Artificial intelligence analysis method and system applied to paper machine press section data model

Abstract

The application provides an AI analysis method and system applied to a paper machine press part data model, a paper machine digital production technical field, first, cross-dimension behavior marks of a paper machine press part data parameter model are acquired, including line pressure, vacuum, vehicle speed and the like parameter adjustment and detection parameters such as water permeability and moisture content of a felt, output response and environmental interaction behavior marks, then, dynamic behavior inversion is carried out on the cross-dimension behavior marks, dynamic reference logic is generated based on the inversion result, then, the inversion result and the dynamic reference logic are closed-loop calibrated, nodes deviating from the reference are positioned and calibration instructions are fed back, finally, the AI analysis process is iteratively optimized according to the result after the closed-loop calibration, an iterative analysis report is integrated, optimization adjustment parameter suggestions are provided for the paper machine, and the adaptability of the paper machine parameters and the felt design is evaluated based on different working conditions, more accurate decision support is provided for felt optimization and parameter upgrading, and the production efficiency and product quality of the paper machine press part are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of paper machine digital production, in particular to an AI analysis method and system applied to a paper machine press section data model. BACKGROUND

[0002] In the production and operation process of the paper machine press section, the data model plays a crucial role in optimizing the production process, improving product quality and reducing production cost. However, the existing analysis method of the paper machine press section data model has many limitations.

[0003] On the one hand, the traditional method often focuses only on single-dimensional data, such as only focusing on parameter adjustment data or output response data, while ignoring the internal relationship and interaction between different dimensional data. The above isolated analysis method cannot comprehensively and accurately grasp the actual running condition of the paper machine press section, and it is difficult to find potential problems and optimization points.

[0004] On the other hand, the existing analysis method lacks dynamic and adaptability. The running environment of the paper machine press section is constantly changing, affected by many factors such as raw material quality, equipment state, process parameters, etc. But the traditional method usually uses fixed analysis mode and benchmark, which cannot dynamically adjust and calibrate according to real-time changing data and behavior, resulting in deviation between analysis results and actual situation. SUMMARY

[0005] In view of the above-mentioned problems, in combination with the first aspect of the present application, the embodiments of the present application provide an AI analysis method applied to a paper machine press section data model, which comprises:

[0006] Obtain the cross-dimensional behavior imprint of the paper machine press section data model, which contains parameter adjustment behavior imprint, output response behavior imprint and environment interaction behavior imprint, to obtain the cross-dimensional behavior imprint of the paper machine press section data model;

[0007] Perform dynamic behavior inversion on the cross-dimensional behavior imprint of the paper machine press section data model, mine the real-time correlation path, trigger conduction mechanism and synergistic action logic between different dimensional imprints, to obtain the dynamic behavior inversion result of the paper machine press section data model;

[0008] Generate dynamic benchmark logic based on the dynamic behavior inversion result of the paper machine press section data model, adjust the correlation adaptation condition, conduction response template and synergistic action boundary of the benchmark according to the real-time change of the inversion result, to obtain the dynamic benchmark logic of the paper machine press section data model;

[0009] The dynamic behavior inversion result of the paper machine press section data model is closed-loop calibrated with the dynamic benchmark logic of the paper machine press section data model, deviation of associated nodes, conduction nodes and synergistic nodes in the inversion result is located, node calibration instructions are generated and fed back to the inversion process, and a dynamic behavior inversion result after closed-loop calibration is obtained;

[0010] An AI analysis flow of the paper machine press section data model is iteratively optimized according to the dynamic behavior inversion result after closed-loop calibration, and an iterative analysis report of the paper machine press section data model is obtained by integrating the inversion result after iteration and the dynamic benchmark logic.

[0011] In still another aspect, the embodiments of the present application also provide an AI analysis system applied to a paper machine press section data model, characterized in that it comprises:

[0012] A processor; a machine readable storage medium for storing machine executable instructions of the processor; wherein the processor is configured to execute the machine executable instructions to perform the AI analysis method applied to the paper machine press section data model.

[0013] In still another aspect, the embodiments of the present application also provide a computer program product, which comprises machine executable instructions stored in a computer readable storage medium, and a processor of a computer device reads the machine executable instructions from the computer readable storage medium, and the processor executes the machine executable instructions to make the computer device execute the AI analysis method applied to the paper machine press section data model.

[0014] Based on the above aspects, by obtaining the cross-dimension behavior imprint of the paper machine press section data model, multiple dimensions such as parameter adjustment, output response and environment interaction are covered, then the cross-dimension behavior imprint is dynamically behavior-inverted, real-time associated paths, trigger conduction mechanisms and synergistic action logic among different dimension imprints are mined, the understanding of the dynamic behavior of the data model is more accurate and detailed, the dynamic benchmark logic is generated based on the dynamic behavior inversion result, and the associated adaptive conditions, conduction response templates and synergistic action boundaries of the benchmark can be adjusted according to the real-time changes of the inversion result, the dynamic adaptability of the benchmark is realized, which can better match the actual running changes of the paper machine press section, through the closed-loop calibration mechanism, the dynamic behavior inversion result is compared with the dynamic benchmark logic, the nodes deviating from the benchmark are located and calibration instructions are generated and fed back to the inversion process, the accuracy and reliability of the analysis are further improved, finally the AI analysis flow is iteratively optimized according to the result after closed-loop calibration, and an iterative analysis report is formed by integration, which provides a comprehensive, accurate and dynamically adaptive analysis basis for production optimization of the paper machine press section, and effectively improves the production efficiency and product quality of the paper machine press section. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is an execution flow diagram of the AI analysis method applied to the paper machine press section data model provided by an embodiment of the present application.

[0016] Figure 2 is a schematic diagram of exemplary hardware and software components of the AI analysis system applied to the paper machine press section data model provided by an embodiment of the present application. DETAILED DESCRIPTION

[0017] The present application will be described in detail below with reference to the accompanying drawings of the specification, Figure 1 is a flow diagram of the AI analysis method applied to the paper machine press section data model provided by an embodiment of the present application, which will be described in detail below.

[0018] Step S110: Obtain the cross-dimension behavior imprint of the paper machine press section data model, which includes the parameter adjustment behavior imprint, the output response behavior imprint, and the environment interaction behavior imprint, to obtain the cross-dimension behavior imprint of the paper machine press section data model.

[0019] In the actual operation process of the paper machine press section, the data model will generate various imprint data reflecting its running state and behavior. The parameter adjustment behavior imprint covers the relevant records of the adjustment of various operating parameters in the production process of the press section, such as the adjustment of the press roll pressure, the change of the vehicle speed, the control of the pulp concentration, and the like. The output response behavior imprint refers to the changes of various output indicators of the paper machine press section after parameter adjustment, such as the response data of the moisture content, thickness, and basis weight of the paper. The environment interaction behavior imprint includes the changes of the operating environment factors of the paper machine press section and the interaction information with the external environment, such as the fluctuations of the workshop temperature and humidity, the changes of the material characteristics, and the like.

[0020] In order to obtain the above-mentioned cross-dimension behavior imprint, real-time data collection needs to be performed through various sensors and data collection devices equipped in the paper machine press section. In the data collection process, for the case involving privacy sensitive data, such as the identity information of the operator, data desensitization technology needs to be used to anonymize the sensitive fields, and the information that can identify the individual is converted into a form that cannot be directly linked to a specific individual. At the same time, encryption transmission technology is used to encrypt the collected data during transmission to prevent the data from being illegally obtained and leaked in the transmission link. In addition, a data access control mechanism needs to be established to strictly manage the permissions of the personnel and systems accessing the above-mentioned behavior imprint data, so that only authorized personnel and systems can access the corresponding data, thereby realizing the privacy protection and leakage prevention of the privacy sensitive data.

[0021] Step S120: Dynamic behavior inversion is performed on the cross-dimension behavior imprint of the paper machine press section data model to mine real-time correlation paths, trigger conduction mechanisms and synergistic logic among different dimension imprints, and a dynamic behavior inversion result of the paper machine press section data model is obtained.

[0022] The dynamic behavior inversion process described above is a key link for in-depth understanding of the running mechanism of the paper machine press section data model. Through analysis of the cross-dimension behavior imprint, the internal relationship and action law between parameter adjustment, output response and environmental interaction are revealed.

[0023] Step S121: The parameter adjustment behavior imprint, output response behavior imprint and environmental interaction behavior imprint in the cross-dimension behavior imprint of the paper machine press section data model are synchronously split according to the running time sequence to obtain parameter adjustment behavior imprint segments, output response behavior imprint segments and environmental interaction behavior imprint segments corresponding to multiple time sequence segments.

[0024] In the continuous running process of the paper machine press section, each behavior imprint is constantly changing with time. In order to more accurately analyze the behavior characteristics in different time periods, it is necessary to synchronously split the above three behavior imprints according to the running time sequence. First, determine the time granularity of splitting, for example, a production shift, a production batch or a fixed time interval (such as every hour) can be used as a time sequence segment. Then, according to the determined time sequence segment, the parameter adjustment behavior imprint, the output response behavior imprint and the environmental interaction behavior imprint are respectively split into segments corresponding to each time sequence segment. In the splitting process, it is necessary to ensure that the time sequence segment division of the three behavior imprints is synchronous, that is, the three imprint segments corresponding to the same time sequence segment reflect the behavior in the same time period. For example, if the running time of a day is split into multiple time sequence segments according to every hour, then the parameter adjustment behavior imprint segment, the output response behavior imprint segment and the environmental interaction behavior imprint segment corresponding to each time sequence segment are the corresponding behavior records in that hour, respectively. Through the above synchronous splitting, the behavior characteristic analysis in different time sequence segments is more accurate and targeted.

