Pre-oxidation furnace temperature real-time control system and method based on data processing

Abstract

The invention discloses a pre-oxidation furnace temperature real-time control system and method based on data processing, and relates to the field of pre-oxidation furnace temperature control, and the pre-oxidation furnace temperature real-time control system based on data processing comprises a data feature acquisition module, a data analysis module, an optimal temperature prediction module and a temperature real-time control module. According to the method, the analysis precision of temperature, transmission speed and tension data is improved, a more reliable temperature change trend and a more reliable stability analysis result are obtained, and a random forest regression model is trained by collecting historical data including a transmission speed stability analysis result, a tension stability analysis result and a corresponding temperature adjustment value and utilizing the data, so that the accuracy of the temperature change trend is improved. The complex interaction between the features is better captured, so that the prediction precision is improved; the temperature set value is dynamically adjusted according to the temperature change trend and quantitative transmission speed and tension stability analysis results, and transmission speed stability and tension stability changes can be responded.

Description

Technical Field

[0001] The present invention relates to the field of pre-oxidation furnace temperature control, and in particular to a real-time pre-oxidation furnace temperature control system and method based on data processing. Background Art

[0002] Pre-oxidation furnaces are mainly used to pre-oxidize certain materials. In the production process of carbon fiber, pre-oxidation is a key step that involves heat treating polyacrylonitrile fiber at a higher temperature in air to partially oxidize and stabilize it. The purpose of this step is to prevent the fiber from melting or decomposing during the subsequent high-temperature carbonization process.

[0003] When the pre-oxidation furnace is in use, the temperature in the furnace is monitored and adjusted in real time to ensure that the pre-oxidation process is carried out within the optimal temperature range. Accurate temperature control is crucial to ensure the success of the pre-oxidation process, because too high or too low a temperature will affect the quality and performance of the final carbon fiber.

[0004] At present, the temperature control of the pre-oxidation furnace usually adopts the following methods.

[0005] Control based on fixed rules: Control is performed according to a pre-set temperature curve, but this method lacks flexibility and cannot cope with dynamic changes in the production process.

[0006] PID control: The temperature is adjusted through proportional, integral and differential control, but the effect of PID control on nonlinear systems is limited, especially when the carbon fiber transmission speed and tension in the pre-oxidation furnace vary greatly, it is difficult to achieve the ideal control effect.

[0007] Expert system-based control: uses expert knowledge and experience to adjust the temperature set point, but due to the lack of data support, this method has low repeatability and reliability.

[0008] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0009] In view of the problems in the related art, the present invention proposes a real-time temperature control system and method for a pre-oxidation furnace based on data processing to overcome the above-mentioned technical problems existing in the existing related art.

[0010] To this end, the specific technical solution adopted by the present invention is as follows: According to one aspect of the present invention, a real-time temperature control system for a pre-oxidation furnace based on data processing is provided. The real-time temperature control system for a pre-oxidation furnace based on data processing includes a data feature acquisition module, a data analysis module, an optimal temperature prediction module and a temperature real-time control module.

[0011] Among them, the data feature acquisition module is used to obtain the pre-oxidation furnace monitoring feature data, and the pre-oxidation furnace monitoring feature data includes the temperature feature data of each heating zone in the pre-oxidation furnace, the transmission speed feature data of the fiber material and the tension feature data of the fiber material.

[0012] The data analysis module is used to analyze the temperature change trend, transmission speed stability and tension stability in the pre-oxidation furnace monitoring characteristic data based on the data analysis algorithm.

[0013] The optimal temperature prediction module is used to utilize the temperature change trend results of the pre-oxidation furnace, the transmission speed stability analysis results and the tension stability analysis results, and combine the modeling method to construct a prediction model for outputting the future optimal temperature of the pre-oxidation furnace.

[0014] The temperature real-time control module is used to control the temperature of the pre-oxidation furnace in real time according to the output results of the prediction model.

[0015] Furthermore, the data feature acquisition module includes an acquisition module, a preprocessing module, a preoxidation furnace temperature feature extraction module, a fiber material transmission speed feature extraction module, and a fiber material tension feature extraction module; The acquisition module is used to collect the temperature data of each heating zone in the pre-oxidation furnace, the transmission speed data of the fiber material and the tension data of the fiber material; A preprocessing module is used to preprocess the temperature data of each heating zone, the transmission speed data of the fiber material and the tension data of the fiber material, and the preprocessing includes data filtering, denoising and abnormal value detection; The pre-oxidation furnace temperature feature extraction module is used to calculate the features of the temperature data, and the temperature feature data includes the rate of change, average value, maximum value and minimum value; A fiber material transmission speed feature extraction module is used to calculate the characteristics of the fiber material transmission speed data, and the fiber material transmission speed feature data includes fluctuation amplitude, average value and instantaneous speed; The fiber material tension feature extraction module is used to calculate the features of the fiber material tension data, and the fiber material tension feature data includes stability, average value and abnormal tension points.

