Intelligent regulation and control method and system of tension equipment for lining cloth production

Abstract

The invention discloses an intelligent regulation and control method and system for tension equipment for lining cloth production, and relates to the technical field of intelligent regulation and control, and the method comprises the steps: constructing a lining cloth production tension database, carrying out twinborn simulation, and generating a digital twinborn body of the tension equipment; tension matching optimization is carried out, and initial tension equipment control parameters are determined; lining cloth production is executed, meanwhile, lining cloth production multi-mode data flow in the lining cloth production process is collected, simulation prediction is conducted, and lining cloth tension prediction parameters are obtained; and if the lining cloth tension prediction parameter exceeds the preset tension fluctuation threshold, triggering a closed-loop processing mechanism to call a tension compensation regulation strategy, and performing compensation correction and dynamic regulation on the initial tension equipment control parameter based on the tension compensation regulation strategy. The technical problems that in the prior art, tension control parameter setting depends on artificial experience, and lining cloth quality fluctuates due to untimely regulation and control are solved, and the technical effects of improving tension control precision, real-time response capacity and lining cloth quality stability are achieved.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, specifically to an intelligent control method and system for tension equipment used in lining production. Background Technology

[0002] In the existing lining production process, the control parameters of tension equipment are mostly set based on the operator's experience, lacking systematic data support and intelligent optimization mechanisms. This manual setting method makes it difficult to accurately match the tension requirements of linings of different specifications, and it is impossible to achieve real-time prediction and automatic adjustment of tension fluctuations during production. This leads to lag in control and untimely response, which in turn causes problems such as lining breakage, deformation, or unstable quality, thus restricting the efficiency and quality improvement of lining production. Summary of the Invention

[0003] This application provides an intelligent control method and system for tension equipment used in lining production, which addresses the technical problem that tension control parameter settings rely on manual experience and that untimely control leads to fluctuations in lining quality in the prior art.

[0004] In view of the above problems, this application provides an intelligent control method and system for tension equipment used in lining production.

[0005] A first aspect of this application provides an intelligent control method for tension equipment used in lining production, the method comprising: Based on production data mining of the tension equipment used for target lining production, a tension database for lining production is constructed. A digital twin simulation is then performed using the structural design information of the tension equipment and the tension database to generate a digital twin of the tension equipment. Lining specifications are input into the digital twin for tension matching and optimization to determine initial tension equipment control parameters. Lining production is executed based on these initial control parameters, while simultaneously collecting multimodal data streams of lining production. The digital twin of the tension equipment is then used to simulate and predict the multimodal data streams, yielding predicted lining tension parameters. If the predicted lining tension parameters exceed a preset tension fluctuation threshold, a closed-loop processing mechanism is triggered, invoking a tension compensation and control strategy. Based on this strategy, the initial tension equipment control parameters are compensated, corrected, and dynamically controlled.

[0006] A second aspect of this application provides an intelligent control system for tension equipment used in lining production, the system comprising: The data mining module is used to mine production data based on the tension equipment used for target lining production, construct a tension database for lining production, and perform twin simulation based on the structural design information of the tension equipment and the tension database to generate a digital twin of the tension equipment. The tension matching module is used to input lining specification parameters into the digital twin of the tension equipment for tension matching optimization, determining the initial tension equipment control parameters. The lining production module is used to execute lining production based on the initial tension equipment control parameters, simultaneously collecting multimodal data streams of lining production during the production process, and loading the tension equipment digital twin to simulate and predict the multimodal data streams to obtain predicted lining tension parameters. The control module is used to trigger a closed-loop processing mechanism to call a tension compensation control strategy if the predicted lining tension parameters exceed a preset tension fluctuation threshold, and to compensate, correct, and dynamically control the initial tension equipment control parameters based on the tension compensation control strategy.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application utilizes production data mining based on tension equipment used in target lining fabric production to construct a tension database for lining fabric production. A digital twin simulation is then performed based on the structural design information of the tension equipment and the tension database to generate a digital twin of the tension equipment. Lining fabric specifications are input into the digital twin for tension matching and optimization to determine initial tension equipment control parameters. Lining fabric production is executed based on these initial control parameters, while simultaneously collecting multimodal data streams of lining fabric production. The digital twin of the tension equipment is then used to simulate and predict these multimodal data streams, yielding predicted tension parameters for the lining fabric. If the predicted tension parameters exceed a preset tension fluctuation threshold, a closed-loop processing mechanism is triggered, invoking a tension compensation and control strategy. Based on this strategy, the initial tension equipment control parameters are compensated, corrected, and dynamically controlled. This invention addresses the technical problem in existing technologies where tension control parameter settings rely on manual experience, leading to untimely adjustments and fluctuations in lining fabric quality. By constructing a tension database and using a digital twin for intelligent simulation, matching, and closed-loop dynamic control, it achieves improved tension control accuracy, real-time response capability, and lining fabric quality stability. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A schematic diagram of the intelligent control method for tension equipment used in lining production provided in this application embodiment; Figure 2 A schematic diagram of the intelligent control system structure of the tension equipment for lining production provided in this application embodiment.

[0010] Figure labeling: Data mining module 11, tension matching module 12, lining production module 13, control module 14. Detailed Implementation

[0011] This application provides an intelligent control method and system for tension equipment used in lining production. It addresses the technical problem in the prior art where tension control parameter settings rely on manual experience and untimely control leads to fluctuations in lining quality. By constructing a tension database and a digital twin for intelligent simulation, matching, and closed-loop dynamic control, it achieves the technical effect of improving tension control accuracy, real-time response capability, and lining quality stability.

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0013] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0014] Example 1, as Figure 1 As shown, this application provides an intelligent control method for tension equipment used in lining production, the method comprising: Step S100: Based on the tension equipment for target lining production, perform production data mining to construct a tension database for lining production, and perform twin simulation based on the structural design information of the tension equipment for target lining production and the tension database for lining production to generate a digital twin of the tension equipment.

[0015] In this embodiment, when performing production data mining based on the tension equipment used for target lining production, data from the historical operation of the tension equipment is first collected to obtain a historical lining production dataset. This dataset is then categorized and labeled into three levels: tension equipment layer production data, lining process layer production data, and lining quality layer production data. Next, through analysis and integration of these three types of data, logical relationships are established, and association mapping alignment is achieved, thereby forming a lining production tension-related dataset. Finally, data cleaning and unified storage integration operations are performed on this dataset to complete the construction of the lining production tension database.