[0025] Step S122: Dimensional behavior inversion is performed on the parameter adjustment behavior imprint segment corresponding to each time sequence segment to analyze the occurrence order, interaction path and linkage trigger relationship of different adjustment behaviors in the parameter adjustment behavior imprint segment, and a parameter adjustment dimensional behavior correlation logic is obtained.

[0026] For each parameter adjustment behavior imprint segment corresponding to a time sequence segment, behavior inversion within the dimension is needed. First, extract each independent adjustment behavior from the parameter adjustment behavior imprint segment, such as roll pressure adjustment behavior, speed adjustment behavior, and pulp concentration adjustment behavior. Then, determine the order of occurrence of these adjustment behaviors, that is, arrange each adjustment behavior according to time. Next, analyze the interaction path between different adjustment behaviors, for example, when the pulp concentration changes, the roll pressure may be adjusted, so there is an interaction path between the pulp concentration adjustment behavior and the roll pressure adjustment behavior. At the same time, the linkage triggering relationship between adjustment behaviors needs to be studied, that is, whether the occurrence of some adjustment behaviors will trigger a chain reaction of other adjustment behaviors. For example, the increase of speed may cause the adjustment of roll pressure and pulp concentration, in which case the speed adjustment behavior is the linkage triggering factor of roll pressure adjustment behavior and pulp concentration adjustment behavior. Through comprehensive analysis of the above occurrence order, interaction path and linkage triggering relationship, the behavior correlation logic within the parameter adjustment dimension is constructed, which clearly shows the internal relationship and action rule between parameter adjustment behaviors within the time sequence segment.

[0027] Step S123: Perform behavior inversion between the parameter adjustment behavior imprint segment and the output response behavior imprint segment corresponding to each time sequence segment, capture the conduction path, time correlation rule and intensity action relationship of parameter adjustment behavior triggering output response behavior, and obtain the behavior correlation path between the parameter and output dimensions.

[0028] Step S1231: Extract the key adjustment behavior and the time of occurrence in the parameter adjustment behavior imprint segment corresponding to each time sequence segment, record the type, implementation amplitude and duration of the key adjustment behavior, and obtain the parameter adjustment key behavior information.

[0029] In the parameter adjustment behavior imprint segment corresponding to each time sequence segment, there are various adjustment behaviors, but not all adjustment behaviors have a significant impact on the output response. Therefore, key adjustment behaviors need to be extracted. The determination of key adjustment behaviors can be based on the influence degree of adjustment behaviors on the main output indicators of the paper machine press section, for example, adjustment behaviors that have a greater impact on key quality indicators such as paper moisture content and thickness can be considered as key adjustment behaviors. For the extracted key adjustment behaviors, record their types, such as roll pressure adjustment, speed adjustment, etc.; implementation amplitude, that is, the size of the adjustment amount, for example, the change amount of roll pressure from an initial value to a target value; and duration, that is, the time length from the start of the adjustment behavior to the completion and stabilization of the adjustment.

[0030] Step S1232: Extract the key response behaviors and occurrence times from the output response behavior imprint segments corresponding to each time segment, record the type, intensity, and duration of the key response behaviors, and obtain the key output response behavior information.

[0031] Similar to extracting key behaviors for parameter adjustment, key response behaviors need to be extracted from the output response behavior imprint segment corresponding to each time segment. Key response behaviors typically refer to changes in the main output indicators of the paper machine's press section, such as changes in paper moisture content, thickness, and basis weight. Record the type of key response behavior, identifying which output indicator it corresponds to; the intensity of the change, i.e., the degree of change in the output indicator, such as the change in moisture content from an initial value to a target value; and the duration, i.e., the time from the onset of the response behavior to its stabilization. Simultaneously, accurately record the occurrence time of the key response behavior for correlation analysis with the occurrence time of key behaviors for parameter adjustment. Through these operations, the key output response behavior information is obtained.

[0032] Step S1233: Match the key behavior information of parameter adjustment with the key behavior information of output response according to the time of occurrence, establish the time correspondence between adjustment behavior and response behavior, and obtain time sequence correspondence data.

[0033] A one-to-one correspondence is established between the occurrence times of key parameter adjustment behaviors and key output response behaviors. For each key parameter adjustment behavior, the key output response behaviors that occur after its occurrence time are identified, and their temporal order is determined. For example, if a key parameter adjustment behavior occurs at time T1, then key output response behaviors that occur after time T1 are likely related to that adjustment behavior. In this way, a temporal correspondence between adjustment behaviors and response behaviors is established, forming temporal correspondence data. This temporal correspondence data clearly shows when the output response behavior occurs after a parameter adjustment behavior.

[0034] Step S1234: Based on the time-series correspondence data analysis, analyze the triggering association between the parameter adjustment key behavior and the output response key behavior, determine the parameter adjustment key behavior corresponding to each output response key behavior, identify the direct transmission path and indirect transmission path between behaviors, and obtain the behavior transmission path set.

[0035] Based on the above timing correspondence data, the triggering association between the parameter adjustment key behaviors and the output response key behaviors is analyzed. For each output response key behavior, according to its occurrence time and the time correspondence with the parameter adjustment key behaviors, the parameter adjustment key behaviors that can trigger the response behavior are determined. If an output response key behavior is directly caused by a certain parameter adjustment key behavior without the intervention of other adjustment behaviors, there is a direct conduction path between the two behaviors. For example, the adjustment of the press roll pressure directly leads to the change of the paper thickness, so the press roll pressure adjustment behavior to the paper thickness response behavior is a direct conduction path. If an output response key behavior is indirectly caused by multiple parameter adjustment key behaviors through a series of intermediate adjustment behaviors, there is an indirect conduction path between the behaviors. For example, the adjustment of the pulp feeding concentration leads to the adjustment of the press roll pressure, which in turn causes the change of the paper moisture content, so the pulp feeding concentration adjustment behavior to the paper moisture content response behavior is an indirect conduction path, which passes through the intermediate link of the press roll pressure adjustment behavior. By identifying the direct conduction path and the indirect conduction path, a behavior conduction path set is obtained, which fully reflects the conduction path situation between the parameter adjustment key behaviors and the output response key behaviors.

[0036] Step S1235: Analyze the time interval between the occurrence time of the adjustment behavior and the occurrence time of the response behavior in the timing correspondence data, summarize the response time rules corresponding to different types of adjustment behaviors, and obtain the time association rule.

[0037] In the timing correspondence data, for each pair of parameter adjustment key behaviors and output response key behaviors with triggering association, the time interval between their occurrence times is calculated. For example, the parameter adjustment key behavior occurs at time T1, and the corresponding output response key behavior occurs at time T2, so the time interval is T2-T1. Then, for different types of parameter adjustment key behaviors, the time intervals of the corresponding output response key behaviors are respectively counted, analyzed and summarized. Through statistical analysis of a large amount of data, it can be found that the response time corresponding to different types of adjustment behaviors has certain rules. For example, the output response time interval corresponding to the vehicle speed adjustment behavior is relatively short, while the output response time interval corresponding to the pulp feeding concentration adjustment behavior is relatively long. The summary of the above time association rule is helpful in predicting the occurrence time of the output response according to the type of the adjustment behavior in actual production process, so as to better control and adjust the production.

[0038] Step S1236: Compare the implementation amplitude of the parameter adjustment key behavior with the performance intensity of the output response key behavior, mine the action relationship between the two, summarize the response intensity rules corresponding to different implementation amplitude adjustment behaviors, and obtain the intensity action relationship.

[0039] The implementation amplitude of the parameter adjustment key behavior is compared and analyzed with the performance intensity of the corresponding output response key behavior. For the same type of parameter adjustment key behavior, different implementation amplitudes can lead to different performance intensities of the output response key behavior. For example, the greater the adjustment amplitude of the roller pressure, the greater the change intensity of the paper thickness. By comparing and analyzing a large amount of data, the relationship between the implementation amplitude of the parameter adjustment key behavior and the performance intensity of the output response key behavior is mined. The change rule of the output response intensity corresponding to the adjustment behavior under different implementation amplitudes is summarized, such as linear relationship, nonlinear relationship, etc., so as to obtain the intensity relationship. The intensity relationship has important guiding significance for determining the appropriate parameter adjustment amplitude according to the expected output response intensity in the production process.

[0040] Step S1237: The behavior transmission path set, time correlation rule and intensity relationship are integrated according to the structure requirement of the inter-dimension correlation path, to obtain the parameter output inter-dimension behavior correlation path.

[0041] The behavior transmission path set, time correlation rule and intensity relationship obtained above are integrated. According to the structure requirement of the inter-dimension correlation path, the behavior transmission path is taken as the path subject, and the time correlation rule and the intensity relationship are taken as the attributes of the path. For example, for a certain behavior transmission path, the corresponding time correlation rule (i.e. the time interval rule of the adjustment behavior to the response behavior on the path) and the intensity relationship (i.e. the relationship between the implementation amplitude of the adjustment behavior and the performance intensity of the response behavior on the path) are determined. Through the above integration, a complete parameter output inter-dimension behavior correlation path is formed, which clearly shows how the parameter adjustment behavior affects the output response behavior through a specific transmission path under certain time correlation rule and intensity relationship.

[0042] Step S124: The environment interaction behavior imprint segment corresponding to each time sequence segment is respectively multi-dimensionally behavior inversed with the parameter adjustment behavior imprint segment and the output response behavior imprint segment, to identify the influence path of the environment interaction behavior on the parameter adjustment behavior, the action mode of the environment interaction behavior on the output response behavior and the synergistic action condition of the three, to obtain the environment parameter output multi-dimension synergistic action logic.