[0016] Further, the data analysis module includes a temperature trend analysis module, a transmission speed trend analysis module and a tension trend analysis module; Among them, the temperature trend analysis module is used to analyze the temperature characteristic data using a regression algorithm to obtain the temperature change trend results; A transmission speed stability analysis module is used to analyze the transmission speed characteristic data of the fiber material using a regression algorithm to obtain the transmission speed stability analysis results of the fiber material; The tension stability analysis module is used to analyze the tension characteristic data of the fiber material using a regression algorithm to obtain the tension stability analysis results.

[0017] Furthermore, the regression algorithm includes a ridge regression algorithm, and the ridge regression algorithm is used to analyze the temperature characteristic data, the transmission speed characteristic data of the fiber material, and the tension characteristic data of the fiber material to obtain the temperature change trend result, the transmission speed stability analysis result, and the tension stability analysis result, including: Standardizing the temperature characteristic data, the transmission speed characteristic data of the fiber material, and the tension characteristic data of the fiber material respectively to obtain a standardized temperature characteristic data set, a standardized transmission speed characteristic data set, and a standardized tension characteristic data set; For the standardized temperature feature data set, the standardized transmission speed feature data set, and the standardized tension feature data set, the features at different time points are respectively selected and integrated into a plurality of temperature feature groups, a plurality of transmission speed feature groups, and a plurality of tension feature groups; For the temperature feature dimension, the transmission speed feature dimension, and the tension feature dimension, calculate the covariance matrix between feature groups and the covariance matrix within the feature group in the temperature feature dimension, the covariance matrix between feature groups and the covariance matrix within the feature group in the transmission speed feature dimension, and the covariance matrix between feature groups and the covariance matrix within the feature group in the tension feature dimension; A loss function including a regularization term is constructed based on the covariance matrix, a normalized vector is calculated by a numerical optimization method, and the normalized features are multiplied by the normalized vector to obtain the normalized variable; Regression analysis was performed on different normative variables to obtain the temperature change trend results, the transmission speed change trend results and the tension change trend results; According to the transmission speed variation trend results and the tension variation trend results, the transmission speed stability analysis results and the tension stability analysis results are obtained.

[0018] Furthermore, the formula of the loss function is: ; In the formula, represents the loss function, Represents feature group x and feature groups y The covariance matrix between Represents feature group y and feature groups x The covariance matrix between Represents feature group y The internal covariance matrix, the features include standardized temperature features, standardized transmission speed features and standardized tension features; u represents a normalized vector, T is the transposition symbol, - tr Indicates negative trace, λ represents the regularization parameter.

[0019] Furthermore, the normalized vector is calculated by numerical optimization method including: Solve generalized eigenvalues ​​using numerical optimization algorithms; The eigenvector corresponding to the largest generalized eigenvalue is selected as the normalized vector.

[0020] Furthermore, the optimal temperature prediction module uses the temperature change trend results of the pre-oxidation furnace, the transmission speed stability analysis results and the tension stability analysis results, and combines the modeling method to construct a prediction model for outputting the future optimal temperature of the pre-oxidation furnace, including: According to the temperature change trend result of the pre-oxidation furnace, the temperature setting value of the pre-oxidation furnace in the next time period is obtained; The transmission speed stability analysis results and the tension stability analysis results are used to optimize the pre-oxidation furnace temperature setting value to obtain the final optimal temperature of the pre-oxidation furnace in the future.

[0021] Furthermore, the transmission speed stability analysis results and tension stability analysis results are used to optimize the pre-oxidation furnace temperature setting value, and the final optimal temperature of the pre-oxidation furnace in the future is obtained, including: Collect historical transmission speed stability analysis results and tension stability analysis results, as well as the corresponding temperature adjustment values, and use the temperature adjustment values ​​as target variables; The transmission speed stability analysis results and the tension stability analysis results are standardized to obtain the transmission speed stability characteristics and the tension stability characteristics; A random forest regression model is used to establish the relationship between the transmission speed stability feature and the tension stability feature and the temperature adjustment value, and the random forest regression model is trained and tested; The influence of the transmission speed stability analysis results and the tension stability analysis results on the temperature regulation value was quantified through feature importance analysis of the random forest regression model; The temperature setting value of the pre-oxidation furnace is adjusted according to the quantitative impact results to obtain the final optimal temperature of the pre-oxidation furnace in the future.

[0022] Furthermore, the temperature real-time control module includes a control logic module, an actuator control module and a real-time monitoring module; Among them, the control logic module is used to formulate the control logic and convert it into control commands according to the final future optimal temperature of the pre-oxidation furnace; an actuator control module, for controlling a heating element in the pre-oxidation furnace according to a control command; The real-time monitoring module is used to monitor the actual temperature of the pre-oxidation furnace in real time to form a closed-loop temperature control of the pre-oxidation furnace.

