AI-driven intelligent optimization method and system for paper production process

Abstract

The present application provides an AI-driven intelligent optimization method and system for paper production processes. By constructing a basic paper production process characterization model that includes basic process knowledge units and real-time change knowledge units, the fixed process information and real-time change information of paper production are effectively integrated. By determining sample learning data, extracting and analyzing the training production process data of each production cycle, the accuracy and adaptability of the basic paper production process characterization model are further improved. The basic process knowledge units and real-time change knowledge units are updated using basic process vector information and real-time change vector information to generate a target paper production process characterization model, which can accurately reflect the fixed processes and dynamic changes of the production links in the real-time change production process. Based on the target paper production process characterization model, adaptive intelligent optimization of the target paper production process is achieved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and more specifically, to an AI-driven intelligent optimization method and system for paper production processes. Background Art

[0002] In the papermaking industry, the complexity and variability of the production process have always been the key factors affecting production efficiency and quality. Traditional production optimization methods often rely on manual experience and fixed rules, which are difficult to adapt to the ever-changing production environment and market demand. Summary of the invention

[0003] In view of this, the purpose of this application is to provide an AI-driven intelligent optimization method and system for paper production process.

[0004] According to the first aspect of the present application, there is provided an AI-driven intelligent optimization method for a papermaking production process, which is applied to a papermaking production optimization system, and the method comprises:

[0005] Determine a basic papermaking production process characterization model, wherein the basic papermaking production process characterization model includes a basic process knowledge unit and a real-time change knowledge unit;

[0006] Determine sample learning data, wherein the sample learning data includes training production process data of multiple production cycles, and the training production process data of each production cycle includes multiple training production links and relationship characteristics between the training production links;

[0007] According to the basic process knowledge unit, basic process vector information of each training production link in the training production process data of each production cycle is obtained respectively; according to the real-time change knowledge unit, real-time change vector information of each training production link in the training production process data of each production cycle is obtained respectively;

[0008] The basic process knowledge unit is updated based on the basic process vector information, and the real-time change knowledge unit is updated based on the basic process vector information and the real-time change vector information to generate a target papermaking production process representation model, wherein the basic process knowledge unit in the target papermaking production process representation model has the same basic process vectors obtained for any production link in the real-time change production process in multiple production cycles, and the real-time change knowledge unit in the target papermaking production process representation model has different real-time change vectors obtained for any production link in the real-time change production process in different production cycles;

[0009] Based on the target papermaking production process characterization model, the target papermaking production process is adaptively and intelligently optimized.

[0010] In a possible implementation manner of the first aspect, the step of updating the basic process knowledge unit based on the basic process vector information includes:

[0011] Based on the basic process vector information of each training production link in the training production process data of each production cycle generated by the basic process knowledge unit, a basic process vector set based on the production cycle of any training production link is obtained, and a scrambled process vector set is generated by scrambling the basic process vector set;

[0012] Calculate a first performance evaluation index related to the basic process knowledge unit based on the basic process vector set; the first performance evaluation index includes: the difference between the basic process vector set and the scrambled process vector set of any training production link, the difference between the basic process vector sets of different training production links, and the difference between the basic process vector sets and the scrambled process vector sets of different training production links;

[0013] The parameter information in the basic process knowledge unit is updated based on the first performance evaluation index.

[0014] In a possible implementation manner of the first aspect, the basic process vector information of each training production link in the training production process data of each production cycle generated based on the basic process knowledge unit, the steps of obtaining a basic process vector set based on the production cycle for any one of the training production links, and scrambling the basic process vector set to generate a scrambled process vector set include:

[0015] Based on the basic process vector information of each training production link in the training production process data of each production cycle generated by the basic process knowledge unit, a basic process vector set based on the production cycle of any training production link is obtained in multiple rounds, and a scrambled process vector set is generated by scrambling the basic process vector set;

[0016] Then the first performance evaluation index also includes: the difference between the basic process vector sets of any training production link generated in multiple rounds.

[0017] In a possible implementation manner of the first aspect, the basic process vector information of each training production link in the training production process data of each production cycle generated based on the basic process knowledge unit, the steps of obtaining a basic process vector set based on the production cycle for any one of the training production links, and scrambling the basic process vector set to generate a scrambled process vector set include:

[0018] The basic process vector information of each production cycle of any training production link is formed into a set based on the production cycle sequence, and the basic process vector information of multiple production cycles is randomly selected from the set to form the basic process vector set based on the production cycle of any training production link;

[0019] The basic process vector information of multiple production cycles in the basic process vector set is scrambled to generate the scrambled process vector set.

[0020] In a possible implementation manner of the first aspect, the step of updating the real-time change knowledge unit based on the basic process vector information and the real-time change vector information includes:

[0021] Based on the basic process vector information of each training production link in the training production process data of each production cycle generated by the basic process knowledge unit, and the real-time change vector information of each training production link of the corresponding production cycle, the global integrated vector information of any one of the training production links in each production cycle is obtained;

[0022] Calculating a second performance evaluation index based on the global integrated vector information of the training production link in each production cycle, the second performance evaluation index including: the difference between the global integrated vector information of any training production link and its associated production link, and the difference between any training production link and its unassociated production link;

[0023] The parameter information in the real-time changing knowledge unit is updated based on the second performance evaluation indicator.

