Digital Twin Optimization System, Method and Equipment for the Granulation Process of Urea-Formaldehyde Particles

Abstract

The present application discloses a digital twin optimization system, method and device for the granulation process of urea-formaldehyde particles. The digital twin optimization system includes: a digital twin simulation unit for constructing a multi-physical field coupling model of the granulation process and setting constraint parameters based on sampling data to fine-tune the feeding data-driven model using physical prior knowledge; an optimization decision unit for setting an optimization function according to the target production requirements to calculate the decision value of the output of the feeding data-driven model and using the feeding-driven model and the decision value to adjust the feeding parameters of the granulation process in real time. Through the solution of the present application, the entire system can be comprehensively simulated and optimized, thereby providing a more accurate feeding optimization scheme.

Description

Technical Field

[0001] This application relates to the field of blanking optimization, and particularly to a digital twin optimization system, method and equipment for the granulation process of urea formaldehyde particles. Background Art

[0002] The blanking optimization in the granulation process of urea formaldehyde particles has become an important research topic in the chemical industry, which is closely related to product quality, production efficiency and resource utilization rate. As an important chemical raw material, urea formaldehyde particles are widely used in agriculture, construction, industry and other fields, and the blanking link in the granulation process directly affects the performance and production cost of the final product. However, not all blanking operations can achieve ideal results, which leads to the difference between optimized blanking and non-optimized blanking. Optimized blanking can achieve uniform particle size, moderate strength and good dissolution performance, while non-optimized blanking may lead to unstable particle quality, increased dust and energy consumption. Therefore, the blanking optimization in the granulation process of urea formaldehyde particles is particularly important for improving product quality, reducing production costs and reducing environmental pollution, which helps to enhance the competitiveness and sustainable development ability of enterprises. Currently, the main technical means of blanking optimization include adjusting the blanking speed, controlling the material layer thickness, optimizing the spraying parameters and improving the equipment structure, etc. These methods achieve the optimization goal by adjusting the fluidity, distribution uniformity and heat transfer efficiency of the material. However, these traditional methods often rely on the experience of operators and repeated tests, are difficult to adapt to complex and changeable production conditions, and cannot achieve real-time dynamic optimization and precise control.

[0003] In recent years, with the development of intelligent manufacturing, the application of digital and intelligent technologies in chemical production has become increasingly widespread. In terms of the blanking optimization in the granulation process of urea formaldehyde particles, some new methods have gradually attracted attention. For example, using an online monitoring system to collect key parameters in real time during the blanking process, and combining data analysis and model prediction to adjust the blanking strategy; or using machine vision technology to analyze the particle morphology in real time, so as to dynamically adjust the blanking parameters. These methods have improved the accuracy and efficiency of blanking optimization to a certain extent, but there are still some limitations. For example, relying solely on online monitoring data is difficult to comprehensively reflect the complex granulation process, and the reliability of machine vision technology in harsh environments such as high temperature and high humidity remains to be improved. In addition, these methods usually can only optimize local parameters and are difficult to achieve global optimization of the entire granulation system.

[0004] Therefore, there is an urgent need for a technical solution that can comprehensively simulate and optimize the entire system, so as to provide a more accurate blanking optimization plan. Summary of the Invention

[0005] To address the deficiencies of the prior art, the embodiments of the present application provide a digital twin optimization system, method, and device for the urea-formaldehyde granulation process. The present application solves the technical problems of the prior art, such as only being able to optimize local parameters and being difficult to achieve global optimization of the entire granulation system.

