Power station boiler heating surface global temperature prediction method and device based on multi-source coupling

Abstract

The present application relates to the technical field of boiler safety intelligent operation and thermal-hydraulic monitoring, and particularly relates to a power plant boiler heating surface global temperature prediction method and device based on multi-source coupling, wherein the method comprises: obtaining at least one of the geometric structure, coal quality, operation parameters of the distributed control system, wall temperature measuring points and soot blowing records of the target power plant boiler; establishing a three-dimensional combustion heat exchange model of the flue gas side, a thermal-hydraulic model of the steam-water side, a key position measuring point data and a field operation data calibration model to obtain multi-source coupling calculation results; constructing a mixed sample library, extracting at least one flue gas side feature, at least one steam-water side feature, at least one operation side feature and at least one historical temperature feature, training a power plant boiler heating surface global temperature prediction model, and outputting the heating surface global tube wall temperature, the working medium temperature, the hot spot position and the over-temperature probability. Thus, the problems of related art, such as local hot spot misjudgment, insufficient generalization ability under all working conditions and difficulty in real-time deployment, are solved.

Description

Technical Field

[0001] This invention relates to the field of boiler safe and intelligent operation and thermal-hydraulic monitoring technology, and in particular to a method and device for predicting the temperature of the entire heating surface of a power plant boiler based on multi-source coupling. Background Technology

[0002] Currently, under conditions of deep peak shaving, wide load operation, frequent coal type switching, and long cycle operation, large coal-fired power plant boilers exhibit significant spatial non-uniformity and dynamic changes in combustion and heat exchange of heating surfaces. Local thermal deviations in heating surfaces such as water-cooled walls, superheaters, reheaters, and economizers may lead to long-term overheating of tube walls, oxide scale shedding, tube rupture, and unplanned shutdowns. Therefore, global temperature prediction and local hot spot identification have become key technologies for the safe and intelligent operation of boilers.

[0003] Among the related technologies, four main technical solutions are used for global temperature prediction and local hotspot identification: pure data-driven prediction, single three-dimensional combustion heat transfer simulation, single thermal-hydraulic simulation, and field measurement at key locations. Among them, the pure data-driven model relies on historical data of DCS (Distributed Control System) for prediction; the single three-dimensional combustion heat transfer simulation describes the three-dimensional flow field and non-uniform heat load on the flue gas side by constructing a three-dimensional model; the single thermal-hydraulic simulation obtains flow rate, pressure, and temperature parameters by solving the steam-water system equations; and the field measurement at key locations mainly uses detection equipment such as sonar and thermal analyzers to conduct on-site detection and analysis of temperature and related parameters at key locations.

[0004] However, in related technologies, the lack of physical constraints such as in-furnace combustion, radiative heat transfer, ash and fouling thermal resistance, and steam-water flow distribution in pure data-driven models makes them prone to inaccuracies when faced with extrapolated operating conditions such as sudden changes in coal quality or deep peak shaving. Single three-dimensional combustion heat transfer simulations are difficult to realistically reflect the parallel pipe flow distribution, header pressure drop, and local working fluid cooling capacity in complex steam-water systems. Because single thermal-hydraulic simulations simplify the flue gas side boundary to the average heat load, it is difficult to identify local high heat loads caused by furnace off-center burning, slagging, ash accumulation, or residual rotation, resulting in missed local hot spots, insufficient generalization ability across all operating conditions, and difficulties in real-time deployment, which urgently need to be improved. Summary of the Invention

[0005] This invention provides a method and device for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling. This addresses the problems in related technologies, such as the lack of physical constraints in purely data-driven models, which makes them prone to inaccuracy when facing extrapolation conditions; the inability of single three-dimensional combustion heat transfer simulation to reflect the complex characteristics of the steam-water system; and the simplification of the flue gas side boundary to the average heat load in single thermal-hydraulic simulation, which makes it impossible to identify local high heat loads, resulting in missed local hot spots, insufficient generalization ability for all operating conditions, and difficulties in real-time deployment.

[0006] A first aspect of this invention provides a method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling, comprising the following steps: acquiring at least one of the following of the target power plant boiler: geometric structure, heating surface arrangement, burner arrangement, steam-water pipeline network, coal quality, operating parameters of the distributed control system, wall temperature measuring points, soot blowing records, and maintenance and scaling information; based on the at least one of these, establishing a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, key location measuring point data, and a calibration model based on field operating data, and exchanging non-uniform heat load of the heating surface between the various models and the measured data through a multi-source coupling interface. The load, working fluid flow rate, working fluid temperature, heat transfer boundary parameters, and measurement point calibration deviation are used to obtain multi-source coupling calculation results after the multi-source coupling convergence condition is met. A hybrid sample library is constructed using the multi-source coupling calculation results and field measured data to extract at least one flue gas side feature, at least one steam-water side feature, at least one operating side feature, and at least one historical temperature feature. A global temperature prediction model for the heating surface of a power plant boiler, including a spatial feature extraction module, a temporal feature modeling module, and a hotspot attention module, is trained to output the global pipe wall temperature, working fluid temperature, hotspot location, and over-temperature probability of the heating surface.

[0007] Through the aforementioned technical means, the embodiments of the present invention can acquire information such as the boiler's geometric structure, coal quality, parameters of the distributed control system, and wall temperature measurement points to establish a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, and a data calibration model. Parameters such as non-uniform heat load, flow rate, and temperature are exchanged through a multi-source coupling interface. After coupling convergence, multi-source coupling calculation results are obtained. A hybrid sample library is constructed using the multi-source coupling calculation results and field measured data to extract flue gas side, steam-water side, operating side, and historical temperature features. Training is then performed to include spatial feature extraction, temporal feature modeling, and hotspot attention modeling. The block's global temperature prediction model outputs pipe wall temperature, working fluid temperature, hot spot location, and over-temperature probability. Through multi-source collaboration among combustion heat transfer mechanism model, steam-water thermal-hydraulic model, key monitoring points, field operation data, and machine learning model, it achieves closed-loop correction of heat load, working fluid cooling capacity, measured deviation, and temperature prediction results. It predicts the global temperature distribution and local hot spots online, thereby improving the accuracy of local hot spot identification and the prediction accuracy and generalization ability under complex operating conditions. This provides a basis for over-temperature early warning, intelligent soot blowing, and combustion adjustment, improving the unit's operational safety, flexibility, and economy.

[0008] Optionally, in one embodiment of the present invention, the output of the global wall temperature of the heating surface, the working fluid temperature, the hot spot location, and the over-temperature probability includes: deploying the global temperature prediction model of the heating surface of the power plant boiler as a real-time proxy model, connecting it to the online monitoring system of the target power plant boiler, so as to output the global wall temperature of the heating surface, the working fluid temperature, the hot spot location, and the over-temperature probability.

[0009] Through the above-mentioned technical means, the embodiments of the present invention can deploy the global temperature prediction model as a real-time proxy model and connect it to the power plant boiler online monitoring system, outputting the global temperature distribution and hotspot information of the heating surface in real time, enabling operators to obtain the dynamic of the full-screen tube wall temperature and the risk of overheating in real time, meeting the timeliness requirements of real-time operation monitoring of power plant boilers and improving response speed.

[0010] Optionally, in one embodiment of the present invention, the method further includes: matching an over-temperature level based on at least one of the global pipe wall temperature of the heated surface, the working fluid temperature, the hot spot location, and the over-temperature probability; generating over-temperature graded early warning information based on the over-temperature level to provide an over-temperature warning; and / or matching corresponding operation control suggestions based on the over-temperature level and prompting the user with the operation control suggestions.

[0011] Through the above-mentioned technical means, the embodiments of the present invention can automatically match the over-temperature level according to the prediction results and generate graded early warning information and operation control suggestions, thereby realizing graded warning and proactive intervention of boiler heating surface over-temperature risk, reducing the risk of over-temperature tube rupture and unplanned shutdown, and improving the safety, flexibility and economy of unit operation.

