Corrugated paper production method and system

Abstract

The invention relates to the technical field of corrugated paper production, and discloses a corrugated paper production method and system.The corrugated paper production method comprises the steps that paper is subjected to raw material pretreatment, and an adhesive is prepared; the core paper obtained after raw material pretreatment is put into a corrugating roller, a lubricating agent is added, and first core paper and second core paper are manufactured through corrugation forming; after the pretreated surface paper is coated with an adhesive, the surface paper and the first core paper are overlapped, and first single-layer corrugated paper is obtained through high-temperature and high-pressure curing; coating the pretreated backing paper with an adhesive, then overlapping the pre-treated backing paper with second core paper, and carrying out high-temperature and high-pressure curing to obtain second single-layer corrugated paper; when high-temperature and high-pressure curing is carried out, paper data of the single-layer corrugated paper are obtained, and the current high-temperature and high-pressure curing state is determined based on the paper data; the first single-layer corrugated paper and the second single-layer corrugated paper pass through middle partition paper, and the corrugated paper is obtained on the basis of cold pressing bonding. The adhesive strength is improved, and the rejection rate is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of corrugated paper production, and in particular, to a corrugated paper production method and system. Background Art

[0002] Corrugated paper is an important packaging material. Due to its excellent cushioning performance, light weight and environmental protection characteristics, it is widely used in fields such as logistics packaging, commodity transportation, and industrial packaging. With the rapid development of the global economy and the booming rise of the logistics industry, the demand for corrugated paper continues to grow, driving the continuous progress and innovation of corrugated paper production technology.

[0003] The production process of corrugated paper usually includes process steps such as paper pretreatment, adhesive preparation, core paper forming, high-temperature and high-pressure curing, and multi-layer lamination. Among them, high-temperature and high-pressure curing is a key link in corrugated paper production, directly affecting the strength, cushioning performance and durability of the paper. Traditional high-temperature and high-pressure curing processes often operate with fixed parameters and are difficult to dynamically adjust according to the actual state of the paper (such as thickness, humidity, etc.), which easily leads to problems of insufficient curing or over-curing, thus affecting the performance and quality stability of the paper.

[0004] Therefore, it is necessary to design a corrugated paper production method and system to solve the problems existing in the current technology. Summary of the Invention

[0005] In view of this, the present invention proposes a corrugated paper production method and system, aiming to solve the problem of poor quality stability in current corrugated paper production.

[0006] On the one hand, the present invention proposes a corrugated paper production method, including:

[0007] Performing raw material pretreatment on the paper and preparing an adhesive; the paper raw materials include face paper, middle partition paper, core paper and bottom paper;

[0008] Putting the core paper after raw material pretreatment into a corrugating roll, adding a lubricant, and forming it into a first core paper and a second core paper by corrugating;

[0009] After coating the pretreated face paper with the adhesive, laminating it with the first core paper, and curing it by high temperature and high pressure to obtain a first single-layer corrugated paper; after coating the pretreated bottom paper with the adhesive, laminating it with the second core paper, and curing it by high temperature and high pressure to obtain a second single-layer corrugated paper;

[0010] When performing high-temperature and high-pressure curing, obtaining single-layer corrugated paper sheet data, and determining the current high-temperature and high-pressure curing state based on the sheet data;

[0011] The first single-layer corrugated paper and the second single-layer corrugated paper are bonded together by cold pressing through the middle separator paper to obtain the corrugated paper.

[0012] Furthermore, when obtaining single-layer corrugated paper data, it includes:

[0013] Acquiring temperature data, pressure data, adhesive gelatinization state and moisture content data of the single-layer corrugated paper, and performing preprocessing;

[0014] The preprocessing includes: data cleaning to eliminate invalid values;

[0015] Time synchronization, used to synchronize timestamps to ensure the time consistency of data;

[0016] The sampling frequency is unified to keep the data frequency consistent.

[0017] Further, when determining the current high temperature and high pressure curing state based on the paper data, it includes:

[0018] Inputting the single-layer corrugated paper data into a pre-trained state model for comparison, and obtaining a paper prediction value;

[0019] The paper prediction value is used to calculate the optimal control parameters through a PLC controller;

[0020] Acquire historical data of single-layer corrugated paper, and input the historical data into an autoencoder for reconstruction;

[0021] The historical data includes historical temperature data, historical pressure data, historical adhesive gelatinization state, historical moisture content data and equipment status data;

[0022] Inputting the single-layer corrugated paper data collected in real time into the autoencoder, and calculating the reconstruction error value;

[0023] The current control parameter state is determined based on the reconstructed error value, and the control parameter is compensated to the PLC controller.