[0016] Next, a twin simulation is performed based on the structural design information of the tension equipment used for target lining production and the lining production tension database. In this process, firstly, based on the structural design information of the target lining production tension equipment, dimensionality reduction is performed according to the equipment modeling accuracy requirements to obtain the dimensionality-reduced equipment structural design parameters. Then, equipment assembly modeling and motion constraints are assigned to construct a 3D model of the tension equipment. Subsequently, the control objectives of the tension equipment are defined, corresponding tasks are extracted, and a tension equipment control task list is formed. Finally, based on this task list and the lining production tension database, a twin simulation is performed on the 3D model to generate a digital twin of the tension equipment.

[0017] Furthermore, in the method provided in the application embodiments, the construction of the lining production tension database further includes: Historical production data is mined based on the tension equipment used in the production of the target lining fabric to obtain a historical production dataset of the lining fabric. The historical production dataset of the lining fabric is classified and labeled to obtain production data of the tension equipment layer, production data of the lining fabric process layer, and production data of the lining fabric quality layer. The production data of the tension equipment layer, the production data of the lining fabric process layer, and the production data of the lining fabric quality layer are correlated and mapped to obtain a tension-related dataset of lining fabric production. The tension-related dataset of lining fabric production is cleaned, stored, and integrated to construct the tension database of lining fabric production.

[0018] In this embodiment, when mining historical production data based on the tensioning equipment used for target lining production, the system first connects the control system of the tensioning equipment to the data acquisition device to extract the original operating records of the equipment under different production batches and operating conditions. The collected data includes tension setpoints, actual tension feedback values, equipment operating status parameters, lining specification parameters, processing time, etc., constituting lining production process data. After being buffered and transferred at the acquisition end, the above information is uniformly summarized to form a historical dataset of lining production.

[0019] Next, the historical dataset of lining production is classified and labeled. In this process, a field attribute recognition method is used to perform structured classification of the historical dataset. By determining the source device, collection location, recorded content, and role of data fields in the production process, data directly reflecting the action parameters of tension equipment are labeled as tension equipment layer production data, such as tension roller speed and electromagnetic tensioner current value. Settings and execution parameters directly related to the process flow are labeled as lining process layer production data, such as fabric weight, tension setpoint, and fabric width. Quality inspection indicators obtained through online or offline testing instruments after product processing are labeled as lining quality layer production data, such as tension uniformity, edge wrinkling, and elongation at break.

[0020] Subsequently, the production data of the tension equipment layer, the production data of the lining process layer, and the production data of the lining quality layer are associated and mapped. By adopting the primary key association method, the timestamp, equipment number, and production batch number are used as unified identification fields to map the three types of data one by one, ensuring that each record can be traced to the actual working conditions, set parameters, and quality results, thereby forming a tension-related dataset for lining production.

[0021] Finally, a systematic data cleaning and format standardization process was performed on the completed tension-related dataset for lining production. Methods such as missing value handling, outlier removal, data deduplication, and unit standardization were used to address noise, redundancy, and inconsistencies in the dataset. The cleaned data was then divided according to structural hierarchy, and logical tables were created based on field relationships. This data was stored in a unified database management system, completing the structured integration of tension-related information and ultimately resulting in the lining production tension database.

[0022] Furthermore, in the method provided in the application embodiments, the generation of a digital twin of the tension device further includes: According to the equipment modeling accuracy requirements, the structural design information of the tension equipment for the target lining production is reduced in dimension to obtain the reduced structural design parameters. Based on the reduced structural design parameters, equipment assembly modeling and motion constraints are assigned to construct a three-dimensional model of the tension equipment. The control target of the tension equipment is obtained, and the control target is extracted to obtain a list of control tasks. Based on the tension database for lining production, a twin simulation of the three-dimensional model of the tension equipment is performed according to the list of control tasks to generate a digital twin of the tension equipment.

[0023] In this embodiment, the structural design information of the tension equipment for producing the target lining fabric is first reduced in dimensionality according to the equipment modeling accuracy requirements. This process uses feature extraction methods to filter and simplify parameters from the original structural design drawings and CAD models of the tension equipment for producing the target lining fabric. Specifically, all parameter items in the structural design information are identified, and geometric details and auxiliary components unrelated to tension control are eliminated based on the actual application requirements of tension control. Only key parameters that directly affect the modeling results are retained, such as the diameter, wheelbase, rotation center of the tension roller group, and the position and installation method of the tension loading device, etc., ultimately yielding the simplified, dimensionality-reduced equipment structural design parameters.

[0024] Subsequently, based on the structural design parameters of the reduced-dimensional equipment, equipment assembly modeling and motion constraints are assigned. This step is implemented using 3D modeling software (such as SolidWorks). Basic component models are established based on the structural parameters after dimensional reduction, and these models are assembled according to equipment assembly relationships. The motion relationships between components are defined by setting connection methods such as revolute joints and prismatic joints. For example, the tension loading arm and the support are set as rotational constraints, and the guide roller and the support are set as sliding constraints to ensure that the motion behavior of each component after modeling is consistent with the actual equipment, thereby constructing a 3D model of the tension equipment.

[0025] Next, the control objectives of the tension equipment are obtained, and tasks are extracted from these objectives to obtain a list of tension equipment control tasks. This step employs document analysis, analyzing the equipment control manual, production process specifications, and PLC program descriptions to clarify the tension equipment control objectives (such as maintaining constant tension, responding to changes in fabric speed, and dynamically adjusting tension output). During the analysis, each control objective is transformed into an executable control task and listed according to the order of operations, such as real-time acquisition of fabric speed signals, calculation of tension error, execution of PID control algorithms, and output of correction signals, ultimately forming the tension equipment control task list.