[0012] Optionally, in one embodiment of the present invention, the step of establishing a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, key location measurement point data, and field operation data calibration model based on at least one of the above, and exchanging non-uniform heat load of the heating surface, working fluid flow rate in the pipe, working fluid temperature, heat transfer boundary parameters, and measurement point calibration deviations between the models and measured data through a multi-source coupling interface to obtain multi-source coupling calculation results, includes: using the heat load mapping module of the multi-source coupling interface to map the three-dimensional flue gas side simulation mesh or surface elements. Radiative heat flux, convective heat flux, and total heat load are mapped to the heat transfer boundary on the steam-water side according to the heating surface tube panel, tube segment, or single tube number. Using the flow and temperature feedback module of the multi-source coupling interface, the mass flow rate, inlet enthalpy, outlet enthalpy, pressure loss, and heat transfer coefficient on the inner side of the tube calculated on the one-dimensional steam-water side are fed back to the three-dimensional combustion heat transfer model on the flue gas side. Using the measured deviation calibration module of the multi-source coupling interface, the model boundary and prediction results are corrected according to the wall temperature measurement point, the operating parameters of the distributed control system, and the coal quality.

[0013] Through the above-mentioned technical means, the embodiments of the present invention can map the non-uniform heat load output by the three-dimensional combustion heat transfer model into the heat input of the steam-water side pipe section through the multi-source coupling interface, and feed back the cooling capacity inside the pipe calculated by the steam-water thermal hydraulic model to the wall boundary of the combustion heat transfer model. At the same time, the model boundary conditions, key parameters and prediction deviations are dynamically corrected by using field measured data, thereby realizing the coordinated coupling of the flue gas side, steam-water side and operating data, forming a closed-loop optimization mechanism that deeply integrates physical mechanism and measured data, and improving the accuracy and reliability of multi-source coupling calculation.

[0014] Optionally, in one embodiment of the present invention, the multi-source coupling convergence condition includes: the relative deviation of the total heat absorption of the heating surface in two adjacent iterations is less than or equal to a preset percentage; the temperature deviation between the working fluid at the outlet of the key tube screen and the temperature of the key tube wall is less than or equal to a preset threshold; the calibration residual of the field measuring point is less than a preset difference; and the location of the local high heat load area reaches a preset condition that no significant drift occurs.

[0015] Through the above-mentioned technical means, the embodiments of the present invention can judge the convergence of multi-source coupling based on multiple dimensions such as the relative deviation of the total heat absorption of the heating surface, the temperature deviation between the working medium and the tube wall of the key tube screen, the calibration residual of the field measuring point, and the location stability of the local high heat load area. This establishes a comprehensive and strict convergence judgment standard, which can ensure that the multi-source coupling calculation meets the engineering application requirements in terms of energy conservation, key parameter accuracy, measured data fitting degree, and local feature stability. This accurately reflects the combustion and heat exchange characteristics of the boiler under actual operating conditions and ensures the engineering credibility of the overall temperature prediction results.

[0016] A second aspect of the present invention provides a device for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling, comprising: an acquisition module for acquiring at least one of the following: the geometric structure of the target power plant boiler, the arrangement of the heating surface, the arrangement of the burners, the steam-water pipeline network, the coal quality, the operating parameters of the distributed control system, wall temperature measuring points, soot blowing records, and maintenance and scaling information; and a modeling module for establishing a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, key location measuring point data, and a calibration model based on the at least one of the above, and exchanging non-uniform heat transfer data of the heating surface between the models and the measured data through a multi-source coupling interface. The load, working fluid flow rate, working fluid temperature, heat transfer boundary parameters, and measurement point calibration deviation are used to obtain the multi-source coupling calculation results after the multi-source coupling convergence condition is met. The prediction module is used to construct a hybrid sample library using the multi-source coupling calculation results and field measured data to extract at least one flue gas side feature, at least one steam-water side feature, at least one operating side feature, and at least one historical temperature feature. The model is then used to train a global temperature prediction model for the heating surface of a power plant boiler, which includes a spatial feature extraction module, a temporal feature modeling module, and a hotspot attention module, to output the global pipe wall temperature, working fluid temperature, hotspot location, and over-temperature probability of the heating surface.

[0017] Through the aforementioned technical means, the embodiments of the present invention can acquire information such as the boiler's geometric structure, coal quality, parameters of the distributed control system, and wall temperature measurement points to establish a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, and a data calibration model. Parameters such as non-uniform heat load, flow rate, and temperature are exchanged through a multi-source coupling interface. After coupling convergence, multi-source coupling calculation results are obtained. A hybrid sample library is constructed using the multi-source coupling calculation results and field measured data to extract flue gas side, steam-water side, operating side, and historical temperature features. Training is then performed to include spatial feature extraction, temporal feature modeling, and hotspot attention modeling. The block's global temperature prediction model outputs pipe wall temperature, working fluid temperature, hot spot location, and over-temperature probability. Through multi-source collaboration among combustion heat transfer mechanism model, steam-water thermal-hydraulic model, key monitoring points, field operation data, and machine learning model, it achieves closed-loop correction of heat load, working fluid cooling capacity, measured deviation, and temperature prediction results. It predicts the global temperature distribution and local hot spots online, thereby improving the accuracy of local hot spot identification and the prediction accuracy and generalization ability under complex operating conditions. This provides a basis for over-temperature early warning, intelligent soot blowing, and combustion adjustment, improving the unit's operational safety, flexibility, and economy.

[0018] Optionally, in one embodiment of the present invention, the prediction module includes: an output unit, configured to deploy the global temperature prediction model of the heating surface of the power plant boiler as a real-time proxy model, and connect it to the online monitoring system of the target power plant boiler to output the global tube wall temperature of the heating surface, the working fluid temperature, the hot spot location, and the over-temperature probability.

[0019] Through the above-mentioned technical means, the embodiments of the present invention can deploy the global temperature prediction model as a real-time proxy model and connect it to the power plant boiler online monitoring system, outputting the global temperature distribution and hotspot information of the heating surface in real time, enabling operators to obtain the dynamic of the full-screen tube wall temperature and the risk of overheating in real time, meeting the timeliness requirements of real-time operation monitoring of power plant boilers and improving response speed.

[0020] Optionally, in one embodiment of the present invention, it further includes: a matching module, configured to match an over-temperature level based on at least one of the global pipe wall temperature of the heated surface, the working fluid temperature, the hot spot location, and the over-temperature probability; a generation module, configured to generate over-temperature graded early warning information based on the over-temperature level for over-temperature warning; and a prompting module, configured to match corresponding operation control suggestions based on the over-temperature level and prompt the user with the operation control suggestions.

[0021] Through the above-mentioned technical means, the embodiments of the present invention can automatically match the over-temperature level according to the prediction results and generate graded early warning information and operation control suggestions, thereby realizing graded warning and proactive intervention of boiler heating surface over-temperature risk, reducing the risk of over-temperature tube rupture and unplanned shutdown, and improving the safety, flexibility and economy of unit operation.

[0022] Optionally, in one embodiment of the present invention, the establishment module includes: a mapping unit, used to map the radiative heat flow, convective heat flow, and total heat load on the three-dimensional flue gas side simulation mesh or surface element to the steam-water side heat transfer boundary according to the heating surface tube screen, tube segment, or single tube number using the heat load mapping module of the multi-source coupling interface; a feedback unit, used to feed back the in-tube mass flow rate, inlet enthalpy, outlet enthalpy, pressure loss, and inner wall heat transfer coefficient calculated on the one-dimensional steam-water side to the three-dimensional combustion heat transfer model on the flue gas side using the flow rate and temperature feedback module of the multi-source coupling interface; and a correction unit, used to correct the model boundary and prediction results based on the wall temperature measurement point, the operating parameters of the distributed control system, and the coal quality using the measured deviation calibration module of the multi-source coupling interface.

[0023] Through the above-mentioned technical means, the embodiments of the present invention can map the non-uniform heat load output by the three-dimensional combustion heat transfer model into the heat input of the steam-water side pipe section through the multi-source coupling interface, and feed back the cooling capacity inside the pipe calculated by the steam-water thermal hydraulic model to the wall boundary of the combustion heat transfer model. At the same time, the model boundary conditions, key parameters and prediction deviations are dynamically corrected by using field measured data, thereby realizing the coordinated coupling of the flue gas side, steam-water side and operating data, forming a closed-loop optimization mechanism that deeply integrates physical mechanism and measured data, and improving the accuracy and reliability of multi-source coupling calculation.

[0024] A third aspect of the present invention provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling as described in the above embodiments.

[0025] A fourth aspect of the present invention provides a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for predicting the global temperature of a power plant boiler heating surface based on multi-source coupling.

[0026] A fifth aspect of the present invention provides a computer program product that stores a computer program that, when executed by a processor, implements the above-described method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling.