[0006] The embodiments of the present application provide a digital twin optimization system for the urea-formaldehyde granulation process, including: a digital twin simulation unit and an optimization decision unit; wherein, the digital twin simulation unit is used to construct a multi-physical field coupling model of the granulation process and set constraint parameters based on sampling data to fine-tune the feeding data-driven model using physical prior knowledge; the optimization decision unit is used to set an optimization function according to the target production requirements to calculate the decision value of the output of the feeding data-driven model, and use the feeding-driven model and the decision value to adjust the feeding parameters of the granulation process in real time; including: setting an optimization function according to the target production requirements; inputting the first physical parameter, feeding parameters, and time series number at the current moment into the feeding data-driven model to obtain the second physical parameter within the target time window; performing a fusion calculation on the second physical parameter, and the result of the fusion calculation is used to substitute into the optimization function to calculate the decision value; iteratively replacing the feeding parameters based on a preset interval to obtain the optimal feeding parameters corresponding to the lowest decision value, and performing real-time adjustment on the granulation process based on the optimal feeding parameters; wherein, performing a fusion calculation on the second physical parameter includes: , where represents the second physical parameter from the 1st to the i-th group within the time window, i represents the number of groups within the time window, and j represents the index of the summation term. represents the natural exponential function. represents the natural logarithm function. represents the second physical parameter of the i-th group.

[0007] In a possible implementation, constructing a multi-physical field coupling model of the granulation process and setting constraint parameters based on sampling data to fine-tune the feeding data-driven model using physical prior knowledge includes: collecting the production data of urea-formaldehyde granulation at the current moment to obtain the first physical parameter; performing simulation calculations on the first physical parameter based on the multi-physical field coupling model of the granulation process to obtain the second physical parameter; and fine-tuning and training the feeding data-driven model using the first physical parameter, the feeding parameters at the current moment, the time series number, and the second physical parameter.

[0008] In one possible implementation, a multi-physics field coupling model of the granulation process is used to simulate and calculate a first physical parameter to obtain a second physical parameter, including: establishing a geometric model of a urea-formaldehyde granulator, and defining the physical fields and coupling relationships of the internal urea-formaldehyde particles; setting initial conditions and boundary conditions according to the first physical parameter, and performing iterative simulation operations on the production parameters within the target time window to obtain the second physical parameter.

[0009] In one possible implementation, defining the physical fields and coupling relationships of the internal urea-formaldehyde particles includes: ,

[0010] where, represents the partial derivative with respect to time t, represents density, represents the differential operator, represents the velocity vector, represents pressure, represents the dynamic viscosity, represents the acceleration due to gravity, represents the specific heat capacity, represents temperature, represents the thermal conductivity, represents the heat source term, represents concentration, represents the diffusion coefficient, represents the reaction source term, represents the mass of particle i, represents the derivative with respect to time t, represents the velocity vector of particle i, represents the contact force between particle i and particle j, represents the drag force between particle i and particle j, represents the moment of inertia of particle i, represents the angular velocity of particle i, represents the lever arm of particle i.

[0011] In one possible implementation, fine-tuning and training the feeding data-driven model using the first physical parameter, the feeding parameter at the current moment, the time series number, and the second physical parameter includes: constructing a fine-tuning data set based on the first physical parameter, the feeding parameter at the current moment, the time series number, and the second physical parameter; determining the number of target hidden layers for updating the feeding data-driven model based on a preset algorithm; using the fine-tuning data set to perform fine-tuning and training on multiple target hidden layers of the pre-trained feeding data-driven model, and freezing the parameter updates of other hidden layers during the fine-tuning and training process.

[0012] In one possible implementation, determining the number of target hidden layers for updating the blanking data-driven model based on a preset algorithm includes: ,

[0013] where, represents the floor function, represents the cyclic adjustment term, and t represents the time window size of the second physical parameter.