[0026] Finally, a twin simulation of the tension equipment's 3D model was performed based on the lining production tension database, according to the tension equipment control task list. In this process, each control task in the tension equipment control task list was first sequentially associated and rectified with the lining production tension database, extracting historical data and control variables corresponding to each control task to form a task-specific lining tension task association dataset. Subsequently, based on the lining tension task association dataset, algorithm model selection and control training optimization were performed sequentially to generate a multi-task lining tension controller set with multi-task control capabilities. Next, the multi-task lining tension controller set was mapped to the tension equipment's 3D model for coupled twin simulation, constructing an initial equipment digital twin. Finally, scene instance injection and simulation verification optimization were performed on the initial equipment digital twin to generate the tension equipment digital twin.

[0027] Furthermore, in the method provided in the application embodiment, the step of performing a twin simulation of the three-dimensional model of the tension equipment based on the lining production tension database according to the tension equipment control task list to generate a digital twin of the tension equipment further includes: Each control task in the tension equipment control task list is sequentially associated and rectified with the lining production tension database to obtain a lining tension task association dataset. Based on the lining tension task association dataset, algorithm model selection and control training optimization are performed sequentially to generate a multi-task lining tension controller set. The multi-task lining tension controller set is mapped to the tension equipment 3D model for coupled twin simulation to obtain an initial equipment digital twin. Scene instance injection and simulation verification optimization are performed on the initial equipment digital twin to generate a tension equipment digital twin.

[0028] In this embodiment, the control tasks in the tension equipment control task list are first sequentially associated and rectified with the lining production tension database. During this process, the control objectives explicitly stated in the tension equipment control task list are used as retrieval criteria, such as real-time tension deviation identification or tension maintenance under high-speed operation. Production data related to the tension equipment layer, lining process layer, and lining quality layer are sequentially extracted from the lining production tension database. By unifying timestamps, equipment numbers, and production batch information, logical association between data from different sources is achieved, completing data structure alignment and semantic integration, ultimately constructing a lining tension task association dataset reflecting the task control logic.

[0029] Subsequently, based on the lining tension task-related dataset, algorithm model selection and control training and optimization were performed sequentially. This process employed supervised learning methods for controller construction. Specifically, based on the input variables (such as tension sensor feedback) and target output (such as the adjusted tension setpoint) for each control task, the type of control strategy model to be used was determined. For example, an adaptive PID control model was selected for dynamic response adjustment tasks, and a feedforward compensation model was selected for complex working condition identification tasks. Then, using the lining tension task-related dataset as training samples, the controller performance was gradually optimized through controller parameter learning and simulation trials, ultimately forming a set of directly callable multi-task lining tension controllers, each with task-specific adjustment logic and response models.

[0030] Next, the multi-task lining tension controller set is mapped onto the 3D model of the tension equipment for coupled twin simulation to achieve dynamic integration of control logic and equipment structural behavior. Specifically, in the completed 3D model of the tension equipment, by setting control input ports and feedback interfaces, the controller's input signals (such as tension deviation) are connected to the tension sensor acquisition points, and the controller's output signals (such as tension adjustment values) are bound to the tension loading device execution unit, realizing data interaction between the virtual controller and the 3D structure. For example, the output port of the constant tension controller directly drives the angular velocity adjustment node of the tension roller assembly in the 3D model to simulate the tension loading change of the virtual fabric. By loading all controllers into the 3D model through a one-to-one mapping method and completing the overall linkage control simulation, an initial digital twin of the equipment with a complete control behavior response mechanism is obtained.

[0031] Finally, scenario instances were injected and simulation verification were performed to optimize the initial equipment digital twin. This process began by pre-setting a set of lining production scenario sets and associating instance data based on these sets to obtain lining tension scenario instance data. This lining tension scenario instance data was then injected into the initial equipment digital twin for tension simulation testing, and the scenario instance simulation data was recorded for analysis. Subsequently, the scenario instance simulation data was verified and evaluated according to the twin's performance indicators to obtain quantifiable twin control performance parameters. Finally, based on the twin's control performance parameters, the initial equipment digital twin underwent iterative parameter optimization, completing the tuning and upgrading of the digital model across multiple scenarios and generating a tension equipment digital twin that meets the target control requirements.

[0032] Furthermore, in the method provided in the application embodiments, the step of injecting scene instances and performing simulation verification optimization on the initial device digital twin to generate a tension device digital twin further includes: A preset set of interlining production condition scenarios is established. Instance data is associated with these scenarios to obtain interlining tension scenario instance data. This interlining tension scenario instance data is then injected into the initial equipment digital twin for tension simulation testing, and the scenario instance simulation data is recorded. The scenario instance simulation data is verified and evaluated according to the twin's performance indicators to obtain twin control performance parameters. Based on these twin control performance parameters, the initial equipment digital twin is iteratively optimized to generate the tension equipment digital twin.

[0033] In this embodiment, a set of interlining production scenario sets is first preset. This set covers common situations in actual production, such as fabric speed fluctuations, fabric width changes, environmental disturbances, material switching, and tension anomalies. The fabric speed fluctuation scenario reflects the tension change requirements caused by acceleration, deceleration, or unstable operation during production. The fabric width change scenario considers the adjustment pressure on the tension control system after different widths of fabric enter the equipment. The environmental disturbance scenario simulates the impact of drastic temperature or humidity changes on the fabric tension performance. The material switching scenario covers the tension adaptation during the alternating production of fabrics with different weights, thicknesses, and yarn counts. The tension anomaly triggering scenario reproduces abnormal states such as tension jumps, control failures, or feedback signal delays that have occurred in historical data.

[0034] Subsequently, based on the set of lining production scenarios, instance data was correlated. By setting process characteristic parameters for each scenario, such as the fabric speed setting range, tension control interval, and ambient temperature and humidity levels, historical batch data meeting the corresponding conditions were extracted from the lining production tension database. The extracted data included production data from the tension equipment layer, lining process layer, and lining quality layer. The data was then integrated through timestamp synchronization, field unification, and structural reorganization to form instance data of lining tension scenarios.

[0035] Next, the fabric tension scenario example data is injected into the initial equipment digital twin for tension simulation testing. The simulation process runs in a virtual simulation environment. The fabric speed curve, target tension value, and environmental conditions from the scenario example data are loaded into the initial equipment digital twin, driving its internal multi-task fabric tension controller set to execute dynamic tension control tasks. During the simulation, information such as the fabric tension response process, changes in control output signals, error feedback, actuator behavior, and virtual fabric tension distribution map are recorded in real time, forming scenario example simulation data.