[0027] This invention can acquire information such as boiler geometry, coal quality, parameters of the distributed control system, and wall temperature measurement points to establish a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, and a data calibration model. It also exchanges parameters such as non-uniform heat load, flow rate, and temperature through a multi-source coupling interface. After coupling convergence, multi-source coupling calculation results are obtained. A hybrid sample library is constructed using the multi-source coupling calculation results and field measured data. Features from the flue gas side, steam-water side, operating side, and historical temperature are extracted. A global domain model including spatial feature extraction, temporal feature modeling, and hotspot attention modules is trained. The temperature prediction model outputs pipe wall temperature, working fluid temperature, hot spot location, and over-temperature probability. Through multi-source collaboration among the combustion heat transfer mechanism model, steam-water thermal-hydraulic model, key monitoring points, field operation data, and machine learning model, it achieves closed-loop correction of heat load, working fluid cooling capacity, measured deviations, and temperature prediction results. It predicts the overall temperature distribution and local hot spots online, thereby improving the accuracy of local hot spot identification and the prediction accuracy and generalization ability under complex operating conditions. This provides a basis for over-temperature early warning, intelligent soot blowing, and combustion adjustment, enhancing the unit's operational safety, flexibility, and economy. This solves problems in related technologies, such as the lack of physical constraints in purely data-driven models leading to inaccuracies when extrapolating to operating conditions; the inability of single three-dimensional combustion heat transfer simulations to reflect the complex characteristics of steam-water systems; and the simplification of the flue gas side boundary to average heat load in single thermal-hydraulic simulations, which fails to identify local high heat loads, resulting in missed local hot spots, insufficient generalization ability across all operating conditions, and difficulties in real-time deployment.

[0028] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0029] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 A flowchart illustrating a method for predicting the global temperature of a power plant boiler heating surface based on multi-source coupling, according to an embodiment of the present invention. Figure 2 A roadmap for boiler heating surface temperature sensing, prediction, and optimization control technology according to an embodiment of the present invention; Figure 3 This is a schematic diagram of data flow in multi-source coupled computing according to an embodiment of the present invention; Figure 4 A block diagram of a multi-feature fusion machine learning model structure provided according to an embodiment of the present invention; Figure 5 A flowchart illustrating the real-time prediction and over-temperature early warning process according to an embodiment of the present invention; Figure 6This is a schematic diagram of a power plant boiler heating surface global temperature prediction device based on multi-source coupling according to an embodiment of the present invention. Figure 7 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present invention.

[0030] Figure label: 10-A device for predicting the temperature of the heating surface of a power plant boiler based on multi-source coupling; 100-Acquisition module, 200-Establishment module, 300-Prediction module; 701-Memory, 702-Processor, 703-Communication interface. Detailed Implementation

[0031] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0032] The following describes, with reference to the accompanying drawings, a method and apparatus for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling, according to embodiments of the present invention. Addressing the issues raised in the background section regarding related technologies, such as the lack of physical constraints in purely data-driven models leading to inaccuracies when extrapolating operating conditions, the inability of single three-dimensional combustion heat transfer simulations to reflect the complex characteristics of the steam-water system, and the simplification of the flue gas side boundary to average heat load in single thermal-hydraulic simulations, which fail to identify local high heat loads, resulting in missed local hotspots, insufficient generalization ability across all operating conditions, and difficulties in real-time deployment, the present invention provides a method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling. This method acquires information such as the boiler's geometry, coal quality, parameters of the distributed control system, and wall temperature measurement points to establish a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, and a data calibration model. Parameters such as non-uniform heat load, flow rate, and temperature are exchanged through a multi-source coupling interface, achieving coupling convergence and yielding... The multi-source coupled calculation results are used to construct a hybrid sample library with field measured data. Temperature features from the flue gas side, steam-water side, operation side, and historical data are extracted. A global temperature prediction model, including spatial feature extraction, temporal feature modeling, and hotspot attention modules, is trained. The model outputs pipe wall temperature, working fluid temperature, hotspot location, and over-temperature probability. Through multi-source collaboration among the combustion heat transfer mechanism model, steam-water thermal-hydraulic model, key monitoring points, field operation data, and machine learning model, closed-loop correction of heat load, working fluid cooling capacity, measured deviation, and temperature prediction results is achieved. The model predicts the global temperature distribution and local hotspots online, thereby improving the accuracy of local hotspot identification and the prediction accuracy and generalization ability under complex operating conditions. This provides a basis for over-temperature early warning, intelligent soot blowing, and combustion adjustment, improving the safety, flexibility, and economy of unit operation. This solves the problems in related technologies, such as the lack of physical constraints in pure data-driven models, which makes them prone to inaccuracy when facing extrapolated operating conditions; the inability of single three-dimensional combustion heat transfer simulation to reflect the complex characteristics of the steam-water system; and the simplification of the flue gas side boundary to the average heat load by single thermal-hydraulic simulation, which makes it impossible to identify local high heat loads, resulting in missed local hot spots, insufficient generalization ability for all operating conditions, and difficulties in real-time deployment.

[0033] Specifically, Figure 1 This is a flowchart illustrating a method for predicting the global temperature of a power plant boiler heating surface based on multi-source coupling, as provided in an embodiment of the present invention.

[0034] like Figure 1 As shown, the method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling includes the following steps: In step S101, at least one of the following is obtained: the geometry of the target power plant boiler, the arrangement of the heating surface, the arrangement of the burners, the steam and water pipeline network, the coal quality, the operating parameters of the distributed control system, the wall temperature measuring points, the soot blowing records, and the maintenance and scaling information.

[0035] It is understood that the geometric structure in the embodiments of the present invention may include the three-dimensional spatial dimensions and configuration of the boiler furnace and flue; the arrangement of the heating surfaces may include the spatial positions of the tube panels of the water-cooled walls, superheaters, reheaters, and economizers; the arrangement of the burners may include the type, number, and spatial position of the burners in the furnace; the steam-water pipeline network may include the topology of the pipeline network for the flow of steam and water inside the boiler; the coal quality may include fuel characteristics such as elemental analysis and calorific value; the operating parameters of the distributed control system may include online monitoring data such as unit load, feedwater flow rate, steam temperature, steam pressure, and air volume; the wall temperature measuring points may include the positions and readings of temperature measuring points arranged on the tube walls of the heating surfaces; the soot blowing records may include the commissioning time and frequency of the soot blowers for each heating surface; and the maintenance and scaling information may include maintenance and inspection data on the oxide layer thickness, scaling thermal resistance, and deposition status of the tube walls of the heating surfaces.

[0036] The embodiments of the present invention can be applied to at least one of the following heating surfaces in pulverized coal boilers, circulating fluidized bed boilers, biomass co-firing boilers, or waste incineration waste heat boilers: water-cooled walls, screen-type superheaters, high-temperature superheaters, high-temperature reheaters, low-temperature reheaters, and economizers.

[0037] In practical implementation, embodiments of the present invention can acquire relevant data about the target power plant boiler, including but not limited to its geometric structure, heating surface arrangement, burner arrangement, steam and water piping network, coal quality, operating parameters of the distributed control system, wall temperature measuring points, soot blowing records, and maintenance and scaling information. For example, embodiments of the present invention can collect boiler design drawings, tube panel layout diagrams, heating surface material specifications, burner elevation, air-coal system parameters, historical operating data, and measuring point lists, and establish an object coding table. Embodiments of the present invention can also conduct on-site measurements by deploying sonar, thermal analyzers, and other detection equipment at key locations to perform on-site detection and analysis of temperatures and related parameters at key locations.

[0038] The geometric structure, heating surface arrangement, burner arrangement, and steam-water pipeline network information are extracted from the boiler design drawings and transformed into geometric parameters recognizable by the 3D model. Coal quality information is acquired in real time through an online coal analyzer or updated periodically through laboratory coal quality analysis reports, including key parameters such as industrial analysis, elemental analysis, and calorific value. The operating parameters of the distributed control system do not depend on a specific manufacturer's DCS and can collect real-time data through OPC (Open Platform Communications), PI (Plant Information System), SIS (Supervisory Information System), or database interfaces, including boiler load, coal feed rate, air volume, furnace pressure, and main steam temperature and pressure. Wall temperature measurement data is obtained from the boiler wall temperature monitoring system, including real-time wall temperature values ​​at key locations on each heating surface. Soot blowing records and maintenance scaling information are obtained by accessing the equipment management system or maintenance record database. For data that cannot be directly obtained, the embodiments of the present invention can supplement with design values ​​or empirical reference values.