[0043] Step S1241: The key environmental factors and changes in the environment interaction behavior imprint segment corresponding to each time sequence segment are extracted, the types, changes and duration of the key environmental factors are recorded, and the environment interaction key factor information is obtained.

[0044] In the environment interaction behavior imprint segment corresponding to each time sequence segment, information of various environmental factors is contained. Key environmental factors need to be extracted from it, which usually refers to environmental factors that have a significant impact on the operation of the press section of the paper machine, such as workshop temperature, humidity, and properties of raw materials (such as fiber length, hardness, etc.). For the extracted key environmental factors, record their types, and specify which environmental factor; change, that is, the numerical change of the environmental factor in the time sequence segment, for example, the temperature rises or falls from a certain value to another value; duration, that is, the length of time that the environmental factor is in a particular change state. Through the above records, environment interaction key factor information is formed.

[0045] Step S1242: associate the environment interaction key factor information with the corresponding parameter adjustment behavior imprint segment, identify the adjustment of the parameter adjustment behavior after the change of the environment interaction key factor, identify the direct and indirect paths of the influence of the environmental factor on the parameter adjustment, and obtain the environmental influence path on the parameter adjustment.

[0046] The environment interaction key factor information is associated with the corresponding parameter adjustment behavior imprint segment for correlation analysis. When the environment interaction key factor changes, observe whether there is a corresponding parameter adjustment behavior in the parameter adjustment behavior imprint segment. If the change of the environmental factor directly leads to the implementation of a certain parameter adjustment behavior, then this is the direct path of the influence of the environmental factor on the parameter adjustment. For example, when the workshop humidity rises to a certain degree, it directly triggers the adjustment behavior of the press roll pressure, and then the change of the workshop humidity to the press roll pressure adjustment behavior is the direct influence path. If the change of the environmental factor first affects other intermediate factors, and then causes the implementation of the parameter adjustment behavior, then this is an indirect path. For example, the increase of the workshop temperature causes the change of the moisture content of the raw material, and then triggers the adjustment behavior of the pulp concentration, and then the change of the workshop temperature to the moisture content of the raw material and then to the adjustment behavior of the pulp concentration is an indirect influence path. By identifying the direct and indirect paths, the environmental influence path on the parameter adjustment is obtained, which reveals how the environmental factor affects the parameter adjustment behavior.

[0047] Step S1243: associate the environment interaction key factor information with the corresponding output response behavior imprint segment, identify the change characteristics of the output response behavior after the change of the environment interaction key factor, summarize the action mode and influence law of the environmental factor on the output response, and obtain the environmental action mode on the output response.

[0048] The environmental interaction key factor information is associated with the corresponding output response behavior imprint segment. The change characteristics of the key response behaviors in the output response behavior imprint segment when the environmental interaction key factor changes are analyzed. For example, whether the change of the workshop temperature leads to the change of the paper moisture content, and how the direction and degree of the change are. Through the analysis of a large amount of data, the action modes of the environmental factors on the output response, such as promotion, inhibition, linear influence, nonlinear influence, and the influence law, such as the difference in the influence degree of the output response in different environmental factor value ranges, are summarized. Thus, the environmental action mode on the output response is formed, which describes how the environmental factors directly or indirectly affect the output response behavior.

[0049] Step S1244: The change timing of the environmental interaction key factor, the parameter adjustment behavior, and the output response behavior is analyzed to determine the effective implementation conditions, the environmental adaptation requirements, and the parameter output coordination boundaries of the three, and the effective implementation conditions, the environmental adaptation requirements, and the parameter output coordination boundaries of the three are obtained.

[0050] The change sequence and the mutual relationship of the environmental interaction key factor, the parameter adjustment behavior, and the output response behavior in time are analyzed. It is determined that under what circumstances the three can synergistically act to achieve the stable operation of the paper machine press section and the desired output indicators. The effective implementation conditions refer to the specific conditions required for the synergistic action of the three, for example, the environmental factors change within a certain range, the parameter adjustment behavior is implemented in a specific manner, and the output response behavior reaches a certain standard. The environmental adaptation requirement refers to the condition that the environmental factor needs to meet to adapt to the synergistic action of the parameter adjustment and the output response. The parameter output coordination boundary refers to the range and limit of the synergistic action of the parameter adjustment behavior and the output response behavior, for example, the amplitude of the parameter adjustment, the index of the output response within what interval, and the like, so as to ensure the synergistic effect of the three. Through the analysis of the above-mentioned contents, the effective implementation conditions, the environmental adaptation requirements, and the parameter output coordination boundaries of the three are determined.

[0051] Step S1245: The environmental parameter output multidimensional synergistic action logic is obtained by integrating the environmental parameter adjustment influence path, the environmental action mode on the output response, the effective implementation conditions of the three, the environmental adaptation requirements, and the parameter output coordination boundaries, and organizing them according to the structural requirements of the multidimensional synergistic action logic.

[0052] The above obtained environmental parameter adjustment influence path, environmental output response action mode, effective implementation condition of the three, environmental adaptation requirement and parameter output coordination boundary are integrated. According to the structure requirement of multi-dimensional coordination logic, the environmental factors, parameter adjustment behavior and output response behavior are taken as three dimensions, and the interaction relationship, coordination implementation condition, environmental adaptation requirement and coordination boundary among them are determined. For example, under a specific environmental adaptation requirement, the environmental factors act on the parameter adjustment behavior through a specific influence path, the parameter adjustment behavior influences the output response behavior through a transmission path, and at the same time the environmental factors also directly influence the output response behavior through an action mode, and the three realize coordinated action within the effective implementation condition and the coordination boundary. Through the above integration, the environmental parameter output multi-dimensional coordination logic is formed, which comprehensively reflects the coordination law among the three dimensions.

[0053] Step S125: The parameter adjustment dimension internal behavior association logic, the parameter output dimension inter-behavior association path and the environmental parameter output multi-dimensional coordination logic are integrated in order of time sequence segments to build a structured inversion model with behavior nodes as the core, associated paths as the link and coordination logic as the constraint, and a preliminary dynamic behavior inversion model is obtained.

[0054] After the analysis of the parameter adjustment dimension internal behavior association logic, the parameter output dimension inter-behavior association path and the environmental parameter output multi-dimensional coordination logic of each time sequence segment, the above logic information is integrated in order of time sequence segments. In the integration process, each behavior (including parameter adjustment behavior, output response behavior and environmental interaction behavior) is taken as a behavior node, the parameter adjustment dimension internal behavior association logic and the parameter output dimension inter-behavior association path are taken as the associated paths connecting each behavior node, and the environmental parameter output multi-dimensional coordination logic is taken as the constraint condition of the whole model. In this way, a structured inversion model, i.e. a preliminary dynamic behavior inversion model, is built. The preliminary dynamic behavior inversion model can clearly show the association relationship and coordination among the behaviors of the paper machine press part data model in different time sequence segments.

[0055] Step S126: The preliminary dynamic behavior inversion model is processed for time sequence continuity, the model parts corresponding to different time sequence segments are connected, the associated paths, transmission mechanisms and coordination logic in the preliminary dynamic behavior inversion model are kept in time sequence, and a paper machine press part data model dynamic behavior inversion result is obtained.

[0056] Since the preliminary dynamic behavior inversion model is integrated in sequence of time segments, there may be certain discontinuity or incoherence between the model parts corresponding to different time segments. Therefore, the preliminary dynamic behavior inversion model needs to be processed for time sequence coherence. Specifically, the associated paths, conduction mechanisms and collaborative logics in the model parts of adjacent time segments are analyzed to ensure their continuity and consistency in time. For example, the output of a behavior node in a previous time segment may be the input of another behavior node in a subsequent time segment, and it is necessary to ensure that the above connection relationship is correctly reflected in the model. For the conduction mechanisms and collaborative logics, it is also necessary to check whether there is contradiction or inconsistency between different time segments, and make corresponding adjustment and optimization, so that the associated paths, conduction mechanisms and collaborative logics in the entire preliminary dynamic behavior inversion model remain consistent in time. After the time sequence coherence processing, the final dynamic behavior inversion result of the paper machine press section data model is obtained, which can more accurately reflect the dynamic behavior characteristics of the paper machine press section data model in continuous operation.

[0057] Step S130: generating dynamic reference logic based on the dynamic behavior inversion result of the paper machine press section data model, adjusting the associated adaptive conditions, conduction response templates and collaborative boundaries of the reference according to the real-time changes of the inversion result, and obtaining the dynamic reference logic of the paper machine press section data model.

[0058] The dynamic reference logic is an important basis for evaluating whether the paper machine press section data model is running normally, and it can be adjusted in real time according to the changes of the model dynamic behavior inversion result, ensuring the adaptability and accuracy of the reference.

[0059] Step S131: extracting the behavior association logic within the parameter adjustment dimension in the dynamic behavior inversion result of the paper machine press section data model, identifying the effective action range, linkage trigger premise and mutual compatibility requirements between the parameter adjustment behaviors, and obtaining the initial basis of the associated adaptive conditions.

[0060] The behavior association logic within the parameter adjustment dimension is extracted from the dynamic behavior inversion result of the paper machine press section data model. This logic is analyzed in depth to identify the effective action range of the association between the parameter adjustment behaviors, i.e., under what circumstances is a certain association effective; the linkage trigger premise, i.e., what conditions need to be met to trigger the linkage between behaviors; and the mutual compatibility requirements, i.e., what compatibility conditions need to be met between different parameter adjustment behaviors to avoid conflicts or adverse effects. Through the identification of the above contents, the initial basis of the associated adaptive conditions is formed.

