Full-process intelligent control system for corrugated packaging whole-line production

Abstract

A full-flow intelligent control system for corrugated packaging whole-line production relates to the technical field of production flow control, and is characterized in that flow information of corrugated packaging whole-line production is acquired, corrugated packaging monitoring point positions are set according to the flow information, and production data of each point position are acquired; the production data of each point location is monitored and controlled in real time; dividing the production monitoring index of each flow sub-sequence into a high-frequency abnormal production monitoring index or a low-frequency abnormal production monitoring index; constructing a high-frequency index pre-estimation model, and obtaining first pre-estimation production data; acquiring a reference production monitoring index of the low-frequency abnormal production monitoring index; constructing a low-frequency index pre-estimation model, and obtaining second pre-estimation production data; and third estimated production data is generated according to the first estimated production data and the second estimated production data, and the initial control item measures of the next collection period are obtained according to the third estimated production data, so that the production efficiency and the product quality of corrugated packaging are remarkably improved.

Description

Technical Field

[0001] The invention relates to the technical field of production process control, and in particular to a full-process intelligent control system for the production of a corrugated packaging whole line. Background Art

[0002] A Chinese patent with publication number CN118789883A discloses a production process for corrugated paper packaging boxes, comprising the following steps: S1, raw material preparation; S2, designing and making a die cutter; S3, cutting; S4, waste removal: placing the cut corrugated paperboard into a waste removal device, and then removing the excess scraps through the waste removal device according to the cutting line formed by die cutting, thereby obtaining semi-finished products and scraps; S5, recycling; S6, gluing boxes.

[0003] A Chinese patent with publication number CN118735192A discloses an intelligent recommendation method and system for a corrugated cardboard packaging production process, including obtaining order feature data of target customers and inputting it into a multi-objective optimization model to obtain an initial recommendation plan, evaluating its first matching degree with the customer demand progress, and adjusting the order priority if the requirements are not met. Otherwise, evaluating the second matching degree between the raw material supply and the customer demand, matching the associated production process for the initial recommendation plan when the conditions are not met, and generating a target recommendation plan; executing and configuring the initial parameters of the production process according to the target recommendation plan, obtaining the actual operating parameters of the production line in real time and drawing a change curve, predicting the production operation status and determining the parameter compensation factor, and updating the initial parameters of the production process.

[0004] Corrugated packaging is a packaging form widely used in the fields of logistics and consumer goods. Its production process involves many complex links. In the existing corrugated packaging production line, production data is scattered in various departments and equipment. There is a lack of a unified management platform, making it difficult to conduct comprehensive analysis. There is a lack of data analysis tools, and the value of production data is not fully utilized. There is a lack of data-based decision support system, making it difficult for management to make scientific decisions. Summary of the invention

[0005] In order to solve the above technical problems, the object of the present invention is to provide a full-process intelligent control system for the production of a whole line of corrugated packaging, including a monitoring center, wherein the monitoring center is communicatively connected with a data acquisition module, a real-time monitoring module, a data analysis module, a high-frequency estimation module, a data processing module, a low-frequency estimation module and a data early warning module;

[0006] The data acquisition module is used to obtain process information of the entire corrugated packaging production line, set corrugated packaging monitoring points according to the process information, and obtain production data of each point;

[0007] The real-time monitoring module is used to monitor and control the production data of each point in real time;

[0008] The data analysis module is used to divide the production monitoring indicators of each process subsequence into high-frequency abnormal production monitoring indicators or low-frequency abnormal production monitoring indicators;

[0009] The high-frequency estimation module is used to construct a high-frequency indicator estimation model to obtain first estimated production data;

[0010] The data processing module is used to obtain reference production monitoring indicators of low-frequency abnormal production monitoring indicators;

[0011] The low-frequency estimation module is used to construct a low-frequency index estimation model to obtain second estimated production data;

[0012] The data early warning module is used to generate third estimated production data according to the first estimated production data and the second estimated production data, and to obtain initial control item measures for the next collection cycle according to the third estimated production data.