[0036] After completing the tension simulation test, the simulated data of the scenario examples were verified and evaluated according to the twin performance indicators. The evaluation process used pre-set tension control targets as the judgment benchmark, including performance indicators such as tension fluctuation not exceeding ±5N, response time not exceeding 2 seconds, and stable control output without oscillation. By analyzing the tension change curves, controller action frequency, and response delay recorded in the scenario example simulation data, the maximum error, average response time, and tension adjustment frequency were extracted to quantify the twin control performance parameters.

[0037] Finally, the initial digital twin of the equipment was iteratively optimized based on the twin's control performance parameters. The optimization process involved analyzing performance deviations in the twin's control performance parameters and adjusting parameters such as the proportional gain, integral time, and derivative coefficients of the multi-task lining tension controller to improve the controller's control accuracy and response speed in various scenarios. After each round of parameter modification, the same lining tension scenario instance data was reloaded for simulation testing, and the control effects before and after optimization were compared. If the performance requirements were still not met, adjustments were continued. After multiple rounds of iterative optimization, a digital twin of the tension equipment that met all performance requirements was finally generated.

[0038] Step S200: Input the lining specifications into the digital twin of the tension device to perform tension matching optimization and determine the initial tension device control parameters.

[0039] In this embodiment, the lining specifications are input into the digital twin of the tension equipment for tension matching and optimization. Specifically, the key process parameters of the lining to be produced are first collected and organized, including but not limited to fabric width, weight, thickness, raw material, and weaving structure, which constitute the complete lining specifications. These specifications are then input into the constructed digital twin of the tension equipment through a human-machine interface or MES system.

[0040] After input is completed, the associated lining production tension database is accessed within the tension equipment digital twin to filter historical sample data that matches the input lining specifications. This database consists of structured production data for the tension equipment layer, lining process layer, and lining quality layer. Through field mapping and parameter similarity calculation, representative historical data sets are identified and used as the input basis for the tension matching optimization stage.

[0041] Next, the digital twin of the tension equipment performs virtual simulation calculations on multiple candidate combinations of tension equipment control parameters, including variables such as tension setpoint, tension loading delay, equipment execution frequency, and controller adjustment amplitude. Each set of control parameters corresponds to a complete tension control process during the simulation, automatically recording key indicators such as tension response curve, tension error change, adjustment delay time, and fluctuation range, forming simulation results for performance analysis.

[0042] To achieve optimal performance, each set of simulation results is scored according to a preset scoring rule. The scoring dimensions include tension error score, tension response time score, tension fluctuation control score, and control output stability score, each assigned a score based on the magnitude of the error in the simulation output, the time required for the tension to reach a stable state, the fluctuation amplitude, and the smoothness of the control curve. Each dimension uses a percentage-based scoring method with assigned weights; for example, the tension error score has a weight of 40%, the tension response time score has a weight of 25%, the tension fluctuation control score has a weight of 20%, and the control stability score has a weight of 15%. The scores for each dimension are weighted and summed to calculate the final comprehensive score for each set of control parameters. For example, when calculating the tension error score, if the target steady-state tension value is 12.0 N, and the simulated output stable tension is 11.6 N, then the absolute tension error is 0.4 N. An error less than or equal to 0.2 N is set as a maximum score of 100 points, with linear deductions for errors between 0.2 N and 1.0 N, and a score above 1.0 N as a failing grade (less than 60 points). Therefore, this item scores 75 points. Regarding the tension response time score, the target is to reach a steady-state range of ±2% within 5 seconds. If simulation data shows that the tension reaches this range within 7 seconds, then according to the principle of deducting 10 points per second for exceeding the set standard, this item scores 80 points. The tension fluctuation control score is evaluated by observing the fluctuation amplitude of tension in the steady-state phase during simulation. A fluctuation amplitude less than ±0.2N is set at a maximum score of 100 points, and points are deducted proportionally for fluctuations between ±0.2N and ±1.0N. For example, if the fluctuation is ±0.4N, the score is 80 points. The control output smoothness score is evaluated based on the smoothness of the control command curve. For example, by analyzing the number of abrupt jumps in the control curve using first-order difference analysis, a maximum score of 100 points is set at no more than 2 jumps, and 10 points are deducted for each additional jump. If there are 5 jumps, the score is 70 points. The scores of each item are weighted and summed to calculate the final comprehensive score for each set of control parameters.

[0043] Based on the scoring results, the set of tension control parameters with the highest score is selected as the optimal configuration under the current lining specifications, thereby determining the initial tension equipment control parameters.

[0044] Step S300: Execute lining production based on the initial tension equipment control parameters, and simultaneously collect the lining production multimodal data stream during the lining production process. Then, load the tension equipment digital twin to simulate and predict the lining production multimodal data stream to obtain the lining tension prediction parameters.

[0045] In this embodiment, lining production is first executed based on initial tension equipment control parameters, while simultaneously collecting multimodal data streams of lining production. During this process, a monitoring requirement analysis is first performed on the tension equipment used for lining production to clarify the scope and types of data to be collected, thus constructing equipment monitoring requirement dimensions. These dimensions cover the geometric and mechanical properties of the lining, equipment status, and environmental characteristics. Subsequently, based on these monitoring requirement dimensions, sensor selection and layout analysis are conducted sequentially, and specific sensor configuration parameters corresponding to each monitoring dimension are determined, ultimately forming a monitoring dimension-sensor parameter set. Based on this, a multimodal sensing network is deployed according to monitoring accuracy and layout requirements. Through this multimodal sensing network, multimodal data streams of lining production are collected and acquired in real time.

[0046] Subsequently, the multimodal data stream of the lining production is input into the established digital twin of the tension equipment for conducting virtual predictive simulation based on real-time data. After loading the current operating data, the digital twin of the tension equipment reconstructs a virtual operating state synchronized with the field within its embedded simulation environment and performs a simulation of the tension response trend. In the simulation calculation, predictive identification is performed on the tension change trend, response delay, equipment control lag, and abnormal tension signals, and finally, a set of predicted lining tension parameters based on the current operating conditions is output.