[0023] According to another aspect of the present invention, a method for real-time temperature control of a pre-oxidation furnace based on data processing is also provided. The method for real-time temperature control of a pre-oxidation furnace based on data processing comprises the following steps: The monitoring characteristic data of the pre-oxidation furnace are obtained, and the monitoring characteristic data of the pre-oxidation furnace include the temperature characteristic data of each heating zone in the pre-oxidation furnace, the transmission speed characteristic data of the fiber material and the tension characteristic data of the fiber material.

[0024] Based on the data analysis algorithm, the temperature change trend, conveying speed stability and tension stability in the pre-oxidation furnace monitoring characteristic data are analyzed.

[0025] The temperature variation trend results of the pre-oxidation furnace, the transmission speed stability analysis results and the tension stability analysis results are used in combination with the modeling method to construct a prediction model for outputting the future optimal temperature of the pre-oxidation furnace.

[0026] The temperature of the pre-oxidation furnace is controlled in real time according to the output results of the prediction model.

[0027] The beneficial effects of the present invention are: (1) The present invention collects three key parameters, namely, the temperature in the pre-oxidation furnace, the fiber material transmission speed, and the fiber material tension, in real time through high-precision sensors, and conducts a comprehensive analysis of these data through a data analysis module, accurately reflecting the actual state in the pre-oxidation furnace. Compared with the traditional method of only monitoring a single temperature parameter, the present invention has stronger systematicity and accuracy. The comprehensive analysis can provide more accurate equipment operation status assessment and early warning information, improve the operation stability of the pre-oxidation furnace, and reduce possible misoperation due to inaccurate monitoring of a single parameter.

[0028] (2) The present invention uses the ridge regression algorithm to effectively improve the analysis accuracy of temperature, transmission speed and tension data, and obtain more reliable temperature change trend and stability analysis results, thereby providing more accurate data support for subsequent real-time control.

[0029] (3) The present invention collects historical data, including the results of the transmission speed stability analysis, the results of the tension stability analysis, and the corresponding temperature adjustment values, and uses these data to train the random forest regression model. The random forest regression model can better capture the complex interactions between features, thereby improving the prediction accuracy. By dynamically adjusting the temperature setting value, the system can respond to changes in the transmission speed stability and tension stability in real time, ensuring the stability of the processing process and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 is a principle block diagram of a real-time control system for the temperature of a pre-oxidation furnace based on data processing according to an embodiment of the present invention; Figure 2 The present invention is a flowchart of a method for real-time temperature control of a pre-oxidation furnace based on data processing according to an embodiment of the present invention.

[0032] In the figure: 1. Data feature acquisition module; 2. Data analysis module; 3. Optimal temperature prediction module; 4. Real-time temperature control module. DETAILED DESCRIPTION

[0033] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0034] According to an embodiment of the present invention, a real-time temperature control system and method for a pre-oxidation furnace based on data processing are provided.

[0035] The present invention will now be further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to one embodiment of the present invention, a real-time control system for the temperature of a pre-oxidation furnace based on data processing is provided. The real-time control system for the temperature of a pre-oxidation furnace based on data processing includes a data feature acquisition module 1, a data analysis module 2, an optimal temperature prediction module 3 and a temperature real-time control module 4.

[0036] Among them, the data feature acquisition module 1 is used to install high-precision temperature sensors, speed sensors, and tension sensors at key positions in the pre-oxidation furnace to obtain the pre-oxidation furnace monitoring feature data in real time, and the pre-oxidation furnace monitoring feature data includes the temperature feature data of each heating zone in the pre-oxidation furnace, the transmission speed feature data of the fiber material, and the tension feature data of the fiber material.

[0037] In one embodiment, the data feature acquisition module includes an acquisition module, a preprocessing module, a pre-oxidation furnace temperature feature extraction module, a fiber material transmission speed feature extraction module, and a fiber material tension feature extraction module.

[0038] The acquisition module is used to collect the temperature data of each heating zone in the pre-oxidation furnace, the transmission speed data of the fiber material and the tension data of the fiber material; The preprocessing module is used to preprocess the temperature data of each heating zone, the transmission speed data of the fiber material and the tension data of the fiber material, and the preprocessing includes data filtering (such as low-pass filtering, high-pass filtering, band-pass filtering, etc.), denoising processing (such as moving average, Kalman filtering, etc.) and outlier detection; in addition, outliers are removed, replaced, and other operations are performed.

[0039] The pre-oxidation furnace temperature feature extraction module is used to calculate the features of the temperature data, and the temperature feature data includes the rate of change, average value, maximum value and minimum value; A fiber material transmission speed feature extraction module is used to calculate the characteristics of the fiber material transmission speed data, and the fiber material transmission speed feature data includes fluctuation amplitude, average value and instantaneous speed; The fiber material tension feature extraction module is used to calculate the features of the fiber material tension data, and the fiber material tension feature data includes stability, average value and abnormal tension points.

[0040] For example, set the temperature, speed and tension data to be collected once per second.