[0039] For example, embodiments of the present invention can acquire relevant data of the target power plant boiler, including but not limited to boiler structure, combustion organization, steam-water system, operating parameters, wall temperature measuring points, coal quality, and soot blowing and maintenance information. A unified object coding system is established for the target boiler, encoding objects such as furnace area, horizontal flue, tail flue, each stage of heating surface, tube panel, single tube, tube section, header, connecting pipe, throttling device, measuring point, soot blower, and burner, so that the three-dimensional flue gas side simulation mesh, one-dimensional steam-water thermal hydraulic model nodes, and on-site DCS measuring points have a traceable correspondence.

[0040] The embodiments of the present invention can systematically acquire multi-source heterogeneous data such as boiler geometry, thermodynamics, fluid dynamics, and operation, providing a complete, consistent, and traceable basic input for the subsequent establishment of a three-dimensional model of the flue gas side, a one-dimensional model of the steam and water side, and a data calibration model. This ensures that the model can truly reflect the actual structure and operating status of the boiler, fundamentally guaranteeing the reliability of global temperature prediction.

[0041] In step S102, based on at least one of the following, a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, key location measurement point data, and field operation data calibration model are established. The non-uniform heat load of the heating surface, the flow rate of the working fluid in the pipe, the working fluid temperature, the heat transfer boundary parameters, and the measurement point calibration deviation are exchanged between the various models and the measured data through a multi-source coupling interface, so as to obtain the multi-source coupling calculation results after the multi-source coupling convergence condition is met.

[0042] It is understood that the three-dimensional combustion heat transfer model on the flue gas side in the embodiments of the present invention can be used to obtain the spatial distribution of flue gas temperature, flow rate, composition, particle concentration, radiative heat flux and convective heat flux in the furnace; the steam-water thermal hydraulic model can be used to calculate the flow distribution, pressure drop, working fluid enthalpy, heat transfer coefficient in the pipe network and outlet temperature of the heating surface; the field data calibration model can be used to integrate key point test data such as laser temperature measurement and acoustic temperature measurement, as well as DCS operation data, wall temperature data, coal quality data, soot blowing records and maintenance records.

[0043] In actual implementation, embodiments of the present invention can construct a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, a calibration model of key location measurement point data and field operation data based on relevant information of the target power plant boiler, including but not limited to geometric structure, heating surface arrangement, burner arrangement, steam-water pipe network, coal quality, operating parameters of the distributed control system, wall temperature measurement points, soot blowing records and maintenance scaling information. This forms a multi-source coupled calculation framework for predicting heating surface temperature and establishes a collaborative mechanism for heat load spatial mapping, working fluid flow / temperature feedback, measured deviation calibration and data-driven proxy prediction. Through the multi-source coupling interface, non-uniform heat load of the heating surface, working fluid flow rate in the pipe, working fluid temperature, heat transfer boundary parameters and measurement point calibration deviations can be exchanged between various models and measured data to achieve collaborative coupling of the flue gas side, steam-water side and operation data.

[0044] Specifically, the three-dimensional combustion heat transfer model is used to obtain the spatial distribution of flue gas temperature, velocity, composition, particle concentration, radiative heat flux, and convective heat flux within the furnace; the steam-water thermal hydraulic model is used to calculate the flow distribution, pressure drop, working fluid enthalpy, in-tube heat transfer coefficient, and outlet temperature of the heating surface pipe network; and the field data calibration model is used to integrate key test data such as laser thermometry and acoustic thermometry, as well as DCS operation data, wall temperature data, coal quality data, soot blowing records, and maintenance records. Through a multi-source coupling interface, the non-uniform heat load output by the three-dimensional combustion heat transfer model is mapped to the heat input of the steam-water side pipe section, and the in-tube cooling capacity calculated by the steam-water thermal hydraulic model is fed back to the wall boundary of the combustion heat transfer model. At the same time, the model boundary conditions, key parameters, and prediction deviations are dynamically corrected using field measured data, achieving coordinated coupling between the flue gas side, the steam-water side, and the operation data.

[0045] Specifically, embodiments of the present invention can perform three-dimensional meshing of the furnace, horizontal flue, tail flue, and key heating surface areas, and establish a three-dimensional combustion heat transfer model using turbulence, combustion, radiation, and particle motion models suitable for pulverized coal combustion, outputting the non-uniform heat load of each heating surface tube panel or tube section. Embodiments of the present invention can construct a steam-water system network including water-cooled walls, superheaters, reheaters, economizers, headers, connecting pipes, and throttling devices, establish a steam-water thermal-hydraulic model, and calculate the flow rate, pressure drop, and working fluid temperature of each loop.

[0046] Among them, the three-dimensional combustion heat transfer simulation software can be a general platform with the ability to solve in-furnace combustion, radiation heat transfer and particle deposition, and the thermal hydraulic software can be a general platform with the ability to solve one-dimensional pipe network, header distribution, pressure drop and working fluid enthalpy.

[0047] Furthermore, in this embodiment of the invention, the heat load on the flue gas side can be mapped to the steam-water thermal hydraulic pipe section; the working fluid side boundary is calculated on the steam-water side and then fed back for three-dimensional combustion heat transfer simulation; the on-site measured data is used to calibrate the model boundary and output deviation; the iteration is repeated until the energy balance, key temperature, measurement point residual and hot spot location meet the multi-source coupling convergence condition.

[0048] The embodiments of this invention can establish a three-dimensional flue gas side model, a one-dimensional steam-water side model, and a measured data calibration model. It uses a multi-source coupling interface to realize the bidirectional closed-loop transfer of heat load, flow rate, temperature, and calibration deviation. It also uses on-site measured data to correct system deviations and obtain multi-source coupling calculation results. This achieves integrated calculation of the entire boiler combustion and heat exchange process, significantly improving the calculation accuracy of heat load on the heating surface and working fluid temperature distribution. It provides a high-quality training benchmark that is spatially accurate and physically consistent for subsequent data-driven models.

[0049] In step S103, a hybrid sample library is constructed using multi-source coupling calculation results and field measured data to extract at least one flue gas side feature, at least one steam-water side feature, at least one operating side feature, and at least one historical temperature feature. A global temperature prediction model for the heating surface of a power plant boiler, including a spatial feature extraction module, a temporal feature modeling module, and a hotspot attention module, is trained to output the global tube wall temperature, working fluid temperature, hotspot location, and over-temperature probability of the heating surface.

[0050] It is understood that, in the embodiments of the present invention, at least one flue gas-side feature may include, but is not limited to, flue gas temperature, flow rate, composition, particle concentration, radiative heat flux and convective heat flux in the furnace; at least one steam-water-side feature may include, but is not limited to, calculating the flow distribution, pressure drop, working fluid enthalpy, heat transfer coefficient in the tube and outlet temperature of the heating surface pipe network; at least one operation-side feature may be operating condition parameters such as unit load, coal type and air volume extracted from the operating parameters of the distributed control system; at least one historical temperature feature may be the historical temperature change trend of the heating surface extracted based on the wall temperature measuring point.

[0051] In actual implementation, embodiments of the present invention can employ orthogonal design and engineering experience-based operating condition combination methods to cover variables such as load, coal quality, oxygen content, air distribution, ash accumulation, soot blowing, burner commissioning mode, and load change rate. A batch of coupled models are run to generate mechanism samples. Then, the on-site DCS, wall temperature, coal quality, and maintenance records are integrated and calibrated to form a hybrid sample library. The operating conditions in the hybrid sample library cover the entire load range of the unit, coal quality or blending ratio, furnace outlet oxygen content, different burnout air / secondary air ratios, ash accumulation thickness, soot blowing cycle, and load increase / decrease rate changes. Deviation calibration of the on-site measured data includes steady-state segment identification, measuring point drift correction, matching of wall temperature measuring points with the spatial position of the tube screen, normalization of the lower heating value of the coal, soot blowing status coding, and missing value interpolation.

[0052] Furthermore, at least one flue gas side feature, at least one steam-water side feature, at least one operating side feature, and at least one historical temperature feature are extracted. A machine learning model containing a spatial feature extraction module, a temporal feature modeling module, and a hotspot attention module is trained. The model input includes the operating side feature, flue gas side feature, steam-water side feature, and historical temperature sequence. The output includes the global pipe wall temperature of the heated surface, the working fluid temperature, the hotspot coordinates, the overheat probability, and the future short-term temperature rise trend.