[0024] In a possible implementation manner of the first aspect, the step of obtaining global integrated vector information of any one of the training production links in each production cycle includes:

[0025] The basic process vector information and the real-time change vector information of any training production link in any production cycle are integrated to generate the global integrated vector information of any training production link in any production cycle.

[0026] In a possible implementation manner of the first aspect, the basic papermaking production process characterization model further includes an observation unit; then the method further includes:

[0027] The observation unit determines whether any basic process vector information and real-time change vector information belong to the same training production link based on the basic process vector information and real-time change vector information of each training production link in the training production process data of each production cycle;

[0028] Based on the observation result of the observation unit, the basic process knowledge unit and the real-time change knowledge unit are updated.

[0029] In a possible implementation manner of the first aspect, the step of updating the basic process knowledge unit and the real-time change knowledge unit based on the observation result of the observation unit includes:

[0030] A third performance evaluation index is calculated based on the observation result of the observation unit, and the third performance evaluation index includes: the observation result of the observation unit based on the integrated vector information of the basic process vector information of any training production link and its real-time change vector information, and the observation result of the observation unit based on the integrated vector information of the basic process vector information of any training production link and the real-time change vector information of other training production links;

[0031] The parameter information of the basic process knowledge unit and the real-time change knowledge unit is updated based on the third performance evaluation index.

[0032] In a possible implementation of the first aspect, the step of adaptively and intelligently optimizing the target papermaking production process based on the target papermaking production process characterization model includes:

[0033] Utilize the sensor network and data acquisition system deployed at the production site to collect data from each production link in the target paper production process in real time;

[0034] The production link data is input into the target papermaking production process characterization model, and the basic process knowledge unit and real-time change knowledge unit of the target papermaking production process characterization model are used to perform real-time analysis on the production process. The basic process vector information and real-time change vector information output by the target papermaking production process characterization model are used to evaluate whether the currently implemented optimization strategy is effective. If the currently implemented optimization strategy is invalid, the currently implemented optimization strategy is adjusted.

[0035] According to the second aspect of the present application, a papermaking production optimization system is provided, which includes a processor and a readable storage medium, wherein the readable storage medium stores a program, which, when executed by the processor, implements the aforementioned AI-driven intelligent optimization method for papermaking production processes.

[0036] According to the third aspect of the present application, a computer-readable storage medium is provided, in which computer-executable instructions are stored. When it is monitored that the computer-executable instructions are executed, the aforementioned AI-driven intelligent optimization method of the papermaking production process is implemented.

[0037] According to any of the above aspects, the embodiment of the present application effectively integrates the fixed process information and real-time change information of papermaking production by constructing a basic papermaking production process characterization model including basic process knowledge units and real-time change knowledge units. By determining the sample learning data, extracting and analyzing the training production process data of each production cycle, the accuracy and adaptability of the basic papermaking production process characterization model are further improved. The basic process knowledge unit and the real-time change knowledge unit are updated using the basic process vector information and the real-time change vector information, and a target papermaking production process characterization model is generated, which can accurately reflect the fixed process and dynamic changes of the production link in the real-time change production process, and provides a scientific basis for the optimization of the papermaking production process. Based on the target papermaking production process characterization model, the adaptive intelligent optimization of the target papermaking production process is realized, and the data changes of the production link can be monitored in real time. According to the basic process vector and real-time change vector information output by the model, the effectiveness of the currently implemented optimization strategy is automatically evaluated, and the optimization strategy is adjusted as needed to ensure the continuous improvement and optimization of the production process, which significantly improves the efficiency and quality of papermaking production, reduces production costs, and reduces human intervention and errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0039] Figure 1 A schematic diagram of a process flow of an AI-driven intelligent optimization method for a papermaking production process provided in an embodiment of the present application is shown;

[0040] Figure 2 A schematic diagram of the component structure of a papermaking production optimization system for implementing the above-mentioned AI-driven intelligent optimization method for papermaking production processes provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0041] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only 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 ordinary technicians in the field without creative work should fall within the scope of protection of the present application.

[0042] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0043] Figure 1 The flowchart of the AI-driven intelligent optimization method for paper production process provided by the embodiment of the present application is shown. It should be understood that in other embodiments, the order of some steps in the AI-driven intelligent optimization method for paper production process can be interchanged according to actual needs, or some steps can be omitted or deleted. The detailed steps of the AI-driven intelligent optimization method for paper production process are described as follows.

[0044] Step S110, determining a basic papermaking production process representation model, wherein the basic papermaking production process representation model includes a basic process knowledge unit and a real-time change knowledge unit.

