Thermodynamic behavior analysis method and system for kiln high-temperature synthesis

Abstract

The invention relates to the technical field of thermal data processing, and discloses a thermodynamic behavior analysis method and system for high-temperature synthesis of a kiln, and the method comprises the steps: collecting kiln operation data through a kiln sensor, building a nonlinear thermodynamic model, and describing the dynamic change of the thermodynamic behavior in the kiln; constructing a thermodynamic behavior map by using the kiln operation data, and predicting thermodynamic behaviors under different operation conditions; adjusting control parameters of a nonlinear kinetic model based on the thermodynamic behavior result; and performing joint iteration on the nonlinear thermodynamic model and the thermodynamic behavior map to realize optimization of kiln operation. The temperature control precision, the energy efficiency and the stability of the kiln are remarkably improved, the energy consumption is reduced, the quality of finished products is improved, and the limitation of a traditional method in the aspects of dynamic adjustment and real-time optimization is overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of hot data processing, and specifically to a method and system for analyzing the thermodynamic behavior of high-temperature synthesis in a kiln furnace. Background Art

[0002] Traditional methods for thermodynamic analysis of kiln furnaces mainly rely on computational fluid dynamics (CFD) models and thermodynamic equilibrium modeling. The CFD model calculates the distribution of various physical quantities in the kiln furnace through numerical simulation. However, its computational complexity is high and it is not easy to apply in real time, usually requiring a large amount of computational resources and time, which is not feasible for the real-time regulation requirements in the actual production process. On the other hand, thermodynamic equilibrium modeling (such as KilnSimu) is usually based on the equilibrium state assumption, ignoring the rapid dynamic changes in the actual operation of the kiln furnace, which makes its effect limited when dealing with the thermodynamic behavior in complex and non-equilibrium states.

[0003] In addition, in recent years, with the development of machine learning and data-driven models, many studies have begun to explore the use of artificial intelligence (AI) algorithms to optimize the thermodynamic behavior of kiln furnaces. However, existing machine learning applications mostly rely on single data-driven models, such as neural networks, support vector machines (SVM), etc. These models often lack constraints on physical laws, resulting in deficiencies in prediction accuracy and reliability. Especially when dealing with complex thermodynamic and chemical reaction processes, simply relying on data-driven models may violate the basic principles of thermodynamics, leading to model outputs that do not conform to actual physical conditions. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the problems of low prediction accuracy, poor optimization efficiency, and difficulty in ensuring physical consistency in the thermodynamic behavior prediction during the high-temperature synthesis process of existing kiln furnaces.

[0006] To solve the above technical problems, the present invention provides the following technical solution: A method for analyzing the thermodynamic behavior of high-temperature synthesis in a kiln furnace, comprising: collecting kiln furnace operation data through kiln furnace sensors, establishing a non-linear thermodynamic model to describe the dynamic changes of the internal thermodynamic behavior of the kiln furnace;

[0007] Using the kiln furnace operation data, constructing a thermodynamic behavior map to predict the thermodynamic behavior under different operating conditions;

[0008] Based on the thermodynamic behavior results, adjusting the control parameters of the non-linear dynamic model;

[0009] Performing joint iteration on the non-linear thermodynamic model and the thermodynamic behavior map to achieve the optimization of kiln furnace operation.

[0010] As a preferred embodiment of the thermodynamic behavior analysis method for high-temperature synthesis in a kiln according to the present invention, wherein: the operating data of the kiln includes real-time operating data and historical operating data;

[0011] The real-time operating data includes real-time temperature, real-time pressure, real-time gas flow, real-time atmosphere composition, real-time reaction rate, real-time energy consumption, and real-time material quality.

[0012] As a preferred embodiment of the thermodynamic behavior analysis method for high-temperature synthesis in a kiln according to the present invention, wherein: a heat transfer equation, a reaction kinetics equation, and a mass transfer equation are established using the preprocessed operating data of the kiln;

[0013] The heat transfer equation, the reaction kinetics equation, and the mass transfer equation are coupled to establish the non-linear thermo-dynamic model.

[0014] As a preferred embodiment of the thermodynamic behavior analysis method for high-temperature synthesis in a kiln according to the present invention, wherein: constructing the thermodynamic behavior map includes defining each region inside the kiln as a node; the edges represent the interaction relationships between the nodes, including heat transfer, reaction substance flow, and gas flow transmission;

[0015] The weight of the edge represents the heat transfer efficiency, the reactant flow rate, and the gas flow intensity; the weight of the edge is updated at each time step through the real-time operating data.

