Corrugating medium water content real-time compensation method based on multi-source data fusion

Abstract

The invention discloses a real-time corrugated base paper moisture content compensation method based on multi-source data fusion, which belongs to the technical field of industrial process control, and comprises the following steps: acquiring equipment operation time, glue permeation parameters, temperature and humidity, corrugated base paper tension change data and production line speed data to generate original monitoring data; performing cross validation cleaning on the original monitoring data to generate dynamic cleaning data; performing dual-mode moisture content prediction on the dynamic cleaning data; dynamically allocating weights to the prediction results to generate dynamic compensation parameters; generating a real-time compensation signal based on the parameter; and the real-time compensation signal is corrected according to the obtained compensated moisture content feedback to obtain a moisture content compensation value. According to the method, a physical model and time sequence model dual-mode fusion prediction technology is adopted, and a closed-loop grading compensation and feedback self-adaptive correction technology is combined, so that accurate steady-state control of the moisture content of the corrugated base paper can be realized, and the coupling influence of glue permeation timeliness and temperature and humidity disturbance is effectively overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial process control, and in particular to a real-time compensation method for moisture content of corrugated base paper based on multi-source data fusion. Background Art

[0002] During the production of corrugated paper, precise control of moisture content directly impacts paper strength, flatness, and subsequent processing performance, making it a core component of automated control systems in the papermaking industry. Existing technologies typically rely on a single sensor to monitor moisture content, combined with a fixed-parameter model for steam injection compensation.

[0003] Current methods primarily use data from temperature, humidity, and velocity sensors to predict moisture content changes using a linear regression model. This model then triggers the steam nozzles based on a preset threshold. Some improved solutions incorporate equipment operating time as a correction factor to mitigate performance drift during long-term production.

[0004] Existing technologies are highly sensitive to transient disturbances such as tension fluctuations, which can easily lead to data distortion. They fail to fully consider the nonlinear attenuation characteristics of glue penetration parameters as equipment ages. The steam injection volume control of the actuator has response lag and overshoot, and lacks closed-loop verification of the actual injection effect. Summary of the Invention

[0005] To solve the above problems, the present invention provides a real-time compensation method for the moisture content of corrugated base paper based on multi-source data fusion. It adopts the dual-mode fusion prediction technology of physical model and time series model, combined with closed-loop graded compensation and feedback adaptive correction technology, which can achieve precise steady-state control of the moisture content of corrugated base paper and effectively overcome the coupling influence of glue penetration timeliness and temperature and humidity disturbances.

[0006] The above objectives can be achieved through the following solutions: A real-time compensation method for the moisture content of corrugated base paper based on multi-source data fusion includes obtaining equipment operating time, glue penetration parameters, temperature and humidity, corrugated base paper tensile change data, and production line speed data to generate original monitoring data; cross-validating and cleaning the original monitoring data to generate dynamic cleaning data; performing dual-mode moisture content prediction on the dynamic cleaning data; dynamically assigning weights to the prediction results to generate dynamic compensation parameters; generating a real-time compensation signal based on the parameters; and correcting the real-time compensation signal based on the obtained compensated moisture content feedback to obtain a moisture content compensation value. The present invention utilizes dual-mode fusion prediction technology of physical models and time series models, combined with closed-loop hierarchical compensation and feedback adaptive correction technology, to achieve precise steady-state control of the moisture content of corrugated base paper, effectively overcoming the coupled effects of glue penetration timeliness and temperature and humidity disturbances.

[0007] Optionally, the cross-validation cleaning processing of the original monitoring data to generate dynamic cleaning data includes: if the corrugated base paper tension change data does not exceed the preset tension fluctuation threshold range, then reviewing and testing the original monitoring data to obtain a moisture content deviation coefficient; correcting the original monitoring data according to the moisture content deviation coefficient to generate dynamic cleaning data.

[0008] Optionally, the moisture content prediction of the dynamic cleaning data to generate a dual-mode prediction result includes: establishing a physical prediction sub-model based on the temperature and humidity data and the corrugated paper production line operating speed data; constructing a time series prediction sub-model based on the equipment operating time data and the glue penetration parameters; and coupling the output results of the physical prediction sub-model and the time series prediction sub-model to generate a dual-mode prediction result.

[0009] Optionally, the dynamic weight allocation of the dual-mode prediction results to generate dynamic compensation parameters includes: calculating the first prediction error value of the physical prediction sub-model and the second prediction error value of the time series prediction sub-model to obtain weight coefficients of the physical prediction sub-model and the time series prediction sub-model; when it is detected that the corrugated paper tension change data exceeds the tension fluctuation threshold range, the weight coefficient of the physical prediction sub-model is increased to obtain a physical dominant weight allocation parameter; when the glue penetration attenuation rate corresponding to the equipment operation time data exceeds the timeliness threshold of the glue penetration parameter, the weight coefficient of the time series prediction sub-model is increased to obtain a time series dominant weight allocation parameter; and weighted linear superposition calculation is performed on the physical dominant weight allocation parameter and the time series dominant weight allocation parameter to generate dynamic compensation parameters.

