Multivariable collaborative polyurethane continuous production intelligent control method and system

Abstract

The invention discloses a multivariable coordinated intelligent control method and system for continuous production of polyurethane, and relates to the technical field of polyurethane production, and the method comprises the steps: carrying out equipment self-driven control based on a polyurethane production chain along with equipment feeding, controlling a reference system in a solver to synchronously carry out motion deduction, carrying out dual-path sampling through a preset time sequence control node, and carrying out multi-variable coordinated intelligent control on continuous production of polyurethane. And triggering the first solver to execute first process judgment based on chemical combination energy, triggering the second solver to execute second process judgment based on a microstructure, performing mutual verification to determine an off-axis control state, reversely deducing a regulation parameter control group, and performing control regulation management on a polyurethane production process. The method and the device are used for solving the technical problem of how to realize optimization and accurate regulation and control of a polyurethane production process through an intelligent control technology on the basis of multivariable coordination in the prior art, and control parameters of each production stage can be accurately adjusted, so that the production efficiency is improved, and the stability of product quality is ensured.

Description

Technical Field

[0001] This invention relates to the field of polyurethane production technology, specifically to a multi-variable collaborative intelligent control method and system for continuous polyurethane production. Background Technology

[0002] In traditional polyurethane production, the complex reaction mechanism makes process control challenging. Existing methods often rely on human experience and single-parameter control, which struggles to respond to complex process variations in real time. This can lead to unstable product quality and low production efficiency. Therefore, precisely controlling parameters at each reaction stage to ensure final product quality remains a pressing issue.

[0003] Currently, although some intelligent control methods based on process monitoring and feedback mechanisms exist, most focus on the regulation of single variables, lacking comprehensive control over the synergistic effects of multiple variables. Existing intelligent control systems emphasize macroscopic aspects and fail to effectively optimize control strategies at each production stage simultaneously.

[0004] Therefore, how to optimize and precisely control the polyurethane production process through intelligent control technology based on multi-variable synergy remains a technical problem that urgently needs to be solved in the industry. Summary of the Invention

[0005] This application provides a multi-variable collaborative intelligent control method and system for continuous polyurethane production, which addresses the technical problem of how to optimize and precisely control the polyurethane production process through intelligent control technology based on multi-variable collaboration in the prior art.

[0006] In view of the above problems, this application provides a multi-variable collaborative intelligent control method and system for continuous polyurethane production.

[0007] In a first aspect, this application provides a multi-variable collaborative intelligent control method for continuous polyurethane production. The method includes: acquiring a polyurethane production chain, wherein the production stages of the production chain include at least prepolymer synthesis, chain extension reaction, and crosslinking reaction, and each production stage is identified by process dynamics; for the polyurethane production chain, determining a reference frame through Wiener process modeling, using chemical energy and microstructure as analogous characteristics, constructing a control solver and embedding it in a production management platform; performing equipment self-driven control based on the polyurethane production chain as the equipment feeds in, with the reference frame in the control solver synchronously performing motion deduction, performing dual-path sampling with preset timing control nodes, triggering a first solver to execute a first process determination based on chemical energy, triggering a second solver to execute a second process determination based on microstructure, mutually verifying to determine the off-axis control state and reverse-engineering the control parameter group; and performing control and regulation management of the polyurethane production process according to the control parameter group.

[0008] Secondly, this application provides a multi-variable collaborative intelligent control system for continuous polyurethane production. The system includes: a production chain acquisition unit for acquiring the polyurethane production chain, wherein the production stages of the production chain include at least prepolymer synthesis, chain extension reaction, and crosslinking reaction, and each production stage is identified by process dynamics; a solver construction unit for constructing a control solver for the polyurethane production chain by determining a reference frame through Wiener process modeling, using chemical energy and microstructure as analogous characteristics, and embedding it in the production management platform; a control solver unit for performing self-driven equipment control based on the polyurethane production chain as the equipment feeds in, wherein the reference frame in the control solver synchronously performs motion deduction, performs dual-path sampling with preset timing control nodes, triggers the first solver to execute a first process determination based on chemical energy, triggers the second solver to execute a second process determination based on microstructure, verifies the off-axis control state, and reverse-engineers the control parameter group; and a control management unit for controlling and regulating the polyurethane production process according to the control parameter group.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages: The multivariable collaborative intelligent control method for continuous polyurethane production provided in this application obtains the polyurethane production chain, determines a reference frame through Wiener process modeling, and constructs a control solver based on chemical energy and microstructure as analogous characteristics, embedding it into the production management platform. As the equipment feeds in, it performs self-driven equipment control based on the polyurethane production chain. The reference frame within the control solver synchronously performs motion deduction, and dual-path sampling is performed using preset timing control nodes. This triggers the first solver to execute a first process determination based on chemical energy, and the second solver to execute a second process determination based on microstructure. Mutual verification determines the off-axis control state and reverse-engineers the control parameter group, thus controlling and regulating the polyurethane production process. This method addresses the technical problem in existing technologies of how to optimize and precisely control the polyurethane production process through intelligent control technology based on multivariable collaboration. Through dynamic process modeling and real-time feedback mechanisms, and by utilizing the collaborative work of different solvers, it can precisely adjust the control parameters of each production stage, thereby improving production efficiency and ensuring product quality stability. Attached Figure Description

[0010] Figure 1 This application provides a schematic diagram of a multi-variable collaborative intelligent control method for continuous polyurethane production. Figure 2 This application provides a schematic diagram of the structure of a multi-variable collaborative intelligent control system for continuous polyurethane production.

[0011] Explanation of reference numerals in the attached figures: Production chain acquisition unit 11, solver construction unit 12, control solver unit 13, regulation and management unit 14. Detailed Implementation

[0012] This application provides a multi-variable collaborative intelligent control method and system for continuous polyurethane production, which addresses the technical problem in the prior art of how to optimize and precisely control the polyurethane production process through intelligent control technology based on multi-variable collaboration.