[0041] When calculating the temperature change rate, use the temperature difference before and after to express it. Calculate the average value of the temperature data during this period of time. Find the maximum and minimum values ​​of the temperature data during this period of time.

[0042] Calculate the fluctuation of the transmission speed and express it as standard deviation. Standard deviation reflects the degree of dispersion of the data. Calculate the average value of the transmission speed data, that is, the sum of all speed data divided by the number of data points. The single speed value collected every minute is the instantaneous speed.

[0043] The stability of the tension is calculated using the coefficient of variation (CV), which is the standard deviation divided by the mean. The coefficient of variation provides a measure of the relative standard deviation, reflecting the stability of the tension data. The mean of the tension data is calculated, which is the sum of all tension data divided by the number of data points. The 3σ principle can be used to detect abnormal tension points, that is, points that exceed the mean plus or minus 3 times the standard deviation are found as abnormal tension points.

[0044] The data analysis module 2 is used to analyze the temperature change trend, transmission speed stability and tension stability in the pre-oxidation furnace monitoring characteristic data based on the data analysis algorithm.

[0045] In one embodiment, the data analysis module includes a temperature trend analysis module, a transmission speed trend analysis module, and a tension trend analysis module.

[0046] Among them, the temperature trend analysis module is used to analyze the temperature characteristic data using a regression algorithm to obtain the temperature change trend results.

[0047] The transmission speed stability analysis module is used to analyze the transmission speed characteristic data of the fiber material using a regression algorithm to obtain the transmission speed stability analysis results of the fiber material.

[0048] The tension stability analysis module is used to analyze the tension characteristic data of the fiber material using a regression algorithm to obtain the tension stability analysis results.

[0049] In one embodiment, the regression algorithm includes a ridge regression algorithm, and the ridge regression algorithm is used to analyze the temperature characteristic data, the transmission speed characteristic data of the fiber material, and the tension characteristic data of the fiber material to obtain the temperature change trend result, the transmission speed stability analysis result, and the tension stability analysis result, including: The temperature characteristic data, the transmission speed characteristic data of the fiber material and the tension characteristic data of the fiber material are standardized respectively to obtain a standardized temperature characteristic data set, a standardized transmission speed characteristic data set and a standardized tension characteristic data set; the standardization is to subtract the mean from the data and then divide it by the standard deviation (z-score standardization).

[0050] For the standardized temperature feature data set, the standardized transmission speed feature data set, and the standardized tension feature data set, the features at different time points are respectively selected and integrated into a number of temperature feature groups, a number of transmission speed feature groups, and a number of tension feature groups. When selecting the features at different time points, a sliding window technology is used, and the feature integration includes integrating the features at different time points into feature groups.

[0051] For the temperature feature dimension, the transmission speed feature dimension, and the tension feature dimension, calculate the covariance matrix between feature groups and the covariance matrix within the feature group in the temperature feature dimension, the covariance matrix between feature groups and the covariance matrix within the feature group in the transmission speed feature dimension, and the covariance matrix between feature groups and the covariance matrix within the feature group in the tension feature dimension; Specifically, the covariance matrix formula is: ; In the formula, Σ represents the covariance matrix, which represents the covariance relationship between feature groups or within feature groups in a specific dimension (temperature feature dimension, transmission speed feature dimension, or tension feature dimension). It describes how two variables (features) vary together. The size of the covariance matrix is ​​the square of the feature dimension.n Indicates the number of samples, the number of feature groups or time points used to calculate covariance. For example, the number of samples (that is, the number of time points or feature groups) contained in the temperature feature dimension.

[0052] x i Indicates i The feature vector of a sample represents the feature data at a certain time point or feature group after standardization, such as the feature vector at a certain time point in the standardized temperature feature data set.

[0053] is the mean of the feature vectors of all samples, that is, the mean vector of all feature groups or time points in the temperature feature dimension, the transmission speed feature dimension, or the tension feature dimension. T is the transpose symbol.

[0054] A loss function including a regularization term is constructed based on the covariance matrix. The regularization term is usually related to the L2 norm in ridge regression. The loss function penalizes complex models through the regularization term, making the model simpler and more general.

[0055] The normalized vector is calculated by a numerical optimization method, and the normalized features are multiplied by the normalized vector to obtain the normalized variate.

[0056] Regression analysis was performed on different standard variables to obtain the temperature change trend results, transmission speed change trend results and tension change trend results. 2 The performance of the regression model was evaluated by the mean square error and the mean square error to ensure that the model had a good fitting effect.

[0057] According to the transmission speed variation trend results and the tension variation trend results, the transmission speed stability analysis results and the tension stability analysis results are obtained.

[0058] Specifically, the transmission speed change trend result: reflects the change pattern of the transmission speed over time. The tension change trend result: reflects the change pattern of the tension over time.

[0059] The stability of the transmission speed is evaluated by the following method: Mean and standard deviation: Calculate the mean and standard deviation of the transmission speed time series. The smaller the standard deviation, the more stable the transmission speed.