[0010] Optionally, generating a real-time compensation signal based on the dynamic compensation parameter includes: obtaining a voltage control signal based on the dynamic compensation parameter; adjusting the opening frequency and injection amount of the steam nozzle according to the voltage control signal to obtain actual injection amount data of the steam nozzle; performing physical characteristic feedback correction on the actual injection amount data of the steam nozzle to generate a real-time compensation signal.

[0011] Optionally, the obtaining of the compensated moisture content data and the feedback correction of the real-time compensation signal to obtain the moisture content compensation value include: calculating the cumulative deviation of the moisture content target value based on the compensated moisture content data; obtaining a deviation perception correction coefficient based on the cumulative deviation of the moisture content target value; adjusting the hidden layer node connection weights of the time series prediction sub-model based on the deviation perception correction coefficient to obtain an adaptive time series weight matrix; and feedback correction of the real-time compensation signal according to the adaptive time series weight matrix to obtain the moisture content compensation value.

[0012] Optionally, the real-time compensation signal is fed back and corrected according to the adaptive timing weight matrix to obtain a moisture compensation value, including: reconstructing the timing prediction model and performing dynamic weight iteration on the adaptive timing weight matrix to obtain optimized compensation parameters; based on the optimized compensation parameters, feedback gain adjustment and dynamic compensation correction are performed on the real-time compensation signal to obtain a moisture compensation value.

[0013] Optionally, the correcting the original monitoring data according to the moisture content deviation coefficient to generate dynamic cleaning data includes: obtaining a multi-channel correction gain parameter based on the moisture content deviation coefficient; and performing multi-factor coupling correction on the temperature and humidity data, the corrugated paper production line operating speed data and the glue penetration parameter based on the multi-channel correction gain parameter to generate dynamic cleaning data.

[0014] Optionally, obtaining the voltage control signal based on the dynamic compensation parameter includes: performing a time sequence correction of the equipment operation attenuation coefficient on the dynamic compensation parameter to obtain a compensation correction parameter; and performing a nonlinear adaptive mapping conversion on the compensation correction parameter to obtain a voltage control signal.

[0015] Based on the same inventive concept, the present invention also provides a real-time compensation system for the moisture content of corrugated base paper based on multi-source data fusion, the system including: a multi-parameter acquisition module for acquiring equipment operating time data, glue penetration parameters, temperature and humidity data, corrugated base paper tensile change data and corrugated base paper production line operating speed data to generate original monitoring data; a dynamic cleaning module for performing cross-validation cleaning processing on the original monitoring data to generate dynamic cleaning data; a dual-mode prediction module for performing moisture content prediction on the dynamic cleaning data to generate a dual-mode prediction result; a weight distribution module for dynamically distributing weights on the dual-mode prediction results to generate dynamic compensation parameters; a signal generation module for generating a real-time compensation signal based on the dynamic compensation parameters; a feedback correction module for acquiring compensated moisture content data, performing feedback correction on the real-time compensation signal to obtain a moisture content compensation value.

[0016] Compared with the prior art, the present invention has the following advantages: Through the multi-source data cross-validation and cleaning mechanism, the sensor measurement noise and abnormal fluctuation interference are effectively eliminated, the accuracy and consistency of the original monitoring data are significantly improved, and a reliable data basis is provided for moisture content control.

[0017] A dual-mode prediction mechanism that combines real-time correlation between physical models and environmental parameters and coupling of time series models with equipment aging trends can comprehensively capture short-term dynamic disturbances and long-term attenuation factors, thereby enhancing the robustness and adaptability of moisture content prediction.

[0018] An adaptive weight allocation strategy based on prediction error and working condition is adopted to dynamically optimize the contribution ratio of the physical model and the time series model to ensure the rationality of the decision-making of the compensation parameters under complex working conditions such as tension mutation or glue aging.

[0019] Through closed-loop control of steam injection volume and graded threshold compensation, the actuator deviation is accurately corrected to achieve rapid response and stable execution of moisture content compensation. At the same time, historical deviation feedback is used to correct the prediction model weights and continuously improve the long-term control accuracy of the system.

[0020] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 The present invention is a flowchart of a method for real-time compensation of moisture content of corrugated paper based on multi-source data fusion according to an embodiment of the present invention.

[0023] Figure 2 2 is a diagram showing tension change and threshold range according to an embodiment of the present invention.

[0024] Figure 3 4 is a diagram showing a correction of a moisture content deviation coefficient according to an embodiment of the present invention.

[0025] Figure 4 It is a physical prediction sub-model diagram of an embodiment of the present invention.

[0026] Figure 5 This is a dynamic compensation parameter generation diagram according to an embodiment of the present invention.

[0027] Figure 6 4 is an adaptive timing weight matrix diagram of an embodiment of the present invention.

[0028] Figure 7 The present invention is a structural diagram of a real-time compensation system for moisture content of corrugated paper based on multi-source data fusion according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0030] Reference Figure 1 One embodiment of the present invention proposes a real-time compensation method for the moisture content of corrugated base paper based on multi-source data fusion. It adopts a dual-mode fusion prediction technology of physical model and time series model, combined with closed-loop graded compensation and feedback adaptive correction technology, which can achieve precise steady-state control of the moisture content of corrugated base paper and effectively overcome the coupling effects of glue penetration timeliness and temperature and humidity disturbances.