[0024] Furthermore, when the single-layer corrugated paper data is input into a pre-trained state model for comparison and a paper prediction value is obtained, it includes:

[0025] Obtain two sets of time series of the single-layer corrugated paper data, wherein the two sets of time series are core time series data and auxiliary time series data;

[0026] In the underlying feature learning stage, the core time series feature learning and auxiliary time series data encoding are obtained, and the core features and auxiliary features are spliced ​​to form enhanced time series features;

[0027] In the high-level feature learning stage, a fully connected network with a two-layer structure is used to extract high-order non-linear features layer by layer based on enhanced temporal features to form a state model;

[0028] Input the single-layer corrugated paper data into the state model for waveform reconstruction and obtain the paper prediction value based on the reconstruction error value.

[0029] Further, when calculating the optimal control parameters by the PLC controller using the paper prediction value, it includes:

[0030] Constrain the PLC controller;

[0031] Determine the objective function based on the prediction sequence of the paper prediction value;

[0032] Convert the rolling optimization solution of the PLC controller into a standard quadratic programming solution to obtain an optimal set of control quantities, and take the first control quantity in the set of control quantities to act on the next moment;

[0033] Convert the first control quantity into a parameter control value, so as to realize optimal parameter control through the PLC controller.

[0034] Further, when obtaining the historical data of the single-layer corrugated paper and inputting the historical data into the autoencoder for reconstruction, it includes:

[0035] Divide the historical data into a training set and a validation set;

[0036] Convert the training set into a time window matrix and construct an encoder based on the time window matrix;

[0037] Train the encoder based on backpropagation and gradient descent to obtain the autoencoder;

[0038] The autoencoder is trained only using normal operating condition data, and the anomaly threshold of the reconstruction error is determined through the validation set.

[0039] Further, when the autoencoder is trained only using normal operating condition data and the anomaly threshold of the reconstruction error is determined through the validation set, it includes:

[0040] Input the validation set into the autoencoder to determine its reconstruction error and construct an error set;

[0041] Arrange the error set in ascending order and determine the anomaly threshold of the reconstruction error;

[0042] When the real-time obtained reconstruction error value is greater than the anomaly threshold, trigger the interference compensation mechanism and trigger an anomaly alarm.

[0043] Further, when the reconstructed error value obtained in real time is greater than the abnormal threshold, triggering the interference compensation mechanism and triggering an abnormal alarm includes:

[0044] Calculating the reconstructed error value of the paper data;

[0045] When the reconstructed error value is greater than the abnormal threshold, marking it as an abnormal point and triggering an abnormal alarm;

[0046] Compensating the optimal parameter control based on the amplitude of the reconstructed error value.

[0047] Further, when determining the current control parameter state based on the reconstructed error value and compensating the control parameters into the PLC controller, it includes:

[0048] When the reconstructed error value is greater than the abnormal threshold, determining the control compensation vector based on proportional feedback;

[0049] Updating the optimal control parameters based on the control compensation vector.

[0050] Compared with the prior art, the beneficial effects of the present invention are as follows: Through the pretreatment of different functional layers (facing paper, core paper, bottom paper, etc.), the cleanliness, dimensional accuracy, and functionality of various papers are enhanced. The adhesive ensures that the glue has appropriate adhesion and permeability, avoiding both weak adhesion caused by too thin glue and uneven coating caused by too thick glue. The corrugating process of the corrugating roll gives the core paper a regular wave structure, which not only improves the cushioning performance of the cardboard but also optimizes the vertical compressive strength through structural design. The real-time data monitoring and status feedback during the high-temperature and high-pressure curing stage dynamically adjust the temperature and pressure parameters to avoid degumming and delamination caused by ungelatinized glue or over-drying. The real-time monitoring during the high-temperature and high-pressure curing process can obtain paper data in real time, thereby dynamically adjusting the curing parameters to ensure full curing of the adhesive while avoiding over-curing, further improving the strength and durability of the paper. Through real-time monitoring, abnormal situations in production can be detected and adjusted in a timely manner, reducing the scrap rate, optimizing resource utilization, and reducing energy consumption. During the high-temperature and high-pressure curing process, real-time monitoring can optimize the curing time and temperature according to the real-time state of the paper by dynamically adjusting the curing parameters, avoiding problems such as a decrease in paper performance caused by over-curing or poor adhesion caused by insufficient curing. This not only improves the physical properties of the paper, such as compressive strength and cushioning performance, but also extends the service life of the corrugated paper, making it perform more excellently in packaging applications.