[0013] Furthermore, the data acquisition module obtains the process information of the whole production line of corrugated packaging, sets the corrugated packaging monitoring points according to the process information, and obtains the production data of each point, including:

[0014] Extracting process unit characteristics of each corrugated packaging production equipment from the process information, and dividing the entire corrugated packaging production line into a number of process sub-sequences according to the process unit characteristics;

[0015] Set up corrugated packaging monitoring points in each process subsequence, and use data retrieval to obtain production monitoring indicators of each corrugated packaging monitoring point according to the functional characteristics in the process unit characteristics of the corresponding process subsequence;

[0016] The corrugated packaging monitoring point acquires production data in real time according to the production monitoring index, marks the acquisition time, and sets the acquisition cycle.

[0017] Furthermore, the process of the real-time monitoring module monitoring and controlling the production data of each point in real time includes:

[0018] Acquire the production data of each corrugated packaging monitoring point, extract various types of production monitoring indicators in the production data, preset the threshold range of each type of production monitoring indicators in each corrugated packaging monitoring point, compare each type of production monitoring indicator in the production data with the corresponding threshold range, if there is a production monitoring indicator that is not in the corresponding threshold range, then mark the production monitoring indicator as an abnormal production monitoring indicator, and generate an abnormal alarm signal of the process subsequence according to the abnormal production monitoring indicator. The monitoring center arranges relevant personnel to detect and analyze the process subsequence according to the abnormal alarm signal, obtains the corresponding control item measures, pre-builds a control database, and matches the production data with the control item measures and stores them in the control database.

[0019] Furthermore, the process of the data analysis module dividing the production monitoring indicators of each process subsequence into high-frequency abnormal production monitoring indicators or low-frequency abnormal production monitoring indicators includes:

[0020] Obtain all historical collection cycles in which each process subsequence generates an abnormal alarm signal, mark the historical collection cycle as an abnormal historical collection cycle, extract each abnormal production monitoring indicator in the abnormal historical collection cycle for statistical analysis, obtain the number of occurrences of each abnormal production monitoring indicator, preset an occurrence threshold, mark the abnormal production monitoring indicator whose occurrence number is greater than the occurrence threshold as a high-frequency abnormal production monitoring indicator, and mark the abnormal production monitoring indicator whose occurrence number is less than or equal to the occurrence threshold as a low-frequency abnormal production monitoring indicator.

[0021] Furthermore, the high-frequency estimation module constructs a high-frequency indicator estimation model, and the process of obtaining the first estimated production data includes:

[0022] Obtain the production data within the historical collection period to which each high-frequency abnormal production monitoring indicator of each process subsequence belongs, use the production data as a training set and a test set, build a high-frequency indicator prediction model based on deep learning, input the training set into the high-frequency indicator prediction model for training until the loss function training is stable, save the model parameters, test the high-frequency indicator prediction model through the test set until it meets the preset requirements, and output the high-frequency indicator prediction model;

[0023] The estimated production data of the next acquisition cycle of each process subsequence is output according to the high-frequency indicator estimation model, and the estimated production data is marked as the first estimated production data.

[0024] Furthermore, the process of the data processing module obtaining the reference production monitoring index of the low-frequency abnormal production monitoring index includes:

[0025] Obtain the historical collection period to which each low-frequency abnormal production monitoring indicator of each process subsequence belongs, extract the high-frequency abnormal production monitoring indicator within the historical collection period, perform statistical analysis on the high-frequency abnormal production monitoring indicator within the historical collection period, and obtain the cumulative number of times each high-frequency abnormal production monitoring indicator and each low-frequency abnormal production monitoring indicator appear in the same historical collection period;

[0026] The high-frequency abnormal production monitoring indicator that appears the highest cumulative number of times in the same historical collection period as the low-frequency abnormal production monitoring indicator shall be marked as the reference production monitoring indicator of the low-frequency abnormal production monitoring indicator.

[0027] Furthermore, the low-frequency estimation module constructs a low-frequency index estimation model, and the process of obtaining the second estimated production data includes:

[0028] Obtain production data within a historical collection period to which a reference production monitoring indicator of a low-frequency abnormal production monitoring indicator belongs, use the production data as training data, build a large sample model based on deep learning, train the large sample model through the training data, obtain the trained large sample model, and obtain model parameters in the large sample model;

[0029] A low-frequency indicator prediction model is constructed based on deep learning, and the model parameters in the large sample model are used as the initial model parameters of the low-frequency indicator prediction model. The production data within the historical collection period to which the low-frequency abnormal production monitoring indicator belongs is obtained, and the production data is used as the training data of the low-frequency indicator prediction model to train the low-frequency indicator prediction model, and the trained low-frequency indicator prediction model is output;

[0030] Output the estimated production data of the next acquisition cycle of each process subsequence according to the low-frequency indicator estimation model, and mark the estimated production data as second estimated production data.