[0014] The embodiments of the present application also provide a digital twin optimization method for the urea formaldehyde granulation process, including: constructing a multi-physical field coupling model of the granulation process, and setting constraint parameters based on sampling data to fine-tune the blanking data-driven model using physical prior knowledge; setting an optimization function according to the target production requirements to calculate the decision value of the output of the blanking data-driven model, and using the blanking drive model and the decision value to adjust the blanking parameters of the granulation process in real time; including: setting an optimization function according to the target production requirements; inputting the first physical parameter, blanking parameter, and time series number at the current moment into the blanking data-driven model to obtain the second physical parameter within the target time window; performing a fusion calculation on the second physical parameter, and the result of the fusion calculation is used to substitute into the optimization function to calculate the decision value; iteratively replacing the blanking parameters based on a preset interval to obtain the optimal blanking parameters corresponding to the lowest decision value, and performing real-time adjustment on the granulation process based on the optimal blanking parameters; where, performing a fusion calculation on the second physical parameter includes: , where, represents the second physical parameter from the 1st to the i-th group within the time window, i represents the number of groups within the time window, and j represents the index of the summation term, represents the natural exponential function, represents the natural logarithm function, represents the second physical parameter of the i-th group.

[0015] The embodiments of the present application also provide a digital twin optimization device for the urea formaldehyde granulation process, including: a processor, a memory, and a system bus; where, the processor and the memory are connected through the system bus; the memory is used to store one or more programs, and the one or more programs include instructions, and when the instructions are executed by the processor, the processor executes the method described in the above embodiments.

[0016] In the digital twin optimization system, method, and equipment for the urea-formaldehyde granulation process provided above, through the construction of a high-precision digital twin model, the embodiments of the present application can improve the feeding optimization process from multiple aspects: First, it can achieve a comprehensive simulation of the entire granulation system, including complex processes such as material flow, heat transfer, and particle formation, thereby obtaining more comprehensive and accurate system state information; Second, the digital twin model can be used for large-scale virtual experiments and optimizations, avoiding the time-consuming and laborious actual experiments in traditional methods and greatly improving the optimization efficiency; Third, the digital twin model can interact with the actual production system in real time, continuously update and optimize model parameters, and achieve dynamic optimization and predictive maintenance. These improvements are expected to bring significant positive impacts, including improving product quality stability, reducing energy consumption, reducing material waste, and extending equipment life. The digital twin feeding optimization system for the urea-formaldehyde granulation process is expected to effectively solve the problems existing in traditional methods, such as slow optimization process, difficulty in adapting to complex working conditions, and inability to achieve global optimization. It will bring revolutionary progress to the granulation process, achieve precise control, intelligent decision-making, and continuous optimization, thereby significantly improving production efficiency and product quality while reducing production costs and environmental impacts. In addition, this digital twin-based optimization method also has good scalability and generality, and is expected to be widely applied in other similar chemical production processes, promoting the entire industry towards intelligent manufacturing and green production. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0018] Figure 1 Schematic block diagram of a digital twin optimization system for the urea-formaldehyde granulation process provided by an embodiment of the present application;

[0019] Figure 2 Flow chart of a feeding optimization method provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] Now, various exemplary embodiments of the present application will be described in detail with reference to the drawings. It should be noted that: Unless otherwise specifically stated, the relative arrangements, numerical expressions, and numerical values of the components and steps described in these embodiments do not limit the scope of the present application.

[0021] Those skilled in the art can understand that terms such as "first" and "second" in the embodiments of the present application are only used to distinguish different steps, devices or modules, etc., and neither represent any specific technical meaning nor indicate an inevitable logical order between them. It should also be understood that in the embodiments of the present application, "a plurality" may refer to two or more, and "at least one" may refer to one, two or more. It should also be understood that for any component, data or structure mentioned in the embodiments of the present application, without clear definition or contrary indication in the context, it can generally be understood as one or more. In addition, the term "and / or" in the present application is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in the present application generally represents an "or" relationship between the associated objects before and after. It should also be understood that the description of each embodiment of the present application emphasizes the differences between the embodiments, and their similarities can be referred to each other. For the sake of brevity, they will not be elaborated one by one.