[0051] On the other hand, the present application also provides a corrugated paper production system for applying the above corrugated paper production method, including:

[0052] An acquisition module configured to obtain single-layer corrugated paper data and perform preprocessing;

[0053] A judgment module, configured to input the single-layer corrugated paper data into a pre-trained state model for comparison and obtain a paper prediction value;

[0054] A management module, configured to calculate optimal control parameters through a PLC controller based on the paper prediction value;

[0055] A compensation module, configured to input the historical data into an autoencoder for reconstruction, and the compensation module is further configured to determine the current control parameter state based on the reconstruction error value and compensate the control parameters into the PLC controller.

[0056] It can be understood that the above-mentioned corrugated paper production system has the same beneficial effects and will not be elaborated here. Description of the Drawings

[0057] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0058] Figure 1 is a flowchart of the corrugated paper production method provided by the embodiment of the present invention;

[0059] Figure 2 is a structural diagram of the corrugated paper provided by the embodiment of the present invention;

[0060] Figure 3 is a functional block diagram of the corrugated paper production system provided by the embodiment of the present invention.

[0061] Among them, 1, face paper; 2, first core paper; 3, middle partition paper; 4, second core paper; 5, bottom paper. Detailed Embodiments

[0062] Hereinafter, exemplary embodiments of the present disclosure will be described in more detail with reference to the drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be completely conveyed to those skilled in the art. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other. Hereinafter, the present invention will be described in detail with reference to the drawings and in combination with the embodiments.

[0063] In some embodiments of the present application, referring to Figure 1-2 as shown, a corrugated paper production method includes:

[0064] S100: Pretreat the raw materials of the paper and prepare the adhesive. The raw materials of the paper include the face paper, the middle partition paper, the core paper, and the bottom paper.

[0065] S200: Put the core paper after raw material pretreatment into the corrugating roll, add lubricant, and form the first core paper and the second core paper through corrugating.

[0066] S300: After coating the pretreated face paper with the adhesive, laminate it with the first core paper and cure it through high temperature and high pressure to obtain the first single-layer corrugated paper. After coating the pretreated bottom paper with the adhesive, laminate it with the second core paper and cure it through high temperature and high pressure to obtain the second single-layer corrugated paper.

[0067] S400: When curing through high temperature and high pressure, obtain the paper data of the single-layer corrugated paper and determine the current high temperature and high pressure curing state based on the paper data.

[0068] S500: Bond the first single-layer corrugated paper and the second single-layer corrugated paper through the middle partition paper to obtain the corrugated paper based on cold pressing.