[0045] In this embodiment, the server first determines a basic paper production process representation model. This basic paper production process representation model is the basis for understanding and optimizing the paper production process, and includes two main parts: basic process knowledge unit and real-time change knowledge unit. The basic process knowledge unit stores fixed process information in paper production, such as the basic manufacturing process of paper, the ratio of raw materials, etc. The real-time change knowledge unit is used to capture dynamic changes in the production process, such as changes in environmental humidity and machine status.

[0046] Step S120, determining sample learning data, wherein the sample learning data includes training production process data of multiple production cycles, and the training production process data of each production cycle includes multiple training production links and relationship features between the training production links.

[0047] Next, the server selected a set of sample learning data from past production records. These sample learning data include multiple complete production cycles, each of which has detailed training production process data. These training production process data not only record the operating parameters of each production link, but also include the relationship characteristics between the links, such as how the adjustment of a certain link affects other links.

[0048] Step S130, based on the basic process knowledge unit, respectively obtain the basic process vector information of each training production link in the training production process data of each production cycle, and based on the real-time change knowledge unit, respectively obtain the real-time change vector information of each training production link in the training production process data of each production cycle.

[0049] In this embodiment, the server begins to use the basic process knowledge unit to extract the basic process vector information of each training production link in each production cycle from the sample learning data. These basic process vector information reflects the fixed process elements in each link. At the same time, the server also uses the real-time change knowledge unit to extract the real-time change vector information of the corresponding link. These real-time change vector information captures the dynamic changes in the production process.

[0050] Step S140, updating the basic process knowledge unit based on the basic process vector information, and updating the real-time change knowledge unit based on the basic process vector information and the real-time change vector information, to generate a target papermaking production process characterization model, wherein the basic process knowledge unit in the target papermaking production process characterization model has the same basic process vectors obtained for any production link in the real-time change production process in multiple production cycles, and the real-time change knowledge unit in the target papermaking production process characterization model has different real-time change vectors obtained for any production link in the real-time change production process in different production cycles.

[0051] With this vector information, the server begins to update the basic paper production process representation model. First, based on the extracted basic process vector information, the basic process knowledge unit is updated to ensure that the model can accurately reflect the basic process in paper production. Then, the server combines the basic process vector information and the real-time change vector information to update the real-time change knowledge unit so that it can more accurately capture the dynamic changes in the production process. This updated model is called the target paper production process representation model. In this target paper production process representation model, for any production link, no matter which production cycle, the basic process vector is the same, while the real-time change vector will vary with changes in production conditions.

[0052] Step S150, performing adaptive intelligent optimization on the target papermaking production process based on the target papermaking production process characterization model.

[0053] Finally, the server uses this target papermaking production process representation model to perform adaptive intelligent optimization of the actual production process. The server collects data from the production process in real time by connecting to the sensor network and data acquisition system at the production site. These data are input into the target papermaking production process representation model, and the basic process knowledge unit and real-time change knowledge unit of the target papermaking production process representation model jointly analyze the production process in real time. The server evaluates whether the currently implemented optimization strategy is effective based on the basic process vector and real-time change vector information output by the target papermaking production process representation model. If it is found that the optimization strategy does not achieve the expected effect, the server will adjust the strategy in a timely manner to ensure continuous improvement and optimization of the production process.

[0054] Based on the above steps, the embodiment of the present application effectively integrates the fixed process information and real-time change information of papermaking production by constructing a basic papermaking production process characterization model including basic process knowledge units and real-time change knowledge units. By determining the sample learning data, extracting and analyzing the training production process data of each production cycle, the accuracy and adaptability of the basic papermaking production process characterization model are further improved. The basic process knowledge unit and the real-time change knowledge unit are updated using the basic process vector information and the real-time change vector information, and a target papermaking production process characterization model is generated, which can accurately reflect the fixed process and dynamic changes of the production link in the real-time change production process, and provides a scientific basis for the optimization of the papermaking production process. Based on the target papermaking production process characterization model, the adaptive intelligent optimization of the target papermaking production process is realized, and the data changes of the production link can be monitored in real time. According to the basic process vector and real-time change vector information output by the model, the effectiveness of the currently implemented optimization strategy is automatically evaluated, and the optimization strategy is adjusted as needed to ensure the continuous improvement and optimization of the production process, which significantly improves the efficiency and quality of papermaking production, reduces production costs, and reduces human intervention and errors.

[0055] In a possible implementation, step S140 includes:

[0056] Step S141, based on the basic process vector information of each training production link in the training production process data of each production cycle generated by the basic process knowledge unit, obtain a basic process vector set based on the production cycle of any one of the training production links, and a scrambled process vector set generated by scrambling the basic process vector set.

[0057] In this embodiment, the server first uses the determined basic process knowledge unit to process each training production link in the training production process data of each production cycle. This basic process knowledge unit can identify and extract the basic process vector information in each production link. For example, in a certain production link of papermaking, the basic process vector may include key process parameters such as pulp ratio, paper thickness, and drying temperature.