[0016] As a preferred embodiment of the thermodynamic behavior analysis method for high-temperature synthesis in a kiln according to the present invention, wherein: predicting the thermodynamic behavior under different operating conditions includes inputting the real-time operating data as node features into a graph neural network, constructing a thermodynamic behavior map model, training the thermodynamic behavior map model using the historical operating data, and capturing the dynamic changes in heat transfer, reactant flow, and gas flow transmission;

[0017] The mean square error is used as the loss function to evaluate the prediction error of the thermodynamic behavior map model, and the Adam optimization algorithm is used to adjust the graph neural network to optimize the thermodynamic behavior map model;

[0018] Based on the trained thermodynamic behavior map model, the thermodynamic behavior under different operating conditions is predicted;

[0019] The thermodynamic behavior includes: the temperature distribution in each region inside the kiln, the change in the reaction rate inside the kiln, and the distribution change of the atmosphere composition in each region of the kiln.

[0020] As a preferred embodiment of the thermodynamic behavior analysis method for high-temperature synthesis in a kiln according to the present invention, wherein: the control parameters of the non-linear dynamics model include the heat transfer coefficient, the reaction rate constant, the diffusion coefficient, and the reactant concentration.

[0021] As a preferred embodiment of the thermodynamic behavior analysis method for high-temperature synthesis in a kiln according to the present invention, wherein: calculating the residual between the thermodynamic behavior and the actual operation data of the kiln; defining the objective function f(θ n ) of the joint iteration to minimize the residual;

[0022] By calculating the gradient and Hessian matrix of the objective function f(θ n ), updating θ n of the objective function f(θ n ) using the improved Newton method, the formula is expressed as:

[0023]

[0024] wherein, represents the gradient of the objective function, and H(θ n ) is the Hessian matrix, representing the second derivative of the objective function; θ n+1 represents the control parameter vector of the updated nonlinear dynamics model; n represents the current iteration step, and n + 1 represents the next iteration step;

[0025] θ n represents the control parameter vector of the current nonlinear dynamics model, and the formula is expressed as:

[0026] θ n = [k n , k r (T) n , D n , C n , w1, w2, …, w m

[0027] wherein, k n represents the heat transfer coefficient at the current iteration step, k r (T) n represents the reaction rate constant at the current iteration step, D n represents the diffusion coefficient at the current iteration step, C n represents the reactant concentration at the current iteration step, w1, w2, …, w m represents the weight of the edge in the thermodynamic behavior atlas model, and m represents the number of edges;

[0028] Set the original learning step size as α0. When the residual increases, the learning step size increases to x times the original step size; when the residual decreases, the learning step size decreases to y times the original step size; wherein, x and y are constants;

[0029] ​The joint iteration includes predicting the thermodynamic behavior through the thermodynamic behavior atlas model by using real-time operation data, outputting the relationships between nodes, and adjusting the control parameters of the non-linear thermodynamics model;

[0030] In the joint iteration, when the residual is less than the set threshold, it is determined that the optimization has converged, and the iteration process is terminated; and after each round of optimization, it is checked whether the control parameters of the updated non-linear dynamics model violate the physical constraints. If they violate the physical constraints, the control parameters of the non-linear dynamics model are adjusted through a penalty function.

[0031] As a preferred solution of the thermodynamic behavior analysis method for high-temperature synthesis in a kiln according to the present invention, wherein: the non-linear thermodynamics model module collects the operation data of the kiln through a kiln sensor, establishes a non-linear thermodynamics model, and describes the dynamic changes of the thermodynamic behavior inside the kiln;

[0032] The thermodynamic behavior atlas module uses the operation data of the kiln to construct a thermodynamic behavior atlas and predict the thermodynamic behavior under different operating conditions;

[0033] The adjustment module adjusts the control parameters of the non-linear dynamics model based on the thermodynamic behavior results;

[0034] The joint iteration module performs joint iteration on the non-linear thermodynamics model and the thermodynamic behavior atlas to achieve the optimization of the kiln operation.

[0035] A computer device includes: a memory and a processor; the memory stores a computer program, and is characterized in that: when the processor executes the computer program, the steps of the method described in any one of the present invention are implemented.

[0036] A computer-readable storage medium stores a computer program thereon, and is characterized in that: when the computer program is executed by a processor, the steps of the method described in any one of the present invention are implemented.