[0047] Furthermore, in the method provided in the application embodiments, the step of collecting the multimodal data stream of lining production during the lining production process further includes: A monitoring requirement analysis is performed on the tension equipment used in the lining production process to determine the monitoring requirement dimensions, which include the geometric and mechanical properties of the lining, the equipment status, and the environmental characteristics. Sensor selection and layout analysis are then performed on the tension equipment according to these monitoring requirement dimensions to obtain a monitoring dimension-sensor parameter set. Based on this monitoring dimension-sensor parameter set, a multimodal sensing network is deployed and constructed, and the multimodal data stream of the lining production process is acquired through this network.

[0048] In this embodiment, a monitoring requirement analysis is first conducted on the tension equipment used in lining production to clarify the scope of key information to be collected during actual operation and to determine the dimensions of equipment monitoring requirements. These dimensions include the geometric and mechanical properties of the lining, the equipment status, and environmental characteristics. Specifically, the geometric properties of the lining refer to its physical structure during production, such as width, edge shape, and amplitude of movement; the mechanical properties refer to the stress parameters, such as tension changes and strain characteristics, experienced by the lining during operation; the equipment status mainly includes the rotational speed, position, loading status, and system response time of the tension control rollers; and the environmental characteristics include external factors affecting tension stability, such as temperature and humidity in the production environment.

[0049] Subsequently, sensor selection and layout analysis were conducted for the tension equipment used in lining production according to the equipment monitoring requirements. During this process, non-contact sensors such as laser rangefinders and line-scan cameras were selected to measure fabric width and edge fluctuations, addressing the fabric's geometric characteristics. Tension sensors and strain gauges were configured to collect real-time fabric tension data, considering mechanical properties. Speed ​​encoders and motor feedback devices were installed to monitor the equipment's drive response, and temperature and humidity sensors and flow meters were installed to measure factors affecting tension stability in the production environment. Technical experts summarized and organized the models, accuracy, installation locations, and output parameters of various sensors, forming a comprehensive monitoring dimension – a sensor parameter set.

[0050] Based on this, and using the monitoring dimension—sensor parameter set—the sensors are deployed in the field, forming a data acquisition system composed of multiple sensor types, i.e., a multimodal sensing network is constructed. This network achieves synchronous, multidimensional data acquisition of lining tension by deploying different types of sensors in various monitoring areas of the equipment. Through this multimodal sensing network, multimodal data streams of lining production are acquired during the continuous production process. These multimodal data streams include real-time tension change information, fabric movement status, equipment drive characteristics, and environmental parameter changes.

[0051] Step S400: If the predicted tension parameter of the lining exceeds the preset tension fluctuation threshold, the closed-loop processing mechanism is triggered to call the tension compensation and control strategy, and the initial tension equipment control parameters are compensated, corrected and dynamically controlled based on the tension compensation and control strategy.

[0052] In this embodiment, the predicted tension parameters of the lining fabric are first compared with the preset tension fluctuation threshold in real time. The preset tension fluctuation threshold is set by technical experts based on the lining fabric material, specifications, and historical control error fluctuation range.

[0053] If the predicted parameters exceed the preset tension fluctuation threshold, a closed-loop processing mechanism is immediately triggered, invoking the tension compensation and control strategy. In this process, the current multimodal data stream of the lining production is first analyzed to identify the specific source of the tension anomaly and extract the cause of the anomaly. Then, the existing lining tension anomaly operation and maintenance strategy library is invoked to match the identified causes of the anomaly, thereby determining the corresponding tension compensation and control strategy.

[0054] Finally, based on the tension compensation and control strategy, the initial tension equipment control parameters are compensated, corrected, and dynamically controlled. Key parameters such as tension setpoint, feedback gain, response period, and loading amplitude are adjusted to achieve adaptive response to abnormal disturbances and maintain the stability of tension control during the lining production process.

[0055] Furthermore, in the method provided in the application embodiments, the triggering closed-loop processing mechanism to invoke the tension compensation control strategy further includes: A closed-loop processing mechanism is triggered to analyze the causes of anomalies in the multimodal data stream of the lining production, thereby obtaining the causes of abnormal lining tension. A lining tension anomaly operation and maintenance strategy library is constructed, and the cause of the abnormal lining tension is matched and called using the lining tension anomaly operation and maintenance strategy library to obtain the tension compensation and control strategy.

[0056] In this embodiment, a closed-loop processing mechanism is triggered to analyze the causes of anomalies in the multimodal data stream of lining production. This process first analyzes the collected multimodal data stream of lining production to identify the root cause of tension anomalies. This multimodal data stream includes data channels from different sensors, specifically including lining tension signals, changes in fabric speed and roller speed, fabric image status, equipment drive feedback signals, and changes in ambient temperature and humidity. For effective analysis, a key variable trend comparison analysis method is used. This involves selecting multiple key variable pairs with causal relationships (e.g., "fabric speed-tension," "temperature and humidity-tension," "loading position-tension," etc.) and comparing their trends through a sliding time window to see if there are obvious response lags, nonlinear fluctuations, or synchronization shifts. For example, if a sudden increase in fabric speed is detected over a period of time, and the tension signal shows a significant lag increase, accompanied by an accelerated fluctuation frequency in the loading roller feedback data, then it is determined that there is a tension anomaly caused by "untimely loading response." If a rapid increase in ambient humidity is observed in another data segment, while the overall fabric tension signal shifts upward and the fluctuations increase, it can be identified as "humidity changes causing a sudden change in fabric stress characteristics." Through joint evolutionary analysis of these variables, a structured description of the causes of abnormal lining tension is finally output.

[0057] Subsequently, the established lining tension anomaly operation and maintenance strategy library was invoked to match strategies for the identified causes of lining tension anomalies. This strategy library has compiled typical anomaly causes and corresponding control measures, such as adjusting buffer settings for unstable fabric speed, correcting response delays for untimely loading response, and calibrating tension baseline values ​​for environmental disturbances. Through conditional comparison, the current anomaly cause was matched one-to-one with the anomaly types defined in the strategy library to determine the control scheme suitable for the current production state. Finally, based on the matching results, a tension compensation control strategy was obtained.