[0031] Furthermore, the process of the data early warning module generating the third estimated production data according to the first estimated production data and the second estimated production data includes:

[0032] For each type of production monitoring indicator in the first estimated production data in the next collection cycle, obtain the production monitoring indicator marked as a low-frequency abnormal production monitoring indicator in the historical collection cycle among the various types of production monitoring indicators, and remove the production monitoring indicator from the first estimated production data;

[0033] For each type of production monitoring indicator in the second estimated production data in the next collection cycle, obtain the production monitoring indicator marked as a high-frequency abnormal production monitoring indicator in the historical collection cycle among the various types of production monitoring indicators, and remove the production monitoring indicator from the second estimated production data;

[0034] Subsequently, the first estimated production data and the second estimated production data are merged to generate third estimated production data, and the estimated numerical time series of various types of production monitoring indicators in the next collection cycle are output based on the third estimated production data.

[0035] Furthermore, the process of the data early warning module acquiring the initial control item measures for the next collection cycle according to the third estimated production data includes:

[0036] Compare the estimated numerical time series sequence of each type of production monitoring indicator in the next acquisition cycle with the corresponding threshold interval. If there is a production monitoring indicator whose corresponding estimated numerical time series sequence is not within the corresponding threshold interval, input the estimated numerical time series sequence of each type of production monitoring indicator into the control database for data retrieval, obtain production data whose similarity meets the preset standard in the control database, obtain control item measures that match the production data, and use the control item measures as the initial control item measures for the next acquisition cycle.

[0037] Compared with the prior art, the beneficial effects of the present invention are: Of course, the following are the detailed beneficial effects of the full-process intelligent control system in the production of corrugated packaging lines:

[0038] 1. By real-time monitoring of data from each production point, the system can quickly detect and resolve abnormal situations in the production process, reduce downtime and production delays, and thus significantly improve production efficiency. Through high-frequency and low-frequency prediction models, the system can predict potential equipment failures in advance, perform preventive maintenance, avoid production interruptions due to equipment failures, and further improve production efficiency. The system can detect the operating status of the equipment in real time, promptly detect and deal with potential safety hazards, and reduce the occurrence of safety accidents. Through the data early warning module, the system can issue an alarm before a dangerous situation occurs, reminding operators to take corresponding measures to ensure production safety.

[0039] 2. Since the training data of low-frequency abnormal production monitoring indicators (such as abnormal bonding speed and abnormal cutting speed) are difficult to obtain, it is difficult to train and learn efficient and accurate prediction models. However, high-frequency abnormal production monitoring indicators (such as abnormal corrugated paper temperature and abnormal folding accuracy) have sufficient training data sets and can effectively learn efficient and accurate prediction models. Therefore, the present invention first uses high-frequency abnormal production monitoring indicators to train a large sample model, and then studies the cumulative number of times each high-frequency abnormal production monitoring indicator and each low-frequency abnormal production monitoring indicator appear in the same historical acquisition cycle. The model parameters of the large sample model are used as the initial model parameters of the low-frequency indicator estimation model, and the low-frequency indicator estimation model is continued to be trained on the training set of the low-frequency indicator estimation model, thereby constructing an efficient and accurate low-frequency indicator estimation model and avoiding the shortcomings of insufficient samples and information.

