prepare a bridge circuit measuring instrument and connect the bridge circuit measuring instrument to a pair of measuring axes through a line;

[0009] 2) The carbon fiber tows in the continuous production line in the carbonization furnace are wound on the pair of measuring shafts in an S-shaped manner;

[0010] 3) Start the bridge circuit measuring instrument, and automatically adjust the value of its adjustable resistance through the bridge circuit measuring instrument to keep the bridge balanced; in the bridge balanced state, the bridge circuit measuring instrument continuously measures the resistance value of each carbon fiber bundle between the two measuring axes;

[0011] 4) Since the resistance value of carbon fiber tows will fluctuate during the production process, a deep learning framework is constructed to extract features related to resistance fluctuations and build a deep learning model. Then, two batches of carbon fiber tows are randomly selected from the qualified samples that have been produced to measure the resistance values, which are used as the training set and validation set of the deep learning model to obtain the predicted resistance waveform. At the same time, the resistance value of the carbon fiber tow measured in step 3) is used as the test set of the deep learning model to obtain the measured resistance waveform.

[0012] 5) Compare the fluctuation patterns of the predicted resistance waveform in step 4) with the measured resistance waveform to determine whether there are quality problems with the carbon fiber tow.

[0013] The carbon fiber carbonization quality monitoring method based on the resistance method of the present invention is further as follows: in the step 1), the measuring shaft includes a shaft body, a conductive slip ring and a brush; a bearing seat is provided on one side of the shaft body; the conductive slip ring is sleeved on the shaft body, and is installed on the bearing seat through a plurality of slip ring fixing screws so that the conductive slip ring can rotate relative to the shaft body; the carbon fiber bundle is wound on the conductive slip ring; one end of the brush abuts against the carbon fiber bundle, and the other end is electrically connected to the circuit.

[0014] The carbon fiber carbonization quality monitoring method based on the resistance method of the present invention is further as follows: a plurality of partitions are provided on the conductive slip ring, and a wire path is formed between adjacent partitions; a carbon fiber bundle passes through each wire path and corresponds to a brush; a plurality of brushes are provided and installed on a brush fixing plate; each brush is provided with pressure by a spring so that it is always in contact with the carbon fiber bundle.

[0015] The carbon fiber carbonization quality monitoring method based on the resistance method of the present invention is further as follows: the step 3) is specifically: the resistance value of the carbon fiber tow is measured by a four-wire system in the bridge circuit measuring instrument, and the calculation formula is: , where the resistance value of the carbon fiber bundle is the resistance to be measured R x , the rest are fixed resistors R1, R2, and adjustable resistor R3;

[0016] The carbon fiber tow runs in a continuous manner in the carbonization furnace, that is, it continues to move forward over time during the processing. When the bridge loses balance, the bridge circuit measuring instrument automatically adjusts the value of the adjustable resistor R3 to keep the bridge balanced. When the bridge is balanced, the bridge circuit measuring instrument measures the resistance value of each carbon fiber tow between the two measuring axes that are continuously transmitted.

[0017] The carbon fiber carbonization quality monitoring method based on the electrical resistance method of the present invention is further as follows: the step 4) is specifically:

[0018] 4-1), input the resistance value time series data based on the sine function simulation, calculate the sine value of each point based on these time points; perform normalization preprocessing on the resistance value data to obtain the training set, validation set and test set;

[0019] 4-2), define the network structure, use a single-layer RNN and a fully connected layer; in each round of training of the RNN neural network, use MSELoss as the loss function; update the model parameters through back propagation, and use the Adam optimizer to perform gradient descent to train the neural network. In each iteration, the model accepts an input of one time step, performs forward propagation and back propagation, and updates the weight parameters;

[0020] 4-3), use the validation set data to evaluate the model performance, adjust the hyperparameters based on the results, and retrain the model to optimize the model; in the test phase, the model uses the previously trained weights to perform recursive predictions starting from the first data point in the sequence, and outputs the final predicted resistance and measured resistance waveforms.

[0021] The carbon fiber carbonization quality monitoring method based on the resistance method of the present invention is further as follows: In step 4-1), a custom normalization process is performed on the resistance value data obtained by measurement, and the calculation formula is: ,in is the normalized resistance value, which is ultimately used as the amplitude of the sine wave. is the actual resistance value, and are the maximum and minimum values ​​of resistance respectively;

[0022] After obtaining the normalized resistance value data, map them onto a sine wave graph. The form of the sine wave is: ,in is the amplitude, f is the frequency, t is the time, It is the phase;

[0023] The sine value of each point is expressed as: data = np.sin(time_steps), where sin() is a function in the numpy library that calculates the sine value of each element in time_steps and stores the result in the variable data.