[0031] The method of this embodiment specifically includes: Obtain equipment operation time data, glue penetration parameters, temperature and humidity data, corrugated base paper tension change data, and corrugated base paper production line operation speed data to generate original monitoring data; Performing cross-validation cleaning processing on the original monitoring data to generate dynamic cleaning data; Predicting the moisture content of the dynamic cleaning data to generate a dual-mode prediction result; Performing dynamic weight allocation on the dual-mode prediction results to generate dynamic compensation parameters; generating a real-time compensation signal based on the dynamic compensation parameter; Based on the compensated moisture content data, the real-time compensation signal is feedback-corrected to obtain a moisture content compensation value.

[0032] Specifically, a programmable logic controller (PLC) is deployed at the industrial site. Its built-in timer collects real-time operating time data, recording the cumulative continuous operating time of the corrugated cardboard production line since startup. An online glue viscometer is also used to monitor glue penetration parameters. This viscometer measures the flow resistance of the glue through a capillary tube and converts it into a standardized glue penetration coefficient based on Newtonian fluid dynamics. Ambient temperature and humidity data are collected by high-precision temperature and humidity sensors installed in the paper feed area of ​​the production line. The temperature sensor measures air temperature based on thermocouple principles, while the humidity sensor uses a capacitive sensing element to obtain relative humidity. Data on changes in corrugated paper tension is collected in real time by a cantilever tension sensor installed on the floating roller bearing seat in the paper web transport path. This sensor uses a strain gauge bridge to detect mechanical changes in paper web tension. Data on the corrugated paper production line operating speed is collected by a rotary encoder mounted on the bearing end of the paper guide roller. This encoder calculates the actual linear speed by counting the number of roller revolutions per unit time and multiplying it by a pre-calibrated roller circumference. The above five types of data are transmitted to the central controller via industrial Ethernet. After data alignment according to millisecond timestamps, they are stored in a structured format in a designated data table of the real-time database. The table fields include five dimensions: equipment operating time, glue permeability coefficient, ambient temperature, ambient humidity, and production line speed, thus forming a set of original monitoring data with time series characteristics.

[0033] Optionally, performing cross-validation cleaning processing on the original monitoring data to generate dynamic cleansed data includes: If the corrugated base paper tension change data does not exceed the preset tension fluctuation threshold range, the original monitoring data is reviewed and tested to obtain a moisture content deviation coefficient; The original monitoring data is corrected according to the moisture content deviation coefficient to generate dynamic cleaning data.

[0034] Specifically, the tension change and threshold range are as follows Figure 2 As shown, first, the tension change data of the corrugated base paper is collected in real time through the tension sensor. This data represents the instantaneous mechanical fluctuations in the paper web during operation. The obtained tension change data is compared with the tension fluctuation threshold range. The threshold range needs to be determined through pre-experimental calibration based on the quantitative parameters of the corrugated base paper and the mechanical characteristics of the production line. If the tension does not exceed the threshold range, the review and detection process is started: the compensated moisture content data in the original monitoring database is synchronously retrieved, and the actual moisture content detection value of the corrugated base paper is obtained in real time through the infrared online moisture meter. Calculate the moisture content deviation coefficient : , in Indicates the actual moisture content detected, collected online by an infrared moisture meter; Indicates the moisture content after compensation, which is read from the original monitoring database. Generate temperature and humidity correction gain parameters, production line operation speed correction gain parameters and glue penetration correction gain parameters. The generation of each parameter is based on the product increment of moisture content deviation coefficient and historical coupling coefficient. The moisture content deviation coefficient correction is as follows: Figure 3 The original multi-source data is then collaboratively adjusted: the ambient temperature and humidity data are multiplied by the temperature and humidity correction gain parameter, the production line speed data are multiplied by the production line speed correction gain parameter, and the glue penetration parameter is multiplied by the glue penetration correction gain parameter to form a dynamic cleaning data set for subsequent processing modules to call.

[0035] For example, during the operation of a production line, the tension sensor detected a value of 0.98 Newtons, which was lower than the lower threshold of 1.5 Newtons, triggering a recheck. The infrared moisture meter measured the actual moisture content of the paper feed area as 6.1%, while the compensated moisture content recorded in the original database was 6.6%, resulting in a calculated deviation coefficient of 0.5%. Based on the historical data regression calibration relationship, a temperature and humidity correction gain of 1.025, a speed correction gain of 1.055, and a glue penetration correction gain of 1.040 were generated. In the original monitoring data, the temperature of 28°C was updated to 28.7°C, the humidity of 70%RH was updated to 71.75%RH, the production line speed of 125 m / min was updated to 131.875 m / min, and the glue penetration parameter 22 was updated to 22.880. This calibrated dynamic cleaning data packet is transmitted to the subsequent prediction module.

[0036] Optionally, the performing moisture content prediction on the dynamic cleaning data to generate a dual-mode prediction result includes: Establishing a physical prediction sub-model based on the temperature and humidity data and the operating speed data of the corrugated paper production line; Building a time series prediction sub-model based on the equipment operation time data and the glue penetration parameters; The output results of the physical prediction sub-model and the time series prediction sub-model are coupled and calculated to generate a dual-mode prediction result.