[0040] Through these detailed beneficial effects, the application of the full-process intelligent control system in the whole-line production of corrugated packaging not only improves production efficiency and product quality, but also reduces production costs, enhances production flexibility and safety, and brings all-round improvement to the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a schematic diagram of a full-process intelligent control system for the production of a corrugated packaging line according to an embodiment of the present application. DETAILED DESCRIPTION

[0042] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0043] like Figure 1 As shown, a full-process intelligent control system for the production of a corrugated packaging line includes a monitoring center, and the monitoring center is communicatively connected to a data acquisition module, a real-time monitoring module, a data analysis module, a high-frequency estimation module, a data processing module, a low-frequency estimation module and a data early warning module;

[0044] The data acquisition module is used to obtain process information of the entire corrugated packaging production line, set corrugated packaging monitoring points according to the process information, and obtain production data of each point;

[0045] The real-time monitoring module is used to monitor and control the production data of each point in real time;

[0046] The data analysis module is used to divide the production monitoring indicators of each process subsequence into high-frequency abnormal production monitoring indicators or low-frequency abnormal production monitoring indicators;

[0047] The high-frequency estimation module is used to construct a high-frequency indicator estimation model to obtain first estimated production data;

[0048] The data processing module is used to obtain reference production monitoring indicators of low-frequency abnormal production monitoring indicators;

[0049] The low-frequency estimation module is used to construct a low-frequency index estimation model to obtain second estimated production data;

[0050] The data early warning module is used to generate third estimated production data according to the first estimated production data and the second estimated production data, and to obtain initial control item measures for the next collection cycle according to the third estimated production data.

[0051] It should be further explained that, in the specific implementation process, the data acquisition module obtains the process information of the whole production line of corrugated packaging, sets the corrugated packaging monitoring points according to the process information, and obtains the production data of each point, including:

[0052] Extracting process unit characteristics of each corrugated packaging production equipment from the process information, and dividing the entire corrugated packaging production line into a number of process sub-sequences according to the process unit characteristics;

[0053] Set up corrugated packaging monitoring points in each process subsequence, and use data retrieval to obtain production monitoring indicators of each corrugated packaging monitoring point according to the functional characteristics in the process unit characteristics of the corresponding process subsequence;

[0054] The corrugated packaging monitoring point acquires production data in real time according to the production monitoring index, marks the acquisition time, and sets the acquisition cycle.

[0055] It should be further explained that, in the specific implementation process, the process of the real-time monitoring module monitoring and controlling the production data of each point in real time includes:

[0056] The production data of each corrugated packaging monitoring point is obtained, the numerical time series sequence corresponding to each type of production monitoring indicator in the production data is extracted, the threshold interval of each type of production monitoring indicator in each corrugated packaging monitoring point is preset, and the numerical time series sequence corresponding to each type of production monitoring indicator in the production data is compared with the corresponding threshold interval. If there is a numerical time series sequence corresponding to a production monitoring indicator that is not located in the corresponding threshold interval, the production monitoring indicator is marked as an abnormal production monitoring indicator, and an abnormal alarm signal of the process subsequence is generated according to the abnormal production monitoring indicator. The monitoring center arranges relevant personnel to detect and analyze the process subsequence according to the abnormal alarm signal, and obtains corresponding control item measures. The control item measures include adjusting equipment parameters, production line operation parameters, etc. A control database is pre-built, and the production data is matched with the control item measures and stored in the control database.

[0057] It should be further explained that, in the specific implementation process, the process in which the data analysis module divides the production monitoring indicators of each process subsequence into high-frequency abnormal production monitoring indicators or low-frequency abnormal production monitoring indicators includes:

[0058] Obtain all historical collection cycles in which each process subsequence generates an abnormal alarm signal, mark the historical collection cycle as an abnormal historical collection cycle, extract each abnormal production monitoring indicator in the abnormal historical collection cycle for statistical analysis, obtain the number of occurrences of each abnormal production monitoring indicator, preset an occurrence threshold, mark the abnormal production monitoring indicator whose occurrence number is greater than the occurrence threshold as a high-frequency abnormal production monitoring indicator, and mark the abnormal production monitoring indicator whose occurrence number is less than or equal to the occurrence threshold as a low-frequency abnormal production monitoring indicator.

[0059] It should be further explained that, in a specific implementation process, the high-frequency estimation module constructs a high-frequency indicator estimation model, and the process of obtaining the first estimated production data includes:

[0060] Obtain the production data within the historical collection period to which each high-frequency abnormal production monitoring indicator of each process subsequence belongs, use the production data as a training set and a test set, build a high-frequency indicator prediction model based on deep learning, input the training set into the high-frequency indicator prediction model for training until the loss function training is stable, save the model parameters, test the high-frequency indicator prediction model through the test set until it meets the preset requirements, and output the high-frequency indicator prediction model;

[0061] The estimated production data of the next acquisition cycle of each process subsequence is output according to the high-frequency indicator estimation model, and the estimated production data is marked as the first estimated production data.