[0024] The carbon fiber carbonization quality monitoring method based on the resistance method of the present invention is further as follows: In step 4-2), the MSELoss loss function is as follows:

[0025] ;

[0026] In the formula, is the number of samples, is the true target value, is the value predicted by the model; the mean squared error is obtained by calculating the square of the difference between the predicted value and the true value of each sample, and then averaging the squared errors of all samples to obtain an average error measure.

[0027] The carbon fiber carbonization quality monitoring method based on the resistance method of the present invention is further as follows: the step 5) is specifically: comparing the fluctuation patterns of the predicted resistance waveform based on the sine function simulation output in step 4) with the measured resistance waveform in terms of waveform shape, periodicity, and amplitude consistency; based on the maximum deviation error not exceeding 0.1%, if the actual performance of the predicted resistance waveform is consistent with the measured resistance waveform and the deviation error is within 0.1%, it indicates that the carbonization quality of the carbon fiber bundle meets expectations and is in a qualified state; conversely, if the performance of the predicted resistance waveform and the measured resistance waveform deviates by more than 0.1%, there may be a potential problem of insufficient carbonization, which requires further investigation and verification.

[0028] The carbon fiber carbonization quality monitoring method based on the resistance method of the present invention is further as follows: the specific method for comparing the fluctuation patterns is as follows: first, check whether the shapes of the predicted waveform and the measured waveform are consistent. Since the waveform is simulated based on a sine function, the predicted resistance waveform should have an extremely similar waveform shape to the measured resistance waveform; secondly, observe whether the fluctuations of the two waveforms are consistent from the beginning of one cycle to the end of the next cycle. The predicted resistance waveform should be highly matched with the cycle of the measured resistance waveform; finally, the amplitude and phase of the predicted resistance and measured resistance waveforms should be close, check whether the maximum and minimum values ​​of the amplitude are close and whether the waveform has an upward or downward offset.

[0029] The carbon fiber carbonization quality monitoring method based on the resistance method of the present invention can also be: the deviation error is calculated as follows: for each corresponding time point in the same cycle, the difference between the predicted waveform resistance value and the measured waveform resistance value is calculated, and the entire waveform is quantitatively compared by calculating the difference at each point, and the maximum value is found from all the calculated relative errors, and the traditional detection method is used to verify whether the produced carbon fiber tow is qualified, and the final maximum deviation error is 0.1%; the relative error is used in the calculation of the method, and its formula is:

[0030] Relative error (t) = 00%;

[0031] In the formula, is the predicted resistance value, is the measured resistance value, based on which the relative error at each moment is calculated.

[0032] Compared with the prior art, the present invention has the following beneficial effects: the carbon fiber carbonization quality monitoring method based on the resistance method of the present invention combines the bridge circuit measurement of resistance and the conductive slip ring winding the carbon fiber bundle, which can efficiently and accurately monitor the resistance value in the carbon fiber production process, and compare the fluctuations of the predicted resistance waveform and the measured resistance waveform according to the deep learning network structure, so as to monitor the quality of the carbon fiber production process and produce high-quality composite materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a front view of the carbon fiber production line of the present invention.

[0034] Figure 2 It is a schematic diagram of the equipment of the carbon fiber carbonization quality monitoring method based on the resistance method of the present invention.

[0035] Figure 3 yes Figure 2 Top view of the measuring axis in .

[0036] Figure 4 It is a waveform diagram showing that the predicted resistance and the actually measured resistance in the method of the present invention are consistent.

[0037] Figure 5 It is a waveform diagram showing the deviation between the predicted resistance and the measured resistance in the method of the present invention.

[0038] Figure 6 It is a waveform diagram showing that the predicted resistance and the measured resistance show large fluctuations in the method of the present invention.

[0039] Figure 7 The present invention is a flow chart for generating predicted resistance and measured resistance waveform diagrams using the method of the present invention. DETAILED DESCRIPTION

[0040] Please refer to the instruction manual Figure 1 To Attachment Figure 7 As shown, the present invention is a method for monitoring carbonization quality of carbon fiber based on electrical resistance method, which comprises the following steps:

[0041] 1), prepare a bridge circuit measuring instrument 10, and connect the bridge circuit measuring instrument 10 to a pair of measuring shafts 1 through a line.