[0061] Step S132: capture the change trend of the paper machine press section data model dynamic behavior inversion result in real time, track the real-time change of the behavior correlation logic in the parameter adjustment dimension, determine the change stability through data comparison of continuous K monitoring periods, adjust the action range and triggering premise of the correlation adaptation condition according to the changed correlation logic, and obtain the dynamic correlation adaptation condition.

[0062] Step S1321: continuously monitor the behavior correlation logic in the parameter adjustment dimension in the paper machine press section data model dynamic behavior inversion result based on the set monitoring period, capture the addition, change and disappearance of the correlation relationship in the behavior correlation logic in the parameter adjustment dimension, and obtain correlation logic change data.

[0063] The monitoring period is set in advance, for example, every hour, every day, etc. According to the set monitoring period, the behavior correlation logic in the parameter adjustment dimension in the paper machine press section data model dynamic behavior inversion result is continuously monitored. During the monitoring process, the focus is on the addition of the correlation relationship, that is, whether a new parameter adjustment behavior correlation relationship appears; the change, that is, whether the existing correlation relationship changes in the action range, triggering premise, etc.; and the disappearance, that is, whether some original correlation relationship no longer exists. The above monitored situations are recorded to form correlation logic change data.

[0064] Step S1322: trend analysis is performed on the correlation logic change data, the change direction and change duration of the behavior correlation logic in the parameter adjustment dimension are identified, the change stability is determined through data comparison of continuous K monitoring periods, and the change attribute is divided.

[0065] The trend analysis is performed on the correlation logic change data, the change direction of the behavior correlation logic in the parameter adjustment dimension is determined, for example, whether the correlation relationship gradually increases or decreases, whether the action range expands or shrinks, etc. At the same time, the change duration is analyzed, that is, how many monitoring periods a certain change trend lasts. Through data comparison of continuous K monitoring periods (K can be set according to actual conditions, such as 3, 5, etc.), if the change trend of the correlation logic remains the same in continuous K monitoring periods, it can be determined that the change has stability. According to the change stability, the change attribute is divided into persistent change and temporary fluctuation. The persistent change refers to a change trend that stably exists in continuous monitoring periods; the temporary fluctuation refers to a change that appears only in individual monitoring periods and then returns to the original state.

[0066] Step S1323: if the change attribute is persistent change, the effective action range, linkage triggering premise and mutual compatibility requirement in the changed behavior correlation logic in the parameter adjustment dimension are extracted as the adjustment basis of the correlation adaptation condition.

[0067] When the change attribute is determined to be a persistent change, the new effective action range, linkage trigger premise, and mutual compatibility requirement are extracted from the changed parameter adjustment dimension behavior association logic. The extracted content serves as the basis for adjusting the association adaptation condition to ensure that the association adaptation condition can adapt to the persistent change of the parameter adjustment dimension behavior association logic.

[0068] Step S1324: According to the adjustment basis, the effective action range of the association adaptation condition is expanded or reduced, and the specific requirements of the linkage trigger premise are optimized.

[0069] According to the above adjustment basis, the effective action range of the association adaptation condition is expanded or reduced. If the effective action range in the changed association logic is expanded, the effective action range of the association adaptation condition should also be expanded; otherwise, it should be reduced. At the same time, the specific requirements of the linkage trigger premise are optimized to match the linkage trigger premise in the changed association logic, ensuring that the association adaptation condition can accurately reflect the association between the parameter adjustment behaviors.

[0070] Step S1325: If the change attribute is temporary fluctuation, the core content of the original association adaptation condition is retained, and the adaptation condition corresponding to the fluctuation part is adjusted locally.

[0071] When the change attribute is temporary fluctuation, it is considered that the change may be caused by accidental factors and will not have a long-term impact on the parameter adjustment dimension behavior association logic. Therefore, the core content of the original association adaptation condition is retained to maintain the stability of the reference logic. For the adaptation condition corresponding to the fluctuation part, only local adjustment is made to cope with temporary changes without changing the overall structure and core requirements of the association adaptation condition.

[0072] Step S1326: Integrate the adjusted effective action range, linkage trigger premise, and local adjustment content to generate a dynamic association adaptation condition.

[0073] The adjusted effective action range, linkage trigger premise, and local adjustment content for temporary fluctuation are integrated. According to the structure requirements of the association adaptation condition, the above adjusted content is organically organized to form a dynamic association adaptation condition. The dynamic association adaptation condition can be adjusted in real time according to the change of the parameter adjustment dimension behavior association logic, ensuring that it adapts to the actual operation of the paper machine press data model.

[0074] Step S133: Analyze the parameter output dimension behavior association path in the paper machine press data model dynamic behavior inversion result, extract the time response range, intensity action standard, and feedback regulation law of parameter output conduction, and obtain the initial basis of the conduction response template.

[0075] The parameter output dimension behavior correlation path in the dynamic behavior inversion result of the paper machine press section data model is analyzed. The time response range of the parameter output conduction is extracted from the correlation path, i.e. the time interval in which the output response behavior may occur after the parameter adjustment behavior occurs; the intensity action standard, i.e. the corresponding relationship standard between the implementation amplitude of the parameter adjustment behavior and the performance intensity of the output response behavior; the feedback adjustment law, i.e. the law of how the parameter adjustment behavior is adjusted when the output response behavior reaches a certain intensity or exceeds a certain range. The above extracted contents are used as the initial basis of the conduction response template to construct the preliminary conduction response template.

[0076] Step S134: Based on the real-time conduction data of the dynamic behavior inversion result of the paper machine press section data model, the running change of the behavior correlation path between the parameter output dimensions is tracked, the change trend is determined by comparing the data differences of adjacent two conduction periods, the time range and the intensity standard of the conduction response template are adjusted according to the changed conduction path, and the dynamic conduction response template is obtained.

[0077] Step S1341: Collecting the real-time conduction data in the dynamic behavior inversion result of the paper machine press section data model, extracting the time interval data, intensity matching data and feedback adjustment data in the parameter output conduction process, and obtaining the real-time conduction characteristic data.

[0078] The real-time conduction data in the dynamic behavior inversion result of the paper machine press section data model is collected, which reflects the actual running condition of the behavior correlation path between the parameter output dimensions. The time interval data in the parameter output conduction process, i.e. the time interval between the parameter adjustment behavior and the output response behavior, is extracted from the real-time conduction data; the intensity matching data, i.e. the matching relationship data between the implementation amplitude of the parameter adjustment behavior and the performance intensity of the output response behavior; the feedback adjustment data, i.e. the feedback adjustment condition data of the output response behavior to the parameter adjustment behavior. The above data is integrated to form the real-time conduction characteristic data.

[0079] Step S1342: Statistical analysis is performed on the real-time conduction characteristic data, the historical conduction characteristic data is compared, the change of the time interval, the intensity matching and the feedback adjustment in the behavior correlation path between the parameter output dimensions is identified, the change trend is determined by comparing the data differences of adjacent two conduction periods, and the change attribute is divided.

[0080] The statistical analysis is performed on the real-time conduction characteristic data to calculate the mean, variance, and correlation coefficient of the time interval, and other statistical indicators. The statistical indicators are compared with the corresponding indicators of the historical conduction characteristic data to identify the changes in the time interval, intensity matching, and feedback adjustment in the behavior correlation path between the parameter output dimensions. The differences in the real-time conduction characteristic data between two adjacent conduction periods (a conduction period can be determined according to the actual situation of the parameter output conduction, such as a complete parameter adjustment-output response process) are compared to determine whether the change trend is gradually increasing, gradually decreasing, or basically stable. According to the stability of the change trend, the change attribute is divided into long-term trend change and short-term accidental change. The long-term trend change refers to a trend that continuously changes in one direction over multiple adjacent conduction periods. The short-term accidental change refers to an accidental change that occurs only in individual adjacent conduction periods.

[0081] Step S1343: If the change attribute is a long-term trend change, the time range of the conduction response is re-determined according to the changed behavior correlation path between the parameter output dimensions, and the time allowance interval is adjusted.

[0082] When the change attribute is a long-term trend change, the time interval rule between the parameter adjustment behavior and the output response behavior is re-analyzed according to the changed behavior correlation path between the parameter output dimensions, and a new conduction response time range is determined. According to the new time range, the time allowance interval in the conduction response template is adjusted so that the time allowance interval can cover the changed time interval.

[0083] Step S1344: The intensity standard of the conduction response is adjusted according to the changed intensity matching data, and the intensity adaptation requirement is optimized.

[0084] At the same time, the corresponding relationship between the implementation amplitude of the parameter adjustment behavior and the performance intensity of the output response behavior is re-evaluated according to the changed intensity matching data, and the intensity standard in the conduction response template is adjusted. The intensity adaptation requirement is optimized to ensure that the intensity standard can accurately reflect the changed intensity matching situation, so that the conduction response template can correctly evaluate whether the intensity of the parameter output conduction meets the requirements.

[0085] Step S1345: If the change attribute is a short-term accidental change, the core parameters of the original conduction response template are maintained, and the adaptation conditions corresponding to the fluctuation part are locally adjusted.

[0086] When the change attribute is determined to be a short-term accidental change, it indicates that the change in the behavior correlation path between the parameter output dimensions is not a long-term stable trend, and may be caused by a sudden, temporary factor. At this time, in order to ensure the overall stability and continuity of the conduction response template, the core parameters in the original conduction response template should be maintained, such as the basic time response range framework and the main intensity action standard. For the adaptive conditions corresponding to the fluctuation part, such as abnormal time intervals or intensity matching deviations that occur within a certain specific conduction period, only local adjustments are made. For example, if a short-term fluctuation causes a small range of deviation in the matching between the parameter adjustment amplitude and the response intensity, a temporary fluctuation interval can be set based on the original intensity standard, or the intensity corresponding relationship of specific types of adjustment behavior and response behavior can be fine-tuned without changing the overall intensity standard system.

