Intelligent control method of natural gas continuous heat treatment equipment based on digital twinning

Abstract

The invention provides an intelligent control method for natural gas continuous heat treatment equipment based on digital twinning, and relates to the technical field of heat treatment equipment control. The method comprises the steps that a digital twinning model of the natural gas continuous heat treatment equipment is established, and production monitoring data collected on line are utilized; the method comprises the following steps: predicting combustion process characteristics and unmeasurable states through a digital twin model, performing multi-layer strategy search by adopting a hierarchical control strategy according to a state prediction result, performing safety constraint simulation verification on the multi-layer strategy in the digital twin model, and predicting unmeasurable state parameters through a soft sensor. And introducing the quality index as a soft constraint condition into optimization control, and generating a control execution parameter meeting the constraint condition. The technical problems that in the prior art, natural gas continuous thermal treatment equipment is poor in control effect and unstable in workpiece quality are solved. The technical effects that the natural gas continuous heat treatment equipment is intelligently, accurately and reliably controlled, and the product quality is improved are achieved.

Description

Technical Field

[0001] This invention relates to the field of heat treatment equipment control technology, and more specifically to an intelligent control method for a continuous natural gas heat treatment equipment based on digital twins. Background Technology

[0002] As a key industrial processing equipment, the control precision of natural gas continuous heat treatment equipment directly affects the heat treatment quality and production efficiency of the workpiece. Traditional control methods mainly rely on empirically set heating curves and simple PID control strategies. While these methods can meet basic heat treatment requirements to a certain extent, they are difficult to adapt to complex and variable process conditions and workpiece characteristics. These methods have limitations in accurately controlling the furnace temperature distribution, combustion efficiency, and workpiece heat exchange process. Especially when processing workpieces of different materials and sizes, they are prone to problems such as uneven temperature control and inconsistent heat treatment effects, affecting the final performance of the workpiece.

[0003] While some existing automated control technologies have improved the accuracy of control, they mostly adopt control strategies based on a single model or fixed algorithm, which cannot fully cover the ever-changing heat treatment process. This further affects the accuracy and reliability of heat treatment control, resulting in poor control performance of heat treatment equipment and unstable workpiece quality.

[0004] Existing natural gas continuous heat treatment equipment suffers from insufficient accuracy and reliability in heat treatment control, resulting in poor control performance and unstable workpiece quality. Summary of the Invention

[0005] The purpose of this application is to provide an intelligent control method for a continuous natural gas heat treatment equipment based on digital twins, in order to solve the technical problems of insufficient accuracy and reliability in the heat treatment control of existing continuous natural gas heat treatment equipment, resulting in poor control effect and unstable workpiece quality.

[0006] In view of the above problems, this application provides an intelligent control method for a natural gas continuous heat treatment equipment based on digital twins, the method comprising: A digital twin model of a continuous natural gas heat treatment equipment is established. This model includes a workpiece thermal process ash box model, a combustion process model, a heat exchange model, and soft sensors. Using online production monitoring data, combustion process characteristics and unmeasurable states are predicted through the digital twin model, yielding state prediction results. Based on the state prediction results, a hierarchical control strategy is employed to search for multi-level strategies, including inner-loop control, middle-loop control, and outer-loop control. Safety constraint simulation verification of the multi-level strategies is performed within the digital twin model. Unmeasurable state parameters are predicted using soft sensors, and quality indicators are introduced as soft constraints into the optimization control to generate control execution parameters that satisfy the constraints.

[0007] Optionally, based on the physical structure and process zoning of the natural gas continuous heat treatment equipment, the furnace body is divided into a preheating zone, a heating zone, a soaking zone, and a heat preservation zone, and a zone energy balance model is constructed. A mapping relationship between monitored values ​​and operating values ​​is established based on the sensor layout and actuator topology, and this mapping relationship is embedded into the zone energy balance model. Based on a mechanism-data fusion strategy, historical production data and online monitoring data are used to identify and calibrate the parameters of the zone energy balance model, update the digital twin, and obtain the digital twin model. Based on historical production data and monitoring data, predictive analyses are performed on the workpiece heat treatment process, the natural gas combustion process, the furnace temperature-workpiece heat transfer relationship, and the workpiece's unmeasurable state parameters. A workpiece thermal process ash box model, a combustion process model, a heat transfer model, and soft sensors are established, and these models are embedded as built-in modules into the digital twin model.

[0008] Optionally, based on the thermal conductivity, specific heat capacity, and density thermophysical parameters of the workpiece material, and combined with the physical structure of the workpiece, the workpiece is divided into multiple heat conduction units; using the historical production data and monitoring data, the heat energy changes of the multiple heat conduction units during the heat treatment process are fitted to obtain the workpiece change relationship during the heat treatment process in the furnace, and the workpiece thermal process gray box model is constructed.

[0009] Optionally, based on natural gas flow rate, air flow rate, and air-fuel ratio, and combined with historical production data and online monitoring data, a time-series heat release rate and thermal efficiency relationship for the natural gas combustion process is established; by using monitoring data of flue gas composition and oxygen content, the time-series heat release rate and thermal efficiency of the natural gas combustion are corrected to obtain the combustion stability evaluation relationship of the natural gas combustion process, and the combustion process model is constructed.

[0010] Optionally, heat transfer boundary conditions between furnace gas and workpiece are established based on furnace temperature distribution, workpiece loading amount, workpiece surface temperature and atmosphere flow parameters; a comprehensive heat transfer relationship under the heat transfer boundary conditions is constructed based on the coupling mechanism of convection and radiation heat transfer; and the comprehensive heat transfer relationship and heat transfer boundary conditions are modified based on historical production data and monitoring data to obtain a heat transfer model that reflects the sequential heating process of the workpiece.

[0011] Optionally, based on historical production and monitoring data of the natural gas continuous heat treatment equipment, an initial data-driven prediction model is constructed, with measurable variables such as furnace temperature, natural gas flow rate, air flow rate, flue gas oxygen content, and workpiece surface temperature as inputs, and unmeasurable parameters such as workpiece core temperature, lower heating value of natural gas, and heat transfer coefficient as outputs. The workpiece thermal process ash box model, combustion process model, and heat transfer model are used as physical constraints and feature generation modules to provide the data-driven prediction model with mechanistic parameters of workpiece thermal conductivity, combustion heat release efficiency, and furnace temperature-workpiece heat transfer coefficient. The data-driven prediction model is trained using the historical production data and compared and corrected with the output of the mechanistic parameters to obtain a fused and corrected data-driven prediction model, which serves as the soft sensor for inference and prediction based on the input measurable variables, and outputs the prediction results of the unmeasurable state parameters.