[0058] For any training production link, the server aggregates the basic process vector information extracted in each production cycle to form a basic process vector set based on the production cycle. This basic process vector set reflects the consistency of the basic process of the production link in different production cycles.

[0059] Step S142: Calculate a first performance evaluation index related to the basic process knowledge unit based on the basic process vector set. The first performance evaluation index includes: the difference between the basic process vector set and the scrambled process vector set of any training production link, the difference between the basic process vector sets of different training production links, and the difference between the basic process vector sets and the scrambled process vector sets of different training production links.

[0060] In this embodiment, in order to evaluate the performance of the basic process knowledge unit, the server needs to perform a scrambling operation on the basic process vector set obtained in the previous step to generate a scrambled process vector set. The scrambling operation can disrupt the order of the basic process vectors in the original set, thereby forming a reference group for comparison.

[0061] Next, the server starts to calculate the first performance evaluation index related to the basic process knowledge unit. This first performance evaluation index includes three parts:

[0062] The difference between the basic process vector set and the scrambled process vector set of any training production link. This difference reflects the stability of the basic process vector in different production cycles. The smaller the difference, the more accurate the information extracted by the basic process knowledge unit.

[0063] The difference between the basic process vector sets of different training production links. This difference can help identify which production links have significant differences in basic processes, which helps optimize specific links in the production process.

[0064] The difference between the basic process vector set and the scrambled process vector set of different training production links. This indicator further verifies the effectiveness of the basic process knowledge unit. If the basic process vectors of different production links are significantly different from the scrambled vectors, it means that the knowledge unit can effectively distinguish the basic process characteristics of different links.

[0065] Step S143: updating parameter information in the basic process knowledge unit based on the first performance evaluation index.

[0066] Based on the first performance evaluation index calculated above, the server updates the parameter information in the basic process knowledge unit. If a certain performance index is not good, it means that there is a problem in the basic process knowledge unit when extracting or processing the basic process vector of the corresponding link, and it is necessary to optimize the performance of the knowledge unit by adjusting the parameters. For example, it may be necessary to adjust the recognition algorithm of the pulp ratio, or optimize the processing method of the drying temperature data.

[0067] Through the above steps, the server can continuously optimize the basic process knowledge unit so that it can more accurately reflect the basic process information in the paper production process, providing a solid foundation for subsequent production process optimization.

[0068] In a possible implementation, step S141 includes: based on the basic process vector information of each training production link in the training production process data of each production cycle generated by the basic process knowledge unit, obtaining a basic process vector set based on the production cycle for any one of the training production links in multiple rounds, and scrambling the basic process vector set to generate a scrambled process vector set.

[0069] Then the first performance evaluation index also includes: the difference between the basic process vector sets of any training production link generated in multiple rounds.

[0070] In this embodiment, the server first uses the basic process knowledge unit to process the training production process data collected from each production cycle. These training production process data include various links of papermaking production, such as pulp preparation, paper forming, drying, etc. For each link, the basic process knowledge unit can identify and extract key basic process vector information.

[0071] Taking the pulp preparation process as an example, the server extracts the basic process vectors of this process in each production cycle through the basic process knowledge unit. These basic process vectors may include key parameters such as the raw material ratio, concentration, and pH value of the pulp.

[0072] Next, the server will collect the basic process vector information extracted from each production cycle for any training production link, such as the paper forming link. In this way, the server forms a basic process vector set of the paper forming link based on the production cycle. This set reflects the consistency and stability of the basic process of the paper forming link in different production cycles.

[0073] In order to further evaluate the performance of the basic process knowledge unit, the server needs to scramble the basic process vector set obtained in the previous step. The scrambling operation is to scramble the order of the basic process vectors in the original set through an algorithm to generate a new scrambled process vector set. This scrambled set will be used as a comparison reference for subsequent performance evaluation index calculations.

[0074] In order to more comprehensively evaluate the stability and accuracy of the basic process knowledge unit, the server will perform multiple rounds of data extraction and set generation operations. In each round of operation, the server will reprocess the data and generate a new set of basic process vectors.

[0075] Subsequently, the server will compare the differences between the basic process vector sets of any training production link generated in multiple rounds. This difference reflects the consistency and stability of the basic process vector set in different rounds of data processing. If the difference is small, it means that the basic process knowledge unit has high repeatability and reliability when processing similar data.

[0076] In a possible implementation, step S141 includes: forming a set of basic process vector information of each production cycle of any training production link based on the production cycle sequence, randomly selecting basic process vector information of multiple production cycles from the set to form the basic process vector set of any training production link based on the production cycle. Scrambling the basic process vector information of multiple production cycles in the basic process vector set generates the scrambled process vector set.

[0077] In this embodiment, the server first selects a specific training production link, such as a paper pressing link. For this training production link, the server has extracted basic process vector information from the training production process data of multiple production cycles. These basic process vector information may include key parameters such as pressing pressure, pressing time, and pressing temperature.