[0087] Step S1346: Integrate the adjusted time range, intensity standard and temporary adjustment information to generate a dynamic conduction response template.

[0088] The time range adjusted through the above steps (including the time interval re-determined under long-term trend changes, or the original time range maintained under short-term accidental changes), the intensity standard (including the intensity adaptation requirement optimized under long-term trend changes, or the core intensity parameter maintained under short-term accidental changes), and the temporary adjustment information for short-term accidental changes are integrated. According to the structured requirements of the conduction response template, the time range and the intensity standard are taken as the main framework of the template, and the temporary adjustment information is embedded as supplementary instructions or conditional clauses, forming a dynamic conduction response template that can be dynamically adjusted according to the real-time changes in the behavior correlation path between the parameter output dimensions.

[0089] Step S135: Analyze the environmental parameter output multidimensional synergy logic in the dynamic behavior inversion result of the paper machine press part data model, identify the effective implementation conditions, environmental adaptation requirements and parameter output synergy boundaries of the three, and obtain the initial basis of the synergy boundary.

[0090] The environmental parameter output multidimensional synergy logic in the dynamic behavior inversion result of the paper machine press part data model is analyzed in depth. This multidimensional synergy logic contains the ways and rules of the synergy of the environmental interaction behavior, the parameter adjustment behavior and the output response behavior. From it, the effective implementation conditions of the three are identified, that is, under what conditions can the three achieve effective synergy, such as the environmental factors being in a specific range, the parameter adjustment behavior being implemented according to a specific mode, etc.; the environmental adaptation requirements, that is, what conditions the environmental factors need to meet to form good synergy with parameter adjustment and output response; and the parameter output synergy boundaries, that is, the range and limit of the parameter adjustment behavior and the output response behavior in the synergy. Through the identification of the above contents, the initial basis of the synergy boundary is obtained.

[0091] Step S136: Capture the real-time changes of the environmental parameter output multi-dimensional synergy logic, determine the influence of changes by monitoring the difference in synergy effect before and after the change of environmental factors, adjust the implementation conditions and adaptation requirements of the synergy boundary according to the changed synergy logic, and obtain the dynamic synergy boundary.

[0092] Continuously capture the real-time changes of the environmental parameter output multi-dimensional synergy logic. By monitoring the difference in synergy effect before and after the change of environmental factors, for example, the stability of the output response index under the synergy effect, the efficiency of parameter adjustment, etc., determine the influence degree and direction of the change of environmental factors on the synergy logic. According to the changed synergy logic, adjust the effective implementation conditions of the synergy boundary, such as expanding or narrowing the effective range of environmental factors, modifying the implementation standards of parameter adjustment behavior, etc.; at the same time, adjust the environmental adaptation requirements, so that the adaptation conditions of environmental factors match the changed synergy logic. Through the above adjustment, the dynamic synergy boundary is obtained, which can be updated in real time according to the changes of the environmental parameter output multi-dimensional synergy logic.

[0093] Step S137: Integrate the dynamic correlation adaptation condition, dynamic conduction response template and dynamic synergy boundary, organize according to the structured requirements of the reference logic, add a real-time update trigger mechanism, and generate a paper machine press section data model dynamic reference logic that can follow the changes of the dynamic behavior inversion result.

[0094] Integrate the dynamic correlation adaptation condition, dynamic conduction response template and dynamic synergy boundary obtained above. According to the structured requirements of the reference logic, take the dynamic correlation adaptation condition as the reference of the behavior correlation in the parameter adjustment dimension, take the dynamic conduction response template as the reference of the behavior conduction between the parameter output dimensions, and take the dynamic synergy boundary as the reference of the multi-dimensional synergy of environmental parameter output. At the same time, add a real-time update trigger mechanism, which can monitor the changes of the paper machine press section data model dynamic behavior inversion result. When the correlation logic, conduction path or synergy logic in the inversion result changes, automatically trigger the update of the corresponding part (dynamic correlation adaptation condition, dynamic conduction response template or dynamic synergy boundary) in the dynamic reference logic. Through the above integration and mechanism addition, a paper machine press section data model dynamic reference logic that can follow the changes of the dynamic behavior inversion result is generated.

[0095] Step S140: Close-loop calibration of the paper machine press section data model dynamic behavior inversion result and the paper machine press section data model dynamic reference logic, locate the correlation nodes, conduction nodes and synergy nodes deviating from the reference in the inversion result, generate node calibration instructions and feedback to the inversion process, and obtain the dynamic behavior inversion result after close-loop calibration.

[0096] The closed-loop calibration is an important link for ensuring the accuracy of the inversion result of the dynamic behavior of the paper machine press section data model. By comparing the inversion result with the dynamic benchmark logic, the deviation is found out and calibrated, forming a closed-loop process of continuous optimization.

[0097] Step S141: The behavior correlation logic in the parameter adjustment dimension in the inversion result of the dynamic behavior of the paper machine press section data model is compared with the dynamic correlation adaptation condition in the dynamic benchmark logic of the paper machine press section data model node by node, and the nodes in the behavior correlation logic in the parameter adjustment dimension that do not meet the dynamic correlation adaptation condition are marked as correlation deviation nodes to obtain a set of correlation deviation nodes.

[0098] Each node of the behavior correlation logic in the parameter adjustment dimension in the inversion result of the dynamic behavior of the paper machine press section data model is compared with the corresponding requirement of the dynamic correlation adaptation condition in the dynamic benchmark logic. It is checked whether the correlation relationship of each node meets the effective action range, linkage trigger premise and mutual compatibility requirement of the dynamic correlation adaptation condition. If the correlation relationship of a certain node exceeds the effective action range of the dynamic correlation adaptation condition, or its linkage trigger premise does not meet the requirement, or there is an incompatibility with other nodes, the node is marked as a correlation deviation node. All the marked correlation deviation nodes are summarized to form a set of correlation deviation nodes.

[0099] Step S142: The behavior correlation path between the parameter output dimensions in the inversion result of the dynamic behavior of the paper machine press section data model is compared with the dynamic conduction response template in the dynamic benchmark logic of the paper machine press section data model path by path, and the nodes in the behavior correlation path between the parameter output dimensions that do not meet the dynamic conduction response template are marked as conduction deviation nodes to obtain a set of conduction deviation nodes.

[0100] Each path of the behavior correlation path between the parameter output dimensions in the inversion result of the dynamic behavior of the paper machine press section data model is analyzed, and each node in the path is compared with the dynamic conduction response template in the dynamic benchmark logic. It is checked whether the conduction time of the node is within the time allowed interval of the dynamic conduction response template and whether the conduction intensity meets the intensity standard. If the conduction time of a certain node exceeds the time allowed interval, or the conduction intensity does not meet the intensity standard, the node is marked as a conduction deviation node. All the conduction deviation nodes are summarized to form a set of conduction deviation nodes.

[0101] Step S143: Perform a logic-by-logic comparison between the environmental parameter output multi-dimensional synergy logic in the dynamic behavior inversion result of the paper machine press section data model and the dynamic synergy boundary in the dynamic benchmark logic of the paper machine press section data model, identify nodes in the environmental parameter output multi-dimensional synergy logic that do not conform to the dynamic synergy boundary, mark them as synergy deviation nodes, and obtain a set of synergy deviation nodes.

[0102] Each node in the environmental parameter output multi-dimensional synergy logic is compared with the dynamic synergy boundary on a logic-by-logic basis. Check whether the synergy of the node conforms to the effective implementation conditions, environmental adaptation requirements, and parameter output synergy boundary of the dynamic synergy boundary. If the synergy of a certain node does not conform to the effective implementation conditions, or the environmental factors do not meet the environmental adaptation requirements, or exceeds the parameter output synergy boundary, mark the node as a synergy deviation node. All synergy deviation nodes are summarized to form a set of synergy deviation nodes.

[0103] Step S144: Trace the source of each deviation node in the set of associated deviation nodes, the set of conduction deviation nodes, and the set of synergy deviation nodes, and mine the upstream node influence path, behavior conduction link, and external action factors generated by the deviation node to obtain a deviation source result.

[0104] Step S1441: Take each deviation node as a starting point, and trace the upstream associated nodes in the dynamic behavior inversion result of the paper machine press section data model in reverse, identify the association relationship and conduction path between the deviation node and the upstream nodes, determine the range of upstream nodes that may affect the deviation node, and obtain the range of upstream associated nodes.

[0105] For each deviation node in the set of associated deviation nodes, the set of conduction deviation nodes, and the set of synergy deviation nodes, the upstream nodes associated with it are found in the dynamic behavior inversion result of the paper machine press section data model. Analyze the association relationship between the deviation node and the upstream nodes, such as direct association or indirect association, and the conduction path between them. Through the above reverse tracing, the range of upstream nodes that may affect the deviation node, i.e., the range of upstream associated nodes, is determined.

[0106] Step S1442: Analyze the behavior state of each node in the range of upstream associated nodes, determine whether the running mode of the upstream node conforms to the model preset logic, identify the direct upstream influence node that causes the deviation node to be generated, and obtain direct influence node information.

[0107] Analyze the behavior state of each node in the upstream associated node range, check whether the running mode of these nodes conforms to the normal logic and rules preset by the model. If the running mode of a certain upstream node does not conform to the preset logic, and the abnormal behavior state of the node can directly lead to the generation of the deviation node, the upstream node is the direct upstream influencing node leading to the generation of the deviation node. Record the type of the direct influencing node, the abnormal behavior state and other information, and obtain the direct influencing node information.

[0108] Step S1443: Identify the behavior transmission link of the direct influencing node to the deviation node, and clearly determine the intermediate nodes, transmission mode and action mechanism in the transmission process, to obtain the details of the deviation transmission link.