[0062] It should be further explained that, in the specific implementation process, the process of the data processing module obtaining the reference production monitoring index of the low-frequency abnormal production monitoring index includes:

[0063] Obtain the historical collection period to which each low-frequency abnormal production monitoring indicator of each process subsequence belongs, extract the high-frequency abnormal production monitoring indicator within the historical collection period, perform statistical analysis on the high-frequency abnormal production monitoring indicator within the historical collection period, and obtain the cumulative number of times each high-frequency abnormal production monitoring indicator and each low-frequency abnormal production monitoring indicator appear in the same historical collection period;

[0064] The high-frequency abnormal production monitoring indicator that appears the highest cumulative number of times in the same historical collection period as the low-frequency abnormal production monitoring indicator shall be marked as the reference production monitoring indicator of the low-frequency abnormal production monitoring indicator.

[0065] It should be further explained that, in a specific implementation process, the low-frequency estimation module constructs a low-frequency index estimation model, and the process of obtaining the second estimated production data includes:

[0066] Obtain production data within a historical collection period to which a reference production monitoring indicator of a low-frequency abnormal production monitoring indicator belongs, use the production data as training data, build a large sample model based on deep learning, train the large sample model through the training data, obtain the trained large sample model, and obtain model parameters in the large sample model;

[0067] A low-frequency indicator prediction model is constructed based on deep learning, and the model parameters in the large sample model are used as the initial model parameters of the low-frequency indicator prediction model. The production data within the historical collection period to which the low-frequency abnormal production monitoring indicator belongs is obtained, and the production data is used as the training data of the low-frequency indicator prediction model to train the low-frequency indicator prediction model, and the trained low-frequency indicator prediction model is output;

[0068] Output the estimated production data of the next acquisition cycle of each process subsequence according to the low-frequency indicator estimation model, and mark the estimated production data as second estimated production data.

[0069] Since the training data of low-frequency abnormal production monitoring indicators (such as abnormal bonding speed and abnormal cutting speed) are difficult to obtain, it is difficult to train and learn efficient and accurate prediction models, but high-frequency abnormal production monitoring indicators (such as abnormal corrugated paper temperature and abnormal folding accuracy) have sufficient training data sets, which can effectively learn efficient and accurate prediction models. Therefore, the present invention first uses high-frequency abnormal production monitoring indicators to train a large sample model, and then studies the cumulative number of times each high-frequency abnormal production monitoring indicator and each low-frequency abnormal production monitoring indicator appear in the same historical acquisition cycle, and uses the model parameters of the large sample model as the initial model parameters of the low-frequency indicator estimation model, and continues to train the low-frequency indicator estimation model on the training set of the low-frequency indicator estimation model, thereby constructing an efficient and accurate low-frequency indicator estimation model and avoiding the shortcomings of insufficient samples and information.

[0070] It should be further explained that, in a specific implementation process, the process of the data early warning module generating the third estimated production data according to the first estimated production data and the second estimated production data includes:

[0071] For each type of production monitoring indicator in the first estimated production data in the next collection cycle, obtain the production monitoring indicator marked as a low-frequency abnormal production monitoring indicator in the historical collection cycle among the various types of production monitoring indicators, and remove the production monitoring indicator from the first estimated production data;

[0072] For each type of production monitoring indicator in the second estimated production data in the next collection cycle, obtain the production monitoring indicator marked as a high-frequency abnormal production monitoring indicator in the historical collection cycle among the various types of production monitoring indicators, and remove the production monitoring indicator from the second estimated production data;

[0073] Subsequently, the first estimated production data and the second estimated production data are merged to generate third estimated production data, and the estimated numerical time series of various types of production monitoring indicators in the next collection cycle are output based on the third estimated production data.