[0012] Optionally, the unmeasurable state parameters include the core temperature of the workpiece, the lower heating value of natural gas, and the heat transfer coefficient; the combustion process characteristics include the furnace temperature distribution, the combustion heat release rate, and the furnace temperature-workpiece heat transfer characteristics, wherein the combustion heat release rate is obtained through a combustion process model, the furnace temperature-workpiece heat transfer characteristics are obtained through a heat transfer model, and the furnace temperature distribution is obtained through both the heat transfer model and the combustion process model.

[0013] Optionally, a multi-layered control strategy framework is established, including an inner-loop control strategy analysis channel, a middle-loop control strategy analysis channel, and an outer-loop control strategy analysis channel, each with input mapping relationships and output labels. The inner-loop control strategy analysis channel uses the combustion heat release rate and furnace temperature-workpiece heat transfer characteristics output by the digital twin model as input to perform analysis and output on adjusting the gas flow rate, air flow rate, and air-fuel ratio. The middle-loop control strategy analysis channel uses the output results of the inner-loop control strategy analysis channel as input, combined with the workpiece core temperature and heat transfer coefficient, to perform analysis and output on adjusting the zoned furnace temperature setpoints and atmosphere parameters. The outer-loop control strategy analysis channel uses the output results of the middle-loop control strategy analysis channel as constraints based on the workpiece's target heat treatment parameters to perform multi-step optimization analysis and output on the zoned target temperature trajectory and gas supply strategy.

[0014] Optionally, the multi-layer strategy is used as an input variable, and the strategy is run through a digital twin model. Simulation is performed using the workpiece thermal process ash box model, combustion process model, heat transfer model, and soft sensors. The results of the combustion process, heat transfer process, workpiece thermal process, and soft sensor prediction are output. The results of the combustion process, heat transfer process, workpiece thermal process, and soft sensor prediction are judged using preset safety constraint indicators. When the safety constraint requirements are met, the multi-layer strategy is determined to have passed the safety constraint simulation verification. The safety constraint requirements are based on the detection of combustion stability, zone furnace temperature deviation, and atmosphere control boundary safety threshold.

[0015] Optionally, the soft sensor is used to predict unmeasurable state parameters, and the workpiece's hardness, microstructure, and residual stress quality indicators are introduced as soft constraints into the optimization control process. By establishing a multi-objective optimization function, under the premise of meeting safety and quality constraints, the zoned target temperature trajectory and gas supply strategy are optimized and solved to generate control execution parameters that meet the constraints, which are then sent to the natural gas continuous heat treatment equipment as the final control command.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages: The method provided in this application establishes a digital twin model of a continuous natural gas heat treatment equipment. This digital twin model includes a workpiece thermal process ash box model, a combustion process model, a heat exchange model, and soft sensors. Using online collected production monitoring data, the method predicts combustion process characteristics and unmeasurable states through the digital twin model, obtaining state prediction results. Based on the state prediction results, a hierarchical control strategy is employed to search for multi-level strategies, including inner-loop control, middle-loop control, and outer-loop control. Safety constraint simulation verification of the multi-level strategies is performed within the digital twin model, and unmeasurable state parameters are predicted using soft sensors. Quality indicators are introduced as soft constraints into the optimization control, generating control execution parameters that satisfy the constraints. This achieves intelligent, accurate, and reliable control of the continuous natural gas heat treatment equipment, improving product quality.

[0017] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the intelligent control method for a natural gas continuous heat treatment equipment based on digital twins provided in this application.

[0020] Figure 2 A schematic diagram illustrating the process of establishing a digital twin model in the intelligent control method for a natural gas continuous heat treatment equipment based on digital twins provided in this application. Detailed Implementation

[0021] This application provides an intelligent control method for a continuous natural gas heat treatment equipment based on digital twins. This method addresses the technical problems of insufficient accuracy and reliability in heat treatment control in existing continuous natural gas heat treatment equipment, leading to poor control performance and unstable workpiece quality. It achieves intelligent, accurate, and reliable control of the continuous natural gas heat treatment equipment, thereby improving product quality.

[0022] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0023] like Figure 1 As shown, this application provides an intelligent control method for a continuous natural gas heat treatment equipment based on digital twins. The intelligent control method for the continuous natural gas heat treatment equipment based on digital twins includes: A digital twin model of a natural gas continuous heat treatment equipment is established. The digital twin model includes a workpiece thermal process ash box model, a combustion process model, a heat exchange model, and soft sensors.

[0024] Specifically, continuous natural gas heat treatment equipment is a large-scale, specialized device used in the industrial field for heat treatment of metallic materials, such as steel or parts. This equipment relies on the heat released from the combustion of natural gas to heat the steel or parts. By precisely controlling the temperature curves during heating and cooling, it alters the internal microstructure of the steel or parts, thereby optimizing their hardness, toughness, and other properties to meet the diverse performance requirements of different industrial applications. However, in actual production scenarios, the high temperatures of continuous natural gas heat treatment equipment make it difficult or costly to measure many critical parameters, such as the core temperature of the workpiece and the lower heating value of natural gas, in real time. Furthermore, the heat content of different batches of natural gas may vary, causing fluctuations in the equipment's performance. This can lead to overheating or brittleness of the steel or parts, making product quality control difficult. Based on this, this application combines the physical structure and process zoning, sensor layout, and actuator topology of the natural gas continuous heat treatment equipment with a mechanism-data fusion strategy. It utilizes CAD software to create a precise 3D model of the equipment's physical structure, presenting the shape, size, and spatial relationship of each component. CAE software, such as ANSYS and ABAQUS, is employed to perform thermodynamic and fluid dynamic simulation analyses based on the operating mechanism of the natural gas continuous heat treatment equipment, simulating the physical processes within the process zones. Combined with professional industrial IoT platform software, such as PTCThingWorx, real-time data is integrated to achieve data-driven dynamic updates. Simultaneously, MATLAB / Simulink is used to construct a digital twin model of the natural gas continuous heat treatment equipment. This digital twin model is a precise mapping of the physical entity of the natural gas continuous heat treatment equipment in virtual space, capable of reflecting the real-time operating status of the equipment, predicting future performance changes, and providing data support and decision-making basis for the formulation of control strategies. The digital twin model includes a workpiece thermal process ash box model, a combustion process model, a heat exchange model, and soft sensors. The workpiece thermal process ash box model is used to simulate the temperature change and heat conduction process of the workpiece during heat treatment. The combustion process model is used to simulate the heat release during natural gas combustion. The heat exchange model is used to simulate the heat transfer process in the furnace. The soft sensors are used to predict state parameters that are difficult to measure directly, such as the core temperature of the workpiece and the lower heating value of natural gas.

[0025] By establishing a digital twin model that includes a workpiece thermal process ash box model, a combustion process model, a heat exchange model, and a soft sensor model, and ensuring high consistency with the physical entity, the comprehensiveness and accuracy of the operation status analysis and prediction of the natural gas continuous heat treatment equipment are improved. This enables intelligent and efficient control of the natural gas continuous heat treatment equipment, thereby improving product quality and production efficiency.