[0013] Example 1: As Figure 1 As shown, this application provides a multi-variable collaborative intelligent control method for continuous polyurethane production, the method comprising: S1: Obtain a polyurethane production chain, wherein the production stages of the production chain include at least prepolymer synthesis, chain extension reaction, and crosslinking reaction, and each production stage is identified with a process dynamic relationship.

[0014] In this embodiment, the polyurethane production process first requires obtaining the polyurethane production chain, which is the series of production stages involved from raw materials to the final product. This process includes at least three main production stages: prepolymer synthesis, chain extension reaction, and crosslinking reaction. Each stage has its specific reaction conditions and objectives, and there are clear process kinetic relationships between the stages.

[0015] In a preferred embodiment, the production stage can be further subdivided to meet the requirements of high-precision control.

[0016] First, in the prepolymer synthesis stage, by selecting the reaction raw materials and controlling the reaction conditions, the monomer raw materials are polymerized to obtain a prepolymer with a certain molecular weight.

[0017] In this process, by introducing raw materials such as isocyanate and polyether polyol in a set ratio, and reacting under specific temperature and pressure conditions, a preliminary polyurethane structure is synthesized. Prepolymer synthesis is the foundation of polyurethane production, and its control parameters, such as reaction time, temperature, and catalyst concentration, directly affect the molecular structure and properties of the final product.

[0018] Next, in the chain extension reaction stage, the prepolymer synthesized material will continue to react with the chain extender to extend its chain length, thereby forming a more complex polymer network structure.

[0019] At this stage, key parameters of the chain extension process include the type and amount of chain extender, as well as the reaction temperature. The main purpose of this reaction stage is to increase the molecular weight of polyurethane, thereby affecting its mechanical properties and chemical stability. Controlling the chain extension reaction requires precise adjustment of the reaction rate to ensure that the final product possesses good physical properties, such as hardness and elasticity.

[0020] Finally, in the crosslinking reaction stage, the polymer chains generated by the chain extension reaction undergo further crosslinking under appropriate conditions to form a more three-dimensional interwoven structure, thereby endowing polyurethane with stronger mechanical properties, thermal stability and chemical resistance.

[0021] At this stage, the type and amount of crosslinking agent, as well as the control parameters of crosslinking temperature and time, are crucial. The degree of completion of the crosslinking reaction directly affects the final properties of polyurethane, such as temperature resistance, oil resistance, and corrosion resistance.

[0022] In the three stages mentioned above, process kinetics refers to the mutual influence and interdependence of reaction rate, conversion rate, and material flow in each stage during the production process.

[0023] For example, the polymer properties obtained in the prepolymer synthesis stage directly affect the efficiency of subsequent chain extension and crosslinking reactions. At the same time, the interaction between control parameters and dynamic reaction processes, i.e. the mechanistic relationship of dynamic processes, can be obtained through big data knowledge mining.

[0024] In summary, the underlying mechanism and related information of polyurethane production have been determined, providing data support for subsequent control.

[0025] Controlling each stage of the above process depends not only on the selection of materials and the setting of reaction conditions, but also on the dynamic adjustment of the process with the support of real-time monitoring and feedback mechanisms, so as to achieve precise control of the production process and quality assurance of the final product.

[0026] S2: For the polyurethane production chain, a reference frame is determined by performing Wiener process modeling, and a control solver is constructed and embedded in the production management platform using chemical energy and microstructure as analogous characteristics.

[0027] In this embodiment, the entire production chain needs to be modeled first in order to effectively control the production process. For this purpose, Wiener process modeling is used to determine the reference frame.

[0028] Specifically, the Wiener process is a mathematical model used to describe stochastic processes, typically employed to simulate uncertainties and random fluctuations present in production processes. In this application, Wiener process modeling allows for the establishment of a reference system encompassing all influencing factors based on the dynamic characteristics of each production stage in the polyurethane production chain, thereby providing a basis for precise control of the production process.

[0029] In one specific implementation, the aforementioned polyurethane production chain, i.e., the predetermined underlying mechanism and related information, is used to quantify and abstract the dynamic relationships such as molecular motion in the polyurethane production process into data relationships, and to construct the temporal sequence of the random motion process.

[0030] For example, an independent increment constraint is set, meaning that changes within any two non-overlapping time intervals are independent of each other; a normal distribution increment is set, meaning that changes within the time interval follow a mean of 0 and a variance of 0. σ 2 The normal distribution of t can be taken as σ = 1; a continuous path is defined, meaning the path is continuous but almost everywhere non-differentiable, reflecting the unpredictability of microscopic motion; initial conditions are set, meaning the reaction process starts from the origin. By setting the above conditions as motion constraints, the temporal nature of random motion is constructed.

[0031] Next, by using chemical energy and microstructure as analogous characteristics, the relationship between the two is introduced into the deviation judgment of the control process.

[0032] Chemical energy refers to the energy released or absorbed by polyurethane during the chemical reaction process, that is, the state characteristics of loss and formation during the dynamic reaction process, reflecting the progress and state of the reaction. In the production process, the change of chemical energy is closely related to the progress of the production stage and can serve as an important indicator for measuring production efficiency and reaction progress.

[0033] Microstructure refers to the microscopic arrangement and structural characteristics of polyurethane molecular chains. Changes in structure directly affect the physical properties and chemical stability of polyurethane. By using chemical energy and microstructure as analogous characteristics, a correlation is established between the two levels, thereby enabling more accurate prediction and regulation of the polyurethane production process.

[0034] Based on this, a control solver is constructed. By embedding the relationship between the Wiener process model, chemical energy, and microstructure, it can automatically adjust production parameters based on real-time feedback production data to achieve precise production control.

[0035] Subsequently, the control solver is embedded in the production management platform. This platform monitors, schedules, and optimizes the entire production process. This allows the solver to not only control individual production stages but also perform multi-variable collaborative optimization across the entire production chain. Ultimately, this ensures optimal coordination among all stages of the production process, improving production efficiency and guaranteeing product quality.