[0042] Among them, the bridge circuit measuring instrument 10 has a multi-channel measurement port to simultaneously detect multiple carbon fiber bundles 2. In this embodiment, the bridge circuit measuring instrument 10 can monitor the resistance of 6 carbon fiber bundles 2 at the same time. The positive and negative poles of the power supply in the bridge circuit measuring instrument 10 must be correctly connected to the bridge body to ensure the power supply and overall operation of the bridge. The galvanometer in the bridge circuit measuring instrument 10 is used to measure the tiny current changes when the bridge is balanced. Its two leads need to be connected to the two galvanometer ports of the bridge body respectively, so as to timely detect and record the current changes of the bridge.

[0043] 2) The carbon fiber tow 2 produced in the continuous production line in the carbonization furnace 5 is wound on the pair of measuring shafts 1 in an S-shaped manner.

[0044] The carbonization furnace 5 is provided with a plurality of pairs of common rotating shafts 3 and the measuring shafts 1 , and the measuring shafts 1 are located on the upper part of the common rotating shafts 3 , through which the carbon fiber tows 2 are continuously transported.

[0045] Furthermore, the measuring shaft 1 is composed of several parts such as a shaft body 12, a conductive slip ring 9 and a brush 6. A bearing seat 8 is provided on one side of the shaft body 12; the conductive slip ring 9 is sleeved on the shaft body 12, and is installed on the bearing seat 8 through a plurality of slip ring fixing screws 7, so that the conductive slip ring 9 can rotate relative to the shaft body 1. The carbon fiber tow 2 is wound on the conductive slip ring 9. One end of the brush 6 abuts against the carbon fiber tow 2, and the other end is electrically connected to the circuit, so that the carbon fiber tow 2, the measuring shaft 1 and the bridge circuit measuring instrument 10 form a loop.

[0046] The conductive slip ring 9 is provided with a plurality of partitions 13, and a wire path 14 is formed between adjacent partitions 13; each wire path 14 passes a carbon fiber bundle 2 and corresponds to a brush 6. The partitions 13 can prevent the carbon fiber bundles 2 from interfering with each other.

[0047] There are several brushes 6 installed on a brush fixing plate 11. Each brush 6 is provided with pressure by a spring so that it is always in electrical contact with the carbon fiber tow 2 and ensures that the brush 6 and the carbon fiber tow 2 can maintain continuous connection during the production process. Figure 2The circuit in the circuit is clear, and the circuit only connects one pair of brushes 6 to the bridge circuit measuring instrument 10. Further, the spring force can be adjusted to adapt to different friction conditions and wear conditions.

[0048] The design principle of the measuring shaft 1 is as follows: the carbon fiber tow 2 is wound on the conductive slip ring 9, and the continuous transportation of the carbon fiber tow 2 is achieved by the rotation of the conductive slip ring 9. The brush 6 abuts against the carbon fiber tow 2 and supplies power to the carbon fiber tow 2 located between a pair of measuring shafts 1, so that the bridge circuit measuring instrument 10 can measure the resistance of the carbon fiber tow 2.

[0049] 3) Start the bridge circuit measuring instrument 10, and automatically adjust the value of its adjustable resistance through the bridge circuit measuring instrument 10 to keep the bridge balanced. In the bridge balanced state, the bridge circuit measuring instrument 10 continuously measures the resistance value of each carbon fiber tow 2 between the two measuring shafts 1.

[0050] Specifically, the bridge circuit measuring instrument 10 uses a four-wire system to measure the resistance value of the carbon fiber tow 2, and the calculation formula is: , where the resistance value of the carbon fiber bundle is the resistance to be measured R x , the rest are fixed resistors R1, R2, and adjustable resistor R3.

[0051] The carbon fiber tow 2 is continuously operated in the carbonization furnace 5, that is, it continuously moves forward with time during the processing. When the bridge loses balance, the bridge circuit measuring instrument 10 automatically adjusts the value of the adjustable resistor R3 to keep the bridge balanced. In the balanced state of the bridge, the bridge circuit measuring instrument 10 measures the resistance value of each carbon fiber tow 2 between the two measuring axes that are continuously transmitted.