[0026] Furthermore, such as Figure 2As shown, a digital twin model of a continuous natural gas heat treatment equipment is established, including: dividing the furnace body into a preheating zone, a heating zone, a soaking zone, and a heat preservation zone according to the physical structure and process zoning of the equipment, and constructing a zoned energy balance model; establishing a mapping relationship between monitored values ​​and operating values ​​based on sensor layout and actuator topology, and embedding it into the zoned energy balance model; using a mechanism-data fusion strategy, using historical production data and online monitoring data to identify and calibrate the parameters of the zoned energy balance model, updating the digital twin, and obtaining the digital twin model; based on historical production data and monitoring data, performing predictive analysis of the workpiece heat treatment process, natural gas combustion process, furnace temperature-workpiece heat transfer relationship, and unmeasurable state parameters of the workpiece, establishing a workpiece thermal process ash box model, combustion process model, heat transfer model, and soft sensors, and embedding the workpiece thermal process ash box model, combustion process model, heat transfer model, and soft sensors as built-in modules into the digital twin model.

[0027] Specifically, when constructing a digital twin model of a continuous natural gas heat treatment equipment, the furnace body is first divided according to its physical structure and process zones. This division includes a preheating zone, a heating zone, a soaking zone, and a holding zone, each with specific temperature control targets and heat treatment tasks. Specifically, the preheating zone is used to initially raise the workpiece temperature, reducing thermal stress upon entering the heating zone; the heating zone heats the workpiece to the target temperature; the soaking zone eliminates temperature gradients within the workpiece, ensuring temperature uniformity; and the holding zone maintains the workpiece at a specific temperature to complete the corresponding microstructural transformation. Within each zone, based on the law of conservation of energy, a balance relationship is established between energy inputs (such as heat generated by natural gas combustion), outputs (such as heat loss and heat absorbed by the workpiece), and the transfer and transformation of energy within the zone. This constitutes a zone energy balance model, which describes the energy changes in each zone under different operating conditions. For example, in the heating zone, energy input mainly comes from the heat generated by natural gas combustion, while energy output includes heat loss to the external environment through the furnace wall, heat absorbed by the workpiece, and heat carried away by flue gas. Energy transfer and conversion within the zone primarily involves heat conduction, convection, and radiation from the combustion zone to the workpiece and furnace wall. The combustion power of natural gas in the heating zone is... The relevant data were obtained by measuring the natural gas flow rate and calorific value using a natural gas flow meter and a calorific value analyzer. The natural gas flow rate V = 10. / h, the lower calorific value of natural gas H=35.5MJ / The heat (energy input) generated by the combustion of natural gas is: =V×H=10×35.5=355MJ / h. The heat flux density q measured on the furnace wall using a heat flow meter is 2kW / h. The furnace wall area of ​​the heating zone is A=5. Then heat loss =q×A×3600÷1000=2×5×3600÷1000=36MJ / h. Multiplying by 3600 converts hours to seconds, and dividing by 1000 converts the unit from W to kW and then to MJ / h. The workpiece mass m=500kg, specific heat capacity c=0.46, and the temperature rise ΔT in the heating zone is adjusted to 400℃. The heat absorbed by the workpiece... =m×c×ΔT=92000kJ=92MJ / h, the heat carried away by the flue gas. =227.25 MJ / h. According to the law of conservation of energy, energy input equals energy output, that is... = + + +ΔE, where ΔE is the change in energy storage within the heating zone. Under steady-state conditions, ΔE = 0. Substituting the above data, we get: 355 ≈ 36 + 92 + 227.25 + 0. This is not completely equal because there are certain errors in actual operation. Other areas, such as the preheating zone, the heat spreader zone, and the insulation zone, are modeled and calculated using a similar method.

[0028] Based on expert experience, historical data, and the layout of the continuous natural gas heat treatment equipment, sensors for measuring parameters such as temperature, pressure, and flow rate are installed at key locations within the equipment. This sensor layout allows for real-time acquisition of operational data. A mapping relationship between monitored and operational values ​​is established based on the sensor layout and actuator topology. Actuator topology refers to the connection methods and control logic of actuators such as valves and fans, which regulate operational parameters such as natural gas flow rate and air supply. The actual monitoring data collected by the sensors is correlated with the actuator actions; for example, the opening and closing of natural gas valves are adjusted based on furnace temperature data monitored by temperature sensors to control heating power. This mapping relationship is embedded into a constructed zoned energy balance model, enabling dynamic calculation and analysis based on real-time monitoring data and operational parameters. The mechanism-data fusion strategy combines physical mechanism models with actual production data, using a data-driven approach to identify and calibrate model parameters. This approach considers both the physical laws of equipment operation, such as heat conduction and combustion chemistry, and fully utilizes the large amount of data accumulated in actual production. Historical production data includes operating records of the natural gas continuous heat treatment equipment under different operating conditions, while online monitoring data reflects the current operating status of the equipment in real time. Based on a mechanism-data fusion strategy, historical production data and online monitoring data are used to identify, calibrate, and optimize the parameters of the zonal energy balance model. This enables the zonal energy balance model to more accurately reflect the actual operating conditions of the equipment, thereby updating the digital twin and obtaining a more accurate digital twin model, ensuring that the digital twin model accurately reflects the actual operating conditions of the equipment.

[0029] Finally, based on historical production and monitoring data from the natural gas continuous heat treatment equipment, the heat treatment process of the workpiece was analyzed. Data on the workpiece's entry and exit states and heating time under different operating conditions were collected. Combining theories such as heat conduction, key factors affecting the workpiece's thermal process were identified, and a gray box model of the workpiece's thermal process was established. For the natural gas combustion process, monitoring data on natural gas flow rate, air flow rate, and air-fuel ratio were used to construct a combustion process model based on combustion principles to reflect the heat generation law of natural gas combustion. Regarding the furnace temperature-workpiece heat transfer relationship, historical data on furnace temperature and workpiece surface and internal temperatures at different times were used. Based on the coupling mechanism of convection and radiation heat transfer, a heat transfer model reflecting the sequential heating process of the workpiece was constructed. For predictive analysis of unmeasurable state parameters of the workpiece, a soft sensor is constructed by integrating the previous three models and measurable data such as furnace temperature and natural gas flow rate. This enables accurate prediction of unmeasurable state parameters of the workpiece, such as internal stress and degree of microstructure transformation. The resulting workpiece thermal process ash box model, combustion process model, heat transfer model, and soft sensor are then embedded as built-in modules into the digital twin model. This allows the digital twin model to more accurately reflect the actual operation of the natural gas continuous heat treatment equipment, thereby achieving efficient, stable, and intelligent control of the heat treatment process and improving product quality and production efficiency.

[0030] Furthermore, a workpiece thermal process gray box model is established, including: dividing the workpiece into multiple heat conduction units based on the workpiece material's thermal conductivity, specific heat capacity, and density thermophysical parameters, combined with the workpiece's physical structure; using the historical production data and monitoring data, fitting the thermal energy changes of the multiple heat conduction units during the heat treatment process to obtain the workpiece change relationship during the furnace heat treatment process, and constructing the workpiece thermal process gray box model.