[0037] Advantages of the present invention: The thermodynamic behavior analysis method for high-temperature synthesis in a kiln provided by the present invention realizes the accurate prediction and optimization of the thermodynamic behavior of the kiln by combining a non-linear thermodynamics model and a thermodynamic behavior atlas. By collecting the operation data of the kiln in real time, a non-linear thermodynamics model is established to describe the dynamic changes of the thermodynamic behavior inside the kiln. The atlas learning method is used to construct a thermodynamic behavior atlas to predict the thermodynamic behavior under different operating conditions and capture the dynamic changes of heat transfer, reaction material flow, and gas flow transmission. Based on the results of the thermodynamic behavior, the control parameters of the non-linear dynamics model are adjusted to optimize the operation of the kiln. The adaptive residual-based Newton method optimization algorithm is adopted, and through joint iterative optimization, it is ensured that the two models are adjusted synchronously to accurately control the thermodynamic behavior of each area of the kiln. By introducing physical constraint conditions, the optimization process ensures compliance with the laws of thermodynamics, improving the energy efficiency, temperature control accuracy, and stability of the kiln. Brief Description of the Drawings

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0039] Figure 1 It is the overall flowchart of a thermodynamic behavior analysis method for high-temperature synthesis in a kiln provided by the first embodiment of the present invention. Detailed Embodiments

[0040] To make the above objects, features, and advantages of the present invention more obvious and understandable, the detailed embodiments of the present invention will be described below with reference to the drawings of the specification. Obviously, the described embodiments are some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0041] Example 1, referring to Figure 1 , which is an embodiment of the present invention, provides a thermodynamic behavior analysis method for high-temperature synthesis in a kiln, including:

[0042] S1: Collect the operation data of the kiln through the kiln sensors, and establish a non-linear thermodynamics model to describe the dynamic changes of the thermodynamic behavior inside the kiln.

[0043] The operation data of the kiln includes real-time operation data and historical operation data.

[0044] The real-time operating data includes real-time temperature, real-time pressure, real-time air flow, real-time atmosphere composition, real-time reaction rate, real-time energy consumption, and real-time material quality.

[0045] Clean and correct the real-time operating data, exclude outliers and missing data, and convert data from different sources or types into a consistent format through standardization processing.

[0046] Couple the heat transfer equation, reaction kinetics equation, and mass transfer equation to establish a non-linear thermo-dynamic model.

[0047] The heat transfer equation describes the heat transfer process in the kiln furnace, and is expressed by the formula:

[0048]

[0049] Among them, ρ represents the material density, c p represents the specific heat capacity, T represents the temperature, k represents the heat transfer coefficient, and Q represents the heat source term.

[0050] The reaction rate is usually strongly affected by temperature, and the reaction kinetics equation needs to describe the relationship between the reaction rate and temperature change.

[0051] The reaction kinetics equation is:

[0052]

[0053] Among them, C is the concentration of the reactant, k r (T) represents the reaction rate constant, and C0 is the initial concentration of the reactant.

[0054] The reaction rate constant k r (T) is expressed by the Arrhenius equation:

[0055]

[0056] Among them, A is the frequency factor, E a is the activation energy, R is the gas constant, and T is the temperature.

[0057] For the diffusion process of the gas components in the kiln furnace, establish a mass transfer equation to simulate the diffusion and flow of substances.

[0058] The mass transfer equation is:

[0059]

[0060] Among them, C is the concentration of the reactant, D is the diffusion coefficient, and R(C) is the reaction rate.

[0061] Furthermore, by accurately collecting real-time operation data of the kiln (such as temperature, pressure, gas flow, etc.) and establishing a non-linear thermodynamics model, the present invention can accurately describe the thermodynamic behavior inside the kiln, especially the dynamic changes under high-temperature and high-pressure environments. Compared with traditional methods, the non-linear thermodynamics model can better handle the rapid temperature control, reaction rate, and changes in atmosphere composition of the kiln, providing a more accurate and real-time simulation of thermodynamic behavior, and laying a solid foundation for subsequent optimization and control of kiln operation.

[0062] S2: Using the operation data of the kiln, construct a thermodynamic behavior map to predict the thermodynamic behavior under different operating conditions.

[0063] In the thermodynamic behavior map of the kiln, each functional area inside the kiln is defined as a node, and each node represents the thermodynamic state of a specific area. The edges represent the interaction relationships between different nodes, including heat transfer, which represents the heat exchange between different areas of the kiln, and the weight represents the efficiency of heat transfer. The flow of reactants represents the flow of chemical reactants and products between different areas, and the weight represents the flow rate of reactants. The gas flow transmission represents the flow of gas between different areas of the kiln, and the weight represents the flow intensity of gas flow transmission.