[0052] 4) Since the rotation of the common rotating shaft 3 during the production process of the carbon fiber tow 2 will cause its resistance value to fluctuate, by constructing Figure 7 The deep learning flowchart framework shown in the figure extracts features related to resistance fluctuations. Two batches of carbon fiber tows with a length of 1 meter are randomly selected from the qualified samples produced to measure the resistance values ​​as the input data of the deep learning model, namely the training set and the validation set. The following is obtained through training: Figures 4 to 6 The predicted resistance waveform based on sine function simulation is shown by the dashed line in the figure. At the same time, based on the resistance value of the carbon fiber tow 2 measured by the instrument in step 3), as the test set of the deep learning model, a deep learning model is constructed as follows: Figures 4 to 6 The solid line in the middle shows the measured resistance waveform based on the sine function simulation. The specific process is:

[0053] 4-1), input the resistance value time series data based on the sine function simulation, calculate the sine value of each point based on these time points; normalize and preprocess the resistance value data to obtain the training set, validation set and test set.

[0054] Among them, the resistance value data obtained by measurement is normalized, and the calculation formula is: , is the normalized resistance value, which is ultimately used as the amplitude of the sine wave. is the actual resistance value, and are the maximum and minimum values ​​of resistance respectively;

[0055] After obtaining the normalized resistance value data, map them onto a sine wave graph. The form of the sine wave is: ,in As the amplitude of the waveform, f is the frequency, t is the time, It is the phase;

[0056] The sine value of each point is expressed as: data = np.sin(time_steps), and the waveform of the predicted resistance and the measured resistance simulated by the sine function is finally output. sin() is a function in the numpy library that calculates the sine value of each element in time_steps and stores the result in the variable data.

[0057] 4-2), define the network structure, use a single-layer RNN and a fully connected layer; in each round of training of the RNN neural network, use MSELoss (mean square error loss function) as the loss function. Update the model parameters through back propagation, and use the Adam optimizer to perform gradient descent to train the neural network. In each iteration, the model accepts an input of one time step, performs forward propagation and back propagation, and updates the weight parameters.

[0058] Among them, the MSELoss loss function is as follows:

[0059] ;

[0060] Where is the number of samples, is the true target value, and is the value predicted by the model. The mean square error is calculated by calculating the square of the difference between the predicted value and the true value of each sample, and then averaging the squared errors of all samples to obtain an average error measure.

[0061] The Adam optimizer maintains two momentums for each parameter, as follows:

[0062] First moment estimate: ;

[0063] In the formula, is the first-order moment estimate of the gradient at the current moment, is the current gradient, is the decay rate of the first-order moment, is the estimated value at the previous moment.

[0064] Second moment estimate: ;

[0065] In the formula, is the second-order moment estimate of the gradient at the current moment, is the decay rate of the second-order moment, is the estimated value at the previous moment.

[0066] Bias correction: ,

[0067] Parameter update: ;

[0068] In the formula, is the model parameter at the current moment, is the learning rate, It is a small constant added to prevent the denominator from being zero.

[0069] 4-3), use the validation set data to evaluate the model performance, adjust the hyperparameters (such as learning rate, hidden layer size, etc.) according to the results, and retrain the model to optimize the model; in the test phase, the model uses the previously trained weights to perform recursive predictions starting from the first data point in the sequence, and outputs the final predicted resistance and measured resistance waveforms.

[0070] The predicted resistance waveform diagram of step 4) is compared with the measured resistance waveform diagram to verify the fluctuation pattern, so as to determine whether there is a quality problem with the carbon fiber tow 2.

[0071] Specifically, the predicted resistance waveform based on the sine function simulation output in step 4) is compared with the measured resistance waveform in terms of waveform shape, periodicity, and amplitude consistency. That is, first, check whether the shapes of the predicted waveform and the measured waveform are consistent. Since the waveform is simulated based on the sine function, the predicted resistance waveform should have an extremely similar waveform shape to the measured resistance waveform; secondly, observe whether the fluctuations of the two waveforms are consistent from the beginning of one cycle to the end of the next cycle. The predicted resistance waveform should be highly matched with the period of the measured resistance waveform; finally, the amplitude and phase of the predicted resistance and measured resistance waveforms should be close, and check whether the maximum and minimum values ​​of the amplitude are close and whether the waveform has an upward or downward offset.

