Intelligent control device and method for furnace temperature

Abstract

The present application relates to the technical field of furnace temperature control, and discloses a furnace temperature intelligent control device and method, which comprises a multi-source sensing temperature field construction module, a hot state process inversion module, a target temperature field mapping generation module, a multi-target dynamic optimization module, a control instruction calculation mapping module and a rolling execution optimization module, obtains a three-dimensional thermal temperature field by interpolating multi-source heterogeneous sensing data, obtains the current material reaction process state by inverting the three-dimensional thermal temperature field, maps it to a preset process knowledge graph to obtain a target temperature benchmark and distribution rule, and constructs a target temperature field distribution, solves the global optimal dynamic transition trajectory by combining the thermodynamics-heat transfer constraints of the furnace, maps the global optimal dynamic transition trajectory to obtain an accurate energy input instruction set, and updates the three-dimensional temperature field distribution by fusing newly collected real-time data to reconstruct the trajectory and obtain an updated control strategy, so that the accuracy of furnace temperature intelligent control can be improved.

Description

Technical Field

[0001] This invention relates to the field of furnace and kiln temperature control technology, and in particular to a furnace and kiln temperature intelligent control device and method. Background Technology

[0002] Existing technologies have significant shortcomings in the data processing and temperature field construction stages of furnace and kiln temperature control. They fail to systematically align and spatially interpolate multi-source heterogeneous sensing data, relying solely on a small amount of discrete temperature measurement data. This makes it impossible to accurately construct a three-dimensional thermo-temperature field covering the entire furnace and kiln area, resulting in incomplete temperature distribution sensing and large local temperature deviations. Furthermore, they fail to incorporate historical fuel control commands to perform thermal process inversion of the temperature field, relying solely on current temperature data to determine the operating status. This makes it difficult to accurately identify the current material reaction process stage, resulting in a lack of targeted temperature control that cannot adapt to the temperature requirements of different stages of material reaction and fails to meet the precise temperature control requirements of furnace and kiln processes.

[0003] Existing technologies do not generate target temperature field distribution based on the material reaction process state and preset process knowledge graph. Instead, they only set a single, fixed target temperature value, ignoring the spatial gradient distribution law of temperature inside the furnace, resulting in a lack of scientific and rational temperature control. Furthermore, they do not perform multi-objective dynamic optimization of the three-dimensional thermodynamic temperature field, target temperature field, and thermodynamic-heat transfer constraints. Instead, they generate control commands through simple adjustment logic, which cannot obtain the globally optimal temperature transition trajectory, resulting in low control accuracy. Finally, they do not establish a rolling execution optimization mechanism, only executing fixed control commands. They cannot dynamically adjust the control strategy based on real-time data, making it difficult to cope with changes in operating conditions and disturbances during furnace operation, leading to poor temperature stability and affecting product quality and production efficiency. Summary of the Invention

[0004] This invention provides an intelligent furnace temperature control device and method to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides an intelligent furnace temperature control device, characterized in that the device includes a multi-source sensing temperature field construction module, a thermal process inversion module, a target temperature field mapping generation module, a multi-objective dynamic optimization module, a control command solution mapping module, and a rolling execution optimization module, wherein:

[0006] The multi-source sensing temperature field construction module is used to perform spatial interpolation processing on the multi-source heterogeneous sensing data during the operation of the furnace and kiln to obtain the three-dimensional thermo-temperature field of the furnace and kiln operation process.

[0007] The thermal process inversion module is used to perform thermal process inversion on the three-dimensional thermo-temperature field based on historical fuel control commands, so as to obtain the current material reaction process state of the three-dimensional thermo-temperature field.

[0008] The target temperature field mapping generation module is used to map the current material reaction process state to a preset process knowledge graph to obtain the target temperature benchmark and distribution rules of the furnace operation process, and to construct a spatial gradient field on the target temperature benchmark and the distribution rules to obtain the target temperature field distribution of the furnace operation process.

[0009] The multi-objective dynamic optimization module is used to perform multi-objective dynamic optimization on the three-dimensional thermo-temperature field, the target temperature field distribution, and the thermodynamic-heat transfer constraints of the furnace, so as to obtain the global optimal dynamic transition trajectory of the target temperature field distribution.

[0010] The control command calculation and mapping module is used to perform control command calculation and mapping on the globally optimal dynamic transition trajectory to obtain the precise energy input command set of the globally optimal dynamic transition trajectory.

[0011] The rolling execution optimization module is used to execute the precise energy input instruction set, assimilate and fuse the newly acquired real-time data to obtain the updated three-dimensional temperature field distribution of the furnace operation process, and based on the updated three-dimensional temperature field distribution, perform rolling time-domain reconstruction of the globally optimal dynamic transition trajectory to obtain the updated control strategy of the furnace operation process.

[0012] In a preferred embodiment, when the multi-source sensing temperature field construction module performs spatial interpolation processing on the multi-source heterogeneous sensing data during furnace operation to obtain the three-dimensional thermo-temperature field of the furnace operation process, it is specifically used for:

[0013] Multi-source heterogeneous sensing data is collected during the operation of the furnace and kiln, and the multi-source heterogeneous sensing data is spatiotemporally aligned and cleaned to obtain a regularized sensing dataset of the furnace and kiln operation process.

[0014] Based on the geometric positions of the data points in the regularized sensing dataset within the three-dimensional space of the furnace, construct the spatial adjacency topology of the data points;

[0015] Based on the spatial adjacency topology, spatial interpolation is performed on the temperature data in the regularized sensing dataset to obtain the preliminary temperature field of the furnace operation process. The formula for calculating the desired temperature value in the preliminary temperature field is as follows:

[0016] ;

[0017] In the formula, The internal coordinates of the furnace are The desired temperature value at the desired spatial location. For the first The target temperature value of each core temperature control point. This represents the spatial influence coefficient of the temperature field distribution. To find the spatial location to the first Spatial Euclidean distance between core temperature control points It is an exponential function;

[0018] The smoothness of the preliminary temperature field is reasonably verified, and abnormal regions in the preliminary temperature field are corrected to obtain the three-dimensional thermo-temperature field of the furnace operation process.

[0019] In a preferred embodiment, when the thermal process inversion module executes thermal process inversion of the three-dimensional thermo-temperature field based on historical fuel control commands to obtain the current material reaction process state of the three-dimensional thermo-temperature field, it is specifically used for:

[0020] By performing feature analysis on historical fuel control commands, a historical control feature sequence of the furnace operation process is obtained;

[0021] Dynamic feature analysis is performed on the three-dimensional thermo-temperature field to obtain the current dynamic characteristics of the thermal field during the operation of the furnace.

[0022] The historical control feature sequence is matched with the current thermal field dynamic features in a time sequence, and the historical operating conditions most similar to the current thermal state and the corresponding stable material reaction stage are identified.

[0023] Based on the real-time change direction and amplitude of the current thermal field dynamic characteristics, a comprehensive reasoning and state determination are performed on the stable material reaction stage to obtain the current material reaction process state of the three-dimensional thermo-temperature field.

[0024] In a preferred embodiment, when the target temperature field mapping generation module maps the current material reaction process state to a preset process knowledge graph to obtain the target temperature benchmark and distribution rules of the furnace operation process, it is specifically used for:

[0025] Multimodal feature decoding is performed on the current material reaction process state to obtain the material type identifier and process stage identifier of the current material reaction process state;

[0026] Based on the process stage identifier, the preset process knowledge graph is traversed to obtain candidate process procedures associated with the process stage identifier.

[0027] According to the preset selection criteria, the target process procedure is selected from the candidate process procedures;

[0028] Extract the core target temperature value corresponding to the process stage identifier in the target process specification, and use the core target temperature value as the target temperature benchmark for the furnace operation process;

[0029] The parameter set describing the spatial distribution of temperature inside the furnace in the target process specification is used as the distribution rule for the furnace operation process.

[0030] In a preferred embodiment, when the target temperature field mapping generation module performs spatial gradient field construction on the target temperature reference and the distribution rule to obtain the target temperature field distribution during the furnace operation process, it is specifically used for:

[0031] The target temperature reference is decomposed into a structured form to obtain the set of master control nodes of the temperature field inside the furnace and the corresponding master control values ​​of the temperature field.

[0032] The distribution rules are tensorized to obtain the temperature gradient constraint tensor in the furnace space domain;

[0033] Using the master control node set and the master temperature field value as hard constraints, and the temperature gradient constraint tensor as a soft optimization objective, a variational optimization problem of the temperature field during the furnace operation process is constructed.

[0034] Solve the variational optimization problem of the temperature field to obtain the target temperature field distribution during the operation of the furnace.

[0035] In a preferred embodiment, when the multi-objective dynamic optimization module performs multi-objective dynamic optimization on the three-dimensional thermo-temperature field, the target temperature field distribution, and the thermodynamic-heat transfer constraints of the furnace to obtain the globally optimal dynamic transition trajectory of the target temperature field distribution, it is specifically used for:

[0036] Based on the three-dimensional thermo-temperature field and the target temperature field distribution, a multi-objective optimization problem is constructed with temperature field convergence, temperature distribution uniformity and process energy efficiency as the core indicators.

[0037] Based on the thermodynamic and heat transfer constraints of the furnace, the dynamic feasible region of the multi-objective optimization problem is established.

[0038] Based on a preset dynamic weight allocation strategy, the core indicators are synergistically weighed within the dynamic feasible domain to obtain a dynamically optimized target weight vector for the target temperature field distribution.

[0039] The dynamic optimization target weight vector is iteratively solved to obtain the set of transition trajectories of the target temperature field distribution;

[0040] Multi-objective optimization decision-making is performed on the trajectories in the set of transition trajectories, and the trajectory with the best comprehensive evaluation is taken as the global optimal dynamic transition trajectory of the target temperature field distribution.