[0060] Coefficient of variation: defined as the standard deviation divided by the mean, also used to measure the relative dispersion of data.

[0061] The stability of tension is evaluated in a similar way to conveyor speed: Mean and Standard Deviation: Calculates the mean and standard deviation of the tension time series.

[0062] Coefficient of variation: used to measure the relative dispersion of tension data.

[0063] In one embodiment, the formula for the loss function is: ; In the formula, represents the loss function, Represents feature group x and feature groups y The covariance matrix between Represents feature group y and feature groups x The covariance matrix between Represents feature group y The internal covariance matrix, the features include standardized temperature features, standardized transmission speed features and standardized tension features; u represents a normalized vector, T is the transposition symbol, - tr Indicates negative trace, λ represents the regularization parameter.

[0064] In one embodiment, calculating the normalized vector by a numerical optimization method includes: Solve the generalized eigenvalues ​​using a numerical optimization algorithm (such as gradient descent, conjugate gradient method, or directly solving the eigenvalue problem); since the largest generalized eigenvalue usually corresponds to the main direction of the data, that is, it can best explain the changes in the data, the eigenvector corresponding to the largest generalized eigenvalue is selected as the normalized vector.

[0065] Specifically, the principle of the loss function is to find the optimal normalized vector by minimizing the loss function formula, which can help the model fit the data better while preventing overfitting. The loss function is solved by numerical optimization methods, such as gradient descent or other optimization algorithms, to find the normalized vector that minimizes the loss function.

[0066] The optimal temperature prediction module 3 is used to utilize the temperature change trend results of the pre-oxidation furnace, the transmission speed stability analysis results and the tension stability analysis results, and to construct a prediction model for outputting the future optimal temperature of the pre-oxidation furnace in combination with a modeling method; In one embodiment, the optimal temperature prediction module uses the temperature change trend results of the pre-oxidation furnace, the transmission speed stability analysis results, and the tension stability analysis results, and combines the modeling method to construct a prediction model for outputting the future optimal temperature of the pre-oxidation furnace, including: According to the temperature change trend result of the pre-oxidation furnace, the temperature setting value of the pre-oxidation furnace in the next time period is obtained; The transmission speed stability analysis results and the tension stability analysis results are used to optimize the pre-oxidation furnace temperature setting value to obtain the final optimal temperature of the pre-oxidation furnace in the future.

[0067] In one embodiment, the temperature setting value of the pre-oxidation furnace is optimized using the results of the conveying speed stability analysis and the tension stability analysis, and the final future optimal temperature of the pre-oxidation furnace is obtained, including: Collect historical transmission speed stability analysis results and tension stability analysis results, as well as the corresponding temperature adjustment values, and use the temperature adjustment values ​​as target variables; The target variable is defined as the temperature adjustment value, that is, the temperature adjustment amount determined based on historical data. For example, if the transmission speed stability or tension stability is poor over a period of time, resulting in the need to adjust the temperature, the temperature change of this adjustment is recorded.

[0068] Tag source: Get the actual temperature change for each adjustment from the historical operation records. For example, if the temperature needs to be increased by 0.2 degrees Celsius due to large fluctuations in the conveyor speed, it will be recorded as +0.2 degrees Celsius.

[0069] The transmission speed stability analysis results and the tension stability analysis results are standardized to obtain the transmission speed stability characteristics and the tension stability characteristics.

[0070] A random forest regression model is used to establish the relationship between the transmission speed stability feature and the tension stability feature and the temperature adjustment value, and the random forest regression model is trained and tested; Specifically, the historical data is divided into a training set and a test set, usually in a ratio of 8:2 or 7:3. The relationship between the transmission speed stability feature and the tension stability feature and the temperature adjustment value is trained using a random forest regression model. The random forest regression model is an ensemble learning method that improves the prediction performance of the model by building multiple decision trees and integrating them. The performance of the model is evaluated using a test data set to ensure that the model has good generalization ability.

[0071] The feature importance analysis of the random forest regression model is used to quantify the impact of the transmission speed stability analysis results and the tension stability analysis results on the temperature adjustment value. Feature importance reflects the importance of each feature to the model prediction results, and is usually measured by counting the number of times a feature is used to split a node or its contribution to improving model performance.

[0072] Adjust the pre-oxidation furnace temperature setting value according to the quantitative impact results to obtain the final optimal temperature of the pre-oxidation furnace in the future. For example, if the analysis results show that the transmission speed stability is poor, increase the temperature according to the temperature adjustment value predicted by the model. If the analysis results show that the tension stability is poor, adjust the temperature appropriately according to the model prediction.

[0073] The temperature real-time control module 4 is used to control the temperature of the pre-oxidation furnace in real time according to the output result of the prediction model.