[0058] Furthermore, in the method provided in the application embodiments, the step of constructing the lining tension anomaly operation and maintenance strategy library further includes: A case library of abnormal lining tension is collected, and the causes of each abnormality in the case library are analyzed for operation and maintenance strategies to obtain a set of abnormality causes-tension operation and maintenance strategies. Based on the set of abnormality causes-tension operation and maintenance strategies, tension compensation training is performed sequentially to construct a lining tension abnormality operation and maintenance strategy library.

[0059] In this embodiment, a database of abnormal tension cases is first compiled by collecting examples of tension fluctuations that occur during actual lining production. During data acquisition, a multi-channel synchronous acquisition method is used to align tension sensor data, fabric speed information, roller speed feedback, environmental temperature and humidity changes, operation records, and image detection data by timestamp, ensuring comparability and temporal consistency of data from different sources. This database of abnormal tension cases covers the entire process before and after the occurrence of tension anomalies.

[0060] Subsequently, operational strategies were analyzed for each anomaly cause in the lining tension anomaly case library. This process employed variable causal inference analysis combined with fault tracing and discrimination. Specifically, the tension anomaly event was first treated as a response variable. The changing trends and abrupt changes of various process variables (such as fabric speed, tension setpoint, loading response time, etc.) before and after its occurrence were analyzed. The cross-correlation and time delay response characteristics between variables were calculated to determine the most likely anomaly cause. For example, if a sudden increase in tension and a loading roller response delay occur simultaneously, with a lag of approximately 2 seconds, it can be preliminarily determined that the anomaly was caused by a loading adjustment lag. After identifying the cause, the control operation records from historical cases (such as the operator's adjustment range of the tension setpoint, modifications to the feedback coefficient, etc.) were combined with the operation log behavior extraction method to extract the correlation patterns between adjustment actions and causes. This resulted in a structured set of anomaly causes and their corresponding control schemes, outputting an anomaly cause-tension operational strategy set.

[0061] Next, tension compensation training is performed sequentially based on the anomaly cause-tension operation and maintenance strategy set. During this process, based on the characteristic information of the anomaly cause-tension operation and maintenance strategy set, a suitable tension compensation strategy algorithm network set is selected, and compensation control training and iterative testing and optimization are carried out for various anomaly causes to obtain a tension compensation strategy controller set with targeted control effects. Subsequently, this controller set is uniformly identified, classified, and integrated, ultimately forming a structured lining tension anomaly operation and maintenance strategy library.

[0062] Furthermore, in the method provided in the application embodiments, the step of sequentially performing tension compensation training based on the anomaly cause-tension maintenance strategy set to construct a lining tension anomaly maintenance strategy library further includes: Based on the characteristic information of the anomaly cause-tension operation and maintenance strategy set, a tension compensation strategy algorithm network set is selected; compensation control training and iterative testing optimization are performed on the tension compensation strategy algorithm network set to obtain a tension compensation strategy controller set, and the tension compensation strategy controller set is identified and integrated to construct the lining tension anomaly operation and maintenance strategy library.

[0063] In this embodiment, the control requirements of each anomaly cause in the anomaly cause-tension operation and maintenance strategy set are first analyzed based on the characteristics of the control requirements. This analysis examines the types of disturbances, control targets, and key influencing variables involved, such as whether loading unit hysteresis response, fabric speed disturbance, or feedback lag is involved. Based on this analysis, a tension compensation strategy algorithm network set with a matching control logic structure is selected. This tension compensation strategy algorithm network set is a collection of various typical industrial control strategies, including proportional-integral control structures, feedforward-feedback coordination structures, dynamic compensation structures, and variable parameter control structures. Different strategies correspond to different types of anomaly control targets. For example, a feedforward-lag compensation control structure can be used for anomalies such as sudden increases in lining tension, while a steady-state offset suppression structure is used for anomalies such as slow tension changes.

[0064] Subsequently, compensation control training was conducted based on the selected tension compensation strategy algorithm network set. The training method involved applying simulated tension anomalies and performing multiple rounds of control parameter training and adjustment on the control structure. During the training process, a virtual simulation environment was constructed, with input variables consistent with the current anomaly cause, such as the fabric speed fluctuation curve, loader delay response data, or temperature and humidity fluctuation parameters. The control strategy structure was embedded in the simulation process and executed, and the tension output change process was collected in real time to evaluate its control capability during dynamic changes. The collected control performance indicators included tension error amplitude, time required to recover stability, and whether the control response overshooted. Based on these indicators, optimization methods such as gradient descent or grid search were used to adjust the control parameters in the strategy structure (such as proportional gain, integral time constant, lag compensation duration, etc.) in multiple rounds to continuously optimize control performance. Through this step, a controller combination that meets the performance requirements was finally obtained, forming a tension compensation strategy controller set.

[0065] After controller training is completed, the obtained tension compensation strategy controller set is uniformly identified and its information integrated. The identification includes controller number, adaptation anomaly type, controller structure type, training parameter results, and applicable working condition boundary conditions. A hierarchical indexing method is used to classify and archive controller entries, forming an efficient and searchable retrieval mechanism. Through this step, a tension compensation strategy controller set with complete information tags, a clear structure, and well-defined adaptations is constructed.

[0066] Finally, all trained and fully labeled controllers are stored together to form a lining tension anomaly operation and maintenance strategy library for actual lining production process applications.

[0067] In summary, the embodiments of this application have at least the following technical effects: This application utilizes production data mining based on tension equipment used in target lining fabric production to construct a tension database for lining fabric production. A digital twin simulation is then performed based on the structural design information of the tension equipment and the tension database to generate a digital twin of the tension equipment. Lining fabric specifications are input into the digital twin for tension matching and optimization to determine initial tension equipment control parameters. Lining fabric production is executed based on these initial control parameters, while simultaneously collecting multimodal data streams of lining fabric production. The digital twin of the tension equipment is then used to simulate and predict these multimodal data streams, yielding predicted tension parameters for the lining fabric. If the predicted tension parameters exceed a preset tension fluctuation threshold, a closed-loop processing mechanism is triggered, invoking a tension compensation and control strategy. Based on this strategy, the initial tension equipment control parameters are compensated, corrected, and dynamically controlled. This invention addresses the technical problem in existing technologies where tension control parameter settings rely on manual experience, leading to untimely adjustments and fluctuations in lining fabric quality. By constructing a tension database and using a digital twin for intelligent simulation, matching, and closed-loop dynamic control, it achieves improved tension control accuracy, real-time response capability, and lining fabric quality stability.