[0041] In a preferred embodiment, when the control command calculation and mapping module performs control command calculation and mapping on the globally optimal dynamic transition trajectory to obtain the precise energy input command set of the globally optimal dynamic transition trajectory, it is specifically used for:

[0042] The global optimal dynamic transition trajectory is discretized in time and space to obtain the temperature field setpoint sequence of the global optimal dynamic transition trajectory.

[0043] Based on the spatial location and thermal influence range of the furnace, the temperature field setpoint sequence is mapped to the basic energy demand command of the furnace.

[0044] The basic energy demand command is adaptively compensated to obtain the compensated energy command for the furnace.

[0045] The compensated energy command is subjected to execution-level quantization to obtain the precise energy input command set.

[0046] In a preferred embodiment, when the rolling execution optimization module executes the precise energy input instruction set and assimilates and fuses the newly acquired real-time data to obtain the updated three-dimensional temperature field distribution of the furnace operation process, it is specifically used for:

[0047] The precise energy input instruction set is input to the energy supply actuator of the furnace, and the status feedback data of the energy supply actuator is collected simultaneously.

[0048] Acquire the newly collected real-time sensing data of the furnace and kiln within the current control cycle, perform multi-source data fusion and cleaning on the real-time sensing data of the furnace and kiln, and obtain the effective real-time dataset of the real-time sensing data of the furnace and kiln.

[0049] The state feedback data, the effective real-time data, and the current state of the three-dimensional thermo-temperature field are used to perform multimodal state reconstruction to generate the assimilated furnace operation state of the furnace operation process.

[0050] Based on the assimilated furnace operating state, the temperature distribution of the furnace temperature field is extrapolated to obtain the updated three-dimensional temperature field distribution of the furnace operation process.

[0051] In a preferred embodiment, when the rolling execution optimization module performs rolling time-domain reconstruction of the globally optimal dynamic transition trajectory based on the updated three-dimensional temperature field distribution to obtain the updated control strategy for the furnace operation process, it is specifically used for:

[0052] Based on the updated three-dimensional temperature field distribution and the target temperature field distribution, a deviation analysis is performed on the global optimal dynamic transition trajectory to obtain the state deviation information between the current temperature field and the target temperature field.

[0053] Based on the state deviation information, the unexecuted portion of the global optimal dynamic transition trajectory is corrected to obtain the corrected transition trajectory of the global optimal dynamic transition trajectory.

[0054] Based on the thermodynamic and heat transfer constraints of the furnace and the current operating conditions, the feasibility of the modified transition trajectory is verified and smoothed to obtain the optimized rolling transition trajectory.

[0055] Based on the optimized rolling transition trajectory, the control command sequence is regenerated to obtain the updated control strategy for the furnace operation process.

[0056] To address the above problems, the present invention also provides a method for intelligent control of furnace temperature, the method comprising:

[0057] S1. Spatial interpolation processing is performed on the multi-source heterogeneous sensing data during the operation of the furnace to obtain the three-dimensional thermo-temperature field of the furnace operation process.

[0058] S2. Based on historical fuel control commands, perform thermal process inversion on the three-dimensional thermo-temperature field to obtain the current material reaction process state of the three-dimensional thermo-temperature field;

[0059] S3. Map the current material reaction process state to a preset process knowledge graph to obtain the target temperature benchmark and distribution rules of the furnace operation process, and construct a spatial gradient field for the target temperature benchmark and the distribution rules to obtain the target temperature field distribution of the furnace operation process.

[0060] S4. Perform multi-objective dynamic optimization on the three-dimensional thermo-temperature field, the target temperature field distribution, and the thermodynamic-heat transfer constraints of the furnace to obtain the global optimal dynamic transition trajectory of the target temperature field distribution.

[0061] S5. Perform control command decomposition and mapping on the globally optimal dynamic transition trajectory to obtain the precise energy input command set of the globally optimal dynamic transition trajectory;

[0062] S6. Execute the precise energy input instruction set and assimilate and fuse the newly acquired real-time data to obtain the updated three-dimensional temperature field distribution of the furnace operation process. Based on the updated three-dimensional temperature field distribution, perform rolling time-domain reconstruction of the global optimal dynamic transition trajectory to obtain the updated control strategy of the furnace operation process.

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] 1. This invention lays a scientific foundation for furnace temperature control through precise temperature field construction and material state perception. It collects multi-source heterogeneous sensing data, processes it through spatiotemporal alignment, cleaning, and spatial interpolation, and constructs a three-dimensional thermodynamic temperature field covering the entire domain, ensuring accurate and comprehensive temperature distribution perception. By combining historical fuel control commands to invert the thermal process, it accurately identifies the current material reaction process state, and then maps it to a preset process knowledge graph to generate a target temperature field distribution suitable for the operating conditions, allowing temperature control to align with the material reaction requirements.

[0065] 2. This invention significantly improves the accuracy and stability of temperature control through multi-objective optimization and dynamic adjustment. Based on a three-dimensional thermodynamic temperature field, a target temperature field, and thermodynamic-heat transfer constraints, multi-objective dynamic optimization obtains the globally optimal transition trajectory; the solution mapping generates a precise energy input instruction set, and during execution, real-time data is assimilated and fused to update the temperature field and continuously reconstruct the trajectory, dynamically adjusting the control strategy to balance temperature uniformity and process energy efficiency, ensuring stable furnace operation and product quality. Attached Figure Description

[0066] Figure 1 This is a system architecture diagram of an intelligent furnace temperature control device provided in an embodiment of the present invention;

[0067] Figure 2 This is a flowchart illustrating an intelligent furnace temperature control method according to an embodiment of the present invention.

[0068] As shown in the figure: 100. A furnace temperature intelligent control device; 101. Multi-source sensing temperature field construction module; 102. Thermal process inversion module; 103. Target temperature field mapping generation module; 104. Multi-objective dynamic optimization module; 105. Control command solution mapping module; 106. Rolling execution optimization module. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments belong to some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0070] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “said” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0071] Depending on the context, the word "if" or "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0072] Furthermore, the timing of the steps in the following method embodiments is merely an example and not a strict limitation.

[0073] In practice, the server-side equipment deployed by a furnace temperature intelligent control device may consist of one or more devices. This furnace temperature intelligent control device can be implemented as a business instance, a virtual machine, or a hardware device. For example, it can be implemented as a business instance deployed on one or more devices in a cloud node. Simply put, it can be understood as software deployed on a cloud node, providing intelligent furnace temperature control to various user terminals. Alternatively, it can be implemented as a virtual machine deployed on one or more devices in a cloud node, with application software installed to manage various user terminals. Or, it can also be implemented as a server composed of numerous identical or different types of hardware devices, with one or more hardware devices configured to provide intelligent furnace temperature control to various user terminals.

[0074] In terms of implementation, the intelligent furnace temperature control device and the user terminal are mutually compatible. That is, if the intelligent furnace temperature control device is implemented as an application installed on a cloud service platform, the user terminal is implemented as a client that establishes a communication connection with the application; or if the intelligent furnace temperature control device is implemented as a website, the user terminal is implemented as a webpage; or if the intelligent furnace temperature control device is implemented as a cloud service platform, the user terminal is implemented as a mini-program in an instant messaging application.

[0075] like Figure 1 The diagram shown is a system architecture diagram of an intelligent furnace temperature control device provided in an embodiment of the present invention.

[0076] The intelligent furnace temperature control device 100 described in this invention can be installed on a cloud server. In terms of implementation, it can function as one or more service devices, or as an application installed on the cloud (e.g., a mobile service operator's server, server cluster, etc.), or it can be developed into a website. Depending on the functions implemented, the intelligent furnace temperature control device 100 may include a multi-source sensing temperature field construction module 101, a thermal process inversion module 102, a target temperature field mapping generation module 103, a multi-objective dynamic optimization module 104, a control command solution mapping module 105, and a rolling execution optimization module 106. The modules described in this invention can also be called units, referring to a series of computer program segments that can be executed by an electronic device's processor and perform a fixed function, stored in the electronic device's memory.

[0077] In this embodiment of the invention, in a furnace temperature intelligent control device, each of the above-mentioned modules can be implemented independently and can be invoked by other modules. Invocation here can be understood as a module connecting to multiple modules of another type and providing corresponding services to those connected modules. In the furnace temperature intelligent control device provided by this embodiment of the invention, without modifying the program code, the applicable scope of the furnace temperature intelligent control device architecture can be adjusted by adding modules and directly invoking them, achieving cluster-based horizontal expansion to quickly and flexibly expand the furnace temperature intelligent control device. In practical applications, the above modules can be set in the same device or different devices, or they can be set in a virtual device, such as a service instance in a cloud server.

[0078] The following describes, with reference to specific embodiments, the various components and specific workflow of an intelligent furnace temperature control device:

[0079] The multi-source sensing temperature field construction module 101 is used to perform spatial interpolation processing on the multi-source heterogeneous sensing data during the operation of the furnace and kiln to obtain the three-dimensional thermo-temperature field during the operation of the furnace and kiln.

[0080] In this embodiment of the invention, when the multi-source sensing temperature field construction module 101 performs spatial interpolation processing on the multi-source heterogeneous sensing data during the furnace operation process to obtain the three-dimensional thermo-temperature field of the furnace operation process, it is specifically used for:

[0081] Multi-source heterogeneous sensing data is collected during the operation of the furnace and kiln, and the multi-source heterogeneous sensing data is spatiotemporally aligned and cleaned to obtain a regularized sensing dataset of the furnace and kiln operation process.

[0082] Based on the geometric positions of the data points in the regularized sensing dataset within the three-dimensional space of the furnace, construct the spatial adjacency topology of the data points;

[0083] Based on the spatial adjacency topology, spatial interpolation is performed on the temperature data in the regularized sensing dataset to obtain the preliminary temperature field of the furnace operation process. The formula for calculating the desired temperature value in the preliminary temperature field is as follows:

[0084] ;

[0085] In the formula, The internal coordinates of the furnace are The desired temperature value at the desired spatial location. For the first The target temperature value of each core temperature control point. This represents the spatial influence coefficient of the temperature field distribution. To find the spatial location to the first Spatial Euclidean distance between core temperature control points It is an exponential function;

[0086] The smoothness of the preliminary temperature field is reasonably verified, and abnormal regions in the preliminary temperature field are corrected to obtain the three-dimensional thermo-temperature field of the furnace operation process.