[0064] Through real-time operation data, the edge weights are dynamically updated at each time step.

[0065] The real-time temperature change affects the heat transfer efficiency. When the temperature increases, the heat transfer efficiency increases, and the weight of the edge increases; when the temperature decreases, the transfer efficiency decreases, and the weight of the edge decreases.

[0066] The reaction rate and concentration changes will affect the flow rate of material flow. Temperature and atmosphere composition factors directly affect the reaction rate, and thus affect the flow rate of reactants. In the reaction zone, when the temperature rises or the oxygen concentration increases, the reaction rate increases, resulting in an accelerated flow rate of reactants and an increase in the edge weight.

[0067] The flow intensity of the gas inside the kiln will affect the flow path and speed of the gas flow. Real-time gas flow data adjusts the weight of gas flow transmission. When the gas flow rate increases, the edge weight increases, and vice versa.

[0068] Taking the real-time operation data as node features, input it into the graph neural network. Through the graph neural network, train the thermodynamic behavior map model to learn the complex relationships between nodes and capture the dynamic changes of heat transfer, reactant flow, and gas flow transmission between different areas. During the training process, the mean square error is used as the loss function to evaluate the prediction error of the thermodynamic behavior map model.

[0069] The Adam optimization algorithm is used to adjust the parameters of the graph neural network, optimizing the thermodynamic behavior atlas model to enable it to more accurately predict the thermodynamic behavior under different operating conditions.

[0070] Based on the trained thermodynamic behavior atlas model, predicting the thermodynamic behavior under different operating conditions includes: predicting the temperature distribution in each region, which represents the prediction results through heat transfer and temperature changes, and providing the dynamic changes of the temperature in different regions inside the kiln.

[0071] Predicting the change in the reaction rate inside the kiln, especially the influence of temperature and atmosphere composition in the reaction zone on the reaction rate.

[0072] Predicting the distribution change of the atmosphere composition in each region of the kiln.

[0073] As the operating conditions of the kiln change, the thermodynamic behavior atlas model will update the prediction results in real time, providing a reference for subsequent optimization and operation adjustment.

[0074] Furthermore, by combining the real-time operation data and historical data of the kiln, a thermodynamic behavior atlas is constructed, enabling each functional area of the kiln (such as the heating area, cooling area, etc.) to accurately represent its thermodynamic state and mutual relationship. This step effectively captures the dynamic changes of heat transfer, reactant flow, and gas flow transmission by learning the complex non-linear interactions between nodes through the graph neural network. Compared with traditional assumption-based models, the thermodynamic behavior atlas can provide real-time and accurate prediction results, thus greatly improving the adaptability and accuracy of kiln operation.

[0075] S3: Based on the thermodynamic behavior results, adjust the control parameters of the non-linear kinetic model.

[0076] The control parameters of the non-linear kinetic model are the key factors affecting the thermodynamic behavior, including:

[0077] The heat transfer coefficient (k), which reflects the heat transfer efficiency in different regions; the reaction rate constant (k r (T)), which describes the non-linear relationship between the reaction rate and temperature change; the diffusion coefficient (D), which controls the diffusion rate of gases and substances, especially the diffusion of the atmosphere composition; the reactant concentration (C), which controls the concentration factor in reaction kinetics.

[0078] According to the temperature distribution in the thermodynamic behavior results, adjust the heat transfer coefficient in the non-linear thermodynamics model. When the temperature of the kiln rises, the heat transfer coefficient and the reaction rate constant increase; when the temperature of the kiln drops, the heat transfer coefficient and the reaction rate constant decrease to meet the requirements under different operating conditions.

[0079] According to the change in the reaction rate of the thermodynamic behavior atlas, adjust the reaction rate constant kr (T); when the reaction rate constant k r (T) increases, the reaction rate speeds up, thereby reducing the concentration of the reactants; when the reaction rate constant decreases, the reaction rate slows down and the reactant concentration increases. The adjustment of the reactant concentration can accurately reflect the change in the reaction rate, thus ensuring the stability and efficiency of the reaction in the kiln.

[0080] This adjustment mechanism ensures that the thermodynamic behavior in the kiln can be precisely controlled under different operating conditions, achieving the dynamic balance and coordination of the control parameters of the non-linear kinetic model.