[0036] In summary, establishing a multi-dimensional, high-precision polyurethane production process control system not only improves the automation level of the production process, but also maintains high stability and accuracy when facing complex and ever-changing production environments, thereby meeting the demand for high-quality and high-efficiency production.

[0037] Furthermore, constructing the control solver, step S2 of this application includes: For the polyurethane production chain, a first variable matrix is ​​determined based on the chemical absorption state, and a second variable matrix is ​​determined based on the chemical production state. A first identifier and a second relation are introduced to integrate the first and second variable matrices as a variable matrix. The first identifier is a classifier based on the absorption or production state, and the second relation is the chemical trend relationship of each variable. The process time sequence based on the reference frame is determined, and the variable matrix is ​​subjected to dynamic process verification training to determine the first solver.

[0038] In this embodiment of the application, in order to further precisely adjust each link of the production process, a variable matrix is ​​established to provide data support for the subsequent control solver.

[0039] First, through a detailed analysis of the reaction states in the polyurethane production chain, the chemical absorption state and the chemical yield state were determined. The chemical absorption state refers to the state exhibited by reactants during energy absorption in the polyurethane production process. Conversely, the chemical yield state refers to the state after the chemical reaction is completed, when the molecular structure of the product changes, energy is released, and a stable state is reached.

[0040] In the dynamic production process at each stage, the absorption and output process occurs at every moment, which can reflect the progress of the chemical reaction.

[0041] In summary, we can identify the characteristics of different stages of the reaction process, thereby obtaining a targeted variable matrix.

[0042] In one embodiment, a first variable matrix is ​​first determined based on the chemical absorption state. This first variable matrix primarily includes variables related to the endothermic process, such as temperature, pressure, reactant concentration, and reaction rate. In this application, the first variable matrix focuses on energy absorption, reaction rate, and reactant consumption.

[0043] Next, based on the chemical outcome states, the second variable matrix was determined. This second variable matrix mainly includes variables related to the exothermic process, such as the reaction temperature, pressure, polymer molecular weight, and degree of crosslinking.

[0044] In the specific reaction process, the chemical absorption state and the chemical production state are inconsistent at different stages. Therefore, each reaction stage can be treated separately.

[0045] Due to the diversity of variables, a first identifier and a second relationship are introduced for effective identification. The first identifier is a classifier based on the absorption or production state, used to distinguish and identify the first and second variable matrices. The first identifier explicitly specifies which reaction state a variable belongs to. The second relationship is the chemotactic relationship between the variables, that is, how the variables are related and change as the reaction progresses. The chemotactic relationship reflects the dependence and inverse influence between variables; for example, as the temperature increases, the reaction rate accelerates, ultimately leading to changes in the properties of the products.

[0046] Subsequently, by integrating the first and second variable matrices under the guidance of the first identifier and the second relationship, an integrated variable matrix is ​​obtained. This provides comprehensive, multi-dimensional input data support for subsequent production process control.

[0047] Building upon this foundation, to ensure dynamic adjustment and control during the production process, a process sequence based on a reference system is further defined. Process sequence refers to a dynamic time series of processes based on the time progression of each stage in the polyurethane production chain, combined with the reaction characteristics of different stages. Using this as a benchmark for synchronous control under diversified analysis of the reaction process allows for better scheduling of operational parameters at each stage of the production process.

[0048] Finally, based on the aforementioned variable matrix and process time series, dynamic process verification training is performed. Specifically, according to the process time series, information mutual mapping of reaction time nodes is executed. Simultaneously, using each variable in the variable matrix as the judgment target, the current reaction process is evaluated for deviations and control requirements through variable identification and judgment. The first solver is constructed based on this mechanism and trained to convergence using a sample-driven training method.

[0049] In summary, the first solver ensures optimal production efficiency and product quality at different stages of the entire production chain by analyzing real-time data during the dynamic process.

[0050] Furthermore, constructing the control solver, step S2 of this application includes: For the polyurethane production chain, a third variable matrix for microstructure evolution is determined, wherein the third variable matrix includes at least pore structure class, pore state class and bond breakage class; a process time sequence based on a reference frame is determined, dynamic process verification training is performed based on the third variable matrix, and a second solver is determined.

[0051] In this embodiment, in addition to considering the energy changes and material transformations of chemical reactions, it is also necessary to conduct in-depth analysis of the microstructure of the product in order to further improve the performance stability of the product.

[0052] Therefore, it is necessary to determine the third variable matrix of microstructure evolution. Among them, changes in microstructure usually directly affect the physical properties of polyurethane materials, such as strength, elasticity, and thermal stability, so accurate understanding of its evolution process is crucial.

[0053] In this application, the third variable matrix includes, but is not limited to, three key categories: pore structure, pore state, and bond breakage.

[0054] One type is the pore structure: the pore structure of polyurethane changes during its production process, affecting the material's density, mechanical properties, and thermal conductivity. The formation, size, and distribution of pores directly affect the final quality of the polyurethane. During production, the formation and distribution of pores can be controlled by adjusting reaction conditions, catalyst type, and reaction temperature, thereby regulating the material's mechanical properties and thermal insulation.

[0055] Secondly, there is the cell state category: the cells in polyurethane are a key factor determining its elasticity and flexibility. The formation, size, and morphological changes of cells are closely related to the polyurethane foaming process. Cell state category includes information such as the degree of cell opening and the formation of closed cells. By controlling the cell formation process, the overall performance of polyurethane materials can be optimized, especially in the production of highly elastic, lightweight materials.

[0056] Thirdly, there are bond breakage variables: During the synthesis of polyurethane, chemical reactions may lead to the breakage or reformation of certain chemical bonds. Bond breakage involves changes in the chemical bonds within the polyurethane molecule, particularly in crosslinking and chain extension reactions. The breaking and recombination of molecular chains are crucial factors determining the final material properties. Monitoring these variables helps predict the mechanical properties and durability of polyurethane.