[0087] By deploying various sensing devices such as temperature sensors and infrared detectors inside and outside the furnace, the system comprehensively collects multi-source heterogeneous sensing data, including temperature data and equipment operating status data, during operation. The timestamps of data collected from different devices are calibrated and aligned using a unified time benchmark to eliminate data deviations in the time dimension. At the same time, based on the physical laws of furnace operation and the normal range of sensing data values, the system verifies the completeness and rationality of each data point, eliminates abnormal data that is missing key information or exceeds the normal range, and integrates them into a standardized sensing dataset with a unified format and reliable data.

[0088] Based on the three-dimensional structural model of the furnace, the specific spatial location inside the furnace corresponding to each data point in the regularized perception dataset is clearly defined. According to the spatial distance and relative orientation between each data point, the adjacency relationship between data points is determined, and adjacent data points are associated and marked. A spatial adjacency topology that can clearly reflect the interconnection relationship between each data point in the three-dimensional space of the furnace is constructed, so that the scattered data points form a spatially related whole.

[0089] Based on the constructed spatial adjacency topology, the adjacent data points around each spatial location to be interpolated are locked. The temperature data of these adjacent data points are used as a reference. Different reference weights are assigned according to the spatial distance between the adjacent data points and the location to be interpolated. The closer the adjacent data points are, the greater their influence on the temperature of the location to be interpolated. The temperature data of all adjacent data points are combined to calculate the temperature value of each location to be interpolated, filling the temperature gaps between discrete data points in the three-dimensional space of the furnace and forming a preliminary temperature field covering the entire area of ​​the furnace.

[0090] By employing a continuous smoothness verification method, all temperature data in the preliminary temperature field are traversed to check whether the temperature changes in adjacent spatial locations conform to the physical laws of heat transfer in the furnace. Abnormal areas with sudden temperature changes, abnormally high or low values ​​are identified. Combining the distribution characteristics of normal temperature data around the abnormal areas and the heat transfer characteristics of the furnace, the temperature values ​​of the abnormal areas are adjusted and corrected to ensure that the temperature of the abnormal areas remains continuous and coordinated with the surrounding temperatures. Ultimately, a three-dimensional thermodynamic temperature field with a continuous and reasonable temperature distribution that can accurately reflect the actual thermal state of the furnace is formed.

[0091] The desired temperature value at the desired spatial location is obtained by calculation using a formula. The calculation is based on the target temperature value of the core temperature control point, the spatial influence coefficient of the temperature field distribution, and the spatial Euclidean distance from the desired spatial location to the core temperature control point. After specific logical operations, the temperature value at that location is finally determined.

[0092] The target temperature value of the core temperature control point comes from the regularized sensing dataset during the operation of the furnace. Data points that are distributed in key locations of the furnace and can reflect the core temperature status are selected from the regularized sensing dataset as core temperature control points. The target temperature value of each core temperature control point is obtained directly by reading the temperature records corresponding to these data points.

[0093] The spatial influence coefficient of temperature field distribution is a fixed value preset based on the structural characteristics, heat transfer law and historical operating temperature data of the furnace. By analyzing the temperature propagation range and influence intensity of different areas of the furnace, and combining the statistical summary of a large amount of actual operating data, a coefficient that can accurately reflect the temperature decay law in the furnace space is determined to ensure that the temperature field calculation conforms to the physical characteristics of the actual operation of the furnace.

[0094] The spatial Euclidean distance between the desired spatial location and the core temperature control point is obtained by determining the coordinates of the desired spatial location in the three-dimensional space of the furnace and the coordinates of each core temperature control point, and by measuring the straight-line distance between the two points in three-dimensional space according to the calculation logic of the distance between two points in space.

[0095] The significance of this formula lies in its ability to accurately calculate the desired temperature value at any desired spatial location within the furnace, providing data support for constructing a complete three-dimensional thermo-temperature field. Based on the target temperature value of the core temperature control point, an exponential function combined with the spatial influence coefficient and spatial Euclidean distance is used to demonstrate the law of temperature decay with increasing spatial distance; the farther away from the core temperature control point, the more significant the temperature decay, ensuring that the temperature calculation conforms to the physical logic of temperature distribution within the furnace. Through the operational logic of the numerator and denominator, the influence of each core temperature control point is weighted and integrated, enabling the desired temperature value at the desired spatial location to comprehensively reflect the temperature influence of the surrounding core temperature control points. The calculation results more closely match the actual temperature distribution of the furnace, providing accurate and reliable temperature data for the subsequent construction of the preliminary temperature field.

[0096] The beneficial effects are as follows: by deploying multiple sensing devices on the furnace, comprehensive multi-source heterogeneous sensing data is collected during operation, covering key information such as temperature and equipment operating conditions, ensuring the comprehensiveness of data collection. The timestamps of data from different devices are aligned using a unified timeline to eliminate time discrepancies. Simultaneously, the integrity and rationality of each data point are verified, and abnormal data lacking key information or exceeding normal value ranges are removed, resulting in a standardized sensing dataset. This operation purifies the data from the source, avoiding interference from fragmented and abnormal data in subsequent processing, and providing high-quality, standardized foundational data for the construction of a three-dimensional thermodynamic temperature field.

[0097] By utilizing the three-dimensional structural model of the furnace, the specific spatial coordinates of each data point in the regularized sensing dataset are clearly defined. Based on the spatial distance and relative orientation between data points, the adjacent relationships are determined and associated markers are applied, constructing a spatial adjacency topology. This relationship clearly presents the adjacent connection status of each data point in three-dimensional space, breaking the isolation of the data and providing a clear spatial association basis for subsequent spatial interpolation calculations, ensuring that the interpolation process can fully utilize the reference value of surrounding data.

[0098] Based on spatial adjacency topology, neighboring data points around the interpolation location are identified. Using the temperature data of these data points as a reference, weights are assigned according to spatial distance, with closer points receiving greater weight. This data is then integrated to calculate the temperature at the interpolation location, forming a preliminary temperature field. Combining the logical operations of the formula, and using the target temperature value of the core temperature control points as a basis, an exponential function is used to represent the temperature decay law with spatial distance. By integrating the influence of each core control point, the calculation results closely match the physical logic of the actual temperature distribution in the furnace. This step fills the temperature gaps between discrete data points, achieving preliminary coverage of the entire furnace temperature range and providing core data support for constructing a complete temperature field.

[0099] A smoothness verification method is employed to traverse the initial temperature field, checking whether temperature changes at adjacent locations are continuous and gradual, and identifying areas with abrupt temperature changes, abnormally high or low values. For abnormal areas, the abnormal temperature values ​​are adjusted by combining temperature data from surrounding normal areas and the physical laws governing furnace operation, ensuring that the abnormal areas are consistent with the surrounding temperature changes, ultimately yielding a three-dimensional thermo-temperature field. This operation further refines the temperature field data, eliminating potential local biases from interpolation calculations, ensuring a continuous and reasonable temperature field distribution, and accurately reflecting the actual temperature state during furnace operation. This provides precise temperature field data for subsequent thermal process inversion, temperature control, and other stages.

[0100] The thermal process inversion module 102 is used to perform thermal process inversion on the three-dimensional thermo-temperature field based on historical fuel control commands, so as to obtain the current material reaction process state of the three-dimensional thermo-temperature field.

[0101] In this embodiment of the invention, when the thermal process inversion module 102 executes thermal process inversion of the three-dimensional thermo-temperature field based on historical fuel control commands to obtain the current material reaction process state of the three-dimensional thermo-temperature field, it is specifically used for:

[0102] By performing feature analysis on historical fuel control commands, a historical control feature sequence of the furnace operation process is obtained;

[0103] Dynamic feature analysis is performed on the three-dimensional thermo-temperature field to obtain the current dynamic characteristics of the thermal field during the operation of the furnace.

[0104] The historical control feature sequence is matched with the current thermal field dynamic features in a time sequence, and the historical operating conditions most similar to the current thermal state and the corresponding stable material reaction stage are identified.

[0105] Based on the real-time change direction and amplitude of the current thermal field dynamic characteristics, a comprehensive reasoning and state determination are performed on the stable material reaction stage to obtain the current material reaction process state of the three-dimensional thermo-temperature field.

[0106] Historical fuel control instructions stored during furnace operation are retrieved. These instructions contain key control information such as the specific method of fuel supply, the duration of supply, and the intensity of supply. The core control elements of each historical instruction are extracted and analyzed, and arranged in chronological order of instruction execution. The scattered historical instructions are transformed into a continuous sequence that clearly reflects the fuel control patterns at different times, thus obtaining the historical control characteristic sequence of the furnace operation process.

[0107] The temperature distribution data of the three-dimensional thermo-temperature field is continuously tracked and dynamically analyzed to extract core dynamic information such as temperature values ​​of key regions, temperature gradient changes between different regions, movement paths of high-temperature regions, and duration of stable temperature. This scattered dynamic information is integrated and summarized to form a feature set that can comprehensively and accurately characterize the current temperature field change state of the furnace, thus obtaining the current thermal field dynamic characteristics of the furnace operation process.

[0108] Using time as the main correlation, the historical control feature sequence is compared with the current thermal field dynamic features one by one along the time dimension. By analyzing the degree of fit between the thermal field change pattern corresponding to the historical control features and the current thermal field dynamic features, the historical operating conditions that best match the historical control mode, thermal field change trend and the current thermal state are selected. At the same time, the stage information of the material reaction in a stable state is extracted from the record of the historical operating conditions, and the historical operating conditions and corresponding stable material reaction stages that are most similar to the current thermal state are identified.