[0069] Specifically, the paper includes kraft paper, linerboard, corrugating medium, etc., and needs to have a certain strength (such as tensile strength, tear strength) and thickness. The adhesive is starch paste or hot melt adhesive. The starch paste needs to be formulated to an appropriate concentration (usually 12%-18%). At the same time, auxiliary materials need to be added to the paper: desiccant (to control humidity), lubricant (to reduce friction), waterproofing agent or plasticizer (to improve performance). Then, raw material pretreatment of the paper is carried out: Dust removal: Use an electrostatic precipitator to remove surface impurities (dust content ≤ 0.1%). Trimming: Use a paper cutter to trim the edges of the base paper within a tolerance of ±0.5 mm. Preparation of starch paste: Mix starch and water in a ratio of 1:8 - 1:10 and heat to 85 - 95°C for gelatinization. Raw materials for the separator paper: Select the corresponding paper and carry out anti-counterfeiting measures (print brand anti-counterfeiting information such as Logo, pattern. Print invisible ink, and the corresponding printed content will be displayed under a specific ultraviolet lamp). Then, process the corrugated paper: Forming and conveying the core paper of the first layer of corrugation and the separator paper: Preheat the roller temperature and adjust the humidity. Corrugation pressing: The concave and convex patterns on the surface of the corrugating roll make the core paper form continuous waves (select the corresponding 45-degree F-type corrugation, etc.), which determines the cushioning strength and compressive performance. Face paper / glue coating: After another layer of paper (such as kraft paper) is coated with starch paste, it is laminated with the corrugation. Hot pressing and bonding: The adhesive is cured through high temperature and high pressure to form a multi-layer structure (core paper + corrugation + face paper). Then, through hot air drying: Use a multi-stage drying cylinder system to gradually remove moisture (water content is controlled at 8%-12%). Cooling: Low-temperature air cooling is used to prevent the cardboard from deforming. The process for the second single-layer corrugated paper is the same as the above steps. Only during corrugation pressing, the concave and convex patterns on the surface of the corrugating roll make the core paper form continuous waves (select the corresponding 90-degree F-type corrugation). Then, laminate the first single-layer corrugated paper and the second single-layer corrugated paper, and gently press to remove air bubbles. Place the separator paper between the first corrugated paper and the second corrugated paper. The arrangement order of the paper is: face paper, first core paper, separator paper, second core paper, bottom paper. During high-temperature and high-pressure curing, if the temperature or pressure fluctuates, it will cause the glue not to gelatinize or excessive drying, which will lead to degumming or delamination. For example, in a high-temperature and low-humidity environment, static electricity accumulates on the paper surface, which will adsorb dust in the workshop, thus contaminating the glue line and resulting in bonding failure. If the water content is out of control, it will cause the water content of the cardboard to be less than 8% after high-temperature curing, which will cause it to absorb moisture and deform, and if it is higher than 12%, it is prone to mildew. Therefore, the control of high-temperature and high-pressure parameters is particularly important. By obtaining the paper parameters under high-temperature and high-pressure conditions, and then judging whether the current high-temperature and high-pressure parameters are normal based on the paper parameters, the risk of paper errors can be reduced.

[0070] It can be understood that through dust removal, trimming, and addition of auxiliary materials (desiccant, lubricant, waterproofing agent, etc.) in the pretreatment process, the cleanliness, dimensional accuracy, and functionality of the paper are improved. Dust removal reduces the interference of surface impurities on bonding, trimming ensures the lamination alignment accuracy, the lubricant reduces the friction damage between the corrugating roll and the paper, and the waterproofing agent enhances the moisture resistance of the cardboard, making the material more adaptable to complex environmental applications.

[0071] The targeted selection and pretreatment of base paper types (kraft paper, boxboard, etc.) optimizes the tensile strength, tear strength and surface flatness of the paper, providing a reliable substrate for subsequent bonding. Starch paste ensures that the glue has appropriate fluidity and adhesion, avoiding insufficient glue penetration due to excessive dilution and uneven coating due to excessive thickness. The corrugation pressing process of the corrugated roller (such as 45-degree F-type corrugation and 90-degree F-type corrugation) gives the core paper a regular wave structure, which not only improves the cushioning performance of the paperboard, but also optimizes the compressive strength by controlling the height and density of the corrugation.

[0072] The addition of lubricant reduces the mechanical stress during corrugation forming, avoids the breakage or crushing of the core paper, and ensures the consistency of the corrugated shape.

[0073] In some embodiments of the present application, obtaining single-layer corrugated paper data includes:

[0074] The temperature data, pressure data, adhesive gelatinization state and moisture content data of single-layer corrugated paper are obtained and pre-processed.

[0075] Preprocessing includes: data cleaning to remove invalid values.

[0076] Time synchronization is used to synchronize timestamps to ensure the time consistency of data.

[0077] The sampling frequency is unified to keep the data frequency consistent.

[0078] It is understandable that temperature and pressure are the core indicators reflecting the stability of the hot pressing process, the gelatinization state of the adhesive directly determines the bonding quality, and the moisture content affects the paper forming and strength performance. By comprehensively collecting these variables, a comprehensive perception of the entire production process can be achieved, ensuring the accuracy of quality analysis and the foresight of process control. Secondly, "data cleaning" in data preprocessing can eliminate invalid or interfering data, avoid meaningless or erroneous information affecting subsequent model training or judgment logic, and thus improve the accuracy of model output and the reliability of system response. Time synchronization is a key link in realizing multi-source data fusion analysis. In actual operation, the sampling devices of different sensors will have inconsistent time bases. If they are not synchronized uniformly, the data will be misaligned in the time dimension, which will seriously affect the logical correspondence between the data. Through time synchronization processing, it is ensured that the parameters of different sources are comparable on the same time axis, thereby ensuring the rationality of data integration, trend analysis and correlation modeling, and improving the modeling quality and timeliness of judgment. The unified sampling frequency is associated with data standardization modeling and model stability. In practical applications, if some high-frequency signals are directly fused with low-frequency signals without being processed, it will lead to inconsistent data structures, which will in turn cause deviations or misunderstandings in the model training stage.