[0057] By integrating variables based on dimensional features across the three dimensions mentioned above, a third variable matrix can be formed, which can reflect the changing trend of the microstructure during the reaction process and provide more refined parameter inputs for controlling the solver.

[0058] Similarly, a process sequence based on a reference frame is established, which is based on the time sequence of the dynamic movement of the reference frame, that is, based on the reaction progress, time and change law of different stages in the polyurethane production chain, to reasonably arrange the sequence and control nodes of each production step.

[0059] Subsequently, dynamic process verification training is performed based on the process timeline of the third variable matrix and the reference frame. Specifically, according to the process timeline, information mutual mapping of reaction time nodes is executed. Simultaneously, using each variable in the third variable matrix as the judgment target, the deviation of the current reaction process from the required control is assessed through variable identification and judgment. A second solver is constructed based on this mechanism. Then, for the third variable matrix, a sample-driven training method is used to train until convergence. Through accurate control of the process timeline and training of the variable matrix, the microstructural evolution in the production process can be predicted in real time, and responses can be made based on detected changes.

[0060] Finally, based on the training results, a second solver was determined. By processing the data in the third variable matrix, key control parameters in the production process can be fed back and adjusted in real time.

[0061] For example, based on changes in the cell state class and pore structure class, the second solver can adjust reaction conditions or equipment settings to optimize the final performance of the polyurethane.

[0062] In summary, microstructural changes during the production process can be precisely controlled and optimized, thereby ensuring that the final product has excellent performance and stability.

[0063] Furthermore, application step S2 includes: A reference frame is determined by modeling the Wiener process, and a first solution channel is established; the first solver and the second solver are run in parallel to establish a second solution channel; the first solution channel and the second solver are connected by a lateral interaction based on process timing, which serves as the control solver.

[0064] In this embodiment, a reference frame is first determined through Wiener process modeling. This frame is used to simulate random changes and uncertainties that may occur during the production process. It can identify the key influencing factors and their changing patterns at each stage of production based on the dynamic characteristics of the production chain. One construction method provided in this application is as described above.

[0065] In a preferred embodiment, a unified reference framework is established across multiple reaction processes to provide a theoretical basis for subsequent control decisions.

[0066] Subsequently, the first solution channel was constructed. Based on the data obtained from Wiener process modeling, the first solution channel performs real-time status assessment of the production process. That is, for real-time stage nodes, it determines whether the actual production state matches the motion state in the reference frame, so as to quickly verify whether the actual production is abnormal.

[0067] Meanwhile, a second solver runs in parallel. This second solver primarily optimizes the microstructure evolution, focusing on details such as pore state, pore structure, and changes in chemical bonds to ensure precise control of the polyurethane's microstructure. During production, the evolution of the microstructure significantly impacts the mechanical properties and stability of the polyurethane. The second solver analyzes changes in reactant molecular chains and the progress of chemical reactions to adjust control strategies, ensuring optimal microstructure evolution during the reaction process. Therefore, the second solver works closely with the first solver to ensure comprehensive optimization of the production process.

[0068] In this solution, the first and second solvers are at the same level and are integrated in parallel to form a second solution channel. That is, the two solvers in the second solution channel start simultaneously and run independently.

[0069] Subsequently, the first and second solver channels are integrated in parallel, and lateral interaction based on process timing is introduced. Specifically, consistency mapping constraints are applied to the production progress of the first and second solver channels. Through lateral interaction, the first and second solvers can exchange information and adjust their respective computational tasks according to timing requirements. In this application, the core of lateral interaction is to achieve real-time synchronization of the production process. That is, at each control node, the control results of the first and second solvers can be calculated and adjusted at the same time, thereby ensuring the coordination and consistency of different control objectives during the production process.

[0070] In summary, the complexity and variability of the polyurethane production process can be effectively managed and controlled, thereby improving the intelligence and precision of the production process.

[0071] S3: The equipment is self-driven and controlled based on the polyurethane production chain as it feeds into the equipment. The reference system in the control solver performs motion simulation synchronously. It performs dual-path sampling with preset timing control nodes, triggers the first solver to perform the first process determination based on chemical energy, and triggers the second solver to perform the second process determination based on microstructure. The mutual verification determines the off-axis control state and reverse-engineers the control parameters.

[0072] S4: Based on the aforementioned control parameters, control and regulate the polyurethane production process.

[0073] In this embodiment, the equipment is self-driven based on the polyurethane production chain as the raw materials are fed into the equipment. That is, when the raw materials are fed into the equipment, the equipment control system will automatically identify the current stage of the production chain and drive the control of the stage parameters.

[0074] Meanwhile, during the self-driving control of the equipment, the reference frame in the control solver synchronously performs motion deduction, that is, the control solver can predict and deduce the motion state of the equipment and each link in the production process, which is synchronized with the random motion deduction of molecular motion and other processes in the production process.

[0075] Based on preset timing control nodes, i.e., key control points in the production process, two different data acquisition methods are used simultaneously to obtain different types of data during the process, specifically including data based on changes in chemical reaction energy and the evolution of microstructure. Through dual-channel sampling, the control system can comprehensively and accurately monitor multiple key parameters in the production process, providing sufficient basis for subsequent decision-making.

[0076] Subsequently, the first solver is triggered to perform a first process determination based on the chemical energy. Specifically, the first solver will determine whether the reaction is proceeding as expected and determine the progress and efficiency of the reaction based on the changes in chemical energy during the reaction process.

[0077] For example, if abnormal energy consumption is detected, the regulatory mechanism is triggered to adjust the reaction conditions to restore them to a normal state.

[0078] Simultaneously, a second solver is triggered to execute a second process determination based on microstructure. Specifically, the second solver focuses on monitoring changes in microstructure, such as the formation of cells in polyurethane and the evolution of pore structure. Through this determination, the changes at the microscopic level of the material during the reaction can be understood, and reaction conditions can be adjusted as necessary to ensure product quality.