[0109] Continuously monitor the real-time changes in the dynamic characteristics of the current thermal field, clarify key changes such as the trend of temperature rise and fall, the degree of increase or decrease in temperature gradient, and the movement trend of high-temperature regions. Combine these real-time change data to comprehensively infer the selected stable material reaction stages, analyze the intrinsic relationship between the current dynamic changes in the thermal field and the stable material reaction stages, determine the degree of progress of the material reaction, whether it is in a transitional state or a stable state, and finally obtain the current material reaction process status of the three-dimensional thermo-temperature field.

[0110] The beneficial effects are that by retrieving historical fuel control commands stored in the furnace, core control elements such as fuel supply quantity, supply frequency, and supply period can be comprehensively extracted and sorted into a historical control characteristic sequence in chronological order. This sequence fully presents the control patterns of different past periods, providing a historical control basis for current operating condition analysis, avoiding isolated judgments divorced from historical operating context, and making subsequent thermal process inversion more relevant to reality.

[0111] Continuous dynamic tracking of the three-dimensional thermo-temperature field extracts key dynamic information such as temperature values, temperature gradient changes, high-temperature region movement trajectories, and temperature stabilization duration in each region, integrating them to form the current dynamic characteristics of the thermal field. This feature comprehensively and accurately characterizes the current temperature distribution and changing trend of the furnace, providing direct thermal field data support for judging the material reaction process and ensuring accurate perception of the current thermal state.

[0112] Using time as a correlation thread, historical control feature sequences are compared one by one with current thermal field dynamic features along a temporal dimension to accurately match the degree of fit between the two, and to select the historical operating conditions with the highest similarity to the current thermal state. At the same time, information on the stage where the material reaction is in a stable state under the historical operating conditions is extracted, providing a historical reference for judging the current material reaction process and reducing the uncertainty of direct judgment.

[0113] Continuous monitoring of real-time changes in the dynamic characteristics of the current thermal field is used to clarify information such as the direction of temperature rise and fall, the magnitude of gradient increase and decrease, and the movement trend of high-temperature regions. This dynamic data is then used to comprehensively infer the reaction stages of selected stable materials. The correlation between current thermal field changes and stable stages is analyzed to accurately determine the progress of the material reaction, whether it is in a transitional or stable state, and ultimately to obtain the current state of the material reaction process. This process achieves the organic integration of historical and real-time data, ensuring that the determination of the material reaction process state is both scientific and accurate, providing precise operating condition basis for the subsequent construction of the target temperature field.

[0114] The target temperature field mapping generation module 103 is used to map the current material reaction process state to a preset process knowledge graph to obtain the target temperature benchmark and distribution rules of the furnace operation process, and to construct a spatial gradient field on the target temperature benchmark and the distribution rules to obtain the target temperature field distribution of the furnace operation process.

[0115] In this embodiment of the invention, when the target temperature field mapping generation module 103 maps the current material reaction process state to a preset process knowledge graph to obtain the target temperature benchmark and distribution rules of the furnace operation process, it is specifically used for:

[0116] Multimodal feature decoding is performed on the current material reaction process state to obtain the material type identifier and process stage identifier of the current material reaction process state;

[0117] Based on the process stage identifier, the preset process knowledge graph is traversed to obtain candidate process procedures associated with the process stage identifier.

[0118] According to the preset selection criteria, the target process procedure is selected from the candidate process procedures;

[0119] Extract the core target temperature value corresponding to the process stage identifier in the target process specification, and use the core target temperature value as the target temperature benchmark for the furnace operation process;

[0120] The parameter set describing the spatial distribution of temperature inside the furnace in the target process specification is used as the distribution rule for the furnace operation process.

[0121] When the target temperature field mapping generation module 103 performs spatial gradient field construction on the target temperature reference and the distribution rule to obtain the target temperature field distribution during the furnace operation process, it is specifically used for:

[0122] The target temperature reference is decomposed into a structured form to obtain the set of master control nodes of the temperature field inside the furnace and the corresponding master control values ​​of the temperature field.

[0123] The distribution rules are tensorized to obtain the temperature gradient constraint tensor in the furnace space domain;

[0124] Using the master control node set and the master temperature field value as hard constraints, and the temperature gradient constraint tensor as a soft optimization objective, a variational optimization problem of the temperature field during the furnace operation process is constructed.

[0125] Solve the variational optimization problem of the temperature field to obtain the target temperature field distribution during the operation of the furnace.

[0126] Multimodal feature decoding is performed on the current material reaction process state to deeply analyze the feature information reflecting the essential properties of the material in this state, clarify the key elements such as the composition and physical properties of the material, and extract the material type identifier that can uniquely identify the type of material. At the same time, the progress of the material reaction and the characteristics of the current reaction stage are analyzed to determine the process stage identifier that can accurately characterize the process progress.

[0127] Based on the obtained process stage identifier, a traversal operation is initiated on the preset process knowledge graph. This knowledge graph pre-stores the association between different process stages and corresponding process procedures. By enumerating all process procedure entries in the knowledge graph that have a mapping relationship with the process stage identifier, these related process procedures are collected and integrated to obtain candidate process procedures associated with the process stage identifier.

[0128] Referring to the preset selection criteria, which cover key evaluation dimensions such as process maturity, production efficiency, energy consumption control, and product quality stability, a comprehensive evaluation is conducted on each of the candidate process procedures. The performance of each candidate process procedure in each evaluation dimension is compared one by one, and the process procedure that achieves the best standard or the best overall performance in all evaluation dimensions is selected. The target process procedure is then selected from the candidate process procedures.

[0129] The target process procedure is analyzed in a structured manner to locate the core temperature control parameter that directly corresponds to the process stage identifier. This parameter is the key temperature standard set to ensure the normal progress of material reaction and the achievement of product quality standards under this process stage. This core temperature parameter is extracted and determined as the target temperature benchmark for the furnace operation process, providing a clear reference standard for subsequent temperature control.

[0130] By deeply exploring the descriptive information about the spatial distribution of temperature inside the furnace in the target process specification, we can extract the parameter set of key contents such as the temperature range of different regions, temperature gradient requirements, and the distribution location of high-temperature regions. This parameter set can be directly used as the distribution rule for the furnace operation process, clarifying the temperature distribution requirements of each spatial location inside the furnace, and providing a basis for constructing a reasonable temperature field.

[0131] The target temperature benchmark is decomposed into a structured form. Based on the three-dimensional spatial structure of the furnace and the distribution of key areas of material reaction, the target temperature benchmark is broken down into several main control nodes covering key locations such as the core reaction zone, edge transition zone, and insulation zone of the furnace. The specific coordinates of each main control node in the furnace space are defined, and a corresponding temperature control standard value is assigned to each main control node. This forms a set of main control nodes and corresponding temperature field control values ​​for the internal temperature field of the furnace, ensuring that the core control position and standard of the temperature field are clear and unambiguous.

[0132] The distribution rules are constructed using tensors. The constraints in the distribution rules, such as the temperature range of different regions, temperature gradient requirements, and spatial distribution ratio, are arranged in a regular manner according to the three-dimensional dimension of the furnace space domain. The scattered rule descriptions are transformed into tensor forms that can fully cover the furnace space and reflect the temperature constraint relationships of each region. This yields the temperature gradient constraint tensor in the furnace space domain, enabling the distribution rules to be directly used for constraint optimization of the spatial temperature field.

[0133] Using the master node set and the master temperature field value as hard constraints, the temperature field constructed subsequently must strictly conform to the corresponding master temperature field value at all master node positions without deviation. At the same time, the temperature gradient constraint tensor is used as a soft optimization objective, clarifying that the temperature field, while satisfying the hard constraints, should conform as closely as possible to the distribution requirements specified by the temperature gradient constraint tensor. By integrating these two types of constraints and objectives, a variational optimization problem of the temperature field in the furnace operation process is established with the rationalization of the temperature field distribution as the core, providing a clear optimization direction for solving the target temperature field.

[0134] The variational optimization problem of the temperature field is solved by a step-by-step iterative approach. First, an initial temperature field that meets the temperature requirements of the master node is initially constructed. Then, the initial temperature field is adjusted and optimized according to the temperature gradient constraint tensor. The temperature values ​​of each region of the furnace are continuously corrected so that the temperature field not only meets the temperature requirements of the master node, but also gradually approaches the distribution law specified by the temperature gradient constraint. After multiple iterations and adjustments, the target temperature field distribution of the furnace operation process that meets all constraints and optimization objectives is obtained.

[0135] The beneficial effects are that it performs multimodal feature decoding on the current material reaction process status, deeply analyzes key information reflecting the essential attributes of the material such as composition and physical properties, and extracts a unique material type identifier; simultaneously, it analyzes the progress of the material reaction and the core characteristics of the current stage, determining a precise process stage identifier to characterize the process progress. This operation achieves accurate extraction of core information about the current operating conditions, providing a clear identifier for subsequent matching of process procedures and ensuring the targeted nature of process adaptation.

[0136] The process begins by traversing a pre-defined process knowledge graph based on process stage identifiers. This knowledge graph pre-stores the relationships between different process stages and their corresponding process procedures. By enumerating all procedure entries that are mapped to the process stage identifier, candidate process procedures are collected and integrated. This approach fully utilizes the accumulated pre-defined process knowledge, avoids unfounded process selection, and provides a rich set of alternatives that are well-suited to the current process stage for selecting the optimal procedure.

[0137] Based on preset selection criteria covering key dimensions such as process maturity, production efficiency, energy consumption control, and product quality stability, each candidate process procedure was comprehensively evaluated. The performance of each procedure across all dimensions was compared, and the procedure with the best overall performance was selected as the target process procedure. This process ensures that the selected process procedure is adaptable to the furnace operation and product production needs in multiple aspects, providing a scientific and reliable technological basis for subsequent temperature control.