[0037] Specifically, firstly, a physical prediction sub-model is established based on the temperature and humidity data and the running speed data of the corrugated paper production line. The physical prediction sub-model is as follows: Figure 4 As shown in the figure, the model is based on the principle of thermodynamic diffusion and calculates the predicted value through the heat balance equation of the relationship between paper moisture absorption and ambient temperature and humidity. The physical prediction sub-model expression is: , The temperature difference between the ambient temperature and the set temperature is obtained in real time through the temperature sensor; The ambient humidity is obtained in real time through the humidity sensor; The speed of the corrugated paper production line is collected through an encoder; and is an empirical coefficient obtained through regression calibration of historical temperature, humidity and speed data.

[0038] The physical prediction sub-model is established in the following steps: First, data collection and preprocessing: Under the steady-state working conditions of the production line, continuously collect 72 hours of temperature and humidity data, production line speed data, and synchronized infrared moisture meter measured moisture content data; perform sliding average filtering (window width 5 minutes) on the original data to eliminate instantaneous noise; construct a data set: the input features are ( = ambient temperature - 25℃ reference temperature, = relative humidity, = production line speed), the output label is the measured moisture content ; Then perform empirical coefficient regression calibration: use regularized multiple linear regression (RidgeRegression) to solve the model: , in is the temperature and humidity coupling influencing factor, is the production line speed sensitivity coefficient, loss function : , is the regularization coefficient, which is set to 0.01 to prevent overfitting. The optimization algorithm is: the least squares closed-form solution: , in For an N×3 feature matrix: Column 1 = , column 2 = , column 3 = 1; for ×1 moisture content matrix; Is the unit matrix; the constructed test set is greater than 0.85 and the maximum absolute error is less than 0.5%. Finally, the coefficients are solidified and deployed: the calibrated 、 The predictions are written to the PLC read-only memory and recalibrated every 30 days based on new data (trigger condition: cumulative prediction error > 10%). A time series prediction sub-model is also constructed based on the equipment operating time data and glue penetration parameters. The network topology of the time series prediction sub-model includes an input layer with a 2D vector (equipment operating time and glue penetration parameters); an LSTM layer with 16 neurons, an activation function, and an initial forget gate bias; a dropout layer with a dropout rate of 0.2 to prevent overfitting; a fully connected layer with 8 neurons and ReLU activation; and an output layer with 1 neuron (moisture content prediction value) and linear activation. The data source is 200 hours of continuous equipment operating time series data (sampling interval 1 minute). Data augmentation is performed by generating sequence samples using a sliding window (window length 60, step size 1). Equipment operating time is normalized to its minimum and maximum values, and glue penetration parameters are normalized to Z-scores. The model uses a long short-term memory (LSTM) network to capture the time series trend of equipment aging and glue penetration attenuation. The input is the equipment operation time data and glue penetration parameters. The forward propagation of the neural network calculates the time series prediction sub-model. , LSTM weights are optimized offline through initial training data. Finally, the two sub-models are output and Perform coupled calculations to generate dual-mode predictions , the coupling adopts initial equal-weight linear superposition, and the expression is: , is the weight coefficient, ensuring that the sum of the weights is 1.

[0039] For example, in a corrugated paper production line, dynamic cleaning data includes temperature and humidity data of 24.5°C and 63%RH, production line speed of 115 meters per minute, equipment operation time of 385 minutes, and glue penetration parameter of 17. When establishing the physical prediction sub-model, according to the pre-calibrated empirical coefficients =0.02 and =0.5, calculate the temperature difference , substituting into the formula we get When constructing the time series prediction sub-model, the device running time of 385 minutes and the glue parameter of 17 are input into the pre-trained LSTM network, and its output is =6.7%. Equal weight superposition is used in coupling calculation to generate dual-mode prediction results. ,This result serves as the input for subsequent dynamic weight allocation.

[0040] Optionally, the dynamically allocating weights on the dual-mode prediction results to generate dynamic compensation parameters includes: Calculating a first prediction error value of the physical prediction sub-model and a second prediction error value of the time series prediction sub-model to obtain weight coefficients of the physical prediction sub-model and the time series prediction sub-model; When it is detected that the corrugated base paper tension change data exceeds the tension fluctuation threshold range, the weight coefficient of the physical prediction sub-model is increased to obtain a physical dominant weight distribution parameter; When the glue penetration attenuation rate corresponding to the equipment operation time data exceeds the timeliness threshold of the glue penetration parameter, the weight coefficient of the time series prediction sub-model is increased to obtain a time series dominant weight distribution parameter; A weighted linear superposition calculation is performed on the physical dominant weight distribution parameter and the temporal dominant weight distribution parameter to generate a dynamic compensation parameter.