[0068] Example 2, based on the same inventive concept as the intelligent control method for tension equipment used in lining production in the foregoing examples, such as... Figure 2 As shown, this application provides an intelligent control system for tension equipment used in lining production. The system and method embodiments in this application are based on the same inventive concept. The system includes: The data mining module 11 is used to mine production data based on the tension equipment for target lining production, construct a tension database for lining production, and perform twin simulation based on the structural design information of the tension equipment and the tension database to generate a digital twin of the tension equipment. The tension matching module 12 is used to input lining specification parameters into the digital twin of the tension equipment for tension matching optimization to determine the initial tension equipment control parameters. The lining production module 13 is used to execute lining production based on the initial tension equipment control parameters, while simultaneously collecting multimodal data streams of lining production during the lining production process, and loading the digital twin of the tension equipment to simulate and predict the multimodal data streams of lining production to obtain predicted lining tension parameters. The control module 14 is used to trigger a closed-loop processing mechanism to call a tension compensation control strategy if the predicted lining tension parameters exceed a preset tension fluctuation threshold, and to compensate, correct, and dynamically control the initial tension equipment control parameters based on the tension compensation control strategy.

[0069] Furthermore, the system is also used to implement the following functions: Historical production data is mined based on the tension equipment used in the production of the target lining fabric to obtain a historical production dataset of the lining fabric. The historical production dataset of the lining fabric is classified and labeled to obtain production data of the tension equipment layer, production data of the lining fabric process layer, and production data of the lining fabric quality layer. The production data of the tension equipment layer, the production data of the lining fabric process layer, and the production data of the lining fabric quality layer are correlated and mapped to obtain a tension-related dataset of lining fabric production. The tension-related dataset of lining fabric production is cleaned, stored, and integrated to construct the tension database of lining fabric production.

[0070] Furthermore, the system is also used to implement the following functions: According to the equipment modeling accuracy requirements, the structural design information of the tension equipment for the target lining production is reduced in dimension to obtain the reduced structural design parameters. Based on the reduced structural design parameters, equipment assembly modeling and motion constraints are assigned to construct a three-dimensional model of the tension equipment. The control target of the tension equipment is obtained, and the control target is extracted to obtain a list of control tasks. Based on the tension database for lining production, a twin simulation of the three-dimensional model of the tension equipment is performed according to the list of control tasks to generate a digital twin of the tension equipment.

[0071] Furthermore, the system is also used to implement the following functions: Each control task in the tension equipment control task list is sequentially associated and rectified with the lining production tension database to obtain a lining tension task association dataset. Based on the lining tension task association dataset, algorithm model selection and control training optimization are performed sequentially to generate a multi-task lining tension controller set. The multi-task lining tension controller set is mapped to the tension equipment 3D model for coupled twin simulation to obtain an initial equipment digital twin. Scene instance injection and simulation verification optimization are performed on the initial equipment digital twin to generate a tension equipment digital twin.

[0072] Furthermore, the system is also used to implement the following functions: A preset set of interlining production condition scenarios is established. Instance data is associated with these scenarios to obtain interlining tension scenario instance data. This interlining tension scenario instance data is then injected into the initial equipment digital twin for tension simulation testing, and the scenario instance simulation data is recorded. The scenario instance simulation data is verified and evaluated according to the twin's performance indicators to obtain twin control performance parameters. Based on these twin control performance parameters, the initial equipment digital twin is iteratively optimized to generate the tension equipment digital twin.

[0073] Furthermore, the system is also used to implement the following functions: A monitoring requirement analysis is performed on the tension equipment used in the lining production process to determine the monitoring requirement dimensions, which include the geometric and mechanical properties of the lining, the equipment status, and the environmental characteristics. Sensor selection and layout analysis are then performed on the tension equipment according to these monitoring requirement dimensions to obtain a monitoring dimension-sensor parameter set. Based on this monitoring dimension-sensor parameter set, a multimodal sensing network is deployed and constructed, and the multimodal data stream of the lining production process is acquired through this network.

[0074] Furthermore, the system is also used to implement the following functions: A closed-loop processing mechanism is triggered to analyze the causes of anomalies in the multimodal data stream of the lining production, thereby obtaining the causes of abnormal lining tension. A lining tension anomaly operation and maintenance strategy library is constructed, and the cause of the abnormal lining tension is matched and called using the lining tension anomaly operation and maintenance strategy library to obtain the tension compensation and control strategy.

[0075] Furthermore, the system is also used to implement the following functions: A case library of abnormal lining tension is collected, and the causes of each abnormality in the case library are analyzed for operation and maintenance strategies to obtain a set of abnormality causes-tension operation and maintenance strategies. Based on the set of abnormality causes-tension operation and maintenance strategies, tension compensation training is performed sequentially to construct a lining tension abnormality operation and maintenance strategy library.

[0076] Furthermore, the system is also used to implement the following functions: Based on the characteristic information of the anomaly cause-tension operation and maintenance strategy set, a tension compensation strategy algorithm network set is selected; compensation control training and iterative testing optimization are performed on the tension compensation strategy algorithm network set to obtain a tension compensation strategy controller set, and the tension compensation strategy controller set is identified and integrated to construct the lining tension anomaly operation and maintenance strategy library.

[0077] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0078] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0079] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. An intelligent control method for tension equipment used in lining production, characterized in that, The method includes: Based on the production data mining of the tension equipment used for target lining production, a tension database for lining production is constructed. Then, based on the structural design information of the tension equipment used for target lining production and the tension database for lining production, a twin simulation is performed to generate a digital twin of the tension equipment. Input the lining specifications into the digital twin of the tension device to perform tension matching optimization and determine the initial tension device control parameters; The lining production is executed based on the initial tension equipment control parameters. At the same time, the multimodal data stream of lining production is collected during the lining production process, and the digital twin of the tension equipment is loaded to simulate and predict the multimodal data stream of lining production to obtain the lining tension prediction parameters. If the predicted tension parameter of the lining exceeds the preset tension fluctuation threshold, the closed-loop processing mechanism is triggered to call the tension compensation and control strategy, and the initial tension equipment control parameters are compensated, corrected and dynamically controlled based on the tension compensation and control strategy.