[0079] Subsequently, the two process decisions are cross-validated, that is, the output results of the first solver and the second solver are cross-validated. By comparing the two decisions, it is determined whether there are any discrepancies or inconsistencies.

[0080] If the results of both tests are consistent and indicate a deviation, the off-axis control state is then determined. The off-axis control state refers to an abnormality or deviation in a certain part of the production process, which may be caused by equipment failure, unstable reaction conditions, or other reasons.

[0081] Subsequently, based on the off-axis control state, the control parameters are further reversed and adjusted. That is, the deviation dimension, deviation direction and scale of the off-axis control state are converted into specific equipment control parameters to restore the normal state of the production process.

[0082] Ultimately, based on the adjusted control parameters, the polyurethane production process is managed and regulated by controlling the production equipment. This ensures efficient and stable operation of the production process and guarantees that the final product quality meets standards.

[0083] In summary, a dynamic and intelligent closed-loop control system is formed, which can respond to various changes and disturbances in real time throughout the entire production chain, ensuring that the polyurethane production process is always in an optimal control state.

[0084] Furthermore, before performing dual-channel sampling with a preset timing control node, the acquisition of the preset timing control node, in step S3 of this application, includes: Based on the stages of prepolymer synthesis, chain extension reaction, and crosslinking reaction, a first type of key node is identified; for each production stage, a second type of key node is identified based on the start, middle, and end of the chemical reaction; a third type of key node is identified based on the unsteady state of each production stage; the first, second, and third types of key nodes are integrated into the process time sequence, and node class identifiers are introduced as the preset time sequence control nodes.

[0085] In this embodiment of the application, it is first necessary to conduct a detailed analysis of the entire production chain to identify the key nodes in the production process, that is, the moments or stages that play a decisive role in the production process, and to focus on ensuring the control precision of the key nodes.

[0086] By identifying key nodes, equipment can be precisely controlled during the production process, avoiding unnecessary waste and quality fluctuations.

[0087] First, the stages of prepolymer synthesis, chain extension reaction, and cross-linking reaction mark the beginning and end of different reaction stages in the production process.

[0088] For example, in the prepolymer synthesis stage, the reactants begin to polymerize to form the prepolymer, which is a critical node in this stage; the start and end of the chain extension reaction and the crosslinking reaction are also critical nodes in their respective stages. These nodes determine the basic framework of the polyurethane production chain and are marked as the first type of critical nodes.

[0089] Next, for each production stage, the initiation, intermediate and final stages of the combination reaction are considered, indicating different progress stages of the reaction process. Based on this, two types of key nodes are further refined and identified.

[0090] For example, in the prepolymer synthesis stage, the initiation node of the chemical reaction refers to the moment when the reactants begin to react; the intermediate node refers to the stage where the reaction is close to equilibrium but not yet complete; and the final node marks the point where the reaction is basically complete and the conversion rate of the reactants reaches the set value. These two types of critical nodes more accurately capture the reaction progress, thus providing more detailed information for subsequent control.

[0091] Furthermore, considering the unsteady-state characteristics of each production stage, the key change moments caused by unsteady-state conditions in the production process were identified by analyzing the dynamic changes in the reaction process, and three types of key nodes were determined.

[0092] For example, sudden changes in temperature, fluctuations in reactant concentration, or a sudden increase in energy release at a certain stage can all trigger unsteady-state phenomena. These are nodes where the reaction is more significant and require special attention.

[0093] Subsequently, the process timing of Category I, Category II, and Category III critical nodes was integrated, that is, different types of critical nodes were arranged in an orderly manner according to the time sequence of the production process. The time relationships between nodes were strictly planned to ensure the coordination and smooth transition of each link.

[0094] Finally, for the key nodes of the above integration, node class identifiers are introduced. These identifiers are labels used to distinguish different types of nodes, for example, by using color or geometric symbols.

[0095] Ultimately, the above steps form a complete sequence of preset timing control nodes. This enables precise control of the polyurethane production process, allowing each link in the production chain to operate efficiently and stably. It ensures that every critical node in the production process can be accurately monitored, thereby optimizing production scheduling, improving product quality, and reducing resource waste.

[0096] Furthermore, the reference frame within the control solver synchronously performs motion deduction. Step S3 of this application includes: For the polyurethane production chain, a process control program is determined, the production equipment is centrally controlled, and the production parameters are automatically controlled. Using the process time sequence as the synchronization basis, the reference frame in the first solution channel of the control solver performs motion deduction of the time random process.

[0097] In this embodiment of the application, the process control procedure needs to be determined first. That is, based on the various stages and reaction processes of polyurethane production, a complete set of operating specifications and control strategies are formulated, and specific parameters and control methods for each production link are defined, such as temperature, pressure, reactant concentration, reaction time, etc.

[0098] In this application, the specific control requirement is a pre-planned control chain, which is converted into a program format and embedded in the central control of the production equipment to achieve automated control of the equipment drive.

[0099] Specifically, the production equipment is configured with central control, which means that the production equipment is centrally managed and scheduled through a central control system, while the process control program serves as the baseline for automated management and scheduling.

[0100] In modern production lines, automated central control systems are typically used to monitor and adjust equipment status in real time, ensuring the coordinated operation of production equipment. The central control system adjusts the operating status of each piece of production equipment according to the instructions of the process control program. For example, it dynamically coordinates equipment startup, shutdown, speed regulation, temperature control, and pressure regulation.

[0101] Synchronous, in automated control, the process time sequence is used as the synchronization basis, that is, as the time frame for production control. In the polyurethane production process, the reaction progress and material conversion rate at different stages change over time. By using the process time sequence as the basis, the progress of each production link can be accurately synchronized, and the motion deduction of the time-stochastic process can be performed on the reference frame in the first solution channel of the control solver. That is, the reference frame and the actual control process are synchronized based on the process time sequence.