[0138] The target process specification is structurally analyzed to identify the core temperature control parameters directly corresponding to the process stage identifiers. These parameters are the key temperature standards for ensuring the normal progress of material reactions and achieving product quality standards at this process stage, and are thus determined as the target temperature benchmark for furnace operation. This process provides a clear and precise core reference standard for temperature control, avoiding the limitations of traditional single fixed target temperatures and ensuring that temperature control meets the specific needs of the current process stage.

[0139] By deeply analyzing the descriptive information regarding the spatial distribution of temperature inside the furnace within the target process specification, and extracting key parameters such as the specified temperature ranges for different regions, temperature gradient requirements, and the location of high-temperature zones, this set of parameters is used as the distribution rule for furnace operation. This rule clarifies the temperature distribution requirements for each spatial location within the furnace, overcoming the deficiency of focusing only on a single target temperature value while ignoring the spatial gradient distribution. It provides a detailed basis for constructing a reasonable and uniform temperature field, ensuring the scientific and rational nature of the temperature distribution.

[0140] Based on the three-dimensional spatial structure of the furnace and the distribution of key material reaction areas, the target temperature benchmark is decomposed into main control nodes covering key locations such as the core reaction zone, edge transition zone, and insulation zone. The specific spatial coordinates of each main control node are clearly defined, and a corresponding temperature control standard value is assigned to each node, forming a set of main control nodes and the main control values ​​for the temperature field. This operation transforms the abstract temperature benchmark into concrete, implementable spatial temperature control targets, ensuring that the core control locations and standards of the temperature field are clear and unambiguous, providing a rigid execution basis for subsequent temperature field construction.

[0141] The constraints in the distribution rules, such as the temperature range, temperature gradient requirements, and spatial distribution ratios for different regions, are systematically arranged according to the three-dimensional dimensions of the furnace space domain. This transforms the scattered rule descriptions into a tensor form that comprehensively covers the furnace space and reflects the correlation of temperature constraints in each region, resulting in the temperature gradient constraint tensor. This tensor transforms the distribution rules from textual descriptions into structured constraints that can be directly used for optimization calculations, ensuring that the spatial distribution of the temperature field meets process requirements and avoiding localized temperature gradient anomalies.

[0142] Using the master control node set and the master temperature field value as hard constraints, the subsequent temperature field must strictly conform to the corresponding temperature standard at all master control node positions, allowing no deviation. The temperature gradient constraint tensor is used as a soft optimization objective, clarifying that the temperature field, while satisfying the hard constraints, should conform as closely as possible to the spatial temperature distribution. By integrating these two types of constraints and objectives, a variational optimization problem for the temperature field is constructed. This ensures both the accurate achievement of the core temperature control points and the rationality of the overall temperature distribution, providing a clear optimization direction for the scientific construction of the target temperature field.

[0143] A step-by-step iterative approach is adopted to solve the variational optimization problem of the temperature field. First, an initial temperature field that meets the temperature requirements of the master control node is constructed. Then, the initial temperature field is repeatedly adjusted and optimized based on the temperature gradient constraint tensor, continuously correcting the temperature values ​​of each region of the furnace. This ensures that the temperature field maintains the core node temperature as required while gradually approximating the spatial gradient distribution. The final target temperature field distribution not only conforms to the temperature benchmark and distribution rules specified in the process but also possesses spatial continuity and rationality, providing a scientific and reliable target basis for subsequent multi-objective dynamic optimization and precise temperature control.

[0144] The multi-objective dynamic optimization module 104 is used to perform multi-objective dynamic optimization on the three-dimensional thermo-temperature field, the target temperature field distribution and the thermodynamic-heat transfer constraints of the furnace, so as to obtain the global optimal dynamic transition trajectory of the target temperature field distribution.

[0145] In this embodiment of the invention, when the multi-objective dynamic optimization module 104 performs multi-objective dynamic optimization on the three-dimensional thermo-temperature field, the target temperature field distribution, and the thermodynamic-heat transfer constraints of the furnace to obtain the globally optimal dynamic transition trajectory of the target temperature field distribution, it is specifically used for:

[0146] Based on the three-dimensional thermo-temperature field and the target temperature field distribution, a multi-objective optimization problem is constructed with temperature field convergence, temperature distribution uniformity and process energy efficiency as the core indicators.

[0147] Based on the thermodynamic and heat transfer constraints of the furnace, the dynamic feasible region of the multi-objective optimization problem is established.

[0148] Based on a preset dynamic weight allocation strategy, the core indicators are synergistically weighed within the dynamic feasible domain to obtain a dynamically optimized target weight vector for the target temperature field distribution.

[0149] The dynamic optimization target weight vector is iteratively solved to obtain the set of transition trajectories of the target temperature field distribution;

[0150] Multi-objective optimization decision-making is performed on the trajectories in the set of transition trajectories, and the trajectory with the best comprehensive evaluation is taken as the global optimal dynamic transition trajectory of the target temperature field distribution.

[0151] By comparing the overall differences between the three-dimensional thermodynamic temperature field and the target temperature field distribution, the temperature field convergence is defined by the degree of fit between the two temperature values ​​in each spatial region, the temperature distribution uniformity is defined by the degree of consistency of the temperature deviation from the overall average temperature in the entire furnace area, and the process energy efficiency is defined by the correspondence between fuel consumption and output effect during the transition of the temperature field from the current state to the target state. These three factors that are directly related to the quality of furnace temperature control and operational efficiency are integrated into core evaluation indicators, and the optimization direction of each indicator is clarified as better temperature field convergence, more uniform temperature distribution, and higher process energy efficiency, thus constructing a multi-objective optimization problem.

[0152] A comprehensive review of the thermodynamic and heat transfer constraints that must be followed during furnace operation is conducted, including the maximum heat resistance limit that the furnace material can withstand, the maximum rate of heat transfer within the furnace, the allowable range of heat dissipation from the furnace body, and the efficiency boundary of fuel combustion into effective heat. These constraints are then transformed into explicit rules, defining the range of temperature values, the rate of temperature change, and the upper limit of energy input during the temperature field change process. This establishes a dynamic feasible domain for the multi-objective optimization problem, ensuring that all optimization explorations are conducted within the physical operating laws of the furnace and within the equipment's carrying capacity.

[0153] The system retrieves a preset dynamic weight allocation strategy. This strategy is formulated based on the process requirements of different operating stages of the furnace, the key nodes of the material reaction process, and the actual impact priority of each core indicator. Within the dynamic feasible domain, it analyzes the importance of temperature field convergence, temperature distribution uniformity, and process energy efficiency under the current operating conditions. It assigns a weight to each core indicator that is precisely matched with the current operating requirements, so that the weight allocation can be flexibly adjusted with changes in operating conditions, achieving a synergistic balance among the indicators, and finally forming a dynamically optimized target weight vector for the target temperature field distribution.

[0154] Guided by the dynamic optimization of the target weight vector, the iterative solution process is initiated. First, an initial transition trajectory that meets the basic optimization requirements is generated based on the initial weight vector. Then, based on the actual performance of this trajectory in the three core indicators of temperature field convergence, temperature distribution uniformity, and process energy efficiency, the weight values ​​of each indicator are adjusted according to the preset dynamic weight allocation strategy. A new transition trajectory is regenerated based on the adjusted weight vector. This process of adjusting weights and generating trajectories is repeated to continuously optimize the degree to which the trajectory meets the core indicators. All effective trajectories generated during the iteration process are collected to form a set of transition trajectories for the target temperature field distribution.

[0155] A multi-objective optimization decision-making standard was formulated, covering dimensions such as temperature field convergence, temperature distribution uniformity deviation, and actual process energy efficiency. Each trajectory in the transition trajectory set was comprehensively and meticulously evaluated, and the specific performance of each trajectory in each evaluation dimension was compared one by one. The balance and optimality of each trajectory in all core indicators were comprehensively considered, and the trajectory that achieved the comprehensive optimal level in terms of temperature field convergence, temperature distribution uniformity, and process energy efficiency was selected and determined as the globally optimal dynamic transition trajectory for the target temperature field distribution.

[0156] The beneficial effects are as follows: by comparing the differences between the three-dimensional thermodynamic temperature field and the target temperature field distribution, a temperature field convergence index is constructed based on the degree of similarity between the two temperature values; a temperature distribution uniformity index is constructed based on the degree of temperature deviation of each region of the furnace from the average temperature; and a process energy efficiency index is constructed based on the ratio of energy consumption to output during the temperature field transition. These are integrated to form a multi-objective optimization problem. This operation clarifies the core optimization direction of temperature control, breaks the limitations of single-index optimization, and ensures that the temperature field not only meets the target requirements but also possesses uniformity and high efficiency, providing a clear objective guide for subsequent optimization.

[0157] The thermodynamic and heat transfer constraints of the furnace are analyzed, including the heat resistance limit of the furnace material, the upper limit of the heat transfer rate, the allowable range of heat loss, and the boundary of fuel combustion thermal efficiency. These constraints are transformed into rules that limit the range and rate of temperature field change, establishing a dynamic feasible region for the multi-objective optimization problem. This feasible region defines the physical boundaries of the optimization process, ensuring that all optimization schemes conform to the objective laws of furnace operation, avoiding safety hazards or equipment damage caused by exceeding equipment or process limits, and guaranteeing the safety and feasibility of the optimization process.

[0158] A preset dynamic weight allocation strategy is invoked. This strategy is formulated based on the current operating status of the furnace, the stage of material reaction, and the changing trends in the importance of each core indicator. Within the dynamic feasible domain, the priority of the three core indicators under different operating conditions is analyzed, and each indicator is assigned a weight value that is appropriate for the current needs, forming a dynamic optimization target weight vector. This vector enables flexible and coordinated balancing of the indicators, and can dynamically adjust the optimization focus according to the actual operating conditions, avoiding optimization imbalances caused by fixed weights, and ensuring that the optimization results conform to real-time operating requirements.