[0078] Next, the server arranges these basic process vector information in the order of the production cycle to form an ordered set. In this way, the server can clearly see how the basic process of paper pressing changes in different production cycles.

[0079] In order to analyze the basic process vector information more deeply, the server randomly selects data from multiple production cycles from the above ordered set. These selected data constitute a new set, namely "the basic process vector set based on the production cycle of any training production link". This random selection method helps the server obtain more representative data samples, thereby more accurately evaluating the performance of the basic process knowledge unit.

[0080] After obtaining the basic process vector set based on the production cycle, the server further processes the set. It uses a scrambling algorithm to scramble the order of the basic process vector information in the set and generate a new "scrambled process vector set".

[0081] The purpose of the scrambling operation is to create a comparison baseline so that the effect of the basic process knowledge unit can be more accurately evaluated when the performance indicators are calculated later. By comparing the original basic process vector set and the scrambled set, the server can analyze the stability and consistency of the basic process in different production cycles.

[0082] Therefore, by organizing, selecting and scrambling the basic process vector information, it provides strong data support for the subsequent performance evaluation index calculation. These operations not only help the server to have a more comprehensive understanding of the basic process conditions in the papermaking production process, but also lay a solid foundation for subsequent optimization work.

[0083] In a possible implementation, step S140 includes:

[0084] Step S144, based on the basic process vector information of each training production link in the training production process data of each production cycle generated by the basic process knowledge unit, and the real-time change vector information of each training production link in the corresponding production cycle, obtain the global integrated vector information of any one of the training production links in each production cycle.

[0085] In this embodiment, the server starts to process the collected training production process data of each production cycle. The training production process data includes basic process vector information generated by the basic process knowledge unit and real-time change vector information captured by the real-time change knowledge unit.

[0086] Taking a specific training production link as an example, such as the paper drying link, the server first integrates the basic process vector information and real-time change vector information of this link in each production cycle. These basic process vector information and real-time change vector information may include drying temperature, drying time, environmental humidity changes, etc.

[0087] The server uses this data to generate a global integrated vector information for the paper drying process of each production cycle. This global integrated vector information comprehensively reflects all important process and real-time changing parameters of this process in a specific production cycle.

[0088] Step S145, calculate the second performance evaluation index based on the global integrated vector information of the training production link in each production cycle, the second performance evaluation index including: the difference between the global integrated vector information of any training production link and its associated production link, and the difference between the global integrated vector information of any training production link and its unassociated production link.

[0089] After obtaining the global integrated vector information, the server starts to calculate the second performance evaluation index, which is intended to evaluate the performance of the real-time change knowledge unit in capturing and reflecting the interrelationships between production links.

[0090] The server first compares the difference between the global integrated vector information of the paper drying process and its directly related production process (such as paper pressing process). This difference reflects the coordination and consistency of these related processes in terms of process and real-time changes.

[0091] The server then compares the differences between the global integrated vector information of the paper drying process and its indirect production process (such as pulp preparation). This comparison helps to identify possible indirect effects or potential production bottlenecks.

[0092] Step S146: updating parameter information in the real-time changing knowledge unit based on the second performance evaluation indicator.

[0093] Based on the calculated second performance evaluation index, the server starts to update the parameter information in the real-time change knowledge unit. If it is found that the global integrated vector information of the paper drying link and its associated links is significantly different, it means that the real-time change knowledge unit may be insufficient in capturing the relationship between these links.

[0094] Therefore, the server will adjust the relevant parameters in the real-time change knowledge unit, such as increasing the sensitivity to changes in ambient humidity or optimizing the processing algorithm for real-time drying temperature data. These adjustments are aimed at improving the accuracy and response speed of the real-time change knowledge unit to better reflect the dynamic changes in the production process.

[0095] In a possible implementation, step S144 includes: integrating basic process vector information and real-time change vector information of any training production link in any production cycle to generate global integrated vector information of any training production link in any production cycle.

[0096] In this embodiment, first, the server selects a specific training production link, such as "paper cutting link", and determines a specific production cycle, such as "week 15". This selection is for subsequent data integration and analysis.

[0097] Next, the server retrieves the basic process vector information of the "paper cutting link" in the 15th week from the database. This information may include preset process parameters such as cutting speed, cutting angle, and blade temperature. At the same time, the server also obtains the real-time change vector information of the same link, which reflects the changes in process parameters caused by various factors (such as raw material quality fluctuations, machine wear, etc.) in the actual production process.

[0098] The server integrates the basic process vector information of the "paper cutting link" in week 15 with the real-time change vector information. Specifically, the server may perform a weighted average of the corresponding parameters in the two types of vector information, or integrate these parameters according to certain algorithmic rules (such as a dynamic adjustment algorithm based on real-time data).

[0099] Finally, the server generated the global integrated vector information of the "paper cutting link" in week 15. This global integrated vector information not only includes the original process settings, but also incorporates the dynamic changes in the actual production process, thus more comprehensively reflecting the real situation of this link in this production cycle.