[0072] After calculation and verification of the deviation error between the predicted resistance waveform and the measured resistance waveform, it is determined that the maximum deviation error should not exceed 0.1%. The specific calculation method of the deviation error is: for each corresponding time point in the same cycle, calculate the difference between the predicted waveform resistance value and the measured waveform resistance value, and quantitatively compare the entire waveform by calculating the difference at each point. From all the calculated relative errors, find the maximum value and use traditional detection methods to verify whether the produced carbon fiber tow is qualified. The final maximum deviation error is 0.1%. The relative error is calculated in this method, and its formula is:

[0073] Relative error (t) = 00%;

[0074] In the formula, is the predicted resistance value, is the measured resistance value, based on which the relative error at each moment is calculated.

[0075] like Figure 4 As shown in FIG. 1 , if the predicted resistance waveform is basically consistent with the actual performance of the measured resistance waveform and the error is within 0.1%, it indicates that the carbonization quality of the carbon fiber tow 2 meets expectations and is in a qualified state; otherwise, according to Figure 5 and Figure 6 As shown, if the measured resistance waveform produces waveform distortion, waveform offset or local waveform distortion, and there is a deviation of more than 0.1% compared with the predicted resistance waveform, there may be a potential problem of insufficient carbonization, which requires further investigation and verification.

[0076] In summary, the carbon fiber carbonization quality monitoring method based on the resistance method of the present invention combines the bridge circuit measurement of resistance and the conductive slip ring winding of the carbon fiber bundle, which can efficiently and accurately monitor the resistance value in the carbon fiber production process, and compare the fluctuations of the predicted resistance waveform with the measured resistance waveform according to the deep learning model, so as to monitor the key parameters of the carbon fiber bundle in the carbonization process in real time and provide an effective evaluation of process stability and product quality.

[0077] The above specific implementation methods are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for monitoring carbonization quality of carbon fiber based on electrical resistance method, characterized in that: The steps include: 1) Prepare a bridge circuit measuring instrument and connect the bridge circuit measuring instrument to a pair of measuring axes through a line; 2) The carbon fiber tows in the continuous production line in the carbonization furnace are wound on the pair of measuring shafts in an S-shaped manner; 3) Start the bridge circuit measuring instrument, and automatically adjust the value of its adjustable resistance through the bridge circuit measuring instrument to keep the bridge balanced; in the bridge balanced state, the bridge circuit measuring instrument continuously measures the resistance value of each carbon fiber bundle between the two measuring axes; 4) Since the resistance value of carbon fiber tows will fluctuate during the production process, a deep learning framework is constructed to extract features related to resistance fluctuations and build a deep learning model. Then, two batches of carbon fiber tows are randomly selected from the qualified samples that have been produced to measure the resistance values, which are used as the training set and validation set of the deep learning model to obtain the predicted resistance waveform. At the same time, the resistance value of the carbon fiber tow measured in step 3) is used as the test set of the deep learning model to obtain the measured resistance waveform. Specifically: 4-1), input the resistance value time series data based on the sine function simulation, calculate the sine value of each point based on these time points; perform normalization preprocessing on the resistance value data to obtain the training set, validation set and test set; That is, a custom normalization process is performed on the measured resistance value data, and the calculation formula is: ,in is the normalized resistance value, which is ultimately used as the amplitude of the sine wave. is the actual resistance value, and are the maximum and minimum values ​​of resistance respectively; After obtaining the normalized resistance value data, they are mapped onto a sine wave graph; the form of the sine wave is: ,in is the amplitude, f is the frequency, t is the time, It is the phase; The sine value of each point is expressed as: data = np.sin(time_steps), where sin() is a function in the numpy library that calculates the sine value of each element in time_steps and stores the result in the variable data; 4-2), define the network structure, use a single-layer RNN and a fully connected layer; in each round of training of the RNN neural network, use MSELoss as the loss function; update the model parameters through back propagation, and use the Adam optimizer to perform gradient descent to train the neural network. In each iteration, the model accepts an input of one time step, performs forward propagation and back propagation, and updates the weight parameters; Among them, the MSELoss loss function is as follows: ; In the formula, is the number of samples, is the true target value, is the value predicted by the model; the mean squared error is obtained by calculating the square of the difference between the predicted value and the true value of each sample, and then averaging the squared errors of all samples to obtain an average error measure; 4-3), use the validation set data to evaluate the model performance, adjust the hyperparameters based on the results, and retrain the model to optimize the model; in the test phase, the model uses the previously trained weights to perform recursive predictions starting from the first data point in the sequence, and outputs the final predicted resistance and measured resistance waveforms; 5) Compare the fluctuation patterns of the predicted resistance waveform in step 4) with the measured resistance waveform to determine whether there are quality problems with the carbon fiber tow.