[0022] At the same time, it should be understood that for the sake of description, the sizes of the various parts shown in the drawings are not drawn in actual proportional relationships. The following description of at least one exemplary embodiment is actually only illustrative and in no way limits the present application, its application or use. Technologies, methods and devices known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, the said technologies, methods and devices should be regarded as part of the specification. It should be noted that similar reference numerals and letters denote similar items in the following drawings, and thus, once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0023] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts fall within the scope of protection of the present application.

[0024] Figure 1Schematic block diagram of a digital twin optimization system for the urea-formaldehyde granulation process provided by an embodiment of the present application. It should be understood that the system shown in the figure is exemplary rather than restrictive. This means that the involved system architecture is not limited to a specific form or design, but is presented as an example. In other words, the architecture shown in the figure can be regarded as a means of expression to clearly describe relevant concepts and relationships, and does not exclude other forms of architecture. Therefore, when interpreting the architecture in the described picture, it should be understood that the model has flexibility and diversity, and its purpose is to provide an exemplary description rather than a restrictive regulation of a specific form. As Figure 1 described, a digital twin optimization system for the urea-formaldehyde granulation process according to an embodiment of the present application includes a digital twin simulation unit 101 and an optimization decision unit 102.

[0025] Furthermore, the digital twin simulation unit 101 constructs a multi-physical field coupling model of the granulation process and sets constraint parameters based on sampling data to fine-tune the feeding data-driven model using physical prior knowledge. Specifically, first, production data of urea-formaldehyde granulation at the current moment is collected to obtain the first physical parameters. It should be understood that various relevant data is collected from the actual production environment, including but not limited to: raw material characteristics, molecular weight, viscosity, solid content, etc. of urea-formaldehyde resin. Process parameters, spray pressure, liquid concentration, nozzle angle, tower temperature, etc. Equipment status, operating status of the spray tower, working status of the feeding device, etc. Product quality, particle size distribution, moisture content, bulk density, etc.

[0026] Data collection can also use a variety of sensors and intelligent devices, such as pressure sensors, temperature sensors, flow meters, online particle size analyzers, etc. The data can be transmitted to the central database through industrial Internet of Things (IIoT) protocols such as OPCUA or MQTT.

[0027] Then, the first physical parameters are simulated and calculated based on the multi-physical field coupling model of the granulation process to obtain the second physical parameters. Specifically, a geometric model of the urea-formaldehyde granulator is established, and the physical fields and coupling relationships of the internal urea-formaldehyde particles are defined. The construction of the geometric model needs to accurately reflect the structure and dimensions of the granulator, including key components such as the spray tower, nozzle, and feeding device. Based on the geometric model, mesh generation is performed. The purpose of mesh generation is to discretize the continuous physical field for numerical calculation. The fineness of mesh generation has an important impact on the accuracy of the simulation results. A multi-physical field coupling model including fluid mechanics, heat and mass transfer, and particle dynamics is constructed to describe the complex physical phenomena in the granulation process. In one embodiment, it may include:

[0028] ,

[0029] Among them, Denotes the partial derivative with respect to time t, Denotes density, Denotes the differential operator, Denotes the velocity vector, Denotes pressure, Denotes the dynamic viscosity, Denotes the acceleration due to gravity, Denotes the specific heat capacity, Denotes temperature, Denotes the thermal conductivity, Denotes the heat source term, Denotes concentration, Denotes the diffusion coefficient, Denotes the reaction source term, Denotes the mass of particle i, Denotes the derivative with respect to time t, Denotes the velocity vector of particle i, Denotes the contact force between particle i and particle j, Denotes the drag force between particle i and particle j, Denotes the moment of inertia of particle i, Denotes the angular velocity of particle i, Denotes the lever arm of particle i.