[0031] Specifically, the process involves obtaining workpieces that require heat treatment using a continuous natural gas heat treatment system, analyzing their materials, and dividing them into multiple heat conduction units based on their thermophysical parameters (e.g., thermal conductivity, specific heat capacity, density) and physical structure. Each heat conduction unit is an independent heat conductor, allowing for more detailed simulation of temperature changes during heat treatment. Historical production and monitoring data are used. Historical production data records temperature changes under different temperatures and times during past production processes, while monitoring data provides real-time temperature and flow rate information. Existing mature curve fitting algorithms, such as the least squares method, are employed to accurately fit the thermal energy changes of each heat conduction unit during heat treatment. This reveals the inherent laws governing the thermal energy changes of each heat conduction unit with time and temperature, thereby determining the relationships between temperature, thermal stress, and other state parameters of different parts of the workpiece during furnace heat treatment. A gray box model of the workpiece's thermal process is then constructed. This gray box model can predict the temperature distribution of the workpiece at different heat treatment stages, providing data support for the formulation of control strategies for the continuous natural gas heat treatment system. For example, during the heating stage, the workpiece thermal process ash box model can predict the temperature differences between the workpiece surface and interior, controlling the heating rate of the natural gas continuous heat treatment equipment to avoid workpiece deformation or cracking caused by thermal stress. During the homogenization stage, the workpiece thermal process ash box model can ensure uniform temperature distribution of the workpiece, improving the consistency of product quality.

[0032] By establishing a gray box model of the workpiece's thermal process, the digital twin model can more accurately reflect the actual state of the workpiece during the heat treatment process, improve the intelligent control of the heat treatment process, and thus improve production efficiency and product quality.

[0033] Furthermore, a combustion process model is established, including: based on natural gas flow rate, air flow rate, and air-fuel ratio, combined with historical production data and online monitoring data, establishing the relationship between the time-series heat release rate and thermal efficiency of the natural gas combustion process; using monitoring data of flue gas composition and oxygen content, correcting the time-series heat release rate and thermal efficiency of the natural gas combustion, obtaining the combustion stability evaluation relationship of the natural gas combustion process, and constructing the combustion process model.

[0034] Specifically, high-precision flow meters are installed on the natural gas transmission pipeline and air supply pipeline to measure the natural gas flow rate and air flow rate in real time. The air-fuel ratio is then calculated based on these measurements. Natural gas flow rate refers to the volume or mass of natural gas passing through a specific cross-section per unit time; air flow rate is the amount of air entering the combustion zone per unit time; and the air-fuel ratio is the ratio of air to natural gas. Based on the obtained natural gas flow rate, air flow rate, and air-fuel ratio, combined with historical production data and online monitoring data from the continuous natural gas heat treatment equipment, data algorithms, such as time series analysis, are used to analyze the changes in natural gas and air flow rates recorded in the historical data, as well as the corresponding furnace temperature changes. This establishes the relationship between the time-series heat release rate and thermal efficiency of the natural gas combustion process. The time-series heat release rate refers to the heat released during the natural gas combustion process per unit time, while the thermal efficiency refers to the proportion of natural gas that is actually converted into effective heat during combustion. Because the actual combustion process is affected by various factors, such as aging combustion equipment and fluctuations in fuel composition, monitoring data on flue gas composition and oxygen content is used to correct the time-series heat release rate and thermal efficiency of natural gas combustion. Specifically, the flue gas composition includes various gases produced after combustion, such as carbon dioxide, carbon monoxide, and nitrogen oxides, reflecting the degree of combustion completeness. Oxygen content reflects whether the amount of air entering the combustion zone is appropriate. By analyzing the flue gas composition and oxygen content, and combining thermodynamic and chemical equilibrium principles, the relationship between the time-series heat release rate and thermal efficiency is adjusted and optimized, thereby obtaining a combustion stability evaluation relationship for the natural gas combustion process. This combustion stability evaluation relationship comprehensively considers the influence of heat release rate, thermal efficiency, and various factors on stability during combustion, accurately judging the stability and reliability of the combustion process under different operating conditions and avoiding temperature fluctuations caused by combustion fluctuations. Finally, through analysis and correction, a combustion process model is constructed. This model can accurately simulate the combustion characteristics of natural gas under different operating conditions, predict the combustion effect under different flow rates and air-fuel ratios, and realize intelligent control to adjust the supply of natural gas and air in real time, ensuring the stability and uniformity of the furnace temperature, thereby improving the heat treatment quality and production efficiency.

[0035] Furthermore, a heat transfer model is established, including: establishing heat transfer boundary conditions between furnace gas and workpiece based on furnace temperature distribution, workpiece loading amount, workpiece surface temperature, and atmosphere flow parameters; constructing a comprehensive heat transfer relationship under the aforementioned heat transfer boundary conditions based on the coupling mechanism of convection and radiation heat transfer; and correcting the parameters of the comprehensive heat transfer relationship and heat transfer boundary conditions based on historical production data and monitoring data to obtain a heat transfer model that reflects the sequential heating process of the workpiece.

[0036] Specifically, temperature distribution data of the furnace is obtained by arranging temperature sensors such as thermocouples at key locations inside the furnace, the amount of workpieces loaded into the furnace is determined by weighing equipment, the surface temperature of the workpieces is measured by infrared thermometers, and the atmosphere flow parameters are monitored by instruments such as anemometers. The furnace temperature distribution, the amount of workpieces loaded into the furnace, the surface temperature of the workpieces, and the atmosphere flow parameters are obtained. The furnace temperature distribution reflects the temperature difference at different locations inside the furnace, the amount of workpieces loaded into the furnace reflects the density of the workpieces in the furnace and affects the furnace gas flow and heat exchange effect, the surface temperature of the workpieces reflects the heat transfer effect, and the atmosphere flow parameters, including flow velocity and flow direction, affect the efficiency of heat transfer between the furnace gas and the workpiece surface. Based on the principles of heat transfer, the direction of heat transfer is determined by the furnace temperature distribution. The influence of furnace gas flow obstruction and circumferential flow on heat transfer is analyzed based on the workpiece loading amount. The temperature difference between the workpiece surface temperature and the furnace gas is used to determine the heat transfer driven by the temperature difference. The strength of convective heat transfer is determined by the atmospheric flow parameters. The path and mode of heat transfer from the furnace gas to the workpiece are determined. The heat transfer boundary conditions between the furnace gas and the workpiece are established. The heat transfer boundary conditions include the convective heat transfer coefficient and the radiative heat transfer coefficient. The convective heat transfer coefficient describes the efficiency of heat transfer from the furnace gas to the workpiece surface through convection, and the radiative heat transfer coefficient describes the efficiency of heat transfer from the furnace gas to the workpiece surface through radiation.