[0074] It should be further explained that, in the specific implementation process, the process of the data early warning module obtaining the initial control item measures for the next collection cycle according to the third estimated production data includes:

[0075] Compare the estimated numerical time series sequence of each type of production monitoring indicator in the next acquisition cycle with the corresponding threshold interval. If there is a production monitoring indicator whose corresponding estimated numerical time series sequence is not within the corresponding threshold interval, input the estimated numerical time series sequence of each type of production monitoring indicator into the control database for data retrieval, obtain production data whose similarity meets the preset standard in the control database, obtain control item measures that match the production data, and use the control item measures as the initial control item measures for the next acquisition cycle.

[0076] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A full-process intelligent control system for the production of corrugated packaging lines, characterized in that: It includes a monitoring center, which is communicatively connected with a data acquisition module, a real-time monitoring module, a data analysis module, a high-frequency estimation module, a data processing module, a low-frequency estimation module and a data early warning module; The data acquisition module is used to obtain process information of the entire corrugated packaging production line, set corrugated packaging monitoring points according to the process information, and obtain production data of each point; The real-time monitoring module is used to monitor and control the production data of each point in real time; The data analysis module is used to divide the production monitoring indicators of each process subsequence into high-frequency abnormal production monitoring indicators or low-frequency abnormal production monitoring indicators; The high-frequency estimation module is used to construct a high-frequency indicator estimation model to obtain first estimated production data; The data processing module is used to obtain reference production monitoring indicators of low-frequency abnormal production monitoring indicators; The low-frequency estimation module is used to construct a low-frequency index estimation model to obtain second estimated production data; The data early warning module is used to generate third estimated production data according to the first estimated production data and the second estimated production data, and to obtain initial control item measures for the next collection cycle according to the third estimated production data.

2. The full-process intelligent control system for the production of corrugated packaging lines according to claim 1 is characterized in that: The data acquisition module obtains the process information of the whole production line of corrugated packaging, sets the corrugated packaging monitoring points according to the process information, and obtains the production data of each point, including: Extracting process unit characteristics of each corrugated packaging production equipment from the process information, and dividing the entire corrugated packaging production line into a number of process sub-sequences according to the process unit characteristics; Set up corrugated packaging monitoring points in each process subsequence, and use data retrieval to obtain production monitoring indicators of each corrugated packaging monitoring point according to the functional characteristics in the process unit characteristics of the corresponding process subsequence; The corrugated packaging monitoring point acquires production data in real time according to the production monitoring index, marks the acquisition time, and sets the acquisition cycle.

3. The full-process intelligent control system for the production of corrugated packaging lines according to claim 2 is characterized in that: The process of real-time monitoring and controlling the production data of each point by the real-time monitoring module includes: Acquire the production data of each corrugated packaging monitoring point, extract various types of production monitoring indicators in the production data, preset the threshold range of each type of production monitoring indicators in each corrugated packaging monitoring point, compare each type of production monitoring indicator in the production data with the corresponding threshold range, if there is a production monitoring indicator that is not in the corresponding threshold range, then mark the production monitoring indicator as an abnormal production monitoring indicator, and generate an abnormal alarm signal of the process subsequence according to the abnormal production monitoring indicator. The monitoring center arranges relevant personnel to detect and analyze the process subsequence according to the abnormal alarm signal, obtains the corresponding control item measures, pre-builds a control database, and matches the production data with the control item measures and stores them in the control database.

4. The full-process intelligent control system for the production of a corrugated packaging line according to claim 3 is characterized in that: The process of the data analysis module dividing the production monitoring indicators of each process subsequence into high-frequency abnormal production monitoring indicators or low-frequency abnormal production monitoring indicators includes: Obtain all historical collection cycles in which each process subsequence generates an abnormal alarm signal, mark the historical collection cycle as an abnormal historical collection cycle, extract each abnormal production monitoring indicator in the abnormal historical collection cycle for statistical analysis, obtain the number of occurrences of each abnormal production monitoring indicator, preset an occurrence threshold, mark the abnormal production monitoring indicator whose occurrence number is greater than the occurrence threshold as a high-frequency abnormal production monitoring indicator, and mark the abnormal production monitoring indicator whose occurrence number is less than or equal to the occurrence threshold as a low-frequency abnormal production monitoring indicator.