[0081] Furthermore, by dynamically adjusting the control parameters of the non-linear thermodynamics model (such as heat transfer coefficient, reaction rate constant, etc.), the model parameters can be optimized in real time according to the actual thermodynamic behavior results of the kiln (such as temperature distribution, reaction rate, etc.). This feedback mechanism ensures that the thermodynamic behavior of the kiln remains stable under different operating conditions and can effectively cope with the complex non-equilibrium kinetic changes inside the kiln. By precisely adjusting the control parameters, the present invention improves the stability, accuracy and efficiency of kiln operation, and avoids the errors and instabilities caused by inaccurate parameters in the traditional methods.

[0082] S4: Jointly iterate the non-linear thermodynamics model and the thermodynamic behavior map to optimize the kiln operation.

[0083] The adaptive residual-based Newton method is used as the optimization algorithm. By iteratively minimizing the residual between the predicted thermodynamic behavior and the actual operation data, and introducing physical constraint conditions and an adaptive parameter adjustment mechanism, the accuracy and physical consistency of the optimization of the kiln thermodynamic behavior are ensured.

[0084] In each round of iterative optimization, the formula for calculating the residual between the predicted thermodynamic behavior and the actual operation data of the kiln is:

[0085] Residua i =|Prediction i -ActualData i |

[0086] where Residua i represents the residual of the i-th data point, that is, the difference between the model prediction value and the actual data; Prediction i represents the predicted thermodynamic behavior of the i-th data point, and ActualData i represents the actual data of the i-th data point, which is the actual operation data collected by the kiln sensor in real time.

[0087] The objective function f(θ n) is the sum of the residuals of all data points in the model. The ultimate goal is to minimize this objective function so that the model prediction results are as close as possible to the actual data, expressed as:

[0088]

[0089] where i represents the data point index and I represents the total number of data points; the data point represents the comparison result between the real-time data and the model prediction data inside the kiln at a specific moment and in a specific area.

[0090] The second derivative of the objective function is calculated using Newton's method to accelerate convergence. The update formula is:

[0091]

[0092] where θ n is the control parameter vector of the current non-linear dynamics model, H(θ n ) is the Hessian matrix, is the gradient of the objective function, and θ n+1 represents the control parameter vector of the updated non-linear dynamics model; n represents the current iteration step, and n + 1 represents the next iteration step.

[0093] θ n is expressed as:

[0094] θ n = [k n , k r (T) n , D n , C n , w1, w2, …, w m

[0095] where k n is the heat transfer coefficient at the current iteration step, k r (T) n is the reaction rate constant at the current iteration step, D n is the diffusion coefficient at the current iteration step, C n is the reactant concentration at the current iteration step, and w1, w2, …, w m are the weights of the edges in the spectral model, and m represents the number of edges.

[0096] Set the original learning step size as α0. When Residua i increases, the learning step size increases to x times the original step size, making the optimization process quickly approach the target solution; when Residua i decreases, the learning step size decreases to y times the original step size, finely adjusting the parameters and avoiding over-adjustment. x and y are set according to empirical values.

[0097] To ensure that the optimization results conform to physical laws, physical constraint conditions are introduced into the optimization process. After each optimization iteration, the penalty function will control the update of the model parameters to ensure that the basic principles of thermodynamics are not violated, expressed as:

[0098] L penalty = λ·max(0, violated constraints) 2

[0099] where λ is the penalty coefficient, L penalty represents the penalty function, and max represents the maximum value.

[0100] In each round of optimization, the thermodynamic behavior map model and the non-linear thermodynamics model will be alternately optimized. According to the real-time operation data, the thermodynamic behavior map is predicted and the relationships between nodes are output.

[0101] According to the thermodynamic behavior map output by the thermodynamic behavior map model, the control parameters of the non-linear dynamics model are adjusted to optimize the thermodynamic behavior of the kiln. Through joint iterative optimization, it is ensured that the optimization processes of the two models are synchronized, and finally the optimal operating parameters are obtained.

[0102] During the iterative process, when the residual is less than the set threshold, it is judged that the optimization has converged and the iterative process ends; and after each round of optimization, it is checked whether the parameter update violates the physical constraints (energy conservation, mass conservation). If the physical laws are violated, the control parameters of the non-linear dynamics model are adjusted through the penalty function.