[0037] Then, based on the coupling mechanism of convection and radiation heat transfer, a comprehensive heat transfer relationship under heat transfer boundary conditions is constructed. Convection heat transfer mainly occurs between the workpiece surface and the surrounding flowing gas, and can be analyzed using Newton's law of cooling. The heat transfer is proportional to the temperature difference between the workpiece surface and the surrounding gas, with the proportionality coefficient being the convective heat transfer coefficient. Radiation heat transfer occurs between the workpiece and the furnace wall or surrounding high-temperature gas in the form of electromagnetic waves, following the Stefan-Boltzmann law. In actual heat transfer processes, convection and radiation do not occur independently. Therefore, by applying Newton's law of cooling and the Stefan-Boltzmann law in heat transfer, combined with the determined heat transfer boundary conditions, the two heat transfer methods are organically combined to construct a comprehensive heat transfer relationship under the heat transfer boundary conditions. This comprehensive heat transfer relationship fully describes the heat transfer process between the furnace gas and the workpiece.

[0038] Finally, based on historical production and monitoring data, the parameters of the comprehensive heat transfer relationship and heat transfer boundary conditions are corrected by comparing model predictions with actual data. The corrected model yields a heat transfer model that accurately reflects the sequential heating process of the workpiece. This heat transfer model precisely reflects the heating status of the workpiece during heat treatment, providing a basis for precise control of the heat treatment process, ensuring uniform heating of the workpiece, and improving product performance, quality, and production efficiency. Simultaneously, by optimizing the heat transfer process, energy consumption can be reduced, achieving the goals of energy conservation and emission reduction.

[0039] Furthermore, a soft sensor is established, including: based on historical production and monitoring data of the natural gas continuous heat treatment equipment, an initial data-driven prediction model is constructed, with measurable variables such as furnace temperature, natural gas flow rate, air flow rate, flue gas oxygen content, and workpiece surface temperature as inputs, and unmeasurable parameters such as workpiece core temperature, lower heating value of natural gas, and heat transfer coefficient as outputs; the workpiece thermal process ash box model, combustion process model, and heat transfer model are used as physical constraints and feature generation modules to provide the data-driven prediction model with the workpiece thermal conductivity characteristics, combustion heat release efficiency, and furnace temperature-workpiece heat transfer coefficient mechanism parameters; the data-driven prediction model is trained using the historical production data, and compared and corrected by combining the mechanism parameter outputs to obtain a fused and corrected data-driven prediction model, which serves as the soft sensor for inference and prediction based on the input measurable variables and outputs the prediction results of the unmeasurable state parameters.

[0040] Specifically, based on historical production data and real-time monitoring data from the continuous natural gas heat treatment equipment, a data-driven prediction model is constructed using machine learning algorithms, such as neural network algorithms. This initial data-driven prediction model takes measurable variables such as furnace temperature, natural gas flow rate, air flow rate, flue gas oxygen content, and workpiece surface temperature as inputs, and takes unmeasurable states such as workpiece core temperature, lower heating value of natural gas, and heat transfer coefficient as outputs. The specific construction process involves dividing the collected historical production data and monitoring data into training and validation sets in an 8:2 ratio. The measurable variables in the training set, such as furnace temperature, natural gas flow rate, air flow rate, flue gas oxygen content, and workpiece surface temperature, are used as input data to the neural network, while the corresponding unmeasurable states, such as workpiece core temperature, lower heating value of natural gas, and heat transfer coefficient, are used as target output data. The weights and bias parameters of the neural network are initialized, a suitable activation function, such as the ReLU function, is selected, nonlinear factors are introduced, and a forward propagation algorithm is used to pass the input data sequentially through the input layer, hidden layer, and finally to the output layer to obtain the model's predicted output. Then, the error between the predicted output and the actual target output is calculated using a loss function, such as the mean squared error loss function. Using the backpropagation algorithm, the gradient is calculated layer by layer from the output layer to the input layer based on the error. Optimization algorithms, such as stochastic gradient descent, are then used to update the weights and bias parameters of the neural network along the gradient descent direction to minimize the loss function. The performance of the neural network model is evaluated using a validation set. If the model's performance no longer improves or overfitting occurs, an initial data-driven prediction model capable of accurately predicting unmeasurable states based on measurable variables is obtained.

[0041] The established workpiece thermal process ash box model, combustion process model, and heat transfer model are used as physical constraints and feature generation modules. These modules provide the data-driven prediction model with mechanistic parameters such as workpiece thermal conductivity, combustion heat release efficiency, and furnace temperature-workpiece heat transfer coefficient. Historical production data is used to train the initial data-driven prediction model. During training, the predicted values ​​of unmeasurable state parameters output by the initial data-driven prediction model based on the input measurable variables are compared with the calculated results of the mechanistic parameters provided by the physical constraint module. If a discrepancy exists, the parameters of the initial data-driven prediction model are adjusted, such as the weights and biases in the neural network, to gradually bring the prediction results closer to the calculated values ​​of the mechanistic parameters, thus correcting the initial data-driven prediction model. After multiple rounds of iterative training and calibration, a data-driven prediction model that integrates physical mechanisms and data pattern corrections was finally obtained. The corrected data-driven prediction model is used as a soft sensor to perform inference and prediction based on real-time input measurable variables. It accurately outputs the prediction results of unmeasurable state parameters such as workpiece core temperature, lower heating value of natural gas, and heat transfer coefficient. This provides comprehensive state information for the intelligent control of the heat treatment process, thereby achieving precise control of the heat treatment process and improving the heat treatment quality of the workpiece.

[0042] By utilizing online production monitoring data, the characteristics of the combustion process and unpredictable states are predicted through the digital twin model, resulting in state prediction results.

[0043] Furthermore, the unmeasurable state parameters include the core temperature of the workpiece, the lower heating value of natural gas, and the heat transfer coefficient. The combustion process characteristics include the furnace temperature distribution, the combustion heat release rate, and the furnace temperature-workpiece heat transfer characteristics. The combustion heat release rate is obtained through a combustion process model, the furnace temperature-workpiece heat transfer characteristics are obtained through a heat transfer model, and the furnace temperature distribution is obtained through both the heat transfer model and the combustion process model.

[0044] Specifically, high-precision sensors installed at key locations within the natural gas continuous heat treatment equipment collect real-time production monitoring data reflecting the current production process, including furnace temperature, natural gas flow rate, air flow rate, flue gas oxygen content, and workpiece surface temperature. This real-time collected production monitoring data is input into a pre-built digital twin model. Upon receiving the online production monitoring data, the digital twin model performs collaborative calculations using its integrated internal models. The combustion process model calculates the combustion heat release rate, the heat transfer model determines the furnace temperature-workpiece heat transfer characteristics, and the combustion process model, in conjunction with the furnace process model, obtains the furnace temperature distribution. Simultaneously, by combining the workpiece thermal process ash box model and the input measurable data, unmeasurable parameters are predicted, including the workpiece core temperature, the lower heating value of natural gas, and the heat transfer coefficient, resulting in predictions of combustion process characteristics and unmeasurable parameters. The combustion process characteristics include furnace temperature distribution, combustion heat release rate, and furnace temperature-workpiece heat transfer characteristics. The combustion heat release rate is obtained through the combustion process model, the furnace temperature-workpiece heat transfer characteristics are obtained through the heat transfer model, and the furnace temperature distribution is obtained through both the heat transfer model and the combustion process model.