[0079] In some embodiments of the present application, when determining the current high-temperature and high-pressure curing state based on paper data, it includes:

[0080] Input the single-layer corrugated paper data into a pre-trained state model for comparison, and obtain the paper prediction value.

[0081] Calculate the optimal control parameters for the paper prediction value through the PLC controller.

[0082] Obtain the historical data of the single-layer corrugated paper, and input the historical data into the autoencoder for reconstruction.

[0083] The historical data includes historical temperature value data, historical pressure data, historical adhesive gelatinization state, historical moisture content data, and equipment status data.

[0084] Input the single-layer corrugated paper data collected in real time into the autoencoder, and calculate the reconstruction error value.

[0085] Determine the current control parameter state based on the reconstruction error value, and compensate the control parameters to the PLC controller.

[0086] It can be understood that the state model is an LSTM-Autoencoder model. By inputting the corrugated paper data into the model for comparison, the future change trend of the paper data can be obtained. For the future change trend, the optimal control parameters are calculated through the PLC controller, and then by adjusting the high-temperature and high-pressure parameters, precise control of the paper is achieved, preventing the occurrence of bonding failure problems. Furthermore, through the feature extraction ability of the autoencoder-decoder for the paper data obtained in real time, it is judged whether the currently calculated optimal control parameters are accurate. If not, compensation is made by compensating the control parameters to the PLC controller to prevent problems in the high-temperature and high-pressure curing of the paper caused by misjudgment of the state model.

[0087] In some embodiments of the present application, when inputting the single-layer corrugated paper data into a pre-trained state model for comparison and obtaining the paper prediction value, it includes:

[0088] Obtain two time series of the single-layer corrugated paper data, and the two time series are respectively core time series data and auxiliary time series data.

[0089] In the bottom layer feature learning stage, obtain the core time series feature learning and auxiliary time series data encoding, and splice the core feature and the auxiliary feature to form an enhanced time series feature.

[0090] In the high-level feature learning stage, use a fully connected network with a two-layer structure based on the enhanced time series feature to extract high-order non-linear features layer by layer to form a state model.

[0091] Input the single-layer corrugated paper data into the state model, perform waveform reconstruction, and obtain the paper prediction value based on the reconstruction error value.

[0092] Specifically, the core time-series data includes temperature, humidity, and paper thickness data. Input the core time-series data into the LSTM network, use two-layer stacked LSTM (with 64 hidden units) to capture long-term dependencies, output the feature vector at each time step, and then apply one-dimensional convolution (with a kernel size of 5 and a stride of 1) to the waveform to extract local fluctuation patterns (such as sharp drops and sharp rises). The pooling layer compresses the feature dimension and retains the key waveform information. Finally, concatenate the sequence features output by the LSTM with the waveform features extracted by the CNN to form the core time-series features. The auxiliary time-series data includes process control and equipment status-related parameters, such as pressure data, corrugating roll speed, ambient humidity, etc. These parameters are usually collected at a lower frequency. Align the auxiliary data with the core data timestamp. Map continuous variables such as pressure and speed to low-dimensional vectors through a fully connected layer (32 nodes, ReLU activation), and model the dependencies between parameters such as wear degree and ambient humidity through the self-attention mechanism to output the embedding vector. Finally, merge the features encoded by the fully connected layer and the self-attention mechanism to form the auxiliary time-series features. Concatenate the core time-series features and the auxiliary time-series features along the feature axis to generate enhanced time-series features. In the high-level learning stage, the first layer: Input the enhanced time-series features (192 dimensions), pass through a fully connected layer with 256 nodes, the activation function is ReLU, and add Dropout (ratio 0.3) to randomly mask some neurons to prevent overfitting. The second layer: Input the 256-dimensional features, compress them to 128 nodes, the activation function is ReLU, and add Dropout (ratio 0.2) to further improve the generalization ability, and then obtain the state model. Input the real-time input core time-series data into the model, output the reconstructed waveform, and calculate the reconstruction error at each time point: where θ is the error value, S is the time-series length, x i is the actual data value, is the reconstructed data value, and the future change trend can be obtained through the reconstruction error.