[0030] After constructing the simulation model, set the initial conditions and boundary conditions according to the first physical parameters, and perform iterative simulation operations on the production parameters within the target time window to obtain the second physical parameters. Input the first physical parameters, including raw material characteristics (such as the molecular weight, viscosity, solid content of urea-formaldehyde resin, etc.), process parameters (such as spray pressure, liquid concentration, nozzle angle, tower temperature, etc.), equipment status (such as the operating status of the spray tower, the working status of the feeding device, etc.) and product quality (such as particle size distribution, moisture content, bulk density, etc.). Run the multi-physics coupling model and perform iterative simulation operations. During the simulation operation process, various physical phenomena and their interactions need to be considered, and numerical solutions are carried out. The specific simulation operation process includes the following steps: Initialization, initialize the simulation model according to the input first physical parameters, including the setting of initial conditions and boundary conditions. Iterative calculation, perform multiple iterative calculations within the target time window. Each iterative calculation updates the physical fields, and according to the current physical parameters, updates the states of each physical field, including the hydrodynamics field, heat transfer and mass transfer field, and particle dynamics field. Parameter adjustment, adjust the parameters of the simulation model according to the results of the iterative calculation to ensure the accuracy and stability of the simulation results. Error evaluation, evaluate the error of the current iterative calculation, and adjust the iteration step size and calculation accuracy according to the error.

[0031] After completing the iterative simulation operation within the target time window, the second physical parameter is output. The second physical parameter may include the velocity field, pressure field, temperature field, concentration field of the fluid, and the movement trajectory of the particles, etc. These parameters can be used to analyze the physical phenomena in the granulation process and optimize the process parameters.

[0032] Next, the blanking data-driven model is fine-tuned and trained using the first physical parameter, the blanking parameter at the current moment, the time series number, and the second physical parameter. Specifically, a fine-tuning data set is constructed based on the first physical parameter, the blanking parameter at the current moment, the time series number, and the second physical parameter. Among them, the first physical parameter, the blanking parameter at the current moment, and the time series number can be used as the input of the model, and the second physical parameter is used as the output of the model. As described in the previous embodiments, the first physical parameter may include raw material characteristics (such as the molecular weight, viscosity, solid content of urea-formaldehyde resin, etc.), process parameters (such as spray pressure, liquid concentration, nozzle angle, tower temperature, etc.), equipment status (such as the operating status of the spray tower, the working status of the blanking device, etc.), and product quality (such as particle size distribution, moisture content, bulk density, etc.). These data can be collected through a variety of sensors and intelligent devices, such as pressure sensors, temperature sensors, flow meters, online particle size analyzers, etc. The blanking parameter at the current moment reflects the operating conditions during the blanking process at the current moment, for example, including the blanking rate: 500 - 2000 kg / h, the rotation speed of the granulation disk: 10 - 30 rpm, the inclination angle of the granulation disk: 30 - 60°, etc. The time series number is used to identify the data at different time points to ensure the timeliness and continuity of the data. The second physical parameter is a parameter obtained by simulating a multi-physical field coupling model, including the velocity field, pressure field, temperature field, concentration field of the fluid, and the movement trajectory of the particles, etc.

[0033] Then, based on a preset algorithm, the number of target hidden layers for updating the blanking data-driven model is determined. In one embodiment, the preset algorithm may include:

[0034] , where, represents the floor function, It represents the cyclic adjustment item. mod7 means taking values cyclically with a period of 7, and t represents the time window size of the second physical parameter. For example, when the time window size t of the second physical parameter is 9, tmod7 = 2. After determining the number of training neural layers, use the fine-tuning dataset to perform fine-tuning training on multiple target hidden layers of the pre-trained blanking data-driven model, and freeze the parameter updates of other hidden layers during the fine-tuning training process. During the training process, use optimization algorithms (such as Adam, SGD, etc.) to adjust the weights and biases of the target hidden layers to minimize the loss function (such as mean square error, cross entropy, etc.). During the fine-tuning training process, freeze the parameters of other hidden layers and keep them unchanged to avoid overfitting of the model. By using a neural network to fit the physical model, the response speed of the system in actual engineering applications can be improved. In addition, by selecting some hidden layers for update during the fine-tuning process, the system response speed can be further improved.