[0045] By collecting data online and using digital twin models for prediction, it is possible to predict key parameters and characteristics that are difficult to measure directly in the production process in real time and accurately. This provides comprehensive status information for real-time monitoring and optimization of the production process, thereby enabling precise control of the heat treatment process and improving product quality and production efficiency.

[0046] Based on the state prediction results, a hierarchical control strategy is adopted to search for multi-level strategies, including inner-loop control, middle-loop control, and outer-loop control.

[0047] Furthermore, based on the state prediction results, a hierarchical control strategy is adopted to perform multi-level strategy search, including inner-loop control, middle-loop control, and outer-loop control. This includes: establishing a multi-level control strategy framework, which includes an inner-loop control strategy analysis channel, a middle-loop control strategy analysis channel, and an outer-loop control strategy analysis channel, each with input mapping relationships and output labels; wherein, the inner-loop control strategy analysis channel uses the combustion heat release rate and furnace temperature-workpiece heat transfer characteristics output by the digital twin model as input to perform analysis and output of adjusting gas flow rate, air flow rate, and air-fuel ratio; the middle-loop control strategy analysis channel uses the output results of the inner-loop control strategy analysis channel as input, combined with the workpiece core temperature and heat transfer coefficient, to perform analysis and output of adjusting the zone furnace temperature setpoint and atmosphere parameters; the outer-loop control strategy analysis channel uses the output results of the middle-loop control strategy analysis channel as constraints based on the workpiece target heat treatment parameters to perform multi-step optimization analysis and output of zone target temperature trajectory and gas supply strategy.

[0048] Specifically, after obtaining the state prediction results, a hierarchical control strategy is adopted to perform a multi-level strategy search within a multi-level control strategy framework. The multi-level strategy search includes three levels: inner-loop control, middle-loop control, and outer-loop control. This multi-level control strategy framework is established based on the control requirements and objectives of the natural gas continuous heat treatment equipment. It includes an inner-loop control strategy analysis channel, a middle-loop control strategy analysis channel, and an outer-loop control strategy analysis channel. Each channel has a specific input mapping relationship and output label. The input mapping relationship is determined by comprehensively considering various key factors and data sources during the operation of the natural gas continuous heat treatment equipment. The output label clearly defines the output control objective of each channel. Specifically, the inner-loop control strategy analysis channel uses the furnace temperature distribution, combustion heat release rate, and furnace temperature-workpiece heat transfer characteristics output from the digital twin model as input. Using control algorithms, such as PID control or fuzzy control algorithms, it accurately analyzes the gas flow rate, air flow rate, and air-fuel ratio to generate adjustment control quantities including gas flow rate, air flow rate, and air-fuel ratio. Through inner-loop control, rapid and stable control of the furnace combustion process is achieved, ensuring efficient and stable combustion. The middle-loop control strategy analysis channel takes the output of the inner-loop control strategy analysis channel as input, combines it with the predicted results of the workpiece core temperature, the lower heating value of natural gas, and the heat transfer coefficient, and uses optimization algorithms, such as genetic algorithms or particle swarm optimization algorithms, to analyze and adjust the zone furnace temperature setpoints and atmosphere parameters. This generates adjustment control quantities for the zone furnace temperature setpoints and atmosphere parameters, achieving dynamic optimization control of the furnace zone temperature and ensuring uniform heating of all parts of the workpiece. The outer-loop control strategy analysis channel, based on the workpiece's target heat treatment parameters and using the output of the middle-loop control strategy analysis channel as constraints, employs multi-step optimization algorithms, such as dynamic programming algorithms, to perform multi-step optimization analysis of the zone target temperature trajectory and gas supply strategy. This generates globally optimal control execution parameters, ensuring the heat treatment process achieves the best results from a global perspective. Through multi-layer strategy search, multiple control strategies, including inner-loop control, middle-loop control, and outer-loop control, are obtained. In the inner-loop control process, the functions and information provided by the combustion process model and heat transfer model in the digital twin model are used to achieve precise control. In the middle-loop control process, the heat transfer model and soft sensor model are invoked to achieve intelligent control of the natural gas continuous heat treatment equipment. During the outer loop control process, the thermal process gray box model and soft sensors are used to optimize the intelligent control of the natural gas continuous thermal treatment equipment.

[0049] Through a hierarchical control strategy, from rapid combustion control in the inner loop to zoned temperature optimization in the middle loop, and then to the generation of globally optimal parameters in the outer loop, fine and precise control of the continuous heat treatment process of natural gas is achieved, ensuring stable and efficient operation of the continuous heat treatment process of natural gas, and improving the heat treatment quality of workpieces and production efficiency.

[0050] In the digital twin model, the safety constraint simulation verification of the multi-layer strategy is carried out, and the unmeasurable state parameters are predicted by soft sensors. The quality index is introduced as a soft constraint condition into the optimization control to generate control execution parameters that meet the constraint conditions.

[0051] Specifically, after obtaining a multi-layered control strategy including inner, middle, and outer loops, this strategy is input into a digital twin model. The impact of the control strategy on the heat treatment process is predicted by simulating its operation in a virtual environment. Safety constraints include key indicators such as combustion stability, furnace temperature control accuracy, and atmosphere control boundaries to ensure the safe and stable operation of the heat treatment process. Simulation verifies whether multiple control strategies meet safety constraints, thus avoiding potential dangerous situations in actual operation. Simultaneously, soft sensors are used to predict unmeasurable state parameters. Quality indicators such as workpiece hardness, microstructure, and residual stress are introduced as soft constraints into the optimization control. Intelligent optimization algorithms adjust and optimize the control parameters, ultimately generating optimal control execution parameters that satisfy both safety constraints and ensure the workpiece heat treatment quality meets standards. Applying these optimal control execution parameters to actual natural gas continuous heat treatment equipment improves the intelligence level of the heat treatment process, enhances the control accuracy and reliability of the heat treatment equipment, and ultimately improves product production efficiency and quality.