[0074] In one embodiment, the temperature real-time control module includes a control logic module, an actuator control module and a real-time monitoring module; The control logic module is used to formulate the control logic and convert it into control commands according to the final optimal temperature of the pre-oxidation furnace in the future; the control strategy usually adopts the PID (proportional-integral-differential) control algorithm. PID control achieves precise temperature control by adjusting the heating power of the heating element.

[0075] The actuator control module is used to control the heating elements or other actuators in the pre-oxidation furnace according to the control command and adjust the heating power according to the control signal.

[0076] The real-time monitoring module is used to monitor the actual temperature of the pre-oxidation furnace in real time to form a closed-loop temperature control of the pre-oxidation furnace. That is, according to the actual temperature feedback, the control logic module adjusts the control parameters to form a closed-loop control to ensure that the temperature error is minimized.

[0077] like Figure 2 According to another embodiment of the present invention, a method for real-time temperature control of a pre-oxidation furnace based on data processing is provided. The method for real-time temperature control of a pre-oxidation furnace based on data processing comprises the following steps: The monitoring characteristic data of the pre-oxidation furnace are obtained, and the monitoring characteristic data of the pre-oxidation furnace include the temperature characteristic data of each heating zone in the pre-oxidation furnace, the transmission speed characteristic data of the fiber material and the tension characteristic data of the fiber material.

[0078] Based on the data analysis algorithm, the temperature change trend, conveying speed stability and tension stability in the pre-oxidation furnace monitoring characteristic data are analyzed.

[0079] The temperature variation trend results of the pre-oxidation furnace, the transmission speed stability analysis results and the tension stability analysis results are used in combination with the modeling method to construct a prediction model for outputting the future optimal temperature of the pre-oxidation furnace.

[0080] The temperature of the pre-oxidation furnace is controlled in real time according to the output results of the prediction model.

[0081] In order to facilitate understanding of the above technical solutions of the present invention, the working principle or operation mode of the present invention in the actual process is described in detail below.

[0082] In the pre-oxidation furnace, temperature, conveying speed and tension are key parameters affecting the quality of carbon fiber. The following data are collected in real time from the pre-oxidation furnace: Temperature characteristic data: Temperature data collected every second at different locations (such as inlet, middle and outlet) in degrees Celsius (°C).

[0083] Conveying speed characteristic data: The conveying speed data of the fiber in the furnace, in meters per second (m / s).

[0084] Tension characteristic data: The tension data of the fiber in the furnace, in Newton (N).

[0085] The above data are standardized so that the mean of each feature data set is 0 and the standard deviation is 1.

[0086] The collected data is analyzed by ridge regression algorithm to obtain the temperature change trend, transmission speed stability analysis and tension stability analysis results. For example, the following temperature characteristic data are collected: ( Tin 1, Tout 1), ( Tin 2, Tout 2), ( Tin 3. Tout 3), ( Tin 4. Tout 4), ( Tin 5. Tout 5). After preprocessing, calculate the temperature change rate, average value, etc., and then standardize the data. Construct feature groups: ( Zin 1, Zout 1), ( Zin 2, Zout 2), ( Zin 3. Zout 3), ( Zin 4. Zout 4), ( Zin 5. Zout 5). Based on these features and feature groups, the covariance matrix between feature groups and the covariance matrix within feature groups are calculated.

[0087] After calculating the covariance matrix, substitute it into the loss function, and then use numerical optimization methods to minimize the loss function formula to find the optimal normalized vector. Specifically, the process of minimizing the above loss function using gradient descent or other numerical optimization algorithms includes: setting the initial weight vector and learning rate. Calculating the loss function about u Update the weights based on the gradient and learning rate. Check if the loss function converges or reaches the maximum number of iterations.

[0088] The standardized feature data is multiplied by the normalized vector to obtain the normalized variable. The temperature change trend result is obtained by performing regression analysis on different normalized variables. The analysis process of transmission speed and tension is the same as the above process.

[0089] Construct a data set. For example, in the training set, the speed standardization results are: zspeed1, zspeed2, zspeed3, zspeed4. The tension standardization results are: ztension1, ztension2, ztension3, ztension4. The temperature adjustment values ​​are: +0.2 degrees Celsius, +0.1 degrees Celsius, −0.1 degrees Celsius, +0.3 degrees Celsius. In the test set, the speed standardization results are: zspeed5. The tension standardization results are: ztension5. The temperature adjustment value is: +0.2 degrees Celsius.

[0090] The relationship between the transmission speed stability feature and the tension stability feature and the temperature adjustment value is trained using a random forest regression model. The influence of the transmission speed stability analysis results and the tension stability analysis results on the temperature adjustment value is quantified through feature importance analysis of the random forest regression model. That is, the number of times each feature in the model is used to split the node or its contribution to the improvement of the model performance is checked.

[0091] When the model is applied, if the new data indicates that the conveying speed stability is poor and the model predicts a temperature adjustment value of +0.2 degrees Celsius, the temperature setting value of the pre-oxidation furnace is adjusted based on this prediction to obtain the optimal future temperature of the pre-oxidation furnace.