2. The intelligent control method for tension equipment used in lining production as described in claim 1, characterized in that, The construction of the interlining production tension database includes: Historical production data mining was performed based on the tension equipment used for the production of the target lining fabric to obtain a historical dataset of lining fabric production. The historical data set of lining production is classified and labeled to obtain production data of tension equipment layer, production data of lining process layer, and production data of lining quality layer. The production data of the tension equipment layer, the production data of the lining process layer, and the production data of the lining quality layer are correlated and mapped to obtain the lining production tension correlation dataset. The data of the lining production tension correlation dataset is cleaned, stored, and integrated to construct the lining production tension database.

3. The intelligent control method for tension equipment used in lining production as described in claim 1, characterized in that, The digital twin of the tension generation device includes: According to the equipment modeling accuracy requirements, the structural design information of the tension equipment for the production of the target lining fabric is reduced in dimension to obtain the reduced equipment structural design parameters. Based on the structural design parameters of the reduced-dimensional equipment, the equipment assembly model and motion constraints are assigned to construct a three-dimensional model of the tension equipment. Obtain the control target of the tension equipment, extract tasks from the control target of the tension equipment, and obtain a list of control tasks for the tension equipment; Based on the tension equipment control task list and the lining production tension database, a digital twin simulation of the tension equipment's three-dimensional model is performed to generate a digital twin of the tension equipment.

4. The intelligent control method for tension equipment used in lining production as described in claim 3, characterized in that, The step of performing a twin simulation of the three-dimensional model of the tension equipment based on the lining production tension database according to the tension equipment control task list to generate a digital twin of the tension equipment includes: According to each control task in the tension equipment control task list, the lining production tension database is sequentially associated and rectified to obtain the lining tension task association dataset. Based on the lining tension task-related dataset, the algorithm model is selected and the control training and optimization are performed sequentially to generate a multi-task lining tension controller set. The multi-task lining tension controller set is mapped to the three-dimensional model of the tension device for coupled twin simulation to obtain an initial digital twin of the device; Scene instance injection and simulation verification optimization are performed on the initial device digital twin to generate a tension device digital twin.

5. The intelligent control method for tension equipment used in lining production as described in claim 4, characterized in that, The step of injecting scene instances and performing simulation verification optimization on the initial device digital twin to generate a tension device digital twin includes: A set of pre-defined interlining production scenario sets is used to obtain interlining tension scenario instance data by associating instance data based on the set of interlining production scenario sets. The tension scenario example data of the lining fabric is injected into the digital twin of the initial device to conduct tension simulation tests, and the scenario example simulation data is recorded; The simulation data of the scenario instance were verified and evaluated according to the twin performance indicators to obtain the twin control performance parameters; Based on the control performance parameters of the twin, the initial digital twin of the equipment is iteratively optimized to generate the digital twin of the tension equipment.

6. The intelligent control method for tension equipment used in lining production as described in claim 5, characterized in that, The data stream collected during the fabric production process includes: A monitoring requirement analysis was conducted on the tension equipment used for lining production to determine the dimensions of equipment monitoring requirements, which include the geometric and mechanical properties of the lining, the equipment status, and the environmental characteristics. According to the equipment monitoring requirements, the sensor selection and layout analysis of the tension equipment for lining production are carried out sequentially to obtain the monitoring dimension-sensor parameter set; Based on the monitoring dimension - sensor parameter set, a multimodal sensing network is deployed and constructed, and multimodal data streams of lining production are collected through the multimodal sensing network.

7. The intelligent control method for tension equipment used in lining production as described in claim 1, characterized in that, The triggering closed-loop processing mechanism invokes the tension compensation control strategy, including: The closed-loop processing mechanism is triggered to analyze the abnormal generation cause of the interlining production multimodal data stream and obtain the cause of the abnormal interlining tension. A lining tension anomaly operation and maintenance strategy library is constructed. The lining tension anomaly operation and maintenance strategy library is used to match and call the causes of the lining tension anomaly to obtain the tension compensation and control strategy.

8. The intelligent control method for tension equipment used in lining production as described in claim 7, characterized in that, The constructed lining tension anomaly operation and maintenance strategy library includes: Collect a case library of abnormal lining tension, analyze the cause of each abnormality in the case library of abnormal lining tension, and obtain a set of abnormality cause-tension operation and maintenance strategies. Based on the aforementioned anomaly cause-tension maintenance strategy set, tension compensation training is performed sequentially to construct a lining tension anomaly maintenance strategy library.

9. The intelligent control method for tension equipment used in lining production as described in claim 8, characterized in that, The tension compensation training is performed sequentially based on the anomaly cause-tension maintenance strategy set to construct a lining tension anomaly maintenance strategy library, including: Based on the characteristic information of the anomaly cause-tension operation and maintenance strategy set, select the tension compensation strategy algorithm network set; Based on the tension compensation strategy algorithm network set, compensation control training and iterative testing optimization are performed to obtain a tension compensation strategy controller set. The tension compensation strategy controller set is then identified and integrated to construct the lining tension anomaly operation and maintenance strategy library.

10. An intelligent control system for tension equipment used in lining production, characterized in that, The system is used to execute the intelligent control method for tension equipment used in lining production as described in any one of claims 1-9, and the system includes: The data mining module is used to mine production data based on the tension equipment for target lining production, construct a tension database for lining production, and perform twin simulation based on the structural design information of the tension equipment for target lining production and the tension database for lining production to generate a digital twin of the tension equipment. The tension matching module is used to input the lining specifications into the digital twin of the tension device for tension matching optimization and to determine the initial tension device control parameters. The lining production module is used to execute lining production based on the initial tension equipment control parameters, while simultaneously collecting multimodal data streams of lining production during the lining production process, and loading the digital twin of the tension equipment to simulate and predict the multimodal data streams of lining production to obtain lining tension prediction parameters. The control module is used to trigger a closed-loop processing mechanism to call the tension compensation control strategy if the predicted tension parameter of the lining exceeds the preset tension fluctuation threshold, and to compensate, correct and dynamically control the initial tension equipment control parameters based on the tension compensation control strategy.

Images