[0102] During the production process, due to many factors (such as fluctuations in raw materials, equipment response time, and changes in reaction conditions), the production process exhibits a certain degree of randomness and uncertainty. A time-stochastic process is employed to ensure a realistic fit between the simulation and actual operation.

[0103] Specifically, the motion simulation of the reference frame simulates the uncertainties in the production process. Using the Wiener process, the possible reaction states and equipment states at different points in time are determined through real-time production progress. That is, the theoretical production state at that stage.

[0104] In summary, it enables automated adjustment and synchronous control in the polyurethane production process, greatly improving production efficiency and product quality stability.

[0105] Furthermore, step S3 of this application includes: Any node based on the preset timing control node is selected as the first node; based on the first node, positioning is performed in the reference system to determine the first reference motion state; according to the first reference motion state, the real-time production status of the production equipment is detected and consistency is determined to determine the status determination result.

[0106] If the determination result is consistent, the subsequent processing of the first node is terminated; if the determination result is inconsistent, a dual-channel sampling instruction is generated to drive the front-end detection array to sample, determine the dual-channel detection data, and send it back to the production management platform.

[0107] In this embodiment of the application, the first node is determined as the production process progresses, that is, any node among the preset timing control nodes that are currently arriving.

[0108] Next, based on the first node, positioning is performed in the reference frame. Specifically, using the process time sequence as a reference, the corresponding position of the first node in the reference frame is determined by mapping based on the time sequence nodes, that is, the specific position of the first node in the production chain, and the first reference motion state is determined based on this.

[0109] The first reference motion state is based on the specific state of the production equipment and reaction process at the first node in the production chain. Motion state refers to the operating status of the equipment, the conversion progress of the reactants, and the current values ​​of production parameters (such as temperature, pressure, concentration, etc.) during the production process.

[0110] In this application, based on the first reference motion state, i.e. the expected motion state of that stage, it is further determined whether the production equipment and reaction process are running on the predetermined trajectory.

[0111] Specifically, the real-time production status of the production equipment is monitored and its consistency is determined. That is, the current production equipment status is compared with the first reference motion status to determine the difference between the two. If the two are consistent, it means that the equipment and production process are in normal operation and no further adjustments are needed; at this time, the subsequent processing of the first node will be terminated, that is, further operation and control of this node will be stopped to ensure that the production process can continue smoothly.

[0112] However, if the determination result is inconsistent, that is, there is a deviation between the current state of the production equipment and the first reference motion state, the fault detection and correction mode will then be entered.

[0113] Specifically, by generating dual-channel sampling instructions, the detection array at the production front end is driven to perform sampling. That is, from two threads, chemical energy and microstructure, the corresponding detection devices at the front end are activated to perform directional acquisition and determine dual-channel detection data.

[0114] Subsequently, the collected dual-channel detection data is transmitted back to the production management platform for further analysis and processing. The platform then further activates the control solver for more detailed analysis and control decisions.

[0115] In summary, this system enables real-time monitoring of every critical stage in the polyurethane production process, allowing for intelligent judgment and control based on real-time data. This ensures the stability of the production process and the reliability of product quality. The dynamic, real-time control method significantly improves the adaptability and intelligence level of the production line.

[0116] Furthermore, determining the off-axis control state and inversely determining the control parameter group, step S3 of this application includes: The dual-path detection data is imported into the second solution channel, and solver matching is performed. By performing vector value self-attention recognition and state determination of the variable matrix, the first control state and the second control state are normalized and determined. The first control state and the second control state are mutually verified to determine the off-axis control state. If the states are inconsistent, mean processing is performed. Based on the off-axis control state, the pre-adjustment state is determined by the off-axis dimension, off-axis direction and off-axis scale. The pre-adjustment state is then converted into a parameter control group to determine the control parameter control group.

[0117] In this embodiment of the application, the data obtained through dual-channel sampling is transmitted to the input channel of the second solver. By performing solver matching, the detection data corresponding to chemical energy and microstructure are respectively imported into the corresponding first solver and second solver.

[0118] Next, self-attention recognition and state determination are performed using vector values ​​based on the variable matrix. Specifically, the first solver, for the imported sampled data, uses the variable matrix as the recognition target to check the variable vector values ​​and further determines whether the variable vector values ​​meet the variable state of that node, which is then used as the first control state.

[0119] Similarly, the second solver uses the third variable matrix as the self-attention recognition target to identify and judge, and determines the second control state.

[0120] Due to the dimensional differences in chemical energy and microstructure, direct cross-verification is difficult. Therefore, a normalization method is used, such as unifying the coefficient form for conversion and verification. For example, all data can be converted to a standard range, such as [0,1], to eliminate differences in dimensions and numerical scales.

[0121] Next, the first and second control states are cross-validated to check for inconsistencies. This ensures that the analysis results across different dimensions are consistent.

[0122] If inconsistencies exist, the values ​​of the cross-mapped quantities are averaged to obtain the final state.

[0123] By fusing the first and second control states, specifically including the state data before and after normalization, the off-axis control state is determined. That is, the off-axis dimension between actual production and the expected deviation, i.e., the corresponding variable, off-axis direction, and scale. The off-axis dimension refers to the specific variable in which the deviation occurs, the off-axis direction describes the trend of the deviation, and the off-axis scale quantifies the magnitude of the deviation.

[0124] Subsequently, based on the off-axis control state, a pre-adjustment state is determined. This involves converting the adjustment requirements based on the off-axis dimension, direction, and scale into control parameters for the production equipment, which serve as the control parameter group. This group specifically includes all control parameters and decision variables that need adjustment, thereby enabling feedback adjustment of the equipment control.

[0125] Based on the above method, dynamic and intelligent adjustment can be achieved throughout the entire production process, thereby ensuring that the polyurethane production process meets the requirements of high efficiency, precision and stability.