[0103] With the optimized parameters, the thermodynamic behavior of each area of the kiln is precisely controlled. By jointly optimizing the model parameters, the operating parameters such as the heating rate, atmosphere composition, and cooling rate of the kiln are precisely optimized to ensure the stable and efficient operation of the kiln under different working conditions. The adaptive parameter adjustment mechanism ensures that the optimization process can dynamically adjust the optimization step size under different working conditions, so that the optimization results can converge stably under various operating conditions.

[0104] Furthermore, through the adaptive residual-based Newton method optimization algorithm, jointly optimizing the non-linear thermodynamics model and the thermodynamic behavior map, the present invention can achieve precise adjustment of the kiln operating parameters. Physical constraint conditions are introduced during the optimization process to ensure that the optimization results conform to the laws of thermodynamics and avoid the defects of pure data-driven models. Joint iterative optimization ensures the synchronous adjustment of the two models, enabling the kiln to achieve the optimal thermodynamic behavior under different operating conditions, significantly improving the energy efficiency, temperature control accuracy, and operating stability of the kiln. In addition, through dynamic adjustment of the learning step size and real-time feedback mechanism, the optimization process has strong adaptability and high efficiency, and can quickly achieve real-time optimization in the actual production environment.

[0105] Embodiment 2, an embodiment of the present invention, provides a thermodynamic behavior analysis system for high-temperature synthesis in a kiln, including:

[0106] A non-linear thermodynamic model module, which collects the operation data of the kiln through kiln sensors, establishes a non-linear thermodynamic model, and describes the dynamic changes of the thermodynamic behavior inside the kiln.

[0107] A thermodynamic behavior atlas module, which uses the operation data of the kiln to construct a thermodynamic behavior atlas and predict the thermodynamic behavior under different operating conditions.

[0108] An adjustment module, which adjusts the control parameters of the non-linear kinetic model based on the thermodynamic behavior results.

[0109] A joint iteration module, which performs joint iteration on the non-linear thermodynamic model and the thermodynamic behavior atlas to achieve the optimization of kiln operation.

[0110] Embodiment 3, an embodiment of the present invention, is different from the previous two embodiments in that:

[0111] If the said function is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that makes contributions to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. And the aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0112] The logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch instructions from the instruction execution system, apparatus, or device and execute the instructions), or in combination with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

[0113] More specific examples (non-exhaustive list) of computer-readable media include the following: electrical connections (electronic devices) having one or more wirings, portable computer disk cartridges (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber devices, and portable compact disc read-only memory (CDROM). Additionally, the computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or otherwise processing as appropriate, and then stored in a computer memory.

[0114] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGA), field programmable gate arrays (FPGA), etc.

[0115] Example 4, an embodiment of the present invention, provides a method and system for analyzing the thermodynamic behavior of high-temperature synthesis in a kiln. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0116] By analyzing the thermodynamic behavior during the high-temperature synthesis process in the kiln, the sintering process of silicon carbide ceramics is optimized. The sintering process of silicon carbide ceramics requires precise temperature control, atmosphere regulation, and reaction rate control. Therefore, this method is of great significance for optimizing the sintering process, improving product quality, and reducing energy consumption.

[0117] Kiln configuration: A standard continuous electric furnace is used, with a furnace body size of 1.5 meters (length) x 0.8 meters (width) x 1.0 meters (height), and a temperature range from room temperature to 1600 °C. This electric furnace can achieve precise temperature control at high temperatures and is suitable for sintering ceramic materials.

[0118] In the experiment, a variety of sensors are used for real-time data acquisition, including: Thermocouple sensors: Installed at different positions inside the kiln to monitor the temperature of each area of the kiln in real time, with a data acquisition frequency of 1 Hz. Pressure sensors: Monitor the air pressure inside the kiln, especially in the reaction zone, with a data acquisition frequency of 1 Hz. Gas analyzers: Used to measure the concentrations of oxygen, nitrogen, and carbon dioxide in the kiln atmosphere, with a data acquisition frequency of 0.5 Hz.

[0119] Before the input of all data into the non-linear thermodynamics model, standardization processing is carried out to ensure the unity of different types of data and exclude missing data and significantly abnormal measurement values. Establishment of the non-linear thermodynamics model: In this experiment, a non-linear model of the thermodynamic behavior inside the kiln is established by coupling the heat transfer, reaction kinetics, and mass transfer equations. The heat transfer equation describes the distribution and change of temperature inside the kiln, the reaction kinetics equation is used to simulate the chemical reactions during the silicon carbide sintering process, and the mass transfer equation describes the change of the atmosphere composition. By coupling these three equations, the thermodynamic state of each region inside the kiln can be accurately simulated.