[0109] After determining the direct influencing node, further identify the behavior transmission link from the direct influencing node to the deviation node. Analyze in detail whether there are intermediate nodes in the transmission process, the type and role of the intermediate nodes; whether the transmission mode is direct transmission or indirect transmission, whether it is transmitted through parameter adjustment, environmental change or other ways; and the action mechanism of the transmission, for example, promotion, inhibition or other types of action. Through the above analysis, obtain the details of the deviation transmission link, which clearly shows how the deviation is transmitted from the direct influencing node to the deviation node.

[0110] Step S1444: Investigate the change of the external environmental factor corresponding to the deviation node, analyze whether there is an abnormal environmental factor and change in the environmental interaction behavior imprint, judge whether the external environmental factor is the inducement of the deviation, and obtain the external action factor information.

[0111] Investigate the change of the external environmental factor corresponding to the deviation node when the deviation node occurs. Analyze whether there is an abnormal environmental factor and change in the environmental interaction behavior imprint, for example, the value of the environmental factor suddenly exceeds the normal range, the change rate is abnormal, etc. Judge whether these abnormal external environmental factors can induce the generation of the deviation node, for example, whether the abnormal change of the environmental factor leads to the behavior abnormality of the direct influencing node, and then transmits to the deviation node. If there is the above external environmental factor, record its type, change and influence mode on the deviation node, and obtain the external action factor information.

[0112] Step S1445: Integrate the upstream associated node range, direct influencing node information, deviation transmission link details and external action factor information to form the deviation traceability result corresponding to each deviation node.

[0113] The upstream associated node range, direct influence node information, bias conduction link details and external influencing factor information are integrated, organized according to the structure requirements of the bias traceability result, and a complete bias traceability result corresponding to each bias node is formed. The multi-dimensional synergistic effect comprehensively reflects the causes, influence paths and related factors of the bias node.

[0114] Step S145: generating a targeted calibration instruction for each bias node according to the bias traceability result, the calibration instruction including node adjustment direction, associated relationship correction requirement and conduction path optimization suggestion, and obtaining a node calibration instruction set.

[0115] According to the bias traceability result, a targeted calibration instruction is formulated for each bias node. The calibration instruction should clearly indicate the adjustment direction of the node, such as whether to enhance or weaken the behavior intensity of the node, whether to lengthen or shorten the conduction time, etc.; the associated relationship correction requirement, such as adjusting the associated manner of the node with the upstream node or downstream node, modifying the linkage trigger premise, etc.; the conduction path optimization suggestion, such as optimizing the direct conduction path, reducing the intermediate nodes in the indirect conduction path, etc. The calibration instructions generated for all bias nodes are collected to obtain a node calibration instruction set.

[0116] Step S146: feeding the node calibration instruction set back to the paper machine press part data model dynamic behavior inversion process to drive the inversion process to adjust the bias nodes, correct the associated relationship of the associated bias nodes, the conduction path of the conduction bias nodes and the coordination logic of the coordination bias nodes, and obtain the dynamic behavior inversion result after closed-loop calibration.

[0117] The node calibration instruction set is fed back to the paper machine press part data model dynamic behavior inversion process. According to the calibration instruction, the inversion process corrects the associated relationship of the associated bias nodes to make them meet the dynamic association adaptation condition; optimizes the conduction path of the conduction bias nodes to adjust the conduction time and intensity to make them meet the dynamic conduction response template; adjusts the coordination logic of the coordination bias nodes to make them meet the dynamic coordination boundary. After adjustment, the dynamic behavior inversion result after closed-loop calibration is obtained, and the accuracy and reliability of the dynamic behavior inversion result after closed-loop calibration are improved.

[0118] Step S150: iteratively optimizing the AI analysis process of the paper machine press part data model according to the dynamic behavior inversion result after closed-loop calibration, integrating the inversion result after iteration and the dynamic benchmark logic, and obtaining a paper machine press part data model iterative analysis report.

[0119] Through the dynamic behavior inversion result after closed-loop calibration, the AI analysis process of the paper machine press part data model is iteratively optimized, the performance and accuracy of the analysis process are continuously improved, and an iterative analysis report is formed.

[0120] Step S151: Analyze the closed-loop calibrated dynamic behavior inversion result, extract the optimized parameter adjustment dimension internal behavior correlation logic, parameter output dimension internal behavior correlation path, and environmental parameter output multi-dimension collaborative action logic in the result, and obtain the model optimization core content.

[0121] The closed-loop calibrated dynamic behavior inversion result is analyzed, and the calibrated and optimized parameter adjustment dimension internal behavior correlation logic, parameter output dimension internal behavior correlation path, and environmental parameter output multi-dimension collaborative action logic are extracted from it. The above-mentioned optimized logic and path are the core embodiment of the model running mechanism, and constitute the model optimization core content.

[0122] Step S152: Adjust the AI analysis process of the paper machine press part data model according to the model optimization core content, optimize the running logic of the data acquisition link, behavior inversion link, benchmark generation link, and calibration link in the analysis process, and obtain the iteratively optimized AI analysis process.

[0123] According to the model optimization core content, each link in the AI analysis process is adjusted. In the data acquisition link, the sampling frequency of the sensor, the data filtering rule, etc. may need to be optimized to improve the accuracy and relevance of data acquisition; in the behavior inversion link, the parameters of the inversion algorithm, the identification method of the correlation path, etc. are adjusted according to the optimized logic; in the benchmark generation link, the update trigger condition of the dynamic benchmark logic is modified, and the adjustment basis is adjusted; in the calibration link, the identification algorithm of the deviation node, the generation rule of the calibration instruction, etc. are optimized. Through the optimization of the running logic of the above-mentioned links, the iteratively optimized AI analysis process is obtained.

[0124] Step S153: Apply the iteratively optimized AI analysis process to the running analysis of the paper machine press part data model, collect model running data, benchmark adjustment data, and calibration implementation data in the analysis process, and obtain iteratively analyzed process data.

[0125] The iteratively optimized AI analysis process is applied to the actual running analysis of the paper machine press part data model. In the analysis process, model running data such as the running state of each behavior node, the activity degree of the correlation path, etc. are collected; benchmark adjustment data such as the number of updates of the dynamic benchmark logic, the adjustment amplitude, etc. are collected; calibration implementation data such as the execution of the calibration instruction, the number of reduced deviation nodes, etc. are collected. The above-mentioned data are summarized to obtain iteratively analyzed process data.

[0126] Step S154: Compare the AI analysis process running data before and after iteration optimization, analyze the influence of process optimization on model inversion accuracy, benchmark adaptability, and calibration implementation effect, and obtain process optimization influence analysis.

[0127] Compare the running data of the AI analysis process before and after iteration optimization, evaluate the effect of process optimization, analyze whether the inversion accuracy of the model is improved, for example, whether the consistency of the inversion result with the actual running situation is improved, whether the benchmark adaptability is enhanced, for example, whether the dynamic benchmark logic can adapt to the changes of the inversion result more quickly and accurately, whether the calibration implementation effect is improved, for example, whether the number of deviation nodes is reduced, whether the stability of the nodes after calibration is improved, etc. Through the above analysis, the process optimization influence analysis is obtained, and the effectiveness and shortcomings of iteration optimization are summarized.

[0128] Step S155: Integrating the dynamic behavior inversion result after closed-loop calibration, the dynamic benchmark logic of the paper machine press part data model, the AI analysis process after iteration optimization, the iteration analysis process data and the process optimization influence analysis, organizing according to the structured requirements of the analysis report, and generating the paper machine press part data model iteration analysis report.

[0129] Integrate the dynamic behavior inversion result after closed-loop calibration, the dynamic benchmark logic of the paper machine press part data model, the AI analysis process after iteration optimization, the iteration analysis process data and the process optimization influence analysis, etc. According to the structure requirements of the analysis report, including introduction, method, result, discussion, conclusion, etc., organize and present the above contents, and generate the paper machine press part data model iteration analysis report. The paper machine press part data model iteration analysis report comprehensively summarizes the iteration optimization process, effect and related data of the model.

[0130] Based on the same inventive concept, please refer to Figure 2 , which shows the structural schematic block diagram of the AI analysis system 100 for the paper machine press part data model provided by the embodiments of the present application for executing the AI analysis method for the paper machine press part data model described above. The AI analysis system 100 for the paper machine press part data model can include a communication unit 110, a machine readable storage medium 120 and a processor 130.

[0131] In this embodiment, the machine readable storage medium 120 and the processor 130 are both located in the AI analysis system 100 for the paper machine press part data model and are separately arranged. However, it should be understood that the machine readable storage medium 120 can also be independent of the AI analysis system 100 for the paper machine press part data model, and can be accessed by the processor 130 through a bus interface. Alternatively, the machine readable storage medium 120 can also be integrated into the processor 130, and can communicate and interact with external systems through the communication unit 110.

[0132] The processor 130 is the control center of the AI analysis system 100 applied to the paper machine press section data model, connects each part of the AI analysis system 100 applied to the paper machine press section data model through various interfaces and lines, executes various functions of the AI analysis system 100 applied to the paper machine press section data model and processes data by running or executing the software programs and / or modules stored in the machine readable storage medium 120 and calling the data stored in the machine readable storage medium 120, thereby monitoring the AI analysis system 100 applied to the paper machine press section data model as a whole. Optionally, the processor 130 can include one or more processing cores; for example, the processor 130 can integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and application program, etc., and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor can also not be integrated into the processor. Among them, the machine readable storage medium 120 is used to store machine executable instructions for executing the scheme of the present application, and the processor 130 is used to execute the machine executable instructions stored in the machine readable storage medium 120 to realize the AI analysis method applied to the paper machine press section data model provided by the foregoing method embodiment.