[0093] In some embodiments of the present application, when calculating the optimal control parameters through the PLC controller using the paper prediction value, it includes:

[0094] Constrain the PLC controller.

[0095] Determine the objective function based on the prediction sequence of the paper prediction value.

[0096] Convert the rolling optimization solution of the PLC controller into a standard quadratic programming solution to obtain an optimal set of control quantities, and take the first control quantity in a set of control quantities to act on the next moment.

[0097] Convert the first control quantity into a parameter control value, so as to achieve optimal parameter control through the PLC controller.

[0098] It can be understood that the maximum and minimum values of the control parameters are determined for the historical normal working conditions, and the control variables of the PLC controller are restricted to be between the maximum and minimum values. When the state model predicts the paper state values in the next N steps as: Then, the objective function (in the form of a typical quadratic form) is: where J is the solution value of the objective function, is the predicted paper state value, is the expected value, Δu k+i is the control increment, Q is the weighted matrix of the output error, used to adjust the importance of different output variables, R is the weighted matrix of the control increment, which is a positive definite matrix, used to avoid frequent adjustment of the control quantity, reflecting the weight of the performance index, N is the prediction time domain length, representing the total number of time steps for forward prediction, i is the prediction step index, the number of steps for forward prediction starting from the current moment, each i corresponds to a future moment, and then an optimal set of control quantities is obtained. Select the first control quantity to act on the next moment, and the first control quantity is the optimal parameter control.

[0099] In some embodiments of the present application, when obtaining the historical data of single-layer corrugated paper and inputting the historical data into the autoencoder for reconstruction, it includes:

[0100] Divide the historical data into a training set and a validation set.

[0101] Convert the training set into a time window matrix, and construct an encoder based on the time window matrix.

[0102] Train the encoder based on backpropagation and gradient descent to obtain the autoencoder.

[0103] The autoencoder is trained only using normal working condition data, and the abnormal threshold of the reconstruction error is determined through the validation set.

[0104] Specifically, extract the paper historical data including fields such as temperature, pressure, moisture content, and gelatinization state, only select the data known to be in normal working conditions, divide it into a training set and a validation set, convert the original time series data into a sliding window format, and then compress the high-dimensional window into a low-dimensional hidden space through the encoder, and then reconstruct the input through the decoder to approach the original sample. Its loss function is where τ is the loss function, used to measure the average level of the reconstruction error, C is the number of training set samples, x (i) is the original input data of the i-th training sample, is the reconstruction result of the i-th sample, and then the abnormal threshold of the reconstruction error is determined through the validation set.

[0105] In some embodiments of the present application, when the autoencoder is trained only using normal operating condition data and the anomaly threshold of the reconstruction error is determined through a validation set, it includes:

[0106] Input the validation set into the autoencoder to determine its reconstruction error, and construct an error set.

[0107] Arrange the error set in ascending order, and determine the anomaly threshold of the reconstruction error.

[0108] When the value of the reconstruction error obtained in real time is greater than the anomaly threshold, trigger the interference compensation mechanism and trigger an anomaly alarm.

[0109] Specifically, input all samples in the validation set into the autoencoder to obtain an error set, sort the error set in ascending order, set the percentile to 95%, calculate the index, finally obtain the anomaly threshold, and then input the data obtained in real time into the autoencoder for reconstruction, determine whether the real-time error is abnormal, when the error is greater than the threshold, determine it as abnormal, and compensate the high-temperature and high-pressure parameters according to the anomaly amplitude.

[0110] In some embodiments of the present application, when the value of the reconstruction error obtained in real time is greater than the anomaly threshold, trigger the interference compensation mechanism and trigger an anomaly alarm, it includes:

[0111] Calculate the reconstruction error value of the paper data.

[0112] When the reconstruction error value is greater than the anomaly threshold, mark it as an abnormal point and trigger an anomaly alarm.

[0113] Compensate the optimal parameter control based on the amplitude of the reconstruction error value.

[0114] In some embodiments of the present application, when determining the current control parameter state based on the reconstruction error value and compensating the control parameters into the PLC controller, it includes:

[0115] When the reconstruction error value is greater than the anomaly threshold, determine the control compensation vector based on proportional feedback.