[0041] Specifically, the first prediction error value of the physical prediction sub-model and the second prediction error value of the time series prediction sub-model are calculated. The prediction error value is defined as the absolute difference between the model output value and the actual moisture content detection value. The initial weight coefficient is determined based on the ratio of the recent error means of the two models. The weight coefficient of the physical prediction sub-model is set to , the weight coefficient of the time series prediction sub-model is , initial satisfaction When the corrugated paper tension change data is detected to exceed the tension fluctuation threshold range, it indicates that the mechanical vibration interference is enhanced. At this time, the weight coefficient of the physical prediction sub-model is increased, and the weighting factor is used to calculate the value of the physical prediction sub-model. Adjust the weight coefficient of the physical prediction sub-model. The weighting factor formula is: , in is the difference in tension beyond the threshold, is the tension reference constant, which is calibrated through production line mechanical testing. The adjusted weight coefficient of the physical prediction sub-model is: , Simultaneously reduce the weight coefficient of the time series prediction sub-model: , When the glue penetration attenuation rate calculated from the equipment operation time variance data exceeds the timeliness threshold, it indicates that the glue performance has degraded, and the percentage of attenuation rate exceeding the limit is expressed. Calculating exponential gain , the formula is: , Improve the time series prediction sub-model and obtain the time series dominant weight distribution parameters: , The physical dominant weight distribution parameters are: .

[0042] Finally, the two dominant state parameters are weighted linearly superimposed, and the dynamic compensation parameters are output when the tension change data exceeds the tension fluctuation threshold range. The dynamic compensation parameters are generated as follows: Figure 5 As shown: , When the glue penetration attenuation rate exceeds the timeliness threshold of the glue penetration parameter, the dynamic compensation parameter is output: .

[0043] For example, in a corrugated paper production line, the physical prediction sub-model outputs , the time series forecast sub-model output , initial weight setting When the tension sensor detects a value of 2.3 Newtons, exceeding the threshold range of 1.5-2.0 Newtons, the calculation , take the calibration constant Weighting factor , then the physical dominant weight If it exceeds 1, the system will automatically truncate it to , the timing weight is updated synchronously as , dynamic compensation parameter output The variance in the equipment's operating time after three hours indicates the glue penetration attenuation rate. Far exceeding the threshold of 80%, triggering the timing weight improvement mechanism to calculate the gain , order-dominant weight allocation parameter The cutoff value is 0.92, and the physical dominant weight distribution parameter is adjusted to 0.08. At this time, the output .

[0044] Optionally, generating a real-time compensation signal based on the dynamic compensation parameter includes: obtaining a voltage control signal based on the dynamic compensation parameter; According to the voltage control signal, the opening frequency and injection amount of the steam nozzle are adjusted to obtain actual injection amount data of the steam nozzle; Physical characteristic feedback correction is performed on the actual injection amount data of the steam nozzle to generate a real-time compensation signal.

[0045] Specifically, the dynamic compensation parameter is first processed by the equipment operation attenuation timing correction unit. The parameter is the percentage value of the moisture content. The attenuation factor calculated by the equipment operation time variance data is defined as , determined by the ratio of the equipment's continuous operating time to its life coefficient, and the correction formula is: , in is the dynamic compensation correction parameter, is the dynamic compensation parameter, Obtained by fitting the equipment history database. Then, nonlinear adaptive mapping transformation is performed. Converted into voltage control signal : , in the formula is the voltage value, is the base moisture content, is the voltage conversion coefficient, obtained through equipment calibration test. Control the steam nozzle drive circuit and adjust the opening frequency and injection volume , the two satisfy the linear relationship:

[0046] , in 、 is the nozzle characteristic constant. Finally, the actual opening data of the steam nozzle is collected. and steam pressure data , calculate the actual injection amount through the fluid mechanics equation : , is the flow coefficient, which is obtained from the factory parameters of the nozzle. and command injection quantity Closed-loop comparison, when the deviation exceeds the threshold The steam pressure compensation increment is generated at the same time, and the final output is a real-time compensation signal including frequency, opening and pressure compensation.

[0047] For example, in a corrugated paper production line, the dynamic compensation parameter , the current device running time variance calculation attenuation factor , after correction Take the base moisture content , calibration coefficient , calculate the voltage signal According to the nozzle characteristic constant Adjust the opening frequency , command injection quantity . Pressure sensor detection in actual operation , the opening sensor feedbacks the actual opening , calculate the actual injection amount . The deviation from the command value is , exceeding the threshold , generate pressure compensation increment signal, and finally output compound compensation instruction to adjust nozzle opening to .

[0048] Optionally, the obtaining of compensated moisture content data and performing feedback correction on the real-time compensation signal to obtain a moisture content compensation value includes: Based on the compensated moisture content data, a cumulative deviation of the moisture content target value is calculated; Obtaining a deviation perception correction coefficient based on the accumulated deviation of the moisture content target value; Based on the deviation perception correction coefficient, adjusting the hidden layer node connection weights of the time series prediction sub-model to obtain an adaptive time series weight matrix; The real-time compensation signal is feedback-corrected according to the adaptive time-series weight matrix to obtain a moisture content compensation value.