[0053] The spatial feature extraction module comprises one or more of convolutional neural networks, graph neural networks, or 3D convolutional networks, used to extract spatial thermal deviation features between the heating surface grid, tube panels, or tube segments. The temporal feature modeling module comprises one or more of LSTM (Long Short-Term Memory), GRU (Gated Recurrent Unit), TCN (Temporal Convolution Network), or Transformer encoders, used to identify the dynamic temperature response caused by load changes, coal quality fluctuations, ash accumulation, and soot blowing disturbances. The training loss function includes an average temperature error term, a hotspot region weighted error term, a temperature rise trend error term, and a physical constraint term. The physical constraint term is at least used to ensure that the heat absorption of the heating surface, the enthalpy rise of the working fluid, and the direction of temperature field change are consistent with the boiler's thermal balance.

[0054] For example, embodiments of the present invention can design full-condition combinations and generate mechanism samples in batches. Then, key point test data, DCS, and wall temperature data are used to calibrate the simulation results to form a trainable dataset. The model is trained using operating side, flue gas side, steam-water side, and historical temperature sequences as inputs, and global pipe wall temperature and working fluid temperature as outputs, with higher weights assigned to hotspot regions. By combining non-uniform heat load with steam-water side cooling capacity, the model can quickly predict the temperature field of the heated surface, hotspot locations, and risk levels, and display the results on the operation monitoring panel or intelligent boiler platform, supporting over-temperature early warning, on-demand soot blowing, and operation optimization.

[0055] This invention can integrate multi-source coupled calculation results with field measured data to construct a hybrid sample library, giving the samples the dual advantages of global physical consistency and measurement point authenticity. Through a machine learning model that includes a spatial feature extraction module, a temporal feature modeling module, and a hotspot attention module, it can accurately capture the spatial non-uniformity and dynamic change patterns of combustion and heat exchange in the furnace. This not only enables high-precision prediction of the global temperature of the heating surface, but also accurately locates local hotspots and quantifies over-temperature risks, providing comprehensive decision support for the safe operation of the boiler.

[0056] Optionally, in one embodiment of the present invention, outputting the global wall temperature of the heating surface, the working fluid temperature, the hot spot location, and the probability of overheating includes: deploying the global temperature prediction model of the heating surface of the power plant boiler as a real-time proxy model, connecting it to the online monitoring system of the target power plant boiler, so as to output the global wall temperature of the heating surface, the working fluid temperature, the hot spot location, and the probability of overheating.

[0057] It is understood that the real-time agent model in the embodiments of the present invention can be understood as deploying the trained prediction model in a lightweight format in the computing environment of the online monitoring system. The real-time agent model can take DCS real-time data as input and can run in real time in the industrial field. The online monitoring system is a boiler distributed control system or a dedicated monitoring platform with data acquisition, processing and display functions.

[0058] In practical implementation, this invention allows the global temperature prediction model for the heating surface of a power plant boiler to be deployed on a local power plant server or industrial edge computing node. It accesses DCS and SIS data, refreshes prediction results on a second-by-second basis, and combines threshold values, temperature rise rate, and material margin to output risk levels and control recommendations. The results are then displayed on the operation monitoring panel or intelligent boiler platform. The real-time proxy model outputs a global temperature resolution at at least one of the following levels: tube panel level, tube segment level, single tube level, or grid level, with an online calculation cycle on the order of seconds.

[0059] For example, embodiments of the present invention can convert the trained spatial feature extraction module, temporal feature modeling module, and hotspot attention module into a format that supports online inference and deploy them on the server or edge computing node of the online monitoring system; the real-time operating parameters of the DCS and the wall temperature measurement point data are continuously pushed to the proxy model according to the prediction cycle through the data interface; the proxy model extracts the flue gas side features, steam and water side features, operating side features, and historical temperature features of the current window in each prediction cycle, and outputs the global pipe wall temperature field, working fluid temperature field, hotspot location, and over-temperature probability prediction value at the current moment through forward inference, and presents them to the operators in the form of cloud map or pipe screen unfolded diagram through a visualization interface.

[0060] This invention can deploy the global temperature prediction model as a real-time proxy model and connect it to the power plant boiler online monitoring system. It can output the global temperature distribution and hotspot information of the heating surface in real time, enabling operators to obtain the dynamic temperature of the entire pipe wall and the risk of overheating in real time, thus meeting the timeliness requirements of real-time operation monitoring of power plant boilers and improving response speed.

[0061] Optionally, in one embodiment of the present invention, the method further includes: matching an over-temperature level based on at least one of the following: the overall pipe wall temperature of the heated surface, the working fluid temperature, the hot spot location, and the probability of over-temperature; generating over-temperature graded early warning information based on the over-temperature level to provide an over-temperature warning; and / or matching corresponding operation control suggestions based on the over-temperature level and prompting the user with operation control suggestions.

[0062] It is understood that the over-temperature level in the embodiments of the present invention can be dynamically determined based on the design allowable wall temperature, material grade, pressure grade, historical temperature rise rate and operating margin, and three levels of warning can be set: attention, warning and danger. The over-temperature level can be set by those skilled in the art according to the actual situation, and no specific restrictions are made here. The over-temperature grade warning information can include the warning level, the scope of impact, the duration and the recommended measures. The operation control suggestions can be operation guidance schemes formulated for different over-temperature situations.

[0063] For example, embodiments of the present invention can set over-temperature level classification standards, such as: Attention (pipe wall temperature exceeds the design value by less than 10°C and the duration is less than 1 hour), Warning (exceeds the design value by 10-20°C or the duration is 1-4 hours), and Danger (exceeds the design value by more than 20°C or the duration is more than 4 hours). When the prediction result shows that a certain area has reached the over-temperature condition, the system automatically matches the corresponding level and generates early warning information including the location, degree, and development trend of the over-temperature, and notifies the operators through audible and visual alarms, pop-up prompts, etc. At the same time, according to the over-temperature level and location, the system matches the corresponding control suggestions from the preset control strategy library. The control suggestions may include starting the soot blowers associated with high-risk pipe screens on demand, adjusting the burnout air opening or swing angle, adjusting the secondary air ratio, limiting the load change rate, and adjusting the coal mill combination or coal blending ratio.

[0064] The embodiments of the present invention can automatically match the over-temperature level based on the prediction results and generate graded early warning information and operation control suggestions, thereby realizing graded warning and proactive intervention of boiler heating surface over-temperature risk, reducing the risk of over-temperature tube rupture and unplanned shutdown, and improving the safety, flexibility and economy of unit operation.

[0065] Optionally, in one embodiment of the present invention, based on at least one of the following, a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, key location measurement point data, and field operation data calibration model are established. The non-uniform heat load of the heating surface, the flow rate of the working fluid in the pipe, the working fluid temperature, heat transfer boundary parameters, and measurement point calibration deviations are exchanged between the various models and the measured data through a multi-source coupling interface to obtain multi-source coupling calculation results. This includes: using the heat load mapping module of the multi-source coupling interface to map the radiative heat flow, convective heat flow, and total heat load on the three-dimensional flue gas side simulation mesh or surface element to the steam-water side heat transfer boundary according to the heating surface tube screen, pipe segment, or single pipe number; using the flow rate and temperature feedback module of the multi-source coupling interface to feed back the pipe mass flow rate, inlet enthalpy, outlet enthalpy, pressure loss, and inner wall heat transfer coefficient calculated on the one-dimensional steam-water side to the three-dimensional combustion heat transfer model on the flue gas side; and using the measured deviation calibration module of the multi-source coupling interface to correct the model boundary and prediction results based on the wall temperature measurement points, the operating parameters of the distributed control system, and the coal quality.

[0066] It is understood that the multi-source coupling interface in this embodiment of the invention includes a heat load mapping module, a flow and temperature feedback module, and a measured deviation calibration module.

[0067] In actual implementation, embodiments of the present invention can use a heat load mapping module to map the radiative heat flow, convective heat flow, and total heat load on the three-dimensional flue gas side simulation mesh or surface element to the heat exchange boundary on the steam-water side according to the heating surface tube screen, tube segment, or single tube number; a flow rate and temperature feedback module can feed back the tube mass flow rate, inlet enthalpy, outlet enthalpy, pressure loss, and inner wall heat transfer coefficient calculated on the one-dimensional steam-water side to the three-dimensional combustion heat transfer model; and a measured deviation calibration module can correct the model boundary and prediction results based on wall temperature measurement points, DCS operating parameters, and coal quality records.