[0116] Update the optimal control parameters based on the control compensation vector.

[0117] Specifically, through the pre-treatment of different functional layers (such as face paper, core paper, bottom paper, etc.), the cleanliness, dimensional accuracy and functionality of various papers are enhanced. The adhesive ensures that the glue has appropriate adhesion and permeability, avoiding weak adhesion caused by too thin glue and uneven coating caused by too thick glue. The corrugating process of the corrugating roll gives the core paper a regular wave structure, which not only improves the cushioning performance of the cardboard, but also optimizes the vertical compressive strength through structural design. The real-time data monitoring and status feedback during the high-temperature and high-pressure curing stage dynamically adjust the temperature and pressure parameters to avoid degumming and delamination caused by ungelatinized glue or over-drying. The real-time monitoring during the high-temperature and high-pressure curing process can obtain paper data in real time, so as to dynamically adjust the curing parameters, ensure the full curing of the adhesive while avoiding over-curing, and further improve the strength and durability of the paper. Through real-time monitoring, abnormal situations in production can be detected and adjusted in time, reducing the scrap rate, optimizing resource utilization and reducing energy consumption. During the high-temperature and high-pressure curing process, real-time monitoring can optimize the curing time and temperature according to the real-time state of the paper by dynamically adjusting the curing parameters, avoiding problems such as the decline in paper performance caused by over-curing or poor adhesion caused by insufficient curing. This not only improves the physical properties of the paper, such as compressive strength and cushioning performance, but also extends the service life of the corrugated paper, making it perform better in packaging applications.

[0118] In another preferred embodiment based on the above embodiments, referring to Figure 3 As shown, this embodiment provides a generative software automatic assembly system based on a contract framework model for applying the above corrugated paper production method and system, including:

[0119] An acquisition module, configured to obtain single-layer corrugated paper data and perform pre-treatment.

[0120] A judgment module, configured to input the single-layer corrugated paper data into a pre-trained state model for comparison and obtain a paper prediction value.

[0121] A management module, configured to calculate the optimal control parameters through a PLC controller with the paper prediction value.

[0122] A compensation module, configured to input historical data into an autoencoder for reconstruction. The compensation module is also configured to determine the current control parameter state based on the reconstruction error value and compensate the control parameters into the PLC controller.

[0123] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0124] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of flows and / or blocks in the flowchart and / or block diagram can also be implemented. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in Figure 1 one or more flows Figure 1 or a plurality of flows and / or blocks.

[0125] These computer program instructions can also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable storage medium generate a manufactured article including an instruction device, and the instruction device implements the functions specified in Figure 1 one or more flows Figure 1 or a plurality of flows and / or blocks.

[0126] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Therefore, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more flows Figure 1 or a plurality of flows and / or blocks.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: still can modify the specific implementation manners of the present invention or make equivalent replacements, and any modification or equivalent replacement that does not deviate from the spirit and scope of the present invention should be covered by the protection scope of the claims of the present invention.

Claims

1. A method for producing corrugated paper, characterized in that: include: Pre-treating the paper raw materials and preparing the adhesive; the paper raw materials include face paper, middle separator paper, core paper and base paper; Putting the pre-treated core paper into a corrugated roller, adding a lubricant, and forming the first core paper and the second core paper by corrugating; After the pretreated surface paper is coated with the adhesive, it is overlapped with the first core paper, and cured by high temperature and high pressure to obtain a first single-layer corrugated paper; after the pretreated bottom paper is coated with the adhesive, it is overlapped with the second core paper, and cured by high temperature and high pressure to obtain a second single-layer corrugated paper; When high temperature and high pressure curing is performed, single-layer corrugated paper data is obtained, and the current high temperature and high pressure curing state is determined based on the paper data; The first single-layer corrugated paper and the second single-layer corrugated paper are bonded together by cold pressing through the middle separator paper to obtain the corrugated paper.

2. The method for producing corrugated paper according to claim 1, characterized in that: When obtaining single-ply corrugated paper data, it includes: Acquiring temperature data, pressure data, adhesive gelatinization state and moisture content data of the single-layer corrugated paper, and performing preprocessing; The preprocessing includes: data cleaning to eliminate invalid values; Time synchronization, used to synchronize timestamps to ensure the time consistency of data; The sampling frequency is unified to keep the data frequency consistent.