[0100] Through such steps, the server can generate a global integrated vector information for each training production link in each production cycle, providing strong data support for subsequent performance analysis, process optimization, etc.

[0101] In a possible implementation, the basic paper production process characterization model further includes an observation unit. Then the method further includes:

[0102] In step A110, the observation unit determines whether any basic process vector information and real-time change vector information belong to the same training production link based on the basic process vector information and real-time change vector information of each training production link in the training production process data of each production cycle.

[0103] Step A120: updating the basic process knowledge unit and the real-time change knowledge unit based on the observation result of the observation unit.

[0104] In this embodiment, in the basic papermaking production process representation model, the observation unit plays an important role in monitoring and verifying data consistency. The server starts the observation unit and starts to check the training production process data of each production cycle.

[0105] The observation unit first obtains a set of data, which includes the basic process vector information and real-time change vector information of each training production link. Taking the "pulp preparation link" as an example, the basic process vector information may include preset process parameters such as pulp concentration and stirring speed, while the real-time change vector information reflects the real-time changes of temperature, humidity and pulp concentration during the actual stirring process.

[0106] The observation unit determines whether these data come from the same training production link by comparing the timestamps, equipment identification and other metadata in the basic process vector information and the real-time change vector information. If the timestamps and equipment identifications match, it means that these data do belong to the "pulp preparation link".

[0107] Once the observation unit completes the data verification and feeds the results back to the server, the server updates the basic process knowledge unit and the real-time change knowledge unit based on the feedback.

[0108] In the case of the “pulp preparation link”, if the observation unit finds that the real-time change vector information is highly consistent with the basic process vector information, it means that the process control of this link is stable, and the server may not make major adjustments to the basic process knowledge unit.

[0109] However, if the observation unit finds that the pulp concentration fluctuation in the real-time data exceeds the preset range, the server will update the basic process knowledge unit, such as adjusting the optimal concentration parameters of pulp preparation to better adapt to actual production conditions. At the same time, the server will also update the real-time change knowledge unit to enable it to more accurately capture and respond to real-time changes in the production process.

[0110] Through this update process, the basic paper production process characterization model can continuously optimize itself and improve production efficiency and product quality. The observation unit plays a key role in data verification and feedback, ensuring that the model can learn and adjust based on accurate data.

[0111] In a possible implementation, step A120 includes:

[0112] Step A121, calculate the third performance evaluation index based on the observation results of the observation unit, and the third performance evaluation index includes: the observation results of the observation unit based on the integrated vector information of the basic process vector information of any training production link and its real-time change vector information, and the observation results of the observation unit based on the integrated vector information of the basic process vector information of any training production link and the real-time change vector information of other training production links.

[0113] Step A122: updating parameter information of the basic process knowledge unit and the real-time change knowledge unit based on the third performance evaluation index.

[0114] In this embodiment, after receiving the observation result of the observation unit, the server starts to calculate the third performance evaluation index. The third performance evaluation index is designed to evaluate the collaborative working performance between the basic process knowledge unit and the real-time change knowledge unit.

[0115] First, the server considers the observation results of the observation unit on "the integrated vector information of the basic process vector information and its real-time change vector information of any training production link". Taking the "paper drying link" as an example, the observation unit will check the integrated data of the basic process vector information (such as the preset drying temperature and time) and the real-time change vector information (such as temperature fluctuations and humidity changes during the actual drying process) of this link. The server will analyze the consistency of this data, such as whether the actual drying temperature is consistent with the preset temperature, and whether the humidity change is within the controllable range.

[0116] Next, the server will also consider the observation results of the observation unit on "the integrated vector information of the basic process vector information of any training production link and the real-time change vector information of other training production links". This means that the server will check the data of the basic process vector information of "paper drying link" combined with the real-time change vector information of another link (such as "paper cutting link"). This cross-link data analysis helps to discover the potential impact between different production links, such as whether the temperature setting of the drying link affects the efficiency of the cutting link.

[0117] Based on these observations, the server will calculate a third performance evaluation indicator, which will reflect the synergy between the basic process and real-time changes, as well as the degree of mutual influence between different links.

[0118] After calculating the third performance evaluation index, the server will update the parameter information of the basic process knowledge unit and the real-time change knowledge unit according to this index.

[0119] If the third performance evaluation indicator shows that there is a large deviation between the basic process and the real-time changes in the "paper drying process", the server may adjust the drying temperature and time settings in the basic process knowledge unit to be closer to the actual production conditions. At the same time, the server may also optimize the algorithm in the real-time change knowledge unit to more accurately capture and respond to real-time changes in the drying process.

[0120] In addition, if the cross-link indicator analysis shows that the temperature setting of the "paper drying link" has a significant impact on the "paper cutting link", the server may add corresponding adjustment rules in the basic process knowledge unit to ensure coordinated operation between the two links.