[0052] Furthermore, the multi-layer strategy is subjected to safety constraint simulation verification in the digital twin model, including: using the multi-layer strategy as input variables, running the strategy through the digital twin model, simulating the workpiece thermal process ash box model, combustion process model, heat transfer model, and soft sensors, and outputting the combustion process operation results, heat transfer process operation results, workpiece thermal process operation results, and soft sensor prediction results; using preset safety constraint indicators to determine the safety of the combustion process operation results, heat transfer process operation results, workpiece thermal process operation results, and soft sensor prediction results, and when the safety constraint requirements are met, the multi-layer strategy is determined to have passed the safety constraint simulation verification, wherein the safety constraint requirements are based on the detection of combustion stability, zone furnace temperature deviation, and atmosphere control boundary safety threshold.

[0053] Specifically, a multi-layered control strategy, including inner-loop, middle-loop, and outer-loop control strategies, is input into the digital twin model. The digital twin model utilizes built-in workpiece thermal process gray box models, combustion process models, heat transfer models, and soft sensors to simulate the process using a multi-layered strategy. The workpiece thermal process gray box model simulates the temperature changes of the workpiece during heat treatment; the combustion process model simulates the heat generated by natural gas combustion and its impact on furnace temperature; the heat transfer model simulates the heat exchange between furnace gas and the workpiece; and the soft sensors predict parameters that cannot be directly measured, such as the core temperature of the workpiece, the lower heating value of natural gas, and the heat transfer coefficient. Through simulation, the results of the combustion process, heat transfer process, workpiece thermal process, and soft sensor predictions are output. Preset safety constraints are used to assess the safety of multiple simulation results. These constraints include combustion stability, zoned furnace temperature deviation, and atmosphere control boundary safety thresholds, ensuring the safety of the heat treatment process. For example, combustion stability indices are used to assess the uniformity and predictability of heat release during combustion, zoned furnace temperature deviation indices are used to ensure the uniformity of temperature distribution in different areas of the furnace, and atmosphere control boundary safety thresholds are used to ensure that the atmosphere composition within the furnace is within a safe range. When the simulation results meet all preset safety constraints, the multi-layer control strategy is considered to have passed the safety constraint simulation verification, ensuring the safety and feasibility of the control strategy. If it fails the safety constraint simulation verification, the multi-layer control strategy needs to be readjusted and optimized until all safety constraints are met to ensure the safety and reliability of the heat treatment process. Through simulation using digital twin models, potential safety risks can be predicted and avoided before practical application, thereby improving the overall safety and reliability of the heat treatment process.

[0054] Furthermore, by predicting unmeasurable state parameters using soft sensors, quality indicators are introduced as soft constraints into the optimization control, generating control execution parameters that meet the constraints. This includes: using the soft sensors to predict unmeasurable state parameters, and introducing the workpiece's hardness, microstructure, and residual stress quality indicators as soft constraints into the optimization control process; by establishing a multi-objective optimization function, under the premise of meeting safety and quality constraints, optimizing the zoned target temperature trajectory and gas supply strategy, generating control execution parameters that meet the constraints, and issuing them as the final control command to the natural gas continuous heat treatment equipment.

[0055] Specifically, by utilizing historical production data and real-time monitoring data, soft sensors predict the heat treatment parameters of the workpiece during the heat treatment process, obtaining unmeasurable state parameters. These unmeasurable state parameters allow for a more accurate assessment of the impact of the current heat treatment state on the workpiece's quality indicators. For example, by predicting the core characteristics of the workpiece, the differences in microstructure evolution in different parts of the workpiece can be obtained, thereby predicting the overall hardness distribution and whether the microstructure meets constraints. Based on the workpiece's performance requirements, processing characteristics, industry standards, and safety and reliability, workpiece quality indicators are set. These indicators include hardness, microstructure, and residual stress. Hardness measures the workpiece's surface or local resistance to plastic deformation, scratches, indentations, and other external forces, affecting its wear resistance, strength, and service life. Microstructure refers to the structural characteristics of the metal grains and phases within the workpiece, including their type, morphology, size, and distribution, determining the workpiece's comprehensive mechanical properties and corrosion resistance. Residual stress is the stress that remains within the workpiece during manufacturing and is in a self-equilibrium state, related to the workpiece's dimensional stability and fatigue performance. The workpiece's hardness, microstructure, and residual stress quality indicators are introduced as soft constraints into the optimization control process. Soft constraints refer to limiting conditions that allow flexible variation within a reasonable range during optimization, serving as a reference direction for improving workpiece quality. Multiple objectives are comprehensively considered in the heat treatment process. These objectives include key indicators ensuring process safety, such as combustion stability, furnace temperature control accuracy, and atmosphere control boundaries, as well as key parameters ensuring product quality, such as workpiece hardness, microstructure, and residual stress. Integrating these multiple objectives, including safety and quality constraints, a multi-objective optimization function is established using methods such as weighted summation or analytic hierarchy process (AHP). Intelligent optimization algorithms, such as genetic algorithms or particle swarm optimization, are employed. Genetic algorithms simulate natural selection and genetic mechanisms in biological evolution, finding the optimal solution through continuous iteration. Particle swarm optimization, based on swarm intelligence, finds the optimal solution through collaboration and information sharing among individuals. An optimization algorithm is used to optimize the target temperature trajectory of each zone and the gas supply strategy, generating control execution parameters that meet the constraints. Optimizing the target temperature trajectory ensures uniform heating of all parts of the workpiece during heat treatment, avoiding quality problems caused by temperature differences. Optimizing the gas supply strategy allows for precise control of gas flow rate and supply time based on the needs of the heat treatment process, achieving efficient energy utilization. The control execution parameters include gas flow rate, air flow rate, furnace temperature setpoint, and atmosphere parameters. These control parameters are sent as final control commands to the natural gas continuous heat treatment equipment, guiding the equipment to precisely adjust the heat treatment process according to the set parameters. This achieves precise control of workpiece quality, improves the efficiency and stability of the production process, and ultimately enhances the quality and consistency of the finished product.