[0159] Guided by the dynamic optimization of the target weight vector, an iterative approach of step-by-step adjustment is adopted to solve the problem. First, an initial transition trajectory is generated based on the initial weights. Then, the weight vector is adjusted according to the trajectory's performance on various core indicators, and a new trajectory is regenerated. This process is repeated iteratively to optimize and collect all valid trajectories, forming a set of transition trajectories. This process fully explores the optimization possibilities under different weight combinations, comprehensively covers potential optimal trajectory solutions, provides rich alternatives for subsequent decisions, and avoids the local optima problem that may result from solving a single trajectory.

[0160] An optimization decision-making standard was established, encompassing dimensions such as temperature field convergence rate, temperature distribution uniformity deviation, and process energy efficiency. Each trajectory in the transition trajectory set was comprehensively evaluated, with each trajectory's performance compared across different dimensions and a comprehensive score calculated. The trajectory with the best overall performance was selected as the globally optimal dynamic transition trajectory. This decision-making process ensured that the ultimately selected trajectory achieved optimal balance across all core indicators, guaranteeing a smooth transition of the temperature field to the target state while balancing temperature uniformity and operational energy efficiency. This provided a scientific and efficient trajectory basis for the subsequent generation of precise control commands.

[0161] The control command calculation and mapping module 105 is used to perform control command calculation and mapping on the globally optimal dynamic transition trajectory to obtain the precise energy input command set of the globally optimal dynamic transition trajectory.

[0162] In this embodiment of the invention, when the control command calculation and mapping module 105 performs control command calculation and mapping on the globally optimal dynamic transition trajectory to obtain the precise energy input command set of the globally optimal dynamic transition trajectory, it is specifically used for:

[0163] The global optimal dynamic transition trajectory is discretized in time and space to obtain the temperature field setpoint sequence of the global optimal dynamic transition trajectory.

[0164] Based on the spatial location and thermal influence range of the furnace, the temperature field setpoint sequence is mapped to the basic energy demand command of the furnace.

[0165] The basic energy demand command is adaptively compensated to obtain the compensated energy command for the furnace.

[0166] The compensated energy command is subjected to execution-level quantization to obtain the precise energy input command set.

[0167] The global optimal dynamic transition trajectory is discretized in time and space. The time dimension is divided according to the control cycle of the furnace operation. The continuous dynamic trajectory is split into multiple sequential time segments. At the same time, based on the spatial structural characteristics of the furnace, such as the furnace zoning and combustion zone distribution, the temperature control requirements of each spatial region in each time segment are defined. The continuous temperature change trajectory is transformed into a discretized data sequence arranged in time order and containing the clear temperature standards of each spatial region, thus obtaining the temperature field setpoint sequence of the global optimal dynamic transition trajectory.

[0168] Based on the specific location attributes of each spatial region of the furnace, as well as the range, intensity of influence and heat transfer path of heat diffusion after energy input in different regions, a correspondence between temperature setpoint and energy input is established. For the temperature requirements of each region in each time segment of the temperature field setpoint sequence, the energy supply form, supply location and basic supply amount required to reach the temperature are analyzed. The discrete temperature field setpoint sequence is transformed into a basic energy demand command that can drive the furnace temperature to change according to the trajectory requirements.

[0169] By combining the real-time operating conditions of the furnace and kiln, including dynamic influencing factors such as the current ambient temperature, the heat dissipation rate of the furnace surface, the initial temperature and moisture content of the material, and the airflow state in the furnace, we analyze the heat loss and temperature conduction delay that may occur during the actual execution of the basic energy demand command. Based on the specific circumstances of these real-time influencing factors, we dynamically adjust the energy supply in the basic energy demand command to supplement the energy that may be lost due to various losses, and obtain the compensated energy command for the furnace and kiln.

[0170] Based on the operating characteristics of the furnace energy supply actuator, including the burner's firepower adjustment range, the feeding device's conveying accuracy, and the energy output response speed, the energy supply quantity, supply duration, and supply frequency in the compensated energy command are quantified. The abstract energy demand in the command is transformed into specific values ​​that the actuator can accurately identify and operate, ensuring that the command meets the equipment's operating specifications and control accuracy requirements. Ultimately, a precise energy input command set with a clear structure, explicit values, and the ability to directly drive the actuator is formed.

[0171] The beneficial effect is that by dividing the globally optimal dynamic transition trajectory into continuous time segments at preset time intervals, and combining this with the three-dimensional spatial partitioning of the furnace, the temperature setpoints for each region within each time segment are clearly defined. This transforms the continuous dynamic trajectory into a discrete data sequence ordered by time and containing the temperature requirements of each region, resulting in a temperature field setpoint sequence. This operation transforms the abstract trajectory into concrete, implementable temperature control nodes, providing clear and accurate temporal and spatial dimensional references for subsequent energy demand calculations, and preventing a disconnect between control commands and trajectory requirements.

[0172] Based on the location characteristics of each spatial region of the furnace and the range and intensity of the thermal influence of energy input on the temperature field in different regions, a correspondence between temperature setpoints and energy inputs is established. According to the regional temperature requirements of each time segment in the temperature field setpoint sequence, the basic energy supply, supply location, and method required to meet that temperature are calculated, transforming the temperature setpoint sequence into basic energy demand instructions. This achieves precise conversion from temperature demand to energy supply, ensuring that energy input can directly match the temperature change requirements specified by the trajectory, laying the foundation for precise control.

[0173] By combining the real-time operating status of the furnace and kiln, including dynamic influencing factors such as current ambient temperature, furnace heat dissipation, and initial material temperature, this study analyzes the potential energy losses and temperature deviations that may occur during the actual execution of the basic energy demand command. Based on this real-time data, the supply of the basic energy demand command is dynamically adjusted and supplemented to make up for potential energy gaps, resulting in a compensated energy command. This process fully considers the dynamic changes in actual operating conditions, corrects the deviation between theoretical calculations and actual execution, and improves the adaptability and reliability of the energy command.

[0174] Referring to the execution accuracy and operating range of the furnace energy input equipment, the parameters such as supply quantity, frequency, and duration in the compensated energy commands are quantified, transforming abstract energy demands into specific operational values ​​that the equipment can precisely execute. This ultimately forms a precise energy input command set with a clear structure, well-defined values, and direct execution capability. This ensures that the commands comply with equipment operating specifications and control accuracy requirements, achieving seamless integration from optimized trajectory to equipment execution and guaranteeing accurate temperature control.

[0175] The rolling execution optimization module 106 is used to execute the precise energy input instruction set, assimilate and fuse the newly acquired real-time data to obtain the updated three-dimensional temperature field distribution of the furnace operation process, and based on the updated three-dimensional temperature field distribution, perform rolling time-domain reconstruction of the globally optimal dynamic transition trajectory to obtain the updated control strategy of the furnace operation process.

[0176] In this embodiment of the invention, when the rolling execution optimization module 106 executes the precise energy input instruction set and assimilates and fuses the newly acquired real-time data to obtain the updated three-dimensional temperature field distribution of the furnace operation process, it is specifically used for:

[0177] The precise energy input instruction set is input to the energy supply actuator of the furnace, and the status feedback data of the energy supply actuator is collected simultaneously.

[0178] Acquire the newly collected real-time sensing data of the furnace and kiln within the current control cycle, perform multi-source data fusion and cleaning on the real-time sensing data of the furnace and kiln, and obtain the effective real-time dataset of the real-time sensing data of the furnace and kiln.

[0179] The state feedback data, the effective real-time data, and the current state of the three-dimensional thermo-temperature field are used to perform multimodal state reconstruction to generate the assimilated furnace operation state of the furnace operation process.

[0180] Based on the assimilated furnace operating state, the temperature distribution of the furnace temperature field is extrapolated to obtain the updated three-dimensional temperature field distribution of the furnace operation process.

[0181] When the rolling execution optimization module 106 performs rolling time-domain reconstruction of the globally optimal dynamic transition trajectory based on the updated three-dimensional temperature field distribution to obtain the updated control strategy for the furnace operation process, it is specifically used for:

[0182] Based on the updated three-dimensional temperature field distribution and the target temperature field distribution, a deviation analysis is performed on the global optimal dynamic transition trajectory to obtain the state deviation information between the current temperature field and the target temperature field.

[0183] Based on the state deviation information, the unexecuted portion of the global optimal dynamic transition trajectory is corrected to obtain the corrected transition trajectory of the global optimal dynamic transition trajectory.

[0184] Based on the thermodynamic and heat transfer constraints of the furnace and the current operating conditions, the feasibility of the modified transition trajectory is verified and smoothed to obtain the optimized rolling transition trajectory.

[0185] Based on the optimized rolling transition trajectory, the control command sequence is regenerated to obtain the updated control strategy for the furnace operation process.

[0186] The precise energy input instruction set is adapted to the control protocol of the actuator to ensure that the instructions can be accurately identified and received by the energy supply actuator of the furnace. Then, the adapted instruction set is input to the actuator to drive the actuator to perform energy supply operation according to the instruction requirements. At the same time, through the status monitoring unit of the actuator, information such as the operating parameters of the mechanism, energy output accuracy, and equipment fault status are collected in real time to form status feedback data of the energy supply actuator, so as to realize the synchronization of instruction execution and status monitoring.

[0187] By deploying multi-source sensing devices such as temperature sensors, pressure sensors, and infrared detectors on the furnace, real-time sensing data of the furnace during the current control cycle is collected. This data covers key information such as temperature, operating pressure, and material status in various areas of the furnace. The collected real-time sensing data undergoes multi-source data fusion and cleaning, and the completeness and rationality of the data are checked one by one. Abnormal data that is missing key information or exceeds the normal range is eliminated. At the same time, similar data collected by different sensing devices are integrated to eliminate data redundancy and conflicts, resulting in an effective real-time dataset of furnace sensing data.