[0049] Specifically, the adaptive timing weight matrix is ​​adjusted as follows Figure 6 As shown, first obtain the cumulative deviation of the moisture content target value The deviation is obtained by subtracting the compensated moisture content set value sequence from the actual detection value sequence point by point and accumulating the absolute values: , In the formula is the cumulative deviation percentage, For the moment The target moisture content is For the moment The actual moisture content detection value, is the number of continuous sampling points. Calculate the deviation perception correction factor : , is the proportional coefficient, which is obtained through regression analysis of the historical deviation database. Input the hidden layer weight adjustment module of the time series prediction sub-model and define three fixed time periods 、 、 ( =1h, =2h, = 4h, calibrated by production line operation characteristics), calculate the time in each period Accumulated value: , For the period The cumulative intensity of 、 、 The maximum value determines the weight adjustment gradient factor. Call the model to initialize the training and save the weight gradient symbol matrix ,in is the gradient of the loss function with respect to the weight (obtained by offline backpropagation), The elements are ±1 or 0. The weights are updated by gradient-sensitive direction compensation: , is the output adaptive timing weight matrix, is the original weight matrix, To set the learning rate, Adjust the gradient factor to determine the maximum weight. After loading the time series prediction sub-model, the dynamic weight allocation is re-executed to generate new compensation parameters, and finally the moisture content compensation value is output through the steam nozzle control link.

[0050] For example, in a corrugated paper production line, the target moisture content sequence after compensation is [6.5%, 6.5%, 6.5%], and the actual detection values ​​at three consecutive sampling points are [6.7%, 6.6%, 6.8%]. Calculate the cumulative deviation =|6.7%-6.5%|+|6.6%-6.5%|+|6.8%-6.5%|=0.6%. Take the proportional coefficient =0.02 0.012. Segmented integral setting =0~1h( =0.008)、 =1~3h( =0.010)、 =3~7h( =0.012), determine the weight adjustment gradient factor is 0.012. Load the original weight matrix and symbol matrix ,set up =0.01 to update the weight matrix. After the new weights were loaded, the time series prediction sub-model input device ran for 400 minutes and a glue parameter of 18. The output moisture content was corrected from 7.0% to 6.65%. Combined with the 6.60% output of the physical sub-model, a compensation parameter of 6.63% was generated through dynamic weight allocation, and the actual moisture content steadily converged to the target range.

[0051] Optionally, performing feedback correction on the real-time compensation signal according to the adaptive time series weight matrix to obtain the moisture content compensation value includes: Performing time series prediction model reconstruction and dynamic weight iteration on the adaptive time series weight matrix to obtain optimized compensation parameters; Based on the optimized compensation parameters, the real-time compensation signal feedback gain is adjusted and dynamically compensated to obtain a water content compensation value.

[0052] Specifically, first obtain the adaptive time series weight matrix, and load the matrix into the time series prediction sub-model to replace the initial weight matrix. Re-execute the model prediction in combination with the current equipment operation time data and glue penetration parameters to obtain the optimized prediction value. Input the optimized prediction value into the dynamic weight allocation module, and recalculate the real-time prediction error value of the physical prediction sub-model and the time series prediction sub-model. The prediction error value is the absolute difference between the model output value and the moisture content measured by the infrared online moisture meter. Adjust the weight coefficient according to the equipment operation status fluctuation data: when it is detected that the corrugated base paper tension change data exceeds the preset tension fluctuation threshold range, increase the weight coefficient of the physical prediction sub-model according to the weight factor; when the glue penetration attenuation rate exceeds the timeliness threshold, increase the weight coefficient of the time series prediction sub-model according to the exponential gain to generate the optimized compensation parameter. Input the optimized compensation parameter into the voltage conversion link to perform attenuation correction and nonlinear mapping, and calculate the theoretical injection quantity instruction according to the nozzle characteristic equation. The formula is as follows: , in is the command injection amount; The frequency of opening the steam nozzle; The nozzle specific constant is obtained through the nozzle factory calibration. The actual opening data of the steam nozzle and the steam pressure data are collected simultaneously, and the actual injection amount is calculated based on the fluid mechanics formula. : , is the flow coefficient, represents the steam pressure difference, Indicates the density of steam. Generate opening compensation increment when the threshold is exceeded Finally, by superimposing the steam pressure control component in the original real-time compensation signal and Forming moisture content compensation value.

[0053] For example, in a corrugated paper production line, an adaptive time series weight matrix is ​​loaded into the time series prediction sub-model, and the input equipment runs for 385 minutes and the glue parameter is 21. The original output moisture content of the model is 7.30%, which is optimized to 7.03%. The physical prediction sub-model output is 7.28%. The current glue penetration attenuation rate is detected to be 45% (lower than the timeliness threshold of 80%) and the tension is 1.8 Newtons (within the threshold). The dynamic weight distribution maintains the equal weight setting to generate the optimized compensation parameter 7.15%. The nozzle frequency instruction is converted to 18Hz by voltage, corresponding to the theoretical injection volume. =152L / min. The actual steam pressure is 0.85MPa and the opening is 88%. =149L / min. The deviation of 3L / min exceeds the threshold and generates an opening compensation increment =+2%, after adjusting the pressure control component, the output moisture content compensation value controls the nozzle opening to increase to 90%, and finally the moisture content of the paper feeding area is stabilized within the range of ±0.3% of the target value.

[0054] Optionally, the correcting the original monitoring data according to the moisture content deviation coefficient to generate dynamic cleaning data includes: Based on the moisture content deviation coefficient, a multi-channel correction gain parameter is obtained; Based on the multi-channel correction gain parameters, multi-factor coupling correction is performed on the temperature and humidity data, the corrugated paper production line operation speed data and the glue penetration parameters to generate dynamic cleaning data.