[0056] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0057] Obviously, those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for intelligent control of a natural gas continuous thermal treatment plant based on digital twinning, characterized in that, The method comprises the following steps: a digital twin model of the natural gas continuous heat treatment equipment is established, the digital twin model comprising a workpiece heat process grey box model, a combustion process model, a heat exchange model, and a soft sensor; combustion process characteristics and unmeasurable state are predicted by the digital twin model using online collected production monitoring data to obtain state prediction results; a multi-layer strategy search is performed according to the state prediction results using a hierarchical control strategy, including inner loop control, middle loop control, and outer loop control; the multi-layer strategy is simulated and verified in the digital twin model under safety constraints, and unmeasurable state parameters are predicted by the soft sensor, and a quality index is introduced as a soft constraint condition into the optimization control to generate control execution parameters meeting the constraint conditions. 2.The digital-twin-based intelligent control method of a natural gas continuous heat treatment device according to claim 1, wherein, The method for establishing the digital twin model of the natural gas continuous heat treatment equipment comprises the following steps: According to the physical structure and process partition of the natural gas continuous heat treatment equipment, the furnace body is divided into a preheating zone, a heating zone, a soaking zone, and a holding zone, and a partition energy balance model is constructed; According to the sensor layout and actuator topological relationship, a mapping relationship between monitoring values and operation values is established, and the partition energy balance model is embedded; Based on a mechanism-data fusion strategy, historical production data and online monitoring data are used for parameter identification and calibration of the partition energy balance model, the digital twin body is updated, and the digital twin model is obtained; According to historical production data and monitoring data, workpiece heat treatment process, natural gas combustion process, furnace temperature-workpiece heat exchange relationship, and workpiece unmeasurable state parameter prediction analysis are performed, a workpiece heat process grey box model, a combustion process model, a heat exchange model, and a soft sensor are established, and the workpiece heat process grey box model, the combustion process model, the heat exchange model, and the soft sensor are embedded in the digital twin model as built-in modules. 3.The intelligent control method of the natural gas continuous heat treatment equipment based on digital twinning according to claim 2, characterized in that, The method for establishing the workpiece heat process grey box model comprises the following steps: Based on the thermal conductivity, specific heat capacity, and density thermal physical parameters of the workpiece material, the workpiece is divided according to the physical structure to obtain a plurality of heat conduction units; Using the historical production data and monitoring data, the thermal energy changes of the plurality of heat conduction units during the heat treatment process are fitted to obtain the workpiece change relationship during the in-furnace heat treatment process, and the workpiece heat process grey box model is constructed. 4.The intelligent control method of the natural gas continuous heat treatment equipment based on digital twinning according to claim 2, characterized in that, The method for establishing the combustion process model comprises the following steps: Based on the natural gas flow, air flow, and air-fuel ratio, the historical production data and online monitoring data are used to establish the time sequence heat release rate and thermal efficiency relationship of the natural gas combustion process; The time sequence heat release rate and thermal efficiency of the natural gas combustion are corrected by monitoring the data of the flue gas composition and oxygen content to obtain the combustion stability evaluation relationship of the natural gas combustion process, and the combustion process model is constructed. 5.The digital-twin-based intelligent control method of a natural gas continuous thermal treatment device according to claim 2, wherein, The method for establishing the heat exchange model comprises the following steps: According to the furnace temperature distribution, workpiece loading amount, workpiece surface temperature, and atmosphere flow parameter, the heat exchange boundary conditions between the furnace gas and the workpiece are established; Based on the convection and radiation heat exchange coupling mechanism, the comprehensive heat exchange relationship under the heat exchange boundary conditions is constructed; The comprehensive heat exchange relationship and heat exchange boundary conditions are parameter corrected based on the historical production data and monitoring data to obtain the heat exchange model reflecting the workpiece time sequence heat treatment process. 6.The intelligent control method of the natural gas continuous heat treatment equipment based on digital twinning according to claim 2, characterized in that, The method for establishing the soft sensor comprises the following steps: Based on historical production data and monitoring data of the natural gas continuous heat treatment equipment, an initial data-driven prediction model is constructed, taking measurable variables of furnace temperature, natural gas flow, air flow, flue gas oxygen content and workpiece surface temperature as inputs, and taking unmeasurable states of workpiece core temperature, natural gas low heat value and heat transfer coefficient as outputs; The workpiece thermal process grey box model, the combustion process model and the heat exchange model are used as physical constraints and feature generation modules to provide the data-driven prediction model with workpiece heat conduction characteristics, combustion heat release efficiency and furnace temperature-workpiece heat exchange coefficient mechanism parameters; The data-driven prediction model is trained using the historical production data, and the mechanism parameter output is compared and corrected to obtain a fusion corrected data-driven prediction model as the soft sensor for reasoning and prediction according to the input measurable variables, and outputting the unmeasurable state parameter prediction results. 7.The digital-twin-based intelligent control method of a natural gas continuous thermal treatment device according to claim 6, characterized in that, The unmeasurable state parameters include workpiece core temperature, natural gas low heat value and heat transfer coefficient, and the combustion process features include furnace temperature distribution, combustion heat release rate and furnace temperature-workpiece heat exchange characteristics, wherein the combustion heat release rate is obtained through the combustion process model, the furnace temperature-workpiece heat exchange characteristics are obtained through the heat exchange model, and the furnace temperature distribution is obtained through the heat exchange model and the combustion process model. 8.The digital-twin-based intelligent control method of a natural gas continuous thermal treatment device according to claim 7, characterized in that, A hierarchical control strategy is adopted for multi-layer strategy search according to the state prediction results, including inner loop control, middle loop control and outer loop control, including: A multi-layer control strategy framework is established, including inner loop control strategy analysis channel, middle loop control strategy analysis channel and outer loop control strategy analysis channel, each having input mapping relationship and output label; The inner loop control strategy analysis channel takes the combustion heat release rate and the furnace temperature-workpiece heat exchange characteristics output by the digital twin model as input to adjust the gas flow, air flow and air-fuel ratio analysis output; The middle loop control strategy analysis channel takes the output results of the inner loop control strategy analysis channel as input, and adjusts the partition furnace temperature set value and atmosphere parameters according to the workpiece core temperature and heat transfer coefficient analysis output; The outer loop control strategy analysis channel takes the workpiece target heat treatment parameters as input, and takes the output results of the middle loop control strategy analysis channel as constraint conditions to perform multi-step optimization analysis of partition target temperature trajectory and gas supply strategy output. 9.The digital-twin-based intelligent control method of a natural gas continuous thermal treatment device according to claim 1, wherein, The multi-layer strategy is simulated and verified under safety constraints in the digital twin model, including: The multi-layer strategy is taken as an input variable, and the strategy is run through the digital twin model, and the workpiece thermal process grey box model, the combustion process model, the heat exchange model and the soft sensor are simulated to output the combustion process running results, the heat exchange process running results, the workpiece thermal process running results and the soft sensor prediction results; The combustion process running results, the heat exchange process running results, the workpiece thermal process running results and the soft sensor prediction results are safety judged by using the preset safety constraint index, and when the safety constraint requirement is met, the multi-layer strategy is determined to pass the safety constraint simulation verification, wherein the safety constraint requirement is based on the detection of combustion stability, partition furnace temperature deviation and atmosphere control boundary safety threshold. 10.The intelligent control method of the natural gas continuous heat treatment equipment based on digital twinning according to claim 9, characterized in that, The unmeasurable state parameters are predicted by the soft sensor, quality indexes are introduced as soft constraint conditions into the optimization control, control execution parameters satisfying the constraint conditions are generated, including: The unmeasurable state parameters are predicted by the soft sensor, and the hardness, structure and residual stress quality indexes of the workpiece are introduced as soft constraint conditions into the optimization control process. By establishing a multi-objective optimization function, the partition target temperature trajectory and the gas supply strategy are optimized and solved under the premise of meeting the safety constraint requirements and the quality constraint conditions, control execution parameters satisfying the constraint conditions are generated as the final control instructions and are issued to the natural gas continuous heat treatment equipment.

Images