[0068] The heat load mapping employs at least one of the following methods: area weighting, pipe length weighting, thermal resistance correction, and ash pollution coefficient correction, to ensure that the heat load on the flue gas side and the heat input on the steam-water side pipe section satisfy energy conservation when the total error is less than 2% of the preset error threshold. The preset error threshold can be set by those skilled in the art according to the actual situation, and no specific restrictions are imposed here.

[0069] For example, in the heat load mapping module, a mapping relationship can be established between the three-dimensional flue gas side simulation mesh and the one-dimensional steam-water side pipe segment. The area-weighted average method is used to distribute the radiative heat flow, convective heat flow, and total heat load on each three-dimensional mesh or surface element to the corresponding pipe screen, pipe segment, or single pipe to obtain the heat load boundary of each heat exchange unit on the steam-water side. In the flow rate and temperature feedback module, the mass flow rate, inlet enthalpy, outlet enthalpy, pressure loss, and inner wall heat transfer coefficient of each pipe segment calculated by the one-dimensional steam-water side model are fed back to the three-dimensional combustion heat transfer model on the flue gas side to correct the wall heat transfer boundary on the flue gas side. In the measured deviation calibration module, the wall temperature measurement point data is acquired in real time, and the deviation between the model prediction value and the measured value at the measurement point location is calculated. This deviation is used as a correction factor to adjust the flue gas side inlet boundary parameters and the steam-water side resistance coefficient, so that the model gradually approaches the measured value in the next iteration.

[0070] This invention can map the non-uniform heat load output by the three-dimensional combustion heat transfer model to the heat input of the steam-water side pipe section through a multi-source coupling interface, and feed back the cooling capacity inside the pipe calculated by the steam-water thermal hydraulic model to the wall boundary of the combustion heat transfer model. At the same time, it uses field measured data to dynamically correct the model boundary conditions, key parameters and prediction deviations, thereby realizing the coordinated coupling of the flue gas side, steam-water side and operating data, forming a closed-loop optimization mechanism that deeply integrates physical mechanisms and measured data, and improving the accuracy and reliability of multi-source coupling calculation.

[0071] Optionally, in one embodiment of the present invention, the multi-source coupling convergence condition includes: the relative deviation of the total heat absorption of the heating surface in two adjacent iterations is less than or equal to a preset percentage; the temperature deviation between the working fluid at the outlet of the key tube screen and the key tube wall is less than or equal to a preset threshold; the calibration residual of the field measuring point is less than a preset difference; and the location of the local high heat load area reaches a preset condition that no significant drift occurs.

[0072] It is understood that the preset percentage in this embodiment of the invention can be 1.0%, and the preset percentage can be set by those skilled in the art according to actual conditions, without specific limitations. The preset threshold can be 5℃, and the preset threshold can be set by those skilled in the art according to actual conditions, without specific limitations. The preset difference can be 3℃, and the preset difference can be set by those skilled in the art according to actual conditions, without specific limitations. The preset condition for no more significant drift can be that the pipe screen number and pipe segment index corresponding to the hot spot identified in two consecutive iterations are consistent, and the offset of the peak position of the heat load on the pipe segment along the pipe length direction does not exceed the length of a single pipe segment, indicating that the topological structure of the thermal deviation has stabilized. The preset condition for no more significant drift can be set by those skilled in the art according to actual conditions, without specific limitations.

[0073] In actual implementation, after each round of multi-source coupling iteration, the embodiments of the present invention can sequentially calculate four convergence indices: calculate the relative deviation of the total heat absorption of the current round from the previous round and compare it with a preset percentage; extract the changes in working fluid temperature and pipe wall temperature at key locations such as the outlet tube screen of the final superheater or high-temperature reheater and compare them with preset thresholds; calculate the root mean square value of the difference between the model prediction value and the measured value at all wall temperature measuring points and compare it with a preset difference value of the measuring point range; extract the pipe segment numbers and peak positions of the top heat loads and compare them with the results of the previous round. When all four indices simultaneously meet the corresponding preset conditions, the iteration is terminated and the parameter set of the current round is output as the final multi-source coupling calculation result; if any one of them is not met, the next round of iteration continues until all conditions are met or the preset maximum number of iterations is reached.

[0074] The embodiments of this invention can determine the convergence of multi-source coupling based on multiple dimensions such as the relative deviation of the total heat absorption of the heating surface, the temperature deviation between the working medium and the tube wall in the key tube panel, the calibration residual of the field measuring points, and the location stability of the local high heat load area. This establishes a comprehensive and strict convergence judgment standard, which can ensure that the multi-source coupling calculation meets the engineering application requirements in terms of energy conservation, key parameter accuracy, measured data fitting degree, and local feature stability. This accurately reflects the combustion and heat exchange characteristics of the boiler under actual operating conditions and ensures the engineering credibility of the overall temperature prediction results.

[0075] Specifically, taking the over-temperature early warning of the high-temperature reheater of a 660MW supercritical four-corner tangential boiler as an example, the working principle of the power plant boiler heating surface global temperature prediction method based on multi-source coupling in this embodiment of the invention is explained in detail.

[0076] A 660MW supercritical tangential coal-fired boiler, operating within a 50%-100% load range, experienced repeated wall temperatures nearing alarm values ​​on the right-side tube panel of the high-temperature reheater after participating in deep peak shaving. Following the method of this invention, a three-dimensional combustion heat transfer model was established, including the furnace, horizontal flue, and tail flue, with approximately 5.2 million mesh elements; a one-dimensional steam-water thermal-hydraulic model was also established, containing 126 tube bundle loops and 32 headers. Through 24 bidirectional iterations, the total heat absorption deviation of the key heating surfaces was less than 0.8%, and the steam temperature deviation at the outlet of the key tube panel was less than 2.5℃.

[0077] The sample library covers 30%-100% BMCR load, two main coal types and one blended coal type, furnace outlet oxygen content of 2.5%-5.5%, ash accumulation thickness of 0-4mm, and soot blowing cycle of 6-24h, forming a total of 126,000 mechanism samples, which are integrated with 18 months of on-site DCS and wall temperature data. The average absolute error of the model test set is 2.2℃, and the hot spot identification accuracy rate is 96.7%. During online trial operation, the system identified 6 local overheating risks in advance, 4 of which were related to the thermal deviation of the flue gas on the right side of the reheater, and 2 of which were related to excessively long soot blowing intervals. After the operators shortened the associated soot blowing cycle and fine-tuned the burnout air opening according to the recommendations, the highest wall temperature of the high-temperature reheater decreased by about 8-12℃, and the coal consumption for power supply decreased by about 0.62g / kWh.

[0078] Specifically, taking the diagnosis of off-center burning of the water-cooled wall of a 1000MW ultra-supercritical counter-firing boiler as an example, the working principle of the multi-source coupling-based global temperature prediction method for the heating surface of a power plant boiler in this embodiment of the invention is explained in detail.

[0079] During low-load stable combustion and rapid load increase processes, a 1000MW ultra-supercritical opposed-fired boiler experienced significant temperature fluctuations in the localized water-cooled wall of the front wall. By employing the method described in this invention, the in-furnace combustion skew heat load output from the three-dimensional combustion heat transfer model was mapped to the parallel loop of the water-cooled wall in the steam-water thermal-hydraulic model. It was found that some front wall loops experienced a temperature rise amplification effect due to low flow rates superimposed with localized high heat loads. After the real-time proxy model was implemented, it could provide an early warning of the temperature rise trend in the upper and middle tube panels of the front wall when the load increase rate exceeded a preset threshold, and suggested limiting the short-term load increase rate and optimizing the secondary air ratio. After application, the temperature rise rate at key measuring points on the front wall decreased by approximately 25% under similar operating conditions, allowing operators to make adjustments approximately 10-15 minutes in advance.

[0080] Specifically, taking the optimization of economizer ash accumulation and on-demand soot blowing in a 300MW subcritical boiler as an example, the working principle of the multi-source coupling-based global temperature prediction method for the heating surface of a power plant boiler in this embodiment of the invention is explained in detail.

[0081] When a 300MW subcritical unit burns high-ash coal, the ash accumulation rate in the economizer area of ​​the tail flue is relatively fast. Traditional timed soot blowing suffers from both insufficient and excessive soot blowing. According to the method of this invention, ash-fouling thermal resistance is used as a feature of the sample library and a correction factor for the coupled model to estimate the ash-fouling status and tube wall temperature in different areas of the economizer in real time. The system correlates the temperature prediction results with the flue gas temperature, feedwater temperature rise, and historical soot blower operation to generate on-demand soot blowing suggestions. Within a maintenance cycle, the fluctuation of the economizer outlet flue gas temperature decreased, soot blowing steam consumption decreased by approximately 5%, and no abnormal local wall temperature rise due to ash accumulation occurred.