5. The full-process intelligent control system for the production of a corrugated packaging line according to claim 4 is characterized in that: The high-frequency estimation module constructs a high-frequency index estimation model, and the process of obtaining the first estimated production data includes: Obtain the production data within the historical collection period to which each high-frequency abnormal production monitoring indicator of each process subsequence belongs, use the production data as a training set and a test set, build a high-frequency indicator prediction model based on deep learning, input the training set into the high-frequency indicator prediction model for training until the loss function training is stable, save the model parameters, test the high-frequency indicator prediction model through the test set until it meets the preset requirements, and output the high-frequency indicator prediction model; The estimated production data of the next acquisition cycle of each process subsequence is output according to the high-frequency indicator estimation model, and the estimated production data is marked as the first estimated production data.

6. The full-process intelligent control system for the production of a corrugated packaging line according to claim 5 is characterized in that: The process of the data processing module obtaining the reference production monitoring index of the low-frequency abnormal production monitoring index includes: Obtain the historical collection period to which each low-frequency abnormal production monitoring indicator of each process subsequence belongs, extract the high-frequency abnormal production monitoring indicator within the historical collection period, perform statistical analysis on the high-frequency abnormal production monitoring indicator within the historical collection period, and obtain the cumulative number of times each high-frequency abnormal production monitoring indicator and each low-frequency abnormal production monitoring indicator appear in the same historical collection period; The high-frequency abnormal production monitoring indicator that appears the highest cumulative number of times in the same historical collection period as the low-frequency abnormal production monitoring indicator shall be marked as the reference production monitoring indicator of the low-frequency abnormal production monitoring indicator.

7. The full-process intelligent control system for the production of a corrugated packaging line according to claim 6 is characterized in that: The low-frequency estimation module constructs a low-frequency index estimation model, and the process of obtaining the second estimated production data includes: Obtain production data within a historical collection period to which a reference production monitoring indicator of a low-frequency abnormal production monitoring indicator belongs, use the production data as training data, build a large sample model based on deep learning, train the large sample model through the training data, obtain the trained large sample model, and obtain model parameters in the large sample model; A low-frequency indicator prediction model is constructed based on deep learning, and the model parameters in the large sample model are used as the initial model parameters of the low-frequency indicator prediction model. The production data within the historical collection period to which the low-frequency abnormal production monitoring indicator belongs is obtained, and the production data is used as the training data of the low-frequency indicator prediction model to train the low-frequency indicator prediction model, and the trained low-frequency indicator prediction model is output; Output the estimated production data of the next acquisition cycle of each process subsequence according to the low-frequency indicator estimation model, and mark the estimated production data as second estimated production data.

8. The full-process intelligent control system for the production of a corrugated packaging line according to claim 7 is characterized in that: The process of the data early warning module generating the third estimated production data according to the first estimated production data and the second estimated production data includes: For each type of production monitoring indicator in the first estimated production data in the next collection cycle, obtain the production monitoring indicator marked as a low-frequency abnormal production monitoring indicator in the historical collection cycle among the various types of production monitoring indicators, and remove the production monitoring indicator from the first estimated production data; For each type of production monitoring indicator in the second estimated production data in the next collection cycle, obtain the production monitoring indicator marked as a high-frequency abnormal production monitoring indicator in the historical collection cycle among the various types of production monitoring indicators, and remove the production monitoring indicator from the second estimated production data; Subsequently, the first estimated production data and the second estimated production data are merged to generate third estimated production data, and the estimated numerical time series of various types of production monitoring indicators in the next collection cycle are output based on the third estimated production data.

9. The full-process intelligent control system for the production of a corrugated packaging line according to claim 8, characterized in that: The process of the data early warning module obtaining the initial control item measures for the next collection cycle according to the third estimated production data includes: Compare the estimated numerical time series sequence of each type of production monitoring indicator in the next acquisition cycle with the corresponding threshold interval. If there is a production monitoring indicator whose corresponding estimated numerical time series sequence is not within the corresponding threshold interval, input the estimated numerical time series sequence of each type of production monitoring indicator into the control database for data retrieval, obtain production data whose similarity meets the preset standard in the control database, obtain control item measures that match the production data, and use the control item measures as the initial control item measures for the next acquisition cycle.

Images