[0126] The multivariable collaborative intelligent control method for continuous polyurethane production provided in this application has the following technical advantages: 1. Multi-stage modeling and reference system construction based on the production chain: For the entire polyurethane production chain, including prepolymer synthesis, chain extension reaction, and cross-linking reaction, a reference system is determined through Wiener process modeling, using chemical energy and microstructure as core analogical characteristics. This enables dynamic quantitative description of the entire production process, providing a unified benchmark for precise control and avoiding control deviations caused by stage fragmentation.

[0127] 2. Dual-Dimensional Control Solver Design: A first solver (based on the first variable matrix of chemical energy, covering absorption and production states) and a second solver (based on the third variable matrix of microstructure, including porosity, pores, and bond states) are constructed, forming a collaborative control solver through lateral interaction. Simultaneously, the production state is captured from both energy change and microstructure evolution dimensions, improving the comprehensiveness of control and reducing the risk of misjudgment based on a single indicator.

[0128] 3. Precise Sampling of Multiple Types of Timing Control Nodes: Integrating stage nodes, reaction cycle nodes, and unsteady-state nodes as preset timing control nodes triggers dual-path sampling. This ensures accurate data collection during critical production stages, reaction inflection points, and abnormal states, avoiding control lag or omissions. For each node, production status is quickly determined based on the synchronicity of the reference system. When states are inconsistent, dual-path verification and off-axis state adjustment are further performed. Specifically, the first and second solvers cross-validate the sampled data to determine the off-axis control state and reverse-engineer the control parameter group. This reduces single-path errors, accurately locates the source of deviations, makes control parameters more targeted, and improves production stability.

[0129] 4. Multi-variable synergistic full-chain regulation: Based on the chemical trend relationships of the variable matrix (including energy and structure-related variables at each stage), production parameters are synergistically adjusted, rather than controlled by a single variable. This avoids chain fluctuations caused by adjustments to local parameters, ensures multi-variable synergistic optimization, and improves product quality stability.

[0130] Example 2: Based on the same inventive concept as the multi-variable collaborative intelligent control method for continuous polyurethane production in the foregoing examples, such as... Figure 2As shown, this application provides a multi-variable collaborative intelligent control system for continuous polyurethane production, the system comprising: Production chain acquisition unit 11 is used to acquire a polyurethane production chain, wherein the production stages of the production chain include at least prepolymer synthesis, chain extension reaction and crosslinking reaction, and each production stage is identified with a process dynamic relationship. The solver construction unit 12 is used to determine the reference frame by performing Wiener process modeling for the polyurethane production chain, using chemical energy and microstructure as analogous characteristics, to construct a control solver and embed it into the production management platform. The control solving unit 13 is used to perform self-driven control of the equipment based on the polyurethane production chain as the equipment feeds in. The reference system in the control solver performs motion deduction synchronously, performs dual sampling with preset timing control nodes, triggers the first solver to execute the first process determination based on chemical energy, triggers the second solver to execute the second process determination based on microstructure, verifies the off-axis control state and reverse-engineers the control parameter group. The control and management unit 14 is used to control and regulate the polyurethane production process according to the control and management parameter group.

[0131] Furthermore, the solver construction unit 12 performs the following steps: for the polyurethane production chain, a first variable matrix is ​​determined based on the chemical absorption state, and a second variable matrix is ​​determined based on the chemical production state; a first identifier and a second relation are introduced to integrate the first variable matrix and the second variable matrix as a variable matrix, wherein the first identifier is a classifier based on the absorption state or production state, and the second relation is the chemical trend relationship of each variable; the process time sequence based on the reference frame is determined, and the variable matrix is ​​subjected to dynamic process verification training to determine the first solver.

[0132] Furthermore, the solver construction unit 12 performs the following steps: for the polyurethane production chain, it determines the third variable matrix of microstructure evolution, wherein the third variable matrix includes at least pore structure class, pore state class and bond breakage class; it determines the process timing based on the reference frame, performs dynamic process verification training based on the third variable matrix, and determines the second solver.

[0133] Furthermore, the solver construction unit 12 performs the following steps: determining the reference frame by performing Wiener process modeling, and building a first solver channel; building a second solver channel by parallelizing the first solver and the second solver; and establishing a lateral interaction between the first solver channel and the second solver based on process timing as the control solver.

[0134] Furthermore, the control solution unit 13 performs the following steps: based on the stage nodes of prepolymer synthesis-chain extension reaction-crosslinking reaction, it is identified as a type of key node; for each production stage, a second type of key node is identified based on the start, middle, and end of the chemical reaction; a third type of key node is identified based on the unsteady state of each production stage; the first, second, and third types of key nodes are integrated into the process time sequence, and node class identifiers are introduced as the preset time sequence control nodes.

[0135] Furthermore, the control solving unit 13 performs the following steps: for the polyurethane production chain, it determines the process control program, performs central control configuration of the production equipment, and drives the automated control of production parameters; using the process timing as the synchronization basis, the reference frame in the first solving channel of the control solver performs motion deduction of the time random process.

[0136] Furthermore, the control solving unit 13 performs the following steps: determining any node based on the preset timing control node as the first node; positioning in the reference frame based on the first node to determine the first reference motion state; and detecting and determining the consistency of the real-time production status of the production equipment according to the first reference motion state to determine the status determination result.

[0137] Furthermore, the control solving unit 13 performs the following steps: if the determination result is consistent, the subsequent processing of the first node is terminated; if the determination result is inconsistent, a dual-channel sampling instruction is generated to drive the front-end detection array to perform sampling, determine the dual-channel detection data, and send it back to the production management platform.

[0138] Furthermore, the control solving unit 13 performs the following steps: importing the dual-path detection data into the second solving channel, performing solver matching, and determining the first control state and the second control state by performing vector value self-attention recognition and state determination of the variable matrix, and normalizing; mutually verifying the first control state and the second control state to determine the off-axis control state, wherein if the states are inconsistent, mean processing is performed; determining the pre-adjustment state based on the off-axis control state using the off-axis dimension, off-axis direction, and off-axis scale, performing parameter control conversion on the pre-adjustment state, and determining the control parameter control group.