[0188] Based on the current state of the three-dimensional thermo-temperature field, the operating status information of the actuator in the state feedback data is integrated with the real-time operating data of the furnace in the effective real-time dataset for multimodal data integration. By analyzing the inherent correlation between different types of data, the state deviation reflected by a single data point is corrected, and a comprehensive state that can fully and accurately characterize the current operating status of the furnace is reconstructed. This generates the assimilated furnace operating state of the furnace operation process, ensuring that the judgment of the furnace operating state is more in line with the actual situation.

[0189] Based on the assimilated furnace and kiln operating status, combined with the furnace and kiln's thermodynamic characteristics, heat transfer laws, and the temperature influence mechanism of material reactions, the changing trend of the furnace and kiln temperature field is deduced. According to the current energy supply status, material reaction process, environmental influencing factors, etc., the direction and magnitude of temperature change in each area of ​​the furnace and kiln are predicted, and the temperature distribution in the entire furnace and kiln space is gradually calculated. The updated three-dimensional temperature field distribution of the furnace and kiln operation process is obtained, providing the latest temperature field data support for the optimization and adjustment of the next round of control commands.

[0190] By comparing the updated 3D temperature field distribution with the target temperature field distribution in various spatial regions, we can analyze the difference between the actual temperature and the target temperature in the same region, the distribution pattern of temperature deviation in different regions, and the time trend of temperature field changes. This will clarify the differences between the current temperature field and the target temperature field in terms of spatial distribution and dynamic changes, and comprehensively summarize the state deviation information that can accurately reflect the difference between the two.

[0191] Based on state deviation information, the system focuses on the unexecuted parts of the globally optimal dynamic transition trajectory. For areas and time periods with large deviations, the system adjusts the temperature setpoints and change rates of the trajectory to adapt the trajectory to the current temperature field deviation, compensate for existing state gaps, and maintain the overall continuity and rationality of the trajectory, thus obtaining the corrected transition trajectory of the globally optimal dynamic transition trajectory.

[0192] Referring to the thermodynamic and heat transfer constraints of the furnace, including the heat resistance limit of the furnace material, the limit of heat transfer rate, and the allowable range of heat loss, the temperature changes in the modified transition trajectory are checked one by one to see if they meet these constraints. Trajectory segments that exceed the constraint boundaries are eliminated. At the same time, a smoothing process is used to adjust the parts of the trajectory with abrupt temperature changes, so that the temperature changes in the trajectory are more gradual and continuous, avoiding the impact of sudden temperature changes on the furnace and materials. The optimized rolling transition trajectory after modification is obtained.

[0193] Based on the optimized rolling transition trajectory, and combined with the control characteristics and operating specifications of the furnace energy supply actuator, the temperature change requirements in the trajectory are transformed into corresponding energy input parameters, including energy supply amount, supply period, and supply location. A series of continuous control instructions are generated in chronological order to form a logically coherent and directly executable control instruction sequence, thereby obtaining the updated control strategy for the furnace operation process and providing precise instruction support for the next round of furnace temperature control.

[0194] The beneficial effects are that by adapting the precise energy input command set to the actuator control protocol format, the energy supply actuator can accurately identify and receive the commands, driving it to complete the energy supply operation according to the instructions. Simultaneously, the actuator status monitoring unit collects information such as operating parameters, energy output accuracy, and equipment fault status in real time, forming status feedback data. This process achieves synchronization between command execution and status monitoring, allowing for real-time monitoring of command implementation, timely detection of execution deviations or equipment anomalies, and providing crucial execution-side data support for subsequent data assimilation and temperature field updates.

[0195] By deploying multi-source sensing devices in the furnace, real-time sensing data such as temperature, pressure, and material status are collected within the current control cycle. The data undergoes multi-source fusion and cleaning, with each data point checked for completeness and rationality. Abnormal data lacking key information or exceeding normal value ranges are removed, and similar data is integrated to eliminate redundancy and conflicts, resulting in a valid real-time dataset. This process purifies the quality of real-time data, ensuring that the data accurately reflects the current operating conditions of the furnace and providing reliable sensing data for subsequent state reconstruction.

[0196] Based on the current state of the three-dimensional thermo-temperature field, this system integrates actuator operation information from state feedback data with operational data from effective real-time datasets. By analyzing the inherent correlations between different types of data, it corrects state biases from individual data points, reconstructing a comprehensive and accurate integrated state characterizing the current operation of the furnace, generating an assimilated furnace operating state. This state integrates data from both the execution and sensing sides, avoiding the limitations of a single data source and ensuring that the judgment of the furnace operating state is accurate and realistic.

[0197] Based on the assimilated furnace operating status, combined with the furnace's thermodynamic characteristics, heat transfer laws, and the influence mechanism of material reaction temperature, the temperature field change trend is deduced. According to the current energy supply status, material reaction process, and environmental influencing factors, the direction and magnitude of temperature changes in each region are predicted, and the overall temperature distribution is gradually calculated, resulting in an updated three-dimensional temperature field distribution. This temperature field reflects the latest temperature state of the furnace in real time, providing accurate and timely temperature data support for the next round of global optimal dynamic transition trajectory reconstruction and control strategy adjustment, ensuring the dynamic optimization of the control closed loop.

[0198] By comparing the updated 3D temperature field distribution with the target temperature field distribution in various spatial regions, the difference between the actual temperature and the target temperature in the same region is analyzed. The distribution pattern of temperature deviation in different regions is identified. Furthermore, by combining the temporal trend of temperature field changes, the core differences between the current temperature field and the target temperature field in terms of spatial distribution and dynamic changes are clarified. This comprehensive analysis of the resulting state deviation information accurately quantifies the gap between the current temperature field and the target requirements, providing a clear target basis for subsequent trajectory correction and avoiding blind adjustments.

[0199] Based on state deviation information, the system focuses on the unexecuted portions of the globally optimal dynamic transition trajectory. For areas and time periods with significant deviations, it adjusts the temperature setpoints and change rates corresponding to the trajectory, ensuring the trajectory accurately adapts to the current temperature field deviation and compensates for existing state discrepancies. Simultaneously, it maintains the overall continuity and logical rationality of the trajectory, resulting in a corrected transition trajectory that consistently matches the actual operating state of the furnace, preventing control failures caused by deviations between the initial trajectory and the actual situation.

[0200] Referring to the thermodynamic and heat transfer constraints of the furnace, including the heat resistance limit of the furnace material, the limit of heat transfer rate, and the allowable range of heat loss, the temperature changes in the modified transition trajectory were verified one by one to ensure that they met these constraints. Trajectory segments exceeding the constraint boundaries were eliminated to ensure that the trajectory remained within the allowable range of the equipment and process. Simultaneously, a smoothing process was employed to adjust the temperature abrupt changes in the trajectory, making the temperature changes more gradual and continuous. This avoided adverse effects on the furnace equipment and material reactions caused by sudden temperature changes, resulting in an optimized rolling transition trajectory that balanced feasibility and safety.

[0201] Based on the optimized rolling transition trajectory, and combined with the control characteristics and operating specifications of the furnace energy supply actuator, the temperature change requirements in the trajectory are transformed into corresponding energy input parameters, including energy supply amount, supply period, and supply location. A series of logically coherent and directly executable control command sequences are generated in chronological order, resulting in the updated control strategy. This strategy is a dynamic optimization of the initial control strategy, capable of accurately responding to changes and disturbances in the furnace's operating conditions, ensuring that the next round of temperature control is more targeted and effective, and guaranteeing that the furnace temperature consistently and smoothly approaches the target temperature field.

[0202] Reference Figure 2 The diagram shown is a flowchart illustrating an intelligent furnace temperature control method according to an embodiment of the present invention. In this embodiment, the intelligent furnace temperature control method includes:

[0203] S1. Spatial interpolation processing is performed on the multi-source heterogeneous sensing data during the operation of the furnace to obtain the three-dimensional thermo-temperature field of the furnace operation process.

[0204] S2. Based on historical fuel control commands, perform thermal process inversion on the three-dimensional thermo-temperature field to obtain the current material reaction process state of the three-dimensional thermo-temperature field;

[0205] S3. Map the current material reaction process state to a preset process knowledge graph to obtain the target temperature benchmark and distribution rules of the furnace operation process, and construct a spatial gradient field for the target temperature benchmark and the distribution rules to obtain the target temperature field distribution of the furnace operation process.

[0206] S4. Perform multi-objective dynamic optimization on the three-dimensional thermo-temperature field, the target temperature field distribution, and the thermodynamic-heat transfer constraints of the furnace to obtain the global optimal dynamic transition trajectory of the target temperature field distribution.

[0207] S5. Perform control command decomposition and mapping on the globally optimal dynamic transition trajectory to obtain the precise energy input command set of the globally optimal dynamic transition trajectory;

[0208] S6. Execute the precise energy input instruction set and assimilate and fuse the newly acquired real-time data to obtain the updated three-dimensional temperature field distribution of the furnace operation process. Based on the updated three-dimensional temperature field distribution, perform rolling time-domain reconstruction of the global optimal dynamic transition trajectory to obtain the updated control strategy of the furnace operation process.

[0209] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0210] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0211] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A furnace / kiln temperature intelligent control device, characterized in that, The device includes a multi-source sensing temperature field construction module, a thermal process inversion module, a target temperature field mapping generation module, a multi-objective dynamic optimization module, a control command solution mapping module, and a rolling execution optimization module, wherein: The multi-source sensing temperature field construction module is used to perform spatial interpolation processing on the multi-source heterogeneous sensing data during the operation of the furnace and kiln to obtain the three-dimensional thermo-temperature field of the furnace and kiln operation process. The thermal process inversion module is used to perform thermal process inversion on the three-dimensional thermo-temperature field based on historical fuel control commands, so as to obtain the current material reaction process state of the three-dimensional thermo-temperature field. The target temperature field mapping generation module is used to map the current material reaction process state to a preset process knowledge graph to obtain the target temperature benchmark and distribution rules of the furnace operation process, and to construct a spatial gradient field on the target temperature benchmark and the distribution rules to obtain the target temperature field distribution of the furnace operation process. The multi-objective dynamic optimization module is used to perform multi-objective dynamic optimization on the three-dimensional thermo-temperature field, the target temperature field distribution, and the thermodynamic-heat transfer constraints of the furnace, so as to obtain the global optimal dynamic transition trajectory of the target temperature field distribution. The control command solution and mapping module is used to perform control command solution and mapping on the globally optimal dynamic transition trajectory to obtain the precise energy input command set of the globally optimal dynamic transition trajectory. The rolling execution optimization module is used to execute the precise energy input instruction set, assimilate and fuse the newly acquired real-time data to obtain the updated three-dimensional temperature field distribution of the furnace operation process, and based on the updated three-dimensional temperature field distribution, perform rolling time-domain reconstruction of the globally optimal dynamic transition trajectory to obtain the updated control strategy of the furnace operation process.