[0035] The optimization decision unit 102 sets an optimization function according to the target production requirements to calculate the decision value of the output of the blanking data-driven model, and uses the blanking-driven model and the decision value to adjust the blanking parameters of the granulation process in real time. In an implementation scenario (as Figure 2 shown, the specific method will be elaborated at Figure 2 ), set the optimization function according to the target production requirements; input the first physical parameter, blanking parameter, and time series number at the current moment into the blanking data-driven model to obtain the second physical parameter within the target time window; perform fusion calculation on the second physical parameter, and the result of the fusion calculation is used to substitute into the optimization function to calculate the decision value; iteratively replace the blanking parameters based on a preset interval to obtain the optimal blanking parameter corresponding to the lowest decision value, and perform real-time adjustment on the granulation process based on the optimal blanking parameter.

[0036] Figure 2 This is a schematic flowchart of a blanking optimization method provided by an embodiment of the present application. As Figure 2 shown, at step S201, set an optimization function according to the target production requirements. This function can include multiple targets, such as product quality, energy consumption, output, etc., and comprehensively consider them through a weighted method: , where, represents the i-th optimization target, is the corresponding weight. For example, in an embodiment, an optimization target function can be constructed: , where, PSD represents the particle size distribution, MC represents the moisture content, represents the energy consumption, represents the temperature deviation, and w1, w2, w3, w4 represent the weight coefficients.

[0037] At step S202, after constructing the optimization function, the first physical parameter, blanking parameter, and time series number at the current moment are input into the blanking data-driven model to obtain the second physical parameter within the target time window. As described in the previous embodiments, the blanking data-driven model is a machine learning model trained based on historical production data for predicting the physical parameters of the granulation process under given input conditions. The inputs of the model include: the first physical parameter, such as raw material particle size distribution, water content, temperature, etc.; the blanking parameter, such as blanking speed, blanking amount, etc.; and the time series number, which is used to represent time information. The output of the model is: the second physical parameter, the predicted physical parameter within the target time window, such as particle size distribution, strength, density, etc. The mathematical expression of the model can be summarized as: , where f represents the blanking data-driven model, is the first physical parameter vector, is the blanking parameter vector, t is the time series number, and Y is the predicted second physical parameter vector.

[0038] At step S203, a fusion calculation is performed on the second physical parameter, and the result of the fusion calculation is used to substitute into the optimization function to calculate the decision value. The first physical parameter , blanking parameter and time series number t at the current moment are input into the blanking data-driven model to obtain the predicted second physical parameter: , a fusion calculation is performed on Y to obtain the comprehensive index Z: Z = g(Y), where g is the fusion function, which can be a simple weighted average or a more complex non-linear function. In one embodiment, it includes: , where represents the second physical parameter from the 1st to the ith group within the time window, i represents the number of groups within the time window, j represents the index of the summation term, represents the natural exponential function, represents the natural logarithm function, represents the second physical parameter of the ith group.

[0039] At step S204, the blanking parameter is iteratively replaced based on a preset interval to obtain the optimal blanking parameter corresponding to the lowest decision value, and the granulation process is adjusted in real time based on the optimal blanking parameter. The blanking parameter is iteratively replaced , and the above steps are repeated to obtain the lowest decision value corresponding to the optimal blanking parameter : , where is the preset search space of the blanking parameter.

[0040] Based on the obtained optimal blanking parameter , perform real-time adjustment on the granulation process. This can include but is not limited to: adjusting the feeding speed and feeding amount, adjusting the operating parameters of the granulation equipment, such as rotation speed, temperature, etc., and adjusting other relevant process parameters. The frequency of real-time adjustment can be set according to production requirements and equipment characteristics, and can be between seconds and minutes.