[0139] Through the foregoing detailed description of the intelligent control method for continuous polyurethane production with multivariable synergy, those skilled in the art can clearly understand the intelligent control method and system for continuous polyurethane production with multivariable synergy in this embodiment. As for the apparatus disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to the description in the method section.

[0140] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-variable collaborative intelligent control method for continuous polyurethane production, characterized in that, The method includes: A polyurethane production chain is obtained, wherein the production stages of the production chain include at least prepolymer synthesis, chain extension reaction, and crosslinking reaction, and each production stage is identified with a process dynamic relationship. For the polyurethane production chain, a reference frame is determined by Wiener process modeling, and a control solver is constructed and embedded in the production management platform using chemical energy and microstructure as analogous characteristics. The equipment is self-driven and controlled based on the polyurethane production chain as it feeds into the equipment. The reference system in the control solver performs motion deduction synchronously. It performs dual sampling with preset timing control nodes, triggers the first solver to execute the first process determination based on chemical energy, and triggers the second solver to execute the second process determination based on microstructure. The mutual verification determines the off-axis control state and reverse-engineers the control parameter group. Based on the aforementioned control parameters, the polyurethane production process is controlled and regulated.

2. The intelligent control method for continuous polyurethane production with multi-variable synergy as described in claim 1, characterized in that, Constructing a control solver includes: For the polyurethane production chain, the first variable matrix is ​​determined based on the chemical absorption state, and the second variable matrix is ​​determined based on the chemical production state. Introducing a first identifier and a second relation, the first variable matrix and the second variable matrix are integrated and combined into a variable matrix, wherein the first identifier is a classifier based on the absorption state or the production state, and the second relation is the chemical trend relationship of each variable; The process timing based on the reference frame is determined, and the variable matrix is ​​subjected to dynamic process verification training to determine the first solver.

3. The intelligent control method for continuous polyurethane production with multi-variable synergy as described in claim 2, characterized in that, Constructing a control solver includes: For the polyurethane production chain, a third variable matrix for microstructure evolution is determined, wherein the third variable matrix includes at least pore structure class, cell state class and bond breakage class; The process timing based on the reference frame is determined, dynamic process verification training is performed based on the third variable matrix, and the second solver is determined.

4. The intelligent control method for continuous polyurethane production with multi-variable synergy as described in claim 3, characterized in that, By modeling the Wiener process, a reference frame is determined, and the first solution channel is established. The first and second solvers are run in parallel to establish a second solution channel; The first solution channel and the second solution channel are connected by a lateral interaction based on process timing, which serves as the control solver.

5. The intelligent control method for continuous polyurethane production with multi-variable synergy as described in claim 1, characterized in that, Before performing dual-channel sampling using a preset timing control node, the acquisition of the preset timing control node includes: Based on the stages of prepolymer synthesis, chain extension reaction, and cross-linking reaction, these are considered as a class of key nodes; For each stage of production, two types of key nodes are identified based on the initiation, middle, and final stages of the combination reaction; Based on the unsteady state of each production stage, three types of key nodes are identified; The process timing of the first-class, second-class, and third-class key nodes is integrated, and node class identifiers are introduced as the preset timing control nodes.

6. The intelligent control method for continuous polyurethane production with multi-variable synergy as described in claim 5, characterized in that, The reference frame within the control solver synchronously performs motion deduction, including: For the polyurethane production chain, a process control procedure is determined, the production equipment is centrally controlled, and the production parameters are automatically controlled. Using process timing as the synchronization basis, the reference frame in the first solution channel of the control solver performs motion deduction of the time-stochastic process.

7. The intelligent control method for continuous polyurethane production with multivariate coordination as described in claim 6, characterized in that, Determine any node based on the preset timing control node as the first node; Based on the first node, positioning is performed in the reference frame to determine the first reference motion state; Based on the first reference motion state, the real-time production status of the production equipment is detected and consistency is determined, and the status determination result is determined.

8. The intelligent control method for continuous polyurethane production with multivariable coordination as described in claim 7, characterized in that, If the determination result is consistent, the subsequent processing of the first node is terminated; If the determination result is inconsistent, a dual-channel sampling instruction is generated to drive the front-end detection array to perform sampling, determine the dual-channel detection data, and send it back to the production management platform.

9. The intelligent control method for continuous polyurethane production with multi-variable synergy as described in claim 8, characterized in that, Determine the eccentricity control state and reverse-engineer the control parameter group, including: The dual-path detection data is imported into the second solution channel, and solver matching is performed. By performing vector value self-attention recognition and state determination of the variable matrix, the first control state and the second control state are determined by normalization. The first control state and the second control state are mutually verified to determine the off-axis control state. If the states are inconsistent, the average value is processed. Based on the off-axis control state, a pre-adjustment state is determined by the off-axis dimension, off-axis direction, and off-axis scale. The pre-adjustment state is then converted into a parameter control group to determine the control parameter control group.

10. A multi-variable collaborative intelligent control system for continuous polyurethane production, characterized in that: The system is used to execute the multivariate collaborative intelligent control method for continuous polyurethane production according to any one of claims 1-9, the system comprising: A production chain acquisition unit is used to acquire a polyurethane production chain, wherein the production stages of the production chain include at least prepolymer synthesis, chain extension reaction, and crosslinking reaction, and each production stage is identified with a process dynamic relationship. The solver construction unit is used to determine the reference frame by performing Wiener process modeling for the polyurethane production chain, using chemical energy and microstructure as analogous characteristics, to construct a control solver and embed it into the production management platform. The control and solving unit is used to perform self-driven control of the equipment based on the polyurethane production chain as the equipment feeds in. The reference system in the control solver performs motion deduction synchronously, performs dual sampling with preset timing control nodes, triggers the first solver to execute the first process determination based on chemical energy, triggers the second solver to execute the second process determination based on microstructure, verifies the off-axis control state and reverse-engineers the control parameter group. The control and management unit is used to control and regulate the polyurethane production process according to the control and management parameters group.

Images