[0121] Through these updates, the server can continuously improve the adaptability and accuracy of basic process knowledge units and real-time change knowledge units, thereby optimizing the entire paper production process.

[0122] In a possible implementation, step S150 includes:

[0123] Step S151, using the sensor network and data acquisition system deployed at the production site to collect data from each production link in the target papermaking production process in real time.

[0124] Step S152, input the production link data into the target papermaking production process characterization model, use the basic process knowledge unit and real-time change knowledge unit of the target papermaking production process characterization model to perform real-time analysis on the production process, and evaluate whether the currently implemented optimization strategy is effective through the basic process vector information and real-time change vector information output by the target papermaking production process characterization model. If the currently implemented optimization strategy is invalid, adjust the currently implemented optimization strategy.

[0125] In this embodiment, at the target paper production site, the server continuously collects data from each production link in real time through the sensor network and data acquisition system deployed on the production line. For example, in the pulp preparation link, the sensor will monitor the concentration, temperature, pH value and other key parameters of the pulp, and transmit these data to the server in real time. Similarly, in the subsequent links such as paper forming, pressing, drying, etc., there are corresponding sensors that continuously capture the production data of each link.

[0126] After receiving these real-time data, the server immediately inputs them into the constructed target paper production process representation model. This model has built-in basic process knowledge units and real-time change knowledge units, which can conduct in-depth analysis of these data.

[0127] Within the model, the basic process knowledge unit will make a preliminary assessment of the current production status based on the preset process parameters and historical data. The real-time change knowledge unit will dynamically analyze changes in the production process in combination with real-time data from the sensor network. For example, in the paper drying process, if the real-time data shows that the drying temperature has increased abnormally, the real-time change knowledge unit will quickly identify this change and analyze its possible impact on paper quality and production efficiency.

[0128] Through real-time analysis of the model, the server will output basic process vector information and real-time change vector information. This information not only reflects the status of the current production process, but also helps the server evaluate whether the currently implemented optimization strategy is effective.

[0129] If it is found that the currently implemented optimization strategies fail to achieve the expected results, such as the quality indicators of the paper do not meet the standards or the production efficiency is low, the server will adjust these strategies in time. For example, if the server finds that the temperature of the drying process is too high, resulting in a decrease in paper quality, it may reduce the drying temperature or adjust the drying time to ensure the quality of the paper.

[0130] This optimization method based on real-time data and model analysis enables the server to continuously and adaptively optimize the papermaking production process, thereby improving production efficiency, reducing energy consumption and ensuring product quality.

[0131] Further, Figure 2 FIG. 1 is a schematic diagram showing a hardware structure of a paper production optimization system 100 for implementing the method provided in the embodiment of the present application. Figure 2 As shown, the paper production optimization system 100 may include at least one processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 and a controller 108 for communication functions. It can be understood by those skilled in the art that Figure 2 The structure shown is only for illustration and does not limit the structure of the paper production optimization system 100. For example, the paper production optimization system 100 may also include Figure 2 More or fewer components as shown, or with Figure 2 Different configurations are shown.

[0132] The memory 104 can be used to store software programs and modules of application software, such as program instructions corresponding to the above-mentioned method embodiments in the embodiments of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, realizing the above-mentioned AI-driven intelligent optimization method for paper production processes. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the paper production optimization system 100 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0133] The transmission device 106 is used to obtain or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the paper production optimization system 100. In one example, the transmission device 106 includes a network adapter, which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency module, which is used to communicate with the Internet wirelessly.

[0134] It should be noted that the order of the above embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. The above describes specific embodiments of the embodiments of the present application. Other embodiments are within the scope of the attached claims. In some cases, the exceptions or steps recorded in the claims can be performed according to an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or can be advantageous.

[0135] Each embodiment in the embodiments of the present application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the above different embodiments, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0136] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program. The above program may be stored in a computer-readable storage medium, and the above storage medium may be a read-only memory, a disk or an optical disk, etc.