[0055] Specifically, first obtain the moisture content deviation coefficient , which is calculated from the difference between the actual moisture content detection value and the compensated moisture content data. Generate multi-channel correction gain parameters: , In the formula is the temperature and humidity correction gain parameter, is the temperature and humidity coupling coefficient; similarly, the speed correction gain parameter of the corrugated paper production line is calculated : , Glue penetration correction gain parameter : is the production line operation speed coupling coefficient, is the glue penetration coupling coefficient, which is obtained by the speed encoder and the glue parameter history drift regression, thereby realizing the multi-factor coupling correction: the ambient temperature and humidity data are multiplied by , production line speed data multiplied by , glue penetration parameter multiplied by , forming a dynamic cleaning data set.

[0056] For example, the tension test value of the corrugated base paper production line is 0.98 Newtons (within the threshold range of 1.5-2.0 Newtons), the infrared moisture meter measures the moisture content at 6.1%, and the compensated moisture content in the original monitoring data is 6.6%. Calculate the moisture content deviation coefficient =|6.1%-6.6%|=0.5%. Take the historical drift regression coefficient , calculate the gain parameter: , , Corrected original monitoring data: Ambient temperature 28°C and humidity 70%RH updated to The production line running speed is updated from 125 meters per minute to Meters per minute, glue penetration parameter 22 is updated to . The final output is dynamic cleaning data for use in subsequent prediction models.

[0057] Optionally, obtaining a voltage control signal based on the dynamic compensation parameter includes: Performing time sequence correction of the equipment operation attenuation coefficient on the dynamic compensation parameter to obtain a compensation correction parameter; Nonlinear adaptive mapping conversion is performed on the compensation correction parameter to obtain a voltage control signal.

[0058] Specifically, first obtain the dynamic compensation parameters , this parameter is the percentage value of water content compensation. Perform timing correction of the equipment operation attenuation coefficient and define the equipment operation attenuation factor as (obtained through fitting the equipment historical operation database, indicating the degree of equipment performance degradation), the correction formula is: , In the formula is the compensation correction parameter. Perform nonlinear adaptive mapping transformation and define the base water content as , the voltage conversion coefficient is (obtained through equipment calibration test), the conversion formula is: , In the formula is the voltage control signal.

[0059] For example, in a corrugated paper production line, the output dynamic compensation parameters . Analyze the current equipment operation time data to obtain the equipment operation attenuation factor Calculate compensation correction parameters . Baseline moisture content , calibration voltage conversion coefficient . Generate voltage control signal This signal is input into the steam nozzle drive circuit, which adjusts the opening frequency to 15Hz to optimize the nozzle injection volume and ensure that the subsequent moisture content compensation is controlled within the process target range.

[0060] Based on the same inventive concept, Figure 7As shown, the present invention also provides a real-time compensation system for moisture content of corrugated paper based on multi-source data fusion, the system comprising: Multi-parameter acquisition module, used to obtain equipment operation time data, glue penetration parameters, temperature and humidity data, corrugated base paper tension change data and corrugated base paper production line operation speed data to generate original monitoring data; A dynamic cleaning module, used to perform cross-validation cleaning processing on the original monitoring data to generate dynamic cleaning data; A dual-mode prediction module, configured to predict the moisture content of the dynamic cleaning data and generate a dual-mode prediction result; A weight allocation module, configured to dynamically allocate weights to the dual-mode prediction results and generate dynamic compensation parameters; A signal generating module, configured to generate a real-time compensation signal based on the dynamic compensation parameter; The feedback correction module is used to obtain the compensated moisture content data, perform feedback correction on the real-time compensation signal, and obtain the moisture content compensation value.

[0061] It should be noted that the electrical connections between the above-mentioned units do not necessarily mean direct connections of lines. Indirect connections are applicable to the embodiments of the present invention as long as the purpose of the present invention is achieved. The above description is only an exemplary embodiment of the present invention and is not intended to limit the scope of the present invention.

[0062] That is, any equivalent changes and modifications made according to the teachings of the present invention are still within the scope of the present invention. Those skilled in the art will readily conceive of other embodiments of the present invention after considering the disclosure of the specification and practical truths. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary technical means in the art not described herein.

Claims

1. A real-time compensation method for moisture content of corrugated paper based on multi-source data fusion, characterized in that: The method comprises: Obtain equipment operation time data, glue penetration parameters, temperature and humidity data, corrugated base paper tension change data, and corrugated base paper production line operation speed data to generate original monitoring data; Performing cross-validation cleaning processing on the original monitoring data to generate dynamic cleaning data; Predicting the moisture content of the dynamic cleaning data to generate a dual-mode prediction result; Performing dynamic weight allocation on the dual-mode prediction results to generate dynamic compensation parameters; generating a real-time compensation signal based on the dynamic compensation parameter; The compensated moisture content data is obtained, and the real-time compensation signal is fed back and corrected to obtain a moisture content compensation value.

2. The method for real-time compensation of moisture content of corrugated paper based on multi-source data fusion according to claim 1, characterized in that: The cross-validation cleaning process is performed on the original monitoring data to generate dynamic cleaning data, including: If the corrugated base paper tension change data does not exceed the preset tension fluctuation threshold range, the original monitoring data is reviewed and tested to obtain a moisture content deviation coefficient; The original monitoring data is corrected according to the moisture content deviation coefficient to generate dynamic cleaning data.