3. The method for producing corrugated paper according to claim 2, characterized in that: When determining the current high temperature and high pressure curing state based on the paper data, it includes: Inputting the single-layer corrugated paper data into a pre-trained state model for comparison, and obtaining a paper prediction value; The paper prediction value is used to calculate the optimal control parameters through a PLC controller; Acquire historical data of single-layer corrugated paper, and input the historical data into an autoencoder for reconstruction; The historical data includes historical temperature data, historical pressure data, historical adhesive gelatinization state, historical moisture content data and equipment status data; Inputting the single-layer corrugated paper data collected in real time into the autoencoder, and calculating the reconstruction error value; The current control parameter state is determined based on the reconstructed error value, and the control parameter is compensated to the PLC controller.

4. The method for producing corrugated paper according to claim 3, characterized in that: When the single-layer corrugated paper data is input into the pre-trained state model for comparison and the paper prediction value is obtained, it includes: Obtaining two sets of time series of the single-layer corrugated paper data, wherein the two sets of time series are core time series data and auxiliary time series data; In the underlying feature learning stage, the core time series feature learning and auxiliary time series data encoding are obtained, and the core features and auxiliary features are spliced ​​to form enhanced time series features; In the high-level feature learning stage, a fully connected network with a two-layer structure is used to extract high-order nonlinear features layer by layer based on enhanced temporal features to form a state model; The single-layer corrugated paper data is input into the state model, waveform reconstruction is performed, and a paper prediction value is obtained based on the reconstruction error value.

5. The method for producing corrugated paper according to claim 4, characterized in that: When the paper prediction value is used to calculate the optimal control parameter through the PLC controller, it includes: Constrain the PLC controller; Determining the objective function based on the prediction sequence of the paper prediction value; The rolling optimization solution of the PLC controller is converted into a standard quadratic programming solution to obtain an optimal set of control quantities, and the first control quantity in the set of control quantities is taken to act at the next moment; The first control quantity is converted into a parameter control value, thereby achieving optimal parameter control through a PLC controller.

6. The method for producing corrugated paper according to claim 5, characterized in that: The method of obtaining the historical data of the single-layer corrugated paper and inputting the historical data into the autoencoder for reconstruction includes: Dividing the historical data into a training set and a validation set; Converting the training set into a time window matrix, and constructing an encoder based on the time window matrix; Training the encoder based on back propagation and gradient descent to obtain the autoencoder; The autoencoder is trained using only normal operating condition data, and an abnormal threshold of a reconstruction error is determined through the validation set.

7. The method for producing corrugated paper according to claim 6, characterized in that: When the autoencoder is trained using only normal operating data and the abnormal threshold of the reconstruction error is determined through the validation set, it includes: Inputting the validation set into the autoencoder to determine its reconstruction error and construct an error set; Arranging the error set in ascending order and determining an abnormal threshold of the reconstruction error; When the reconstruction error value acquired in real time is greater than the abnormal threshold, an interference compensation mechanism is triggered, and an abnormal alarm is triggered.

8. The method for producing corrugated paper according to claim 7, characterized in that: When the reconstruction error value obtained in real time is greater than the abnormal threshold, the interference compensation mechanism is triggered, and the abnormal alarm is triggered, including: Calculating the reconstruction error value of the paper data; When the reconstruction error value is greater than the abnormal threshold, it is marked as an abnormal point and an abnormal alarm is triggered; The optimal parameter control is compensated based on the magnitude of the reconstruction error value.

9. The method for producing corrugated paper according to claim 8, characterized in that: Determining the current control parameter state based on the reconstructed error value and compensating the control parameter to the PLC controller includes: When the reconstruction error value is greater than the abnormal threshold, determining a control compensation vector based on proportional feedback; The optimal control parameter is updated based on the control compensation vector.

10. A corrugated paper production system, applied to the corrugated paper production method according to any one of claims 1 to 9, characterized in that: include: The acquisition module is configured to acquire single-layer corrugated paper data and perform pre-processing; A judgment module is configured to input the single-layer corrugated paper data into a pre-trained state model for comparison and obtain a paper prediction value; A management module is configured to calculate optimal control parameters through a PLC controller using the paper prediction value; The compensation module is configured to input the historical data into the autoencoder for reconstruction. The compensation module is also configured to determine the current control parameter state based on the reconstruction error value and compensate the control parameter into the PLC controller.

Images