2. The method for monitoring carbon fiber carbonization quality based on electrical resistance method according to claim 1, characterized in that: In the step 1), the measuring shaft includes a shaft body, a conductive slip ring and a brush; a bearing seat is provided on one side of the shaft body; the conductive slip ring is sleeved on the shaft body and mounted on the bearing seat through a plurality of slip ring fixing screws so that the conductive slip ring can rotate relative to the shaft body; the carbon fiber bundle is wound on the conductive slip ring; one end of the brush abuts against the carbon fiber bundle, and the other end is electrically connected to the circuit.

3. The method for monitoring carbonization quality of carbon fiber based on electrical resistance method according to claim 2, characterized in that: The conductive slip ring is provided with a plurality of partitions, and a wire path is formed between adjacent partitions; a carbon fiber bundle passes through each wire path and corresponds to a brush; a plurality of brushes are provided and installed on a brush fixing plate; each brush is provided with pressure by a spring so that it is always in contact with the carbon fiber bundle.

4. The method for monitoring carbonization quality of carbon fiber based on electrical resistance method according to claim 1, characterized in that: The step 3) is specifically as follows: a four-wire system is used inside the bridge circuit measuring instrument to measure the resistance value of the carbon fiber tow, and the calculation formula is: , where the resistance value of the carbon fiber bundle is the resistance to be measured R x , the rest are fixed resistors R1, R2, and adjustable resistor R3; The carbon fiber tow runs in a continuous manner in the carbonization furnace, that is, it continues to move forward over time during the processing. When the bridge loses balance, the bridge circuit measuring instrument automatically adjusts the value of the adjustable resistor R3 to keep the bridge balanced. When the bridge is balanced, the bridge circuit measuring instrument measures the resistance value of each carbon fiber tow between the two measuring axes that are continuously transmitted.

5. The method for monitoring carbon fiber carbonization quality based on electrical resistance method according to claim 1, characterized in that: The step 5) is specifically: comparing the predicted resistance waveform based on the sine function simulation output in step 4) with the measured resistance waveform in terms of waveform shape, periodicity, and amplitude consistency; Based on the fact that the maximum deviation error should not exceed 0.1%, if the predicted resistance waveform is consistent with the actual performance of the measured resistance waveform and the deviation error is within 0.1%, it indicates that the carbonization quality of the carbon fiber tow meets expectations and is in a qualified state; conversely, if the predicted resistance waveform deviates from the measured resistance waveform by more than 0.1%, there may be a potential problem of insufficient carbonization, which requires further investigation and verification.

6. The method for monitoring carbonization quality of carbon fiber based on electrical resistance method according to claim 5, characterized in that: The specific method for comparing the fluctuation patterns is as follows: first, check whether the shapes of the predicted waveform and the measured waveform are consistent. Since the waveform is simulated based on a sine function, the predicted resistance waveform should have an extremely similar waveform shape to the measured resistance waveform; secondly, observe whether the fluctuations of the two waveforms are consistent from the beginning of one cycle to the end of the next cycle. The predicted resistance waveform should be highly matched with the cycle of the measured resistance waveform; finally, the amplitude and phase of the predicted resistance and measured resistance waveforms should be close, check whether the maximum and minimum amplitude values ​​are close and whether the waveform has an upward or downward offset.

7. The method for monitoring carbonization quality of carbon fiber based on electrical resistance method according to claim 5, characterized in that: The deviation error is calculated as follows: for each corresponding time point in the same cycle, the difference between the predicted waveform resistance value and the measured waveform resistance value is calculated, and the entire waveform is quantitatively compared by calculating the difference at each point. From all the calculated relative errors, the maximum value is found and the traditional detection method is used to verify whether the produced carbon fiber tow is qualified. The final maximum deviation error is 0.1%; the relative error is used in the calculation of this method, and its formula is: Relative error (t) = In the formula, is the predicted resistance value, is the measured resistance value, based on which the relative error at each moment is calculated.

Images