[0133] It should be noted that, in order to simplify the description of the present application and to facilitate the understanding of one or more embodiments of the present application, in the foregoing description of the embodiments of the present application, various features are sometimes combined into one embodiment, figure or description thereof.

Claims

1. An AI analysis method applied to a data model of a paper machine press section, characterized in that, The method includes: The cross-dimensional behavioral imprint of the paper machine press section data model is obtained. The cross-dimensional behavioral imprint includes parameter adjustment behavioral imprint, output response behavioral imprint and environmental interaction behavioral imprint. Dynamic behavior inversion is performed on the cross-dimensional behavioral imprints of the paper machine press section data model to explore the real-time correlation paths, triggering and transmission mechanisms and synergistic logic between different dimensional imprints, and obtain the dynamic behavior inversion results of the paper machine press section data model. Based on the dynamic behavior inversion results of the paper machine press section data model, a dynamic baseline logic is generated. The correlation adaptation conditions, transmission response templates and cooperative action boundaries of the baseline are adjusted according to the real-time changes of the inversion results to obtain the dynamic baseline logic of the paper machine press section data model. The dynamic behavior inversion results of the paper machine press section data model are compared with the dynamic baseline logic of the paper machine press section data model in a closed loop. The associated nodes, transmission nodes and cooperative nodes that deviate from the baseline in the inversion results are located, node calibration instructions are generated and fed back to the inversion process, and the dynamic behavior inversion results after closed loop calibration are obtained. The AI ​​analysis process of the paper machine press section data model is iteratively optimized based on the dynamic behavior inversion results after closed-loop calibration. The inversion results and dynamic benchmark logic are integrated to obtain the iterative analysis report of the paper machine press section data model. The dynamic baseline logic generated based on the dynamic behavior inversion results of the paper machine press section data model, which adjusts the correlation adaptation conditions, transmission response templates, and cooperative action boundaries of the baseline according to the real-time changes of the inversion results, yields the dynamic baseline logic of the paper machine press section data model, including: Extract the behavior correlation logic within the parameter adjustment dimension from the dynamic behavior inversion results of the paper machine press section data model, identify the effective scope of the correlation between parameter adjustment behaviors, the triggering premise of linkage and mutual compatibility requirements, and obtain the initial basis for the correlation adaptation conditions. The dynamic behavior inversion results of the data model of the paper machine press section are captured in real time. The real-time changes of the behavior association logic within the parameter adjustment dimension are tracked. The stability of the change is determined by comparing the data of K consecutive monitoring cycles. The scope of action and triggering conditions of the association adaptation conditions are adjusted according to the changed association logic to obtain the dynamic association adaptation conditions. The behavioral correlation paths between parameter output dimensions in the dynamic behavior inversion results of the paper machine press section data model are analyzed, and the time response range, intensity standard and feedback adjustment law of parameter output transmission are extracted to obtain the initial basis of the transmission response template. Based on the real-time transmission data of the dynamic behavior inversion result of the paper machine press section data model, the operation changes of the behavior correlation path between parameter output dimensions are tracked. The change trend is determined by comparing the data differences between two adjacent transmission cycles. The time range and intensity standard of the transmission response template are adjusted according to the changed transmission path to obtain the dynamic transmission response template. The dynamic behavior inversion results of the paper machine press section data model are analyzed to identify the multi-dimensional synergistic logic of environmental parameter output in the data model, and the effective implementation conditions, environmental adaptation requirements and parameter output synergistic boundaries of the three are identified to obtain the initial basis for the synergistic boundary. The system captures real-time changes in multi-dimensional synergistic logic of environmental parameters, determines the impact of changes by monitoring the differences in synergistic effects before and after changes in environmental factors, and adjusts the implementation conditions and adaptation requirements of the synergistic boundary according to the changed synergistic logic to obtain the dynamic synergistic boundary. The dynamic correlation adaptation conditions, dynamic transmission response templates, and dynamic collaborative action boundaries are integrated and organized according to the structured requirements of the baseline logic. A real-time update triggering mechanism is added to generate a dynamic baseline logic for the paper machine press section data model that can follow the changes in the dynamic behavior inversion results.

2. The AI ​​analysis method applied to the data model of the press section of a paper machine according to claim 1, characterized in that, The dynamic behavior inversion of the cross-dimensional behavioral imprints of the paper machine press section data model is performed to mine the real-time correlation paths, triggering and transmission mechanisms, and synergistic logic between different dimensional imprints, and to obtain the dynamic behavior inversion results of the paper machine press section data model, including: The parameter adjustment behavior imprint, output response behavior imprint, and environmental interaction behavior imprint in the cross-dimensional behavior imprint of the paper machine press section data model are synchronously split according to the runtime sequence to obtain parameter adjustment behavior imprint segments, output response behavior imprint segments, and environmental interaction behavior imprint segments corresponding to multiple time segments. For each time segment, the behavior inversion within the dimension of the parameter regulation behavior imprint segment is performed. The occurrence order, interaction path and linkage triggering relationship of different regulation behaviors in the parameter regulation behavior imprint segment are analyzed to obtain the behavior association logic within the parameter regulation dimension. For each time segment, the parameter adjustment behavior imprint segment and the output response behavior imprint segment are subjected to inter-dimensional behavior inversion to capture the transmission path, temporal correlation law and intensity relationship of parameter adjustment behavior triggering output response behavior, and obtain the inter-dimensional behavior correlation path of parameter output. For each time segment, the environmental interaction behavior imprint segment is compared with the parameter adjustment behavior imprint segment and the output response behavior imprint segment in a multi-dimensional behavior inversion. The influence path of environmental interaction behavior on parameter adjustment behavior, the mode of action on output response behavior and the synergistic effect conditions of the three are identified, and the multi-dimensional synergistic effect logic of environmental parameter output is obtained. The behavioral correlation logic within the parameter adjustment dimension, the behavioral correlation path between parameter output dimensions, and the multi-dimensional synergistic logic of environmental parameter output are integrated according to the time sequence to construct a structured inversion model with behavioral nodes as the core, correlation paths as the link, and synergistic logic as the constraint, thus obtaining a preliminary dynamic behavioral inversion model. The preliminary dynamic behavior inversion model is processed to ensure temporal consistency, connecting the model parts corresponding to different time segments. This ensures that the associated paths, transmission mechanisms, and collaborative logic in the preliminary dynamic behavior inversion model remain temporally consistent, resulting in the dynamic behavior inversion results of the paper machine press section data model.

3. The AI ​​analysis method applied to the data model of the press section of a paper machine according to claim 1, characterized in that, The process involves performing closed-loop calibration of the dynamic behavior inversion results of the paper machine press section data model with the dynamic baseline logic of the paper machine press section data model. This includes locating related nodes, transmission nodes, and collaborative nodes in the inversion results that deviate from the baseline, generating node calibration instructions, and feeding them back to the inversion process. The result is the closed-loop calibrated dynamic behavior inversion results, including: The behavior correlation logic within the parameter adjustment dimension in the dynamic behavior inversion result of the paper machine press section data model is compared node by node with the dynamic correlation adaptation conditions in the dynamic baseline logic of the paper machine press section data model. Nodes in the behavior correlation logic within the parameter adjustment dimension that do not meet the dynamic correlation adaptation conditions are identified and marked as correlation deviation nodes, thus obtaining a set of correlation deviation nodes. The behavior association path between parameter output dimensions in the dynamic behavior inversion result of the paper machine press section data model is compared with the dynamic transmission response template in the dynamic baseline logic of the paper machine press section data model path one by one. Nodes that do not conform to the dynamic transmission response template in the behavior association path between parameter output dimensions are identified and marked as transmission deviation nodes, thus obtaining a set of transmission deviation nodes. The multi-dimensional synergistic logic of environmental parameter output in the dynamic behavior inversion result of the paper machine press section data model is compared with the dynamic synergistic boundary in the dynamic baseline logic of the paper machine press section data model. Nodes in the multi-dimensional synergistic logic of environmental parameter output that do not conform to the dynamic synergistic boundary are identified and marked as synergistic deviation nodes, thus obtaining a set of synergistic deviation nodes. For each deviation node in the set of associated deviation nodes, the set of transmitted deviation nodes, and the set of collaborative deviation nodes, the source is traced to mine the upstream node influence path, behavior transmission link, and external factors that caused the deviation node, and the deviation source tracing result is obtained. Based on the deviation tracing results, a targeted calibration instruction is generated for each deviation node. The calibration instruction includes the node adjustment direction, correlation correction requirements, and transmission path optimization suggestions, resulting in a set of node calibration instructions. The set of node calibration instructions is fed back to the dynamic behavior inversion process of the data model of the paper machine press section, driving the inversion process to adjust the deviation nodes, correct the correlation of related deviation nodes, the transmission path of the transmission deviation nodes, and the collaborative logic of the collaborative deviation nodes, so as to obtain the dynamic behavior inversion result after closed-loop calibration.