[0041] By introducing the granulation process optimization method and system based on the data-driven model, this application can improve the product quality stability. By predicting and optimizing the feeding parameters in real time, the fluctuations in the granulation process are reduced, making the product quality more stable. Reduce energy consumption. The optimization decision unit takes into account the energy consumption factor. By precisely controlling the feeding parameters, unnecessary energy consumption is reduced. Improve production efficiency. The real-time adjustment mechanism enables the granulation process to quickly adapt to changes in raw materials and the environment, reduces the downtime for adjustment, and improves the overall production efficiency. Reduce human intervention. The system has a high degree of automation, reduces the dependence on the experience of operators, and reduces the risk of human errors. Support personalized production. By flexibly setting the optimization function, the system can adapt to different production requirements and support personalized production of multiple varieties and small batches. In addition, a knowledge base can be accumulated. The system continuously accumulates data and experience during operation, forming a valuable knowledge base, which is helpful for process improvement and new product development. Improve equipment utilization rate. Through optimized control, the idling and overloading of equipment are reduced, the service life of the equipment is extended, and the equipment utilization rate is improved.

[0042] Furthermore, the embodiment of this application also provides a digital twin optimization device for the urea-formaldehyde granulation process, including: a processor, a memory, and a system bus; the processor and the memory are connected through the system bus; the memory is used to store one or more programs, and the one or more programs include instructions, and when the instructions are executed by the processor, the processor executes any one of the above methods.

[0043] Furthermore, the embodiment of this application also provides a computer program product, which, when running on a terminal device, enables the terminal device to execute any one of the above processing methods.

[0044] As can be seen from the description of the above embodiments, those skilled in the art can clearly understand that all or part of the steps in the above-described embodiment methods can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions for causing a computer device (which can be a personal computer, a server, or a network communication device such as a media gateway, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of the present application.

[0045] It should be noted that the various embodiments in this specification are described in a progressive manner, and the key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0046] It should also be noted that in this text, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including the said element.

[0047] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A digital twin optimization system for the granulation process of urea-formaldehyde particles, characterized in that, Including: A digital twin simulation unit and an optimization decision-making unit; wherein, The digital twin simulation unit is used to construct a multi-physical field coupling model of the granulation process, and set constraint parameters based on sampling data to fine-tune the feeding data-driven model using physical prior knowledge; including: collecting production data of urea-formaldehyde particle granulation at the current moment to obtain the first physical parameters; performing simulation calculations on the first physical parameters based on the multi-physical field coupling model of the granulation process to obtain the second physical parameters; the first physical parameters include raw material characteristics, process parameters, equipment status, and product quality; the second physical parameters include the velocity field, pressure field, temperature field, concentration field of the fluid, and the movement trajectory of the particles; Construct a fine-tuning data set based on the first physical parameters, the feeding parameters at the current moment, the time series number, and the second physical parameters; the feeding parameters include the feeding rate, the rotation speed of the granulation disk, and the inclination angle of the granulation disk; Determine the number of target hidden layers for updating the feeding data-driven model based on a preset algorithm; Use the fine-tuning data set to perform fine-tuning training on multiple target hidden layers of the pre-trained feeding data-driven model, and freeze the parameter updates of other hidden layers during the fine-tuning training process; The determining the number of target hidden layers for updating the feeding data-driven model based on a preset algorithm includes: , where represents the floor function, represents the cyclic adjustment term, mod7 represents taking values cyclically with a period of 7. When t = 9, t mod 7 = 2, and t represents the time window size of the second physical parameter; the optimization decision unit is used to set an optimization function according to the target production requirements; input the first physical parameter, blanking parameter, and time series number at the current moment into the blanking data-driven model to obtain the second physical parameter within the target time window; perform fusion calculation on the second physical parameter within the target time window, and the result of the fusion calculation is used to substitute into the optimization function to calculate the decision value; iteratively replace the blanking parameter based on the preset interval to obtain the lowest decision value corresponding optimal blanking parameter , and perform real-time adjustment on the granulation process based on the optimal blanking parameter ; where ; where is the preset search space for the blanking parameter; the first physical parameter X1, the blanking parameter X2, and the time series number ; , Y is the second physical parameter within the target time window; g is the fusion function Among them, the fusion calculation of the second physical parameters within the target time window includes: , where represents the second physical parameter of the 1st to the i-th groups within the target time window, i represents the number of groups within the target time window, and j represents the index of the summation term represents the natural exponential function represents the second physical parameter of the i-th group within the target time window 2. The digital twin optimization system according to claim 1, characterized in that Wherein, Performing simulation calculations on the first physical parameters using the multi-physical field coupling model of the granulation process to obtain the second physical parameters, including: Establish a geometric model of the urea-formaldehyde particle granulator, and define the physical fields and coupling relationships of the contained urea-formaldehyde particles; Set the initial conditions and boundary conditions according to the first physical parameters, and perform iterative simulation operations on the production parameters within the target time window to obtain the second physical parameters.