[0092] In summary, the present invention collects three key parameters, namely, the temperature in the pre-oxidation furnace, the fiber material transmission speed and the fiber material tension, in real time through high-precision sensors, and comprehensively analyzes these data through a data analysis module, accurately reflecting the actual state in the pre-oxidation furnace, and has stronger systematicity and accuracy than the traditional method of monitoring only a single temperature parameter. Comprehensive analysis can provide more accurate equipment operation status evaluation and early warning information, improve the operation stability of the pre-oxidation furnace, and reduce the possible misoperation caused by inaccurate monitoring of a single parameter. The present invention uses a ridge regression algorithm to effectively improve the analysis accuracy of temperature, transmission speed and tension data, obtain more reliable temperature change trends and stability analysis results, and thus provide more accurate data support for subsequent real-time control. The present invention collects historical data, including transmission speed stability analysis results, tension stability analysis results and corresponding temperature adjustment values, and uses these data to train a random forest regression model. Through the random forest regression model, the complex interactions between features can be better captured, thereby improving the prediction accuracy. By dynamically adjusting the temperature setting value, the system can respond to changes in transmission speed stability and tension stability in real time to ensure the stability of the processing process and product quality.

[0093] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A real-time control system for the temperature of a pre-oxidation furnace based on data processing, characterized in that: The real-time temperature control system of the pre-oxidation furnace based on data processing includes a data feature acquisition module, a data analysis module, an optimal temperature prediction module and a temperature real-time control module; The data feature acquisition module is used to obtain the monitoring feature data of the pre-oxidation furnace, and the monitoring feature data of the pre-oxidation furnace includes the temperature feature data of each heating zone in the pre-oxidation furnace, the transmission speed feature data of the fiber material, and the tension feature data of the fiber material; The data analysis module is used to analyze the temperature change trend, transmission speed stability and tension stability in the pre-oxidation furnace monitoring characteristic data based on the data analysis algorithm; The optimal temperature prediction module is used to utilize the temperature change trend results of the pre-oxidation furnace, the transmission speed stability analysis results and the tension stability analysis results, and to construct a prediction model for outputting the future optimal temperature of the pre-oxidation furnace in combination with a modeling method; The temperature real-time control module is used to control the temperature of the pre-oxidation furnace in real time according to the output result of the prediction model; The data feature acquisition module includes an acquisition module, a preprocessing module, a pre-oxidation furnace temperature feature extraction module, a fiber material transmission speed feature extraction module and a fiber material tension feature extraction module; Wherein, the acquisition module is used to collect temperature data of each heating zone in the pre-oxidation furnace, transmission speed data of the fiber material and tension data of the fiber material; The preprocessing module is used to preprocess the temperature data of each heating zone, the transmission speed data of the fiber material and the tension data of the fiber material, and the preprocessing includes data filtering, denoising and abnormal value detection; The pre-oxidation furnace temperature feature extraction module is used to calculate the features of temperature data, and the temperature feature data includes the rate of change, average value, maximum value and minimum value; The fiber material transmission speed feature extraction module is used to calculate the characteristics of the fiber material transmission speed data, and the fiber material transmission speed feature data includes fluctuation amplitude, average value and instantaneous speed; The fiber material tension feature extraction module is used to calculate the features of the fiber material tension data, and the fiber material tension feature data includes stability, average value and abnormal tension points.

2. The real-time control system for pre-oxidation furnace temperature based on data processing according to claim 1 is characterized in that: The data analysis module includes a temperature trend analysis module, a transmission speed trend analysis module and a tension trend analysis module; Wherein, the temperature trend analysis module is used to analyze the temperature characteristic data using a regression algorithm to obtain the temperature change trend result; The transmission speed stability analysis module is used to analyze the transmission speed characteristic data of the fiber material using a regression algorithm to obtain the transmission speed stability analysis result of the fiber material; The tension stability analysis module is used to analyze the tension characteristic data of the fiber material using a regression algorithm to obtain a tension stability analysis result.

3. The real-time control system for pre-oxidation furnace temperature based on data processing according to claim 2 is characterized in that: The regression algorithm includes a ridge regression algorithm, and the ridge regression algorithm is used to analyze the temperature characteristic data, the transmission speed characteristic data of the fiber material, and the tension characteristic data of the fiber material to obtain the temperature change trend result, the transmission speed stability analysis result, and the tension stability analysis result, including: Standardizing the temperature characteristic data, the transmission speed characteristic data of the fiber material, and the tension characteristic data of the fiber material respectively to obtain a standardized temperature characteristic data set, a standardized transmission speed characteristic data set, and a standardized tension characteristic data set; For the standardized temperature feature data set, the standardized transmission speed feature data set, and the standardized tension feature data set, the features at different time points are respectively selected and integrated into a plurality of temperature feature groups, a plurality of transmission speed feature groups, and a plurality of tension feature groups; For the temperature feature dimension, the transmission speed feature dimension, and the tension feature dimension, calculate the covariance matrix between feature groups and the covariance matrix within the feature group in the temperature feature dimension, the covariance matrix between feature groups and the covariance matrix within the feature group in the transmission speed feature dimension, and the covariance matrix between feature groups and the covariance matrix within the feature group in the tension feature dimension; A loss function including a regularization term is constructed based on the covariance matrix, a normalized vector is calculated by a numerical optimization method, and the normalized features are multiplied by the normalized vector to obtain the normalized variable; Regression analysis was performed on different normative variables to obtain the temperature change trend results, the transmission speed change trend results and the tension change trend results; According to the transmission speed variation trend results and the tension variation trend results, the transmission speed stability analysis results and the tension stability analysis results are obtained.