4. The AI ​​analysis method applied to the data model of the press section of a paper machine according to claim 2, characterized in that, The step of performing inter-dimensional behavior inversion on the parameter adjustment behavior imprint fragments and output response behavior imprint fragments corresponding to each time segment, capturing the transmission path, temporal correlation rules, and intensity relationship of parameter adjustment behavior triggering output response behavior, and obtaining the inter-dimensional behavior correlation path of parameter output includes: Extract the key adjustment behaviors and their occurrence times from the parameter adjustment behavior imprint segments corresponding to each time segment, and record the type, implementation magnitude and duration of the key adjustment behaviors to obtain key parameter adjustment behavior information; Extract the key response behaviors and their occurrence times from the output response behavior imprint segments corresponding to each time segment, record the type, intensity, and duration of the key response behaviors, and obtain the key output response behavior information. The key behavior information of parameter adjustment and the key behavior information of output response are mapped according to the time of occurrence to establish the time correspondence between adjustment behavior and response behavior, and the time sequence correspondence data is obtained. Based on the time-series correspondence data analysis, the triggering association between the key behavior of parameter adjustment and the key behavior of output response is determined, the key behavior of parameter adjustment corresponding to each key behavior of output response is determined, the direct and indirect transmission paths between behaviors are identified, and a set of behavior transmission paths is obtained. By analyzing the time interval between the occurrence of regulatory behavior and the occurrence of response behavior in the time-series correspondence data, the response time patterns corresponding to different types of regulatory behavior are summarized, and the time correlation patterns are obtained. By comparing the implementation magnitude of key parameter adjustment behaviors with the performance intensity of key output response behaviors, we can explore the relationship between the two, summarize the response intensity patterns corresponding to different implementation magnitude adjustment behaviors, and obtain the intensity-effect relationship. The behavioral transmission path set, time correlation rules, and intensity relationship are integrated according to the structural requirements of inter-dimensional correlation paths to obtain the behavioral correlation paths between parameter output dimensions.

5. The AI ​​analysis method applied to the data model of the press section of a paper machine according to claim 2, characterized in that, The process involves performing multi-dimensional behavior inversion on the environmental interaction behavior imprint fragments corresponding to each time segment, along with the parameter adjustment behavior imprint fragments and the output response behavior imprint fragments. This identifies the influence path of environmental interaction behavior on parameter adjustment behavior, the mode of action on output response behavior, and the synergistic conditions among the three, resulting in the multi-dimensional synergistic logic of environmental parameter output, including: Extract key environmental factors and their changes from the environmental interaction behavior imprint segments corresponding to each time segment, record the type, changes and duration of key environmental factors, and obtain information on key environmental interaction factors; The key environmental interaction factors are associated with the corresponding parameter adjustment behavior imprint fragments to identify the adjustment of parameter adjustment behavior after changes in key environmental interaction factors, and to identify the direct and indirect paths of environmental factors affecting parameter adjustment, thus obtaining the path of environmental influence on parameter adjustment. By associating the key environmental interaction factors with the corresponding output response behavior imprint fragments, we can identify the characteristics of output response behavior changes after changes in key environmental interaction factors, summarize the ways and laws of how environmental factors affect output response, and obtain the mode of environmental effect on output response. By analyzing the time sequence of changes in key environmental interaction factors, parameter regulation behavior, and output response behavior, the time range of their synergistic effect is determined, and the synergistic effect time range is obtained. By combining the influence path of the environment on parameter adjustment, the mode of the environment on the output response, and the time range of the synergistic effect, the implementation conditions of the synergistic effect of the three are explored, including the changes in environmental factors, the implementation methods of parameter adjustment, and the adaptation standards of the output response, so as to obtain the implementation conditions of the synergistic effect. By integrating the environmental influence path on parameter adjustment, the environmental response mode to output, the time range of synergistic effect, and the conditions for synergistic effect implementation, and organizing them according to the structural requirements of multi-dimensional synergistic effect logic, a multi-dimensional synergistic effect logic for environmental parameter output is obtained.

6. The AI ​​analysis method applied to the data model of the press section of a paper machine according to claim 1, characterized in that, The system captures the changing trend of the dynamic behavior inversion results of the paper machine press section data model in real time, tracks the real-time changes of the behavior correlation logic within the parameter adjustment dimension, determines the stability of the changes through data comparison of three consecutive monitoring periods, and adjusts the scope and triggering conditions of the correlation adaptation conditions according to the changed correlation logic to obtain dynamic correlation adaptation conditions, including: Based on the set monitoring period, the behavior correlation logic within the parameter adjustment dimension in the dynamic behavior inversion result of the paper machine press section data model is continuously monitored, and the addition, change and disappearance of the correlation relationship in the behavior correlation logic within the parameter adjustment dimension are captured to obtain correlation logic change data. Trend analysis is performed on the data of the changes in the related logic to identify the direction and duration of changes in the behavioral related logic within the parameter adjustment dimension. The stability of the changes is determined by comparing data from K consecutive monitoring periods, and the change attributes are classified. If the changing attribute is a continuous change, extract the effective scope of action, linkage triggering premises and mutual compatibility requirements in the behavior association logic within the parameter adjustment dimension after the change, and use them as the basis for adjusting the association adaptation conditions; Based on the adjustment criteria, the effective scope of the associated adaptation conditions may be expanded or narrowed, and the specific requirements for the triggering premise of linkage may be optimized. If the change attribute is temporary fluctuation, retain the core content of the original associated adaptation conditions and make local adjustments to the adaptation conditions corresponding to the fluctuation part; The effective scope of action, triggering conditions, and local adjustments after integration and adjustment are used to generate dynamic correlation and adaptation conditions. Furthermore, the real-time transmission data based on the dynamic behavior inversion results of the paper machine press section data model tracks the operational changes of the behavioral correlation path between parameter output dimensions, determines the change trend by comparing the data differences between two adjacent transmission cycles, and adjusts the time range and intensity standard of the transmission response template according to the changed transmission path to obtain a dynamic transmission response template, including: Real-time transmission data is collected from the dynamic behavior inversion results of the data model of the paper machine press section. The time interval data, intensity matching data and feedback adjustment data of the parameter output transmission process are extracted to obtain real-time transmission characteristic data. Statistical analysis is performed on the real-time transmission characteristic data, and compared with historical transmission characteristic data to identify changes in time intervals, intensity matching, and feedback adjustment in the behavioral correlation paths between parameter output dimensions. The trend of change is determined by comparing the data differences between two adjacent transmission cycles, and the change attributes are classified. If the change attribute is a long-term trend change, redetermine the time range of the transmission response based on the behavioral correlation path between the output dimensions of the changed parameters, and adjust the allowable time interval. Adjust the strength standard of the conduction response based on the changed strength matching data, and optimize the strength adaptation requirements; If the change is a short-term, accidental change, maintain the core parameters of the original conduction response template, and temporarily adjust the time range and intensity standard corresponding to the changed part; By integrating and adjusting the time range, intensity standards, and temporary adjustment information, a dynamic transmission response template is generated.

7. The AI ​​analysis method applied to the data model of the press section of a paper machine according to claim 3, characterized in that, The process involves tracing the source of each deviation node in the set of associated deviation nodes, the set of transmitted deviation nodes, and the set of coordinated deviation nodes, mining the upstream node influence path, behavioral transmission link, and external factors that caused the deviation node, and obtaining the deviation source tracing results, including: Starting from each deviation node, the upstream related nodes in the dynamic behavior inversion results of the paper machine press section data model are traced back in reverse. The relationship and transmission path between the deviation node and the upstream node are identified, and the range of upstream nodes that may affect the deviation node is determined to obtain the range of upstream related nodes. Analyze the behavior state of each node in the upstream associated node range, determine whether the operation mode of the upstream node conforms to the model's preset logic, identify the directly upstream influencing nodes that cause the deviation nodes, and obtain information on the directly influencing nodes; Identify the behavioral transmission links from directly influencing nodes to deviation nodes, clarify the intermediate nodes, transmission methods, and mechanisms of action in the transmission process, and obtain details of the deviation transmission links; Investigate the changes in external environmental factors corresponding to deviation nodes, analyze whether there are abnormal environmental factors and changes in environmental interaction behavior imprints, determine whether external environmental factors are the cause of deviation, and obtain information on external factors. By integrating the upstream associated node range, directly influencing node information, deviation transmission link details, and external factor information, a deviation tracing result corresponding to each deviation node is formed.

8. The AI ​​analysis method applied to the data model of the press section of a paper machine according to claim 1, characterized in that, The AI ​​analysis process for iteratively optimizing the paper machine press section data model based on the dynamic behavior inversion results after closed-loop calibration integrates the iterative inversion results and dynamic baseline logic to obtain an iterative analysis report of the paper machine press section data model, including: The dynamic behavior inversion results after closed-loop calibration are analyzed, and the behavior correlation logic within the optimized parameter adjustment dimension, the behavior correlation path between parameter output dimensions, and the multi-dimensional synergistic logic of environmental parameter output are extracted from the results to obtain the core content of model optimization. Based on the core optimization of the model, the AI ​​analysis process of the paper machine press section data model is adjusted, and the operation logic of the data acquisition, behavior inversion, benchmark generation and calibration links in the analysis process is optimized to obtain the iteratively optimized AI analysis process. The iteratively optimized AI analysis process is applied to the operation analysis of the data model of the paper machine press section. Model operation data, benchmark adjustment data and calibration implementation data are collected during the analysis process to obtain iterative analysis process data. By comparing the AI ​​analysis process operation data before and after iterative optimization, the impact of process optimization on model inversion accuracy, benchmark adaptability and calibration implementation effect is analyzed, resulting in a process optimization impact analysis. The dynamic behavior inversion results after closed-loop calibration, the dynamic baseline logic of the paper machine press department data model, the AI ​​analysis process after iterative optimization, the data of the iterative analysis process and the impact analysis of process optimization are integrated and organized according to the structured requirements of the analysis report to generate the iterative analysis report of the paper machine press department data model.

9. An AI analysis system applied to the data model of the press section of a paper machine, characterized in that, include: processor; A machine-readable storage medium for storing machine-executable instructions of the processor; The processor is configured to execute the AI ​​analysis method for the data model of the paper machine press section as described in any one of claims 1 to 8 by executing the machine-executable instructions.

Images