3. The method for real-time compensation of moisture content of corrugated paper based on multi-source data fusion according to claim 1, characterized in that: The performing of moisture content prediction on the dynamic cleaning data to generate a dual-mode prediction result includes: Establishing a physical prediction sub-model based on the temperature and humidity data and the operating speed data of the corrugated paper production line; Building a time series prediction sub-model based on the equipment operation time data and the glue penetration parameters; The output results of the physical prediction sub-model and the time series prediction sub-model are coupled and calculated to generate a dual-mode prediction result.

4. The method for real-time compensation of moisture content of corrugated paper based on multi-source data fusion according to claim 3, characterized in that: The dynamically weighting the dual-mode prediction results to generate dynamic compensation parameters includes: Calculating a first prediction error value of the physical prediction sub-model and a second prediction error value of the time series prediction sub-model to obtain weight coefficients of the physical prediction sub-model and the time series prediction sub-model; When it is detected that the corrugated base paper tension change data exceeds the tension fluctuation threshold range, the weight coefficient of the physical prediction sub-model is increased to obtain a physical dominant weight distribution parameter; When the glue penetration attenuation rate corresponding to the equipment operation time data exceeds the timeliness threshold of the glue penetration parameter, the weight coefficient of the time series prediction sub-model is increased to obtain a time series dominant weight distribution parameter; A weighted linear superposition calculation is performed on the physical dominant weight distribution parameter and the temporal dominant weight distribution parameter to generate a dynamic compensation parameter.

5. The method for real-time compensation of moisture content of corrugated paper based on multi-source data fusion according to claim 1, characterized in that: The generating a real-time compensation signal based on the dynamic compensation parameter includes: obtaining a voltage control signal based on the dynamic compensation parameter; According to the voltage control signal, the opening frequency and injection amount of the steam nozzle are adjusted to obtain actual injection amount data of the steam nozzle; Physical characteristic feedback correction is performed on the actual injection amount data of the steam nozzle to generate a real-time compensation signal.

6. The method for real-time compensation of moisture content of corrugated paper based on multi-source data fusion according to claim 3, characterized in that: The obtaining of the compensated moisture content data and the feedback correction of the real-time compensation signal to obtain the moisture content compensation value include: Based on the compensated moisture content data, a cumulative deviation of the moisture content target value is calculated; Obtaining a deviation perception correction coefficient based on the accumulated deviation of the moisture content target value; Based on the deviation perception correction coefficient, adjusting the hidden layer node connection weights of the time series prediction sub-model to obtain an adaptive time series weight matrix; The real-time compensation signal is feedback-corrected according to the adaptive time-series weight matrix to obtain a moisture content compensation value.

7. The method for real-time compensation of moisture content of corrugated paper based on multi-source data fusion according to claim 6, characterized in that: The step of performing feedback correction on the real-time compensation signal according to the adaptive time series weight matrix to obtain a moisture content compensation value includes: Performing time series prediction model reconstruction and dynamic weight iteration on the adaptive time series weight matrix to obtain optimized compensation parameters; Based on the optimized compensation parameters, the real-time compensation signal feedback gain is adjusted and dynamically compensated to obtain a water content compensation value.

8. The method for real-time compensation of moisture content of corrugated paper based on multi-source data fusion according to claim 2, characterized in that: The correcting the original monitoring data according to the moisture content deviation coefficient to generate dynamic cleaning data includes: Based on the moisture content deviation coefficient, a multi-channel correction gain parameter is obtained; Based on the multi-channel correction gain parameters, multi-factor coupling correction is performed on the temperature and humidity data, the corrugated paper production line operation speed data and the glue penetration parameters to generate dynamic cleaning data.

9. The method for real-time compensation of moisture content of corrugated paper based on multi-source data fusion according to claim 5, characterized in that: The obtaining of a voltage control signal based on the dynamic compensation parameter includes: Performing time sequence correction of the equipment operation attenuation coefficient on the dynamic compensation parameter to obtain a compensation correction parameter; Nonlinear adaptive mapping conversion is performed on the compensation correction parameter to obtain a voltage control signal.

10. A real-time compensation system for moisture content of corrugated base paper based on multi-source data fusion, applied to a real-time compensation method for moisture content of corrugated base paper based on multi-source data fusion as claimed in any one of claims 1 to 9, characterized in that: The system comprises: Multi-parameter acquisition module, used to obtain equipment operation time data, glue penetration parameters, temperature and humidity data, corrugated base paper tension change data and corrugated base paper production line operation speed data to generate original monitoring data; A dynamic cleaning module, used to perform cross-validation cleaning processing on the original monitoring data to generate dynamic cleaning data; A dual-mode prediction module, configured to predict the moisture content of the dynamic cleaning data and generate a dual-mode prediction result; A weight allocation module, configured to dynamically allocate weights to the dual-mode prediction results and generate dynamic compensation parameters; A signal generating module, configured to generate a real-time compensation signal based on the dynamic compensation parameter; The feedback correction module is used to obtain the compensated moisture content data, perform feedback correction on the real-time compensation signal, and obtain the moisture content compensation value.

Images