Claims

1. An AI-driven intelligent optimization method for paper production process, characterized in that: Applied to a paper production optimization system, the method comprises: Determine a basic papermaking production process characterization model, wherein the basic papermaking production process characterization model includes a basic process knowledge unit and a real-time change knowledge unit; Determine sample learning data, wherein the sample learning data includes training production process data of multiple production cycles, and the training production process data of each production cycle includes multiple training production links and relationship characteristics between the training production links; According to the basic process knowledge unit, basic process vector information of each training production link in the training production process data of each production cycle is obtained respectively; according to the real-time change knowledge unit, real-time change vector information of each training production link in the training production process data of each production cycle is obtained respectively; The basic process knowledge unit is updated based on the basic process vector information, and the real-time change knowledge unit is updated based on the basic process vector information and the real-time change vector information to generate a target papermaking production process representation model, wherein the basic process knowledge unit in the target papermaking production process representation model has the same basic process vectors obtained for any production link in the real-time change production process in multiple production cycles, and the real-time change knowledge unit in the target papermaking production process representation model has different real-time change vectors obtained for any production link in the real-time change production process in different production cycles; Based on the target papermaking production process characterization model, the target papermaking production process is adaptively and intelligently optimized; The step of updating the basic process knowledge unit based on the basic process vector information includes: Based on the basic process vector information of each training production link in the training production process data of each production cycle generated by the basic process knowledge unit, a basic process vector set based on the production cycle of any training production link is obtained, and a scrambled process vector set is generated by scrambling the basic process vector set; Calculate a first performance evaluation index related to the basic process knowledge unit based on the basic process vector set; the first performance evaluation index includes: the difference between the basic process vector set and the scrambled process vector set of any training production link, the difference between the basic process vector sets of different training production links, and the difference between the basic process vector sets and the scrambled process vector sets of different training production links; Update parameter information in the basic process knowledge unit based on the first performance evaluation index; The steps of obtaining basic process vector information of each training production link in the training production process data of each production cycle generated based on the basic process knowledge unit, obtaining a basic process vector set based on the production cycle of any training production link, and scrambling the basic process vector set to generate a scrambled process vector set include: Based on the basic process vector information of each training production link in the training production process data of each production cycle generated by the basic process knowledge unit, a basic process vector set based on the production cycle of any training production link is obtained in multiple rounds, and a scrambled process vector set is generated by scrambling the basic process vector set; Then the first performance evaluation index also includes: the difference between the basic process vector sets of any training production link generated in multiple rounds; The steps of obtaining basic process vector information of each training production link in the training production process data of each production cycle generated based on the basic process knowledge unit, obtaining a basic process vector set based on the production cycle of any training production link, and scrambling the basic process vector set to generate a scrambled process vector set include: The basic process vector information of each production cycle of any training production link is formed into a set based on the production cycle sequence, and the basic process vector information of multiple production cycles is randomly selected from the set to form the basic process vector set based on the production cycle of any training production link; Scrambling basic process vector information of multiple production cycles in the basic process vector set to generate the scrambled process vector set; The step of updating the real-time change knowledge unit based on the basic process vector information and the real-time change vector information includes: Based on the basic process vector information of each training production link in the training production process data of each production cycle generated by the basic process knowledge unit, and the real-time change vector information of each training production link of the corresponding production cycle, the global integrated vector information of any one of the training production links in each production cycle is obtained; Calculating a second performance evaluation index based on the global integrated vector information of the training production link in each production cycle, the second performance evaluation index including: the difference between the global integrated vector information of any training production link and its associated production link, and the difference between any training production link and its unassociated production link; The parameter information in the real-time changing knowledge unit is updated based on the second performance evaluation indicator.

2. The AI-driven intelligent optimization method for papermaking production process according to claim 1 is characterized in that: The step of obtaining the global integrated vector information of any one of the training production links in each production cycle includes: The basic process vector information and the real-time change vector information of any training production link in any production cycle are integrated to generate the global integrated vector information of any training production link in any production cycle.

3. The AI-driven intelligent optimization method for papermaking production process according to claim 1 or 2, characterized in that: The basic papermaking production process characterization model also includes an observation unit; the method also includes: The observation unit determines whether any basic process vector information and real-time change vector information belong to the same training production link based on the basic process vector information and real-time change vector information of each training production link in the training production process data of each production cycle; Based on the observation result of the observation unit, the basic process knowledge unit and the real-time change knowledge unit are updated.

4. The AI-driven intelligent optimization method for paper production process according to claim 3 is characterized in that: The step of updating the basic process knowledge unit and the real-time change knowledge unit based on the observation result of the observation unit includes: A third performance evaluation index is calculated based on the observation result of the observation unit, and the third performance evaluation index includes: the observation result of the observation unit based on the integrated vector information of the basic process vector information of any training production link and its real-time change vector information, and the observation result of the observation unit based on the integrated vector information of the basic process vector information of any training production link and the real-time change vector information of other training production links; The parameter information of the basic process knowledge unit and the real-time change knowledge unit is updated based on the third performance evaluation index.

5. The AI-driven intelligent optimization method for papermaking production process according to claim 1 or 2, characterized in that: The step of adaptively and intelligently optimizing the target papermaking production process based on the target papermaking production process characterization model comprises: Utilize the sensor network and data acquisition system deployed at the production site to collect data from each production link in the target paper production process in real time; The production link data is input into the target papermaking production process characterization model, and the basic process knowledge unit and real-time change knowledge unit of the target papermaking production process characterization model are used to perform real-time analysis on the production process. The basic process vector information and real-time change vector information output by the target papermaking production process characterization model are used to evaluate whether the currently implemented optimization strategy is effective. If the currently implemented optimization strategy is invalid, the currently implemented optimization strategy is adjusted.

6. A paper production optimization system, characterized in that: The papermaking production optimization system includes a processor and a readable storage medium, wherein the readable storage medium stores a program, which, when executed by the processor, implements the AI-driven intelligent optimization method for papermaking production processes described in any one of claims 1 to 5.

Images