3. A digital twin optimization method for the urea-formaldehyde granulation process applying the system according to any one of claims 1-2, characterized in that, Including: Construct a multi-physical field coupling model of the granulation process, and set constraint parameters based on sampling data to fine-tune the feeding data-driven model using physical prior knowledge; including: collecting production data of urea-formaldehyde particle granulation at the current moment to obtain the first physical parameters; performing simulation calculations on the first physical parameters based on the multi-physical field coupling model of the granulation process to obtain the second physical parameters; the first physical parameters include raw material characteristics, process parameters, equipment status, and product quality; the feeding parameters include the feeding rate, the rotation speed of the granulation disk, and the inclination angle of the granulation disk; the second physical parameters include the velocity field, pressure field, temperature field, concentration field of the fluid, and the movement trajectory of the particles; Construct a fine-tuning data set based on the first physical parameters, the feeding parameters at the current moment, the time series number, and the second physical parameters; Determine the number of target hidden layers for updating the feeding data-driven model based on a preset algorithm; Use the fine-tuning data set to perform fine-tuning training on multiple target hidden layers of the pre-trained feeding data-driven model, and freeze the parameter updates of other hidden layers during the fine-tuning training process; The determining the number of target hidden layers for updating the feeding data-driven model based on a preset algorithm includes: , where represents the floor function, represents the cyclic adjustment term, mod7 means taking values cyclically with a period of 7. When t = 9, t mod 7 = 2, and t represents the time window size of the second physical parameter; Set an optimization function according to the target production requirements; input the first physical parameter, blanking parameter, and time series number at the current moment into the blanking data-driven model to obtain the second physical parameter within the target time window; perform fusion calculation on the second physical parameter within the target time window, and the result of the fusion calculation is used to substitute into the optimization function to calculate the decision value; iteratively replace the blanking parameter based on the preset interval to obtain the optimal blanking parameter corresponding to the lowest decision value , and based on the optimal blanking parameter perform real-time adjustment on the granulation process; where ; where is the preset search space for blanking parameters; minJ is the lowest decision value; the first physical parameter X1, blanking parameter X2, and time series number ; , Y is the second physical parameter within the target time window; g is the fusion function; Among them, the fusion calculation of the second physical parameters within the target time window includes: , where, represents the second physical parameter of the 1st to the i-th groups within the target time window, i represents the number of groups within the target time window, and j represents the index of the summation term, represents the natural exponential function, represents the second physical parameter of the i-th group within the target time window.

4. A digital twin optimization device for the granulation process of urea-formaldehyde particles, characterized in that, Including: A processor, a memory, and a system bus; wherein the processor and the memory are connected through the system bus; The memory is configured to store one or more programs, the one or more programs including instructions that, when executed by the processor, cause the processor to execute the method recited in claim 3.

Images