[0082] Specifically, taking the auxiliary assessment of high-temperature corrosion risk in biomass co-firing boilers as an example, the working principle of the multi-source coupling-based full-domain temperature prediction method for the heating surface of power plant boilers in this embodiment of the invention is explained in detail.

[0083] During a biomass co-firing test of a coal-fired boiler, the risk of corrosion and ash accumulation on high-temperature heating surfaces increased due to the influence of chlorine, alkali metal, and ash fusion characteristics of the fuel. This invention addresses this issue by adding chlorine content, alkali metal content, ash fusion point, and co-firing ratio to the coal quality characteristics, and adding the ash cleaning cycle to the maintenance characteristics. The model can assess the risk level of the coupling between the output temperature of the high-temperature superheater and reheater tube panels and the ash thermal resistance, providing auxiliary basis for the upper limit of the co-firing ratio, soot blowing strategies, and maintenance inspection locations.

[0084] Specifically, it can be combined with Figures 2 to 5 As shown, a specific embodiment is used to elaborate in detail on the working principle of the multi-source coupling-based global temperature prediction method for the heating surface of a power plant boiler in this invention.

[0085] like Figure 2As shown, this embodiment of the invention can integrate multi-dimensional combustion flow field and heat transfer calculations, thermal-hydraulic system calculations, and on-site operation and coal quality data through a multi-source input layer. Based on this, multi-source coupled modeling and sample construction are carried out. Bidirectional iterative calculations on the flue gas side and the steam-water side are achieved through heat load spatial mapping and flow / temperature feedback, simultaneously constructing a hybrid sample library containing simulation and measured data. After sample construction, flue gas side features, steam-water side features, and operation side features are extracted respectively and input into a machine learning model integrating spatial feature extraction, temporal feature extraction, and attention-weighted fusion mechanisms for training. The trained model is then lightweighted and transformed into a real-time proxy model. After deployment, it outputs global pipe wall temperature distribution, working fluid temperature prediction, local hotspot location, and over-temperature risk warning. Based on the prediction results, closed-loop control measures such as precise soot blowing, operation adjustment, and safety management are implemented. The control effect is fed back to the multi-source input layer to achieve continuous iterative optimization of the system.

[0086] like Figure 3 As shown, embodiments of the present invention can solve the furnace temperature field, flue gas flow field, radiation / convective heat transfer, and heat load distribution of the heating surface through multidimensional combustion flow field and heat transfer calculations. The working fluid flow distribution, pressure loss, inlet / outlet temperature, and pipe wall temperature response are calculated through a thermal hydraulic system. The multidimensional combustion flow field and heat transfer calculations and the thermal hydraulic system calculations achieve data interaction through a bidirectional iterative coupling mechanism. Specifically, the heat load calculated on the flue gas side is transferred to the steam-water side as a heat transfer boundary condition through heat load spatial mapping, while the flow rate, temperature, and other parameters calculated on the steam-water side are fed back to the flue gas side through flow rate and temperature feedback to correct the wall heat transfer conditions. The iterative process is uniformly controlled by a convergence criterion. When all convergence conditions are met, the coupled calculation results, including global wall temperature, working fluid temperature, local hot spots, and risk assessment, are jointly output.

[0087] like Figure 4 As shown, the data input layer integrates three types of multi-dimensional data: flue gas-side spatial features, steam-water-side operating condition features, and operational-side temporal features. Flue gas-side spatial features include local flue gas temperature, flow rate, radiative heat flux, and ash thermal resistance; steam-water-side operating condition features include flow rate, pressure, inlet temperature, and endothermic deviation; and operational-side temporal features include load, coal quality, oxygen content, air distribution, and soot blowing status. The input data first undergoes preprocessing steps including standardization, timestamp alignment, and feature selection. After preprocessing, the feature data is fed into a CNN (Convolutional Neural Network) and an LSTM, respectively. The extracted features enter the feature fusion layer and undergo attention weighting processing. Finally, the prediction results are obtained through a fully connected output layer. The prediction output includes global pipe wall temperature, working fluid temperature, local hotspot coordinates, and overheat risk level. Model training and optimization continuously optimize the model's prediction accuracy and generalization ability through loss function calculation, parameter updates, and model validation.

[0088] like Figure 5 As shown, operational data is collected in real time from the DCS or online monitoring system, cleaned and preprocessed, and multi-source features are constructed. A real-time proxy model is used to calculate the predicted temperature field of the heated surface, which is then used for hotspot identification and risk assessment. The system determines whether over-temperature conditions exist: if no over-temperature occurs, normal monitoring and result storage are implemented, and the cycle either ends or continues; if over-temperature occurs, a graded over-temperature warning is issued, generating targeted operational control suggestions including precise soot blowing, air distribution adjustment, and load optimization. Control and feedback measures are then implemented, and the control effects are fed back to the data acquisition stage, forming a complete closed-loop control process.

[0089] The multi-source coupling-based global temperature prediction method for power plant boiler heating surfaces proposed in this embodiment of the invention can acquire information such as boiler geometry, coal quality, parameters of the distributed control system, and wall temperature measurement points to establish a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, and a data calibration model. It exchanges parameters such as non-uniform heat load, flow rate, and temperature through a multi-source coupling interface, achieving coupling convergence to obtain multi-source coupling calculation results. A hybrid sample library is constructed using the multi-source coupling calculation results and field measured data to extract temperature features from the flue gas side, steam-water side, operating side, and historical data. Training is then performed to include spatial feature extraction and temporal feature extraction. The global temperature prediction model, based on the modeling and hotspot attention module, outputs pipe wall temperature, working fluid temperature, hotspot location, and over-temperature probability. Through multi-source collaboration among the combustion heat transfer mechanism model, the steam-water thermal-hydraulic model, key monitoring points, field operation data, and machine learning models, it achieves closed-loop correction of heat load, working fluid cooling capacity, measured deviations, and temperature prediction results. This enables online prediction of the global temperature distribution and local hotspots, thereby improving the accuracy of local hotspot identification and the prediction accuracy and generalization ability under complex operating conditions. This provides a basis for over-temperature early warning, intelligent soot blowing, and combustion adjustment, enhancing the unit's operational safety, flexibility, and economy. This solves the problems in related technologies where purely data-driven models lack physical constraints, leading to inaccuracies when extrapolating to other operating conditions; single 3D combustion heat transfer simulations are insufficient to reflect the complex characteristics of steam-water systems; and single thermal-hydraulic simulations simplify the flue gas side boundary to average heat load, failing to identify local high heat loads, resulting in missed local hotspot detection, insufficient generalization ability across all operating conditions, and difficulties in real-time deployment.

[0090] Next, referring to the accompanying drawings, we describe the global temperature prediction device for the heating surface of a power plant boiler based on multi-source coupling, according to an embodiment of the present invention.

[0091] Figure 6 This is a schematic diagram of the structure of a power plant boiler heating surface global temperature prediction device based on multi-source coupling according to an embodiment of the present invention.

[0092] like Figure 6As shown, the power plant boiler heating surface global temperature prediction device 10 based on multi-source coupling includes: acquisition module 100, establishment module 200 and prediction module 300.

[0093] The acquisition module 100 is used to acquire at least one of the following: the geometric structure of the target power plant boiler, the arrangement of the heating surface, the arrangement of the burners, the steam and water pipeline network, the coal quality, the operating parameters of the distributed control system, the wall temperature measuring points, the soot blowing records, and the maintenance and scaling information.

[0094] Module 200 is established to create a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, key location measurement point data, and field operation data calibration model based on at least one of them. It also exchanges non-uniform heat load of the heating surface, working fluid flow rate in the pipe, working fluid temperature, heat transfer boundary parameters, and measurement point calibration deviations between the various models and measured data through a multi-source coupling interface, so as to obtain the multi-source coupling calculation results after the multi-source coupling convergence condition is met.

[0095] The prediction module 300 is used to construct a hybrid sample library using multi-source coupled calculation results and field measured data to extract at least one flue gas side feature, at least one steam-water side feature, at least one operating side feature and at least one historical temperature feature. It trains a global temperature prediction model for the heating surface of a power plant boiler, which includes a spatial feature extraction module, a temporal feature modeling module and a hotspot attention module, to output the global tube wall temperature, working fluid temperature, hotspot location and over-temperature probability of the heating surface.