2. The intelligent furnace temperature control device as described in claim 1, characterized in that, The multi-source sensing temperature field construction module, when performing spatial interpolation processing on the multi-source heterogeneous sensing data during furnace operation to obtain the three-dimensional thermo-temperature field of the furnace operation process, is specifically used for: Multi-source heterogeneous sensing data is collected during the operation of the furnace and kiln, and the multi-source heterogeneous sensing data is spatiotemporally aligned and cleaned to obtain a regularized sensing dataset of the furnace and kiln operation process. Based on the geometric positions of the data points in the regularized sensing dataset within the three-dimensional space of the furnace, construct the spatial adjacency topology of the data points; Based on the spatial adjacency topology, spatial interpolation is performed on the temperature data in the regularized sensing dataset to obtain the preliminary temperature field of the furnace operation process. The formula for calculating the desired temperature value in the preliminary temperature field is as follows: ； In the formula, The internal coordinates of the furnace are The desired temperature value at the desired spatial location. For the first The target temperature value for each core temperature control point. This represents the spatial influence coefficient of the temperature field distribution. To find the spatial location to the first Spatial Euclidean distance between core temperature control points It is an exponential function; The smoothness of the preliminary temperature field is reasonably verified, and the abnormal regions in the preliminary temperature field are corrected to obtain the three-dimensional thermo-temperature field of the furnace operation process.

3. The intelligent furnace temperature control device as described in claim 1, characterized in that, When the thermal process inversion module executes historical fuel control commands to perform thermal process inversion on the three-dimensional thermo-temperature field and obtains the current material reaction process state of the three-dimensional thermo-temperature field, it is specifically used for: By performing feature analysis on historical fuel control commands, a historical control feature sequence of the furnace operation process is obtained; Dynamic feature analysis is performed on the three-dimensional thermo-temperature field to obtain the current dynamic characteristics of the thermal field during the operation of the furnace. The historical control feature sequence is matched with the current thermal field dynamic features in a time sequence, and the historical operating conditions most similar to the current thermal state and the corresponding stable material reaction stage are identified. Based on the real-time change direction and amplitude of the current thermal field dynamic characteristics, a comprehensive reasoning and state determination are performed on the stable material reaction stage to obtain the current material reaction process state of the three-dimensional thermo-temperature field.

4. The intelligent furnace temperature control device as described in claim 1, characterized in that, When the target temperature field mapping generation module maps the current material reaction process state to a preset process knowledge graph to obtain the target temperature benchmark and distribution rules of the furnace operation process, it is specifically used for: Multimodal feature decoding is performed on the current material reaction process state to obtain the material type identifier and process stage identifier of the current material reaction process state; Based on the process stage identifier, the preset process knowledge graph is traversed to obtain candidate process procedures associated with the process stage identifier. According to the preset selection criteria, the target process procedure is selected from the candidate process procedures; Extract the core target temperature value corresponding to the process stage identifier in the target process specification, and use the core target temperature value as the target temperature benchmark for the furnace operation process; The parameter set describing the spatial distribution of temperature inside the furnace in the target process specification is used as the distribution rule for the furnace operation process.

5. The intelligent furnace temperature control device as described in claim 1, characterized in that, When the target temperature field mapping generation module performs spatial gradient field construction on the target temperature benchmark and the distribution rules to obtain the target temperature field distribution during the furnace operation process, it is specifically used for: The target temperature reference is decomposed into a structured form to obtain the set of master control nodes of the temperature field inside the furnace and the corresponding master control values ​​of the temperature field. The distribution rules are tensorized to obtain the temperature gradient constraint tensor in the furnace space domain; Using the master control node set and the master temperature field value as hard constraints, and the temperature gradient constraint tensor as a soft optimization objective, a variational optimization problem of the temperature field during the furnace operation process is constructed. Solve the variational optimization problem of the temperature field to obtain the target temperature field distribution during the operation of the furnace.

6. The intelligent furnace temperature control device as described in claim 1, characterized in that, When the multi-objective dynamic optimization module performs multi-objective dynamic optimization on the three-dimensional thermo-temperature field, the target temperature field distribution, and the thermodynamic-heat transfer constraints of the furnace, and obtains the globally optimal dynamic transition trajectory of the target temperature field distribution, it is specifically used for: Based on the three-dimensional thermo-temperature field and the target temperature field distribution, a multi-objective optimization problem is constructed with temperature field convergence, temperature distribution uniformity and process energy efficiency as the core indicators. Based on the thermodynamic and heat transfer constraints of the furnace, the dynamic feasible region of the multi-objective optimization problem is established. Based on a preset dynamic weight allocation strategy, the core indicators are synergistically weighed within the dynamic feasible domain to obtain a dynamically optimized target weight vector for the target temperature field distribution. The dynamic optimization target weight vector is iteratively solved to obtain the set of transition trajectories of the target temperature field distribution; Multi-objective optimization decision-making is performed on the trajectories in the set of transition trajectories, and the trajectory with the best comprehensive evaluation is taken as the global optimal dynamic transition trajectory of the target temperature field distribution.

7. The intelligent furnace temperature control device as described in claim 1, characterized in that, When the control command solution mapping module performs control command solution mapping on the globally optimal dynamic transition trajectory to obtain the precise energy input command set of the globally optimal dynamic transition trajectory, it is specifically used for: The global optimal dynamic transition trajectory is discretized in time and space to obtain the temperature field setpoint sequence of the global optimal dynamic transition trajectory. Based on the spatial location and thermal influence range of the furnace, the temperature field setpoint sequence is mapped to the basic energy demand command of the furnace. The basic energy demand command is adaptively compensated to obtain the compensated energy command for the furnace. The compensated energy command is subjected to execution-level quantization to obtain the precise energy input command set.

8. The intelligent furnace temperature control device as described in claim 1, characterized in that, When the rolling execution optimization module executes the precise energy input instruction set and assimilates and fuses the newly acquired real-time data to obtain the updated three-dimensional temperature field distribution of the furnace operation process, it is specifically used for: The precise energy input instruction set is input to the energy supply actuator of the furnace, and the status feedback data of the energy supply actuator is collected simultaneously. Acquire the newly collected real-time sensing data of the furnace and kiln within the current control cycle, perform multi-source data fusion and cleaning on the real-time sensing data of the furnace and kiln, and obtain the effective real-time dataset of the real-time sensing data of the furnace and kiln. The state feedback data, the effective real-time data, and the current state of the three-dimensional thermo-temperature field are reconstructed in a multi-modal manner to generate the assimilated furnace operation state of the furnace operation process. Based on the assimilated furnace operating state, the temperature distribution of the furnace temperature field is extrapolated to obtain the updated three-dimensional temperature field distribution of the furnace operation process.

9. The intelligent furnace temperature control device as described in claim 1, characterized in that, When the rolling execution optimization module performs rolling time-domain reconstruction of the globally optimal dynamic transition trajectory based on the updated three-dimensional temperature field distribution to obtain the updated control strategy for the furnace operation process, it is specifically used for: Based on the updated three-dimensional temperature field distribution and the target temperature field distribution, a deviation analysis is performed on the global optimal dynamic transition trajectory to obtain the state deviation information between the current temperature field and the target temperature field. Based on the state deviation information, the unexecuted portion of the global optimal dynamic transition trajectory is corrected to obtain the corrected transition trajectory of the global optimal dynamic transition trajectory. Based on the thermodynamic and heat transfer constraints of the furnace and the current operating conditions, the feasibility of the modified transition trajectory is verified and smoothed to obtain the optimized rolling transition trajectory. Based on the optimized rolling transition trajectory, the control command sequence is regenerated to obtain the updated control strategy for the furnace operation process.

10. A method for intelligent control of furnace temperature, characterized in that, The method for using the intelligent furnace temperature control device according to claim 1: S1. Spatial interpolation processing is performed on the multi-source heterogeneous sensing data during the operation of the furnace to obtain the three-dimensional thermo-temperature field of the furnace operation process. S2. Based on historical fuel control commands, perform thermal process inversion on the three-dimensional thermo-temperature field to obtain the current material reaction process state of the three-dimensional thermo-temperature field; S3. Map the current material reaction process state to a preset process knowledge graph to obtain the target temperature benchmark and distribution rules of the furnace operation process, and construct a spatial gradient field for the target temperature benchmark and the distribution rules to obtain the target temperature field distribution of the furnace operation process. S4. Perform multi-objective dynamic optimization on the three-dimensional thermo-temperature field, the target temperature field distribution, and the thermodynamic-heat transfer constraints of the furnace to obtain the global optimal dynamic transition trajectory of the target temperature field distribution. S5. Perform control command decomposition and mapping on the globally optimal dynamic transition trajectory to obtain the precise energy input command set of the globally optimal dynamic transition trajectory; S6. Execute the precise energy input instruction set and assimilate and fuse the newly acquired real-time data to obtain the updated three-dimensional temperature field distribution of the furnace operation process. Based on the updated three-dimensional temperature field distribution, perform rolling time-domain reconstruction of the global optimal dynamic transition trajectory to obtain the updated control strategy of the furnace operation process.

Images