[0120] Based on the real-time collected operation data of the kiln, a thermodynamic behavior map of the kiln is constructed. The nodes in the map represent the thermodynamic states of different regions of the kiln, and the edges represent the relationships of heat transfer, reactant flow, and gas flow transmission between these regions. Through the graph neural network (GNN), the model learns the interaction relationships between these regions, thereby being able to predict the thermodynamic behavior under different operating conditions.

[0121] In each round of iterative optimization, the non-linear kinetics model and the thermodynamic behavior map are optimized by the adaptive residual-based Newton method. Specifically, the residual between the model prediction result and the actual kiln operation data is calculated, and this residual is minimized through the optimization algorithm. Physical constraint conditions (such as energy conservation and mass conservation) are checked after each iteration to ensure that the optimization result conforms to the laws of thermodynamics.

[0122] A total of 500 optimization iterations were carried out in this experiment, and the running time for each iteration was 5 minutes. Data such as temperature, reaction rate, and atmosphere composition were collected during the experiment, and all data were accurate to two decimal places. The following are some of the experimental data:

[0123] Comparison of predicted temperature and actual value: Predicted temperature (°C): 1500.05, 1499.75, 1500.20; Actual temperature (°C): 1500.00, 1500.00, 1500.00; Residual: 0.05 °C, 0.25 °C, 0.20 °C

[0124] Comparison of predicted reaction rate and actual value: Predicted reaction rate: 0.80, 0.85, 0.78; Actual reaction rate: 0.81, 0.86, 0.79; Residual: 0.01, 0.01, 0.01

[0125] Comparison of energy consumption: Energy consumption before original optimization: 1200 kWh; Energy consumption after optimization: 1145 kWh; Energy consumption reduction: 4.58%

[0126] Based on the above data, the advantages and disadvantages of the present invention and traditional methods were compared. In traditional methods, the temperature control of the kiln usually relies on empirical formulas and cannot be dynamically adjusted according to actual data. In contrast, the present invention combines a non-linear thermodynamics model and a spectral learning method, enabling real-time optimization of thermodynamic behavior and precise adjustment of operating parameters in each iteration. The residual value gradually decreases, indicating continuous improvement in prediction accuracy; while the energy consumption is reduced by approximately 4.58%, showing a significant energy-saving effect compared with traditional methods.

[0127] The joint iteration method of the present invention effectively improves the yield and product quality by continuously optimizing key parameters such as temperature, reaction rate, and atmosphere composition. In addition, the introduction of physical constraints (such as energy and mass conservation) ensures that the optimization process complies with the basic principles of thermodynamics and avoids problems of deviating from physical laws commonly seen in traditional methods. Different from traditional methods that rely on static adjustment of parameters, the present invention can dynamically learn and real-time adjust control parameters, stably optimizing the kiln operation under different working conditions, demonstrating its obvious advantages in industrial applications.

[0128] The practical application of establishing a non-linear thermodynamics model and a thermodynamic behavior spectrum through real-time sensor data and realizing the optimization of kiln operation through joint iteration optimization. Experimental results show that the present invention has significant advantages in temperature control accuracy, reaction rate regulation, and energy consumption optimization, providing an innovative solution for improving the operation efficiency and product quality of the kiln.

[0129] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A thermodynamic behavior analysis method for high-temperature synthesis in a kiln, characterized in that Including: Collecting the operation data of the kiln through kiln sensors, establishing a non-linear thermo-dynamic model to describe the dynamic changes of the thermodynamic behavior inside the kiln; Using the operation data of the kiln to construct a thermodynamic behavior map to predict the thermodynamic behavior under different operating conditions; Based on the results of the thermodynamic behavior, adjusting the control parameters of the non-linear dynamic model; Performing joint iteration on the non-linear thermo-dynamic model and the thermodynamic behavior map to achieve the optimization of the kiln operation.

2. The thermodynamic behavior analysis method for high-temperature synthesis in a kiln as described in claim 1, characterized in that: The operation data of the kiln includes real-time operation data and historical operation data; The real-time operation data includes real-time temperature, real-time pressure, real-time air flow, real-time atmosphere composition, real-time reaction rate, real-time energy consumption, and real-time material quality.

3. The thermodynamic behavior analysis method for high-temperature synthesis in a kiln according to claim 2, characterized in that: Using the preprocessed operation data of the kiln to establish a heat transfer equation, a reaction kinetics equation, and a mass transfer equation; Coupling the heat transfer equation, the reaction kinetics equation, and the mass transfer equation to establish the non-linear thermo-dynamic model.