4. The real-time temperature control system of a pre-oxidation furnace based on data processing according to claim 3 is characterized in that: The formula of the loss function is: ； In the formula, represents the loss function, Represents feature group x and feature groups y The covariance matrix between Represents feature group y and feature groups x The covariance matrix between Represents feature group y The internal covariance matrix, the features include standardized temperature features, standardized transmission speed features and standardized tension features; u represents a normalized vector, T is the transposition symbol, - tr Indicates negative trace, λ represents the regularization parameter.

5. The real-time control system for pre-oxidation furnace temperature based on data processing according to claim 3 is characterized in that: The method of calculating the normalized vector by numerical optimization includes: Solve generalized eigenvalues ​​using numerical optimization algorithms; The eigenvector corresponding to the largest generalized eigenvalue is selected as the normalized vector.

6. The real-time temperature control system of a pre-oxidation furnace based on data processing according to claim 1 is characterized in that: The optimal temperature prediction module uses the temperature change trend results of the pre-oxidation furnace, the transmission speed stability analysis results and the tension stability analysis results, and combines the modeling method to construct a prediction model for outputting the future optimal temperature of the pre-oxidation furnace, including: According to the temperature change trend result of the pre-oxidation furnace, the temperature setting value of the pre-oxidation furnace in the next time period is obtained; The transmission speed stability analysis results and the tension stability analysis results are used to optimize the pre-oxidation furnace temperature setting value to obtain the final optimal temperature of the pre-oxidation furnace in the future.

7. A real-time temperature control system for a pre-oxidation furnace based on data processing according to claim 6, characterized in that: The above-mentioned optimization of the pre-oxidation furnace temperature setting value by using the transmission speed stability analysis results and the tension stability analysis results to obtain the final future optimal temperature of the pre-oxidation furnace includes: Collect historical transmission speed stability analysis results and tension stability analysis results, as well as the corresponding temperature adjustment values, and use the temperature adjustment values ​​as target variables; The transmission speed stability analysis results and the tension stability analysis results are standardized to obtain the transmission speed stability characteristics and the tension stability characteristics; A random forest regression model is used to establish the relationship between the transmission speed stability feature and the tension stability feature and the temperature adjustment value, and the random forest regression model is trained and tested; The influence of the transmission speed stability analysis results and the tension stability analysis results on the temperature regulation value was quantified through feature importance analysis of the random forest regression model; The temperature setting value of the pre-oxidation furnace is adjusted according to the quantitative impact results to obtain the final optimal temperature of the pre-oxidation furnace in the future.

8. The real-time temperature control system of a pre-oxidation furnace based on data processing according to claim 7 is characterized in that: The temperature real-time control module includes a control logic module, an actuator control module and a real-time monitoring module; Wherein, the control logic module is used to formulate control logic and convert it into control commands according to the final future optimal temperature of the pre-oxidation furnace; The actuator control module is used to control the heating elements in the pre-oxidation furnace according to the control command; The real-time monitoring module is used to monitor the actual temperature of the pre-oxidation furnace in real time to form a closed-loop temperature control of the pre-oxidation furnace.

9. A method for real-time temperature control of a pre-oxidation furnace based on data processing, applied to the real-time temperature control system of a pre-oxidation furnace based on data processing according to any one of claims 1 to 8, characterized in that: The real-time control method of the pre-oxidation furnace temperature based on data processing comprises the following steps: Acquiring pre-oxidation furnace monitoring characteristic data, wherein the pre-oxidation furnace monitoring characteristic data includes temperature characteristic data of each heating zone in the pre-oxidation furnace, transmission speed characteristic data of the fiber material, and tension characteristic data of the fiber material; Analyze the temperature change trend, conveying speed stability and tension stability in the pre-oxidation furnace monitoring characteristic data based on the data analysis algorithm; Using the temperature change trend results of the pre-oxidation furnace, the transmission speed stability analysis results, and the tension stability analysis results, combined with the modeling method, a prediction model for outputting the future optimal temperature of the pre-oxidation furnace is constructed; The temperature of the pre-oxidation furnace is controlled in real time according to the output results of the prediction model.

Images