4. The thermodynamic behavior analysis method for high-temperature synthesis in a kiln as described in claim 3, characterized in that: The construction of the thermodynamic behavior map includes defining each region inside the kiln as a node; the edges represent the interaction relationships between the nodes, including heat transfer, reaction substance flow, and air flow transmission; The weight of the edge represents the heat transfer efficiency, the reactant flow rate, and the air flow intensity; through the real-time operation data, the weight of the edge is updated at each time step.

5. The thermodynamic behavior analysis method for high-temperature synthesis in a kiln as described in claim 4, characterized in that: Predicting the thermodynamic behavior under different operating conditions includes inputting the real-time operation data as node features into a graph neural network, constructing a thermodynamic behavior map model, training the thermodynamic behavior map model using the historical operation data to capture the dynamic changes of heat transfer, reactant flow, and air flow transmission; Using the mean square error as the loss function to evaluate the prediction error of the thermodynamic behavior map model, and using the Adam optimization algorithm to adjust the graph neural network to optimize the thermodynamic behavior map model; Based on the trained thermodynamic behavior map model, predicting the thermodynamic behavior under different operating conditions; The thermodynamic behavior includes: the temperature distribution in each region inside the kiln, the change of the reaction rate inside the kiln, and the distribution change of the atmosphere composition in each region of the kiln.

6. The thermodynamic behavior analysis method for high-temperature synthesis in a kiln as described in claim 5, characterized in that: The control parameters of the non-linear dynamic model include heat transfer coefficient, reaction rate constant, diffusion coefficient, and reactant concentration.

7. The thermodynamic behavior analysis method for high-temperature synthesis in a kiln as described in claim 6, characterized in that: Calculate the residuals between the thermodynamic behavior and the actual operating data of the kiln; define the objective function f(θ n ) of the joint iteration to minimize the residuals; By calculating the gradient and Hessian matrix of the objective function f(θ n ), and using the improved Newton's method to update θ n of the objective function f(θ n ), the formula is expressed as: Among them, represents the gradient of the objective function, and H(θ n ) is the Hessian matrix, representing the second derivative of the objective function; θ n+1 represents the control parameter vector of the updated nonlinear dynamic model; n represents the current iteration step, and n + 1 represents the next iteration step; θ n represents the control parameter vector of the current non-linear dynamics model, and is expressed by the formula as follows: θ n = [k n , k r (T) n , D n , C n , w1, w2, …, w m ​ Among them, k n represents the heat transfer coefficient at the current iteration step, k r (T) n represents the reaction rate constant at the current iteration step, D n represents the diffusion coefficient at the current iteration step, C n represents the concentration of reactants at the current iteration step, w1, w2, …, w m represents the weight of the edge in the thermodynamic behavior atlas model, and m represents the number of edges; Setting the original learning step size as α0, when the residual increases, the learning step size increases to x times the original step size; when the residual decreases, the learning step size decreases to y times the original step size; where x and y are constants; The joint iteration includes, through the real-time operation data, the thermodynamic behavior map model predicts the thermodynamic behavior, outputs the relationship between the nodes, and adjusts the control parameters of the non-linear thermo-dynamic model; In the joint iteration, when the residual is less than the set threshold, it is judged that the optimization has converged, and the iteration process ends; and after each round of optimization, it is checked whether the updated control parameters of the non-linear dynamic model violate the physical constraints, and if they violate the physical constraints, the control parameters of the non-linear dynamic model are adjusted through a penalty function.

8. A thermodynamic behavior analysis system for high-temperature synthesis in a kiln using the method according to any one of claims 1-7, characterized in that: Non-linear thermo-dynamic model module, collecting the operation data of the kiln through kiln sensors, establishing a non-linear thermo-dynamic model to describe the dynamic changes of the thermodynamic behavior inside the kiln; A thermodynamic behavior map module that uses the kiln operation data to construct a thermodynamic behavior map and predict the thermodynamic behavior under different operating conditions; An adjustment module that adjusts the control parameters of the non-linear dynamics model based on the thermodynamic behavior results; A joint iteration module that performs joint iteration on the non-linear thermodynamic model and the thermodynamic behavior map to optimize the kiln operation.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the thermodynamic behavior analysis method for high-temperature synthesis of a kiln according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of the thermodynamic behavior analysis method for high-temperature synthesis of a kiln according to any one of claims 1 to 7.

Images