[0096] Optionally, in one embodiment of the present invention, the prediction module 300 includes an output unit.

[0097] The output unit is used to deploy the global temperature prediction model of the heating surface of the power plant boiler as a real-time proxy model and connect it to the online monitoring system of the target power plant boiler to output the global tube wall temperature, working fluid temperature, hot spot location and over-temperature probability of the heating surface.

[0098] Optionally, in one embodiment of the present invention, the power plant boiler heating surface global temperature prediction device 10 based on multi-source coupling further includes: a matching module, a generation module, and a prompting module.

[0099] The matching module is used to match the over-temperature level based on at least one of the following: the overall pipe wall temperature of the heated surface, the working fluid temperature, the hot spot location, and the over-temperature probability.

[0100] The generation module is used to generate over-temperature level warning information based on the over-temperature level, so as to provide over-temperature alerts.

[0101] The prompt module is used to match corresponding operation and control suggestions based on the over-temperature level and prompt the user with the operation and control suggestions.

[0102] Optionally, in one embodiment of the present invention, the establishment module 200 includes: a mapping unit, a feedback unit, and a correction unit.

[0103] The mapping unit is used to map the radiative heat flow, convective heat flow and total heat load on the three-dimensional flue gas side simulation mesh or surface element to the steam-water side heat exchange boundary according to the numbering of the heated surface tube screen, tube segment or single tube.

[0104] The feedback unit is used to feed back the mass flow rate, inlet enthalpy, outlet enthalpy, pressure loss, and inner wall heat transfer coefficient of the pipe obtained from the one-dimensional steam-water side calculation to the three-dimensional combustion heat transfer model on the flue gas side using the flow and temperature feedback module of the multi-source coupling interface.

[0105] The correction unit is used to correct the model boundary and prediction results based on the measured deviation calibration module of the multi-source coupling interface, according to the wall temperature measurement point, the operating parameters of the distributed control system and the coal quality.

[0106] It should be noted that the foregoing explanation of the embodiment of the method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling also applies to the device for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling in this embodiment, and will not be repeated here.

[0107] The multi-source coupling-based global temperature prediction device for power plant boiler heating surfaces proposed in this embodiment of the invention can acquire information such as boiler geometry, coal quality, parameters of the distributed control system, and wall temperature measurement points to establish a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, and a data calibration model. It exchanges parameters such as non-uniform heat load, flow rate, and temperature through a multi-source coupling interface, achieving coupling convergence to obtain multi-source coupling calculation results. Using the multi-source coupling calculation results and field measured data, a hybrid sample library is constructed to extract flue gas side, steam-water side, operating side, and historical temperature features. Training is then performed to include spatial feature extraction and temporal feature extraction. The global temperature prediction model, based on the modeling and hotspot attention module, outputs pipe wall temperature, working fluid temperature, hotspot location, and over-temperature probability. Through multi-source collaboration among the combustion heat transfer mechanism model, the steam-water thermal-hydraulic model, key monitoring points, field operation data, and machine learning models, it achieves closed-loop correction of heat load, working fluid cooling capacity, measured deviations, and temperature prediction results. This enables online prediction of the global temperature distribution and local hotspots, thereby improving the accuracy of local hotspot identification and the prediction accuracy and generalization ability under complex operating conditions. This provides a basis for over-temperature early warning, intelligent soot blowing, and combustion adjustment, enhancing the unit's operational safety, flexibility, and economy. This solves the problems in related technologies where purely data-driven models lack physical constraints, leading to inaccuracies when extrapolating to other operating conditions; single 3D combustion heat transfer simulations are insufficient to reflect the complex characteristics of steam-water systems; and single thermal-hydraulic simulations simplify the flue gas side boundary to average heat load, failing to identify local high heat loads, resulting in missed local hotspot detection, insufficient generalization ability across all operating conditions, and difficulties in real-time deployment.

[0108] Figure 7 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. The electronic device may include: The memory 701, the processor 702, and the computer program stored on the memory 701 and executable on the processor 702.

[0109] When the processor 702 executes the program, it implements the method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling provided in the above embodiments.

[0110] Furthermore, electronic devices also include: Communication interface 703 is used for communication between memory 701 and processor 702.

[0111] The memory 701 is used to store computer programs that can run on the processor 702.

[0112] The memory 701 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0113] If the memory 701, processor 702, and communication interface 703 are implemented independently, then the communication interface 703, memory 701, and processor 702 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized into address buses, data buses, control buses, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0114] Optionally, in a specific implementation, if the memory 701, processor 702, and communication interface 703 are integrated on a single chip, then the memory 701, processor 702, and communication interface 703 can communicate with each other through an internal interface.

[0115] The processor 702 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0116] This invention also provides a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for predicting the global temperature of a power plant boiler heating surface based on multi-source coupling.

[0117] This invention also provides a computer program product storing a computer program that, when executed by a processor, implements the above-described method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling.

[0118] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0119] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0120] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0121] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0122] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0123] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0124] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0125] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for predicting the global temperature of a power plant boiler heating surface based on multi-source coupling, characterized in that, Includes the following steps: Obtain at least one of the following from the target power plant boiler: geometry, heating surface arrangement, burner arrangement, steam and water piping network, coal quality, operating parameters of the distributed control system, wall temperature measurement points, soot blowing records, and maintenance and scaling information; Based on at least one of the above, a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, and a calibration model using key location measurement point data and field operation data are established. A multi-source coupling interface is used to exchange non-uniform heat load on the heating surface, working fluid flow rate in the pipe, working fluid temperature, heat transfer boundary parameters, and measurement point calibration deviations between the models and measured data. Multi-source coupling calculation results are obtained after the multi-source coupling convergence conditions are met. These multi-source coupling convergence conditions include: the relative deviation of the total heat absorption of the heating surface between two adjacent iterations is less than or equal to a preset percentage; the temperature deviation between the working fluid at the outlet of the key pipe screen and the temperature deviation of the key pipe wall are less than or equal to a preset threshold; the calibration residual at the field measurement points is less than a preset difference; and the location of the local high heat load area reaches a preset condition where significant drift no longer occurs. Specifically, based on at least one of the above, a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, key location measurement point data, and field operation data calibration model are established. A multi-source coupling interface is used to exchange non-uniform heat load on the heating surface, working fluid flow rate in the pipe, working fluid temperature, heat transfer boundary parameters, and measurement point calibration deviations between the models and measured data to obtain multi-source coupling calculation results. This includes: using the heat load mapping module of the multi-source coupling interface to map the radiative heat flow, convective heat flow, and total heat load on the three-dimensional flue gas side simulation mesh or surface element to the steam-water side heat transfer boundary according to the heating surface tube screen, pipe segment, or single pipe number; using the flow rate and temperature feedback module of the multi-source coupling interface to feed back the pipe mass flow rate, inlet enthalpy, outlet enthalpy, pressure loss, and inner wall heat transfer coefficient calculated on the one-dimensional steam-water side to the three-dimensional combustion heat transfer model on the flue gas side; and using the measured deviation calibration module of the multi-source coupling interface to correct the model boundary and prediction results based on the wall temperature measurement points, the operating parameters of the distributed control system, and the coal quality. Using the multi-source coupling calculation results and field measured data, a hybrid sample library is constructed to extract at least one flue gas side feature, at least one steam-water side feature, at least one operation side feature, and at least one historical temperature feature. A global temperature prediction model for the heating surface of a power plant boiler, including a spatial feature extraction module, a temporal feature modeling module, and a hotspot attention module, is trained to output the global tube wall temperature, working fluid temperature, hotspot location, and over-temperature probability of the heating surface.

2. The method for predicting the global temperature of a power plant boiler heating surface based on multi-source coupling according to claim 1, characterized in that, The output includes the global pipe wall temperature of the heated surface, the working fluid temperature, the hot spot location, and the probability of overheating, including: The global temperature prediction model of the heating surface of the power plant boiler is deployed as a real-time proxy model and connected to the online monitoring system of the target power plant boiler to output the global tube wall temperature of the heating surface, the working fluid temperature, the hot spot location, and the over-temperature probability.

3. The method for predicting the global temperature of a power plant boiler heating surface based on multi-source coupling according to claim 2, characterized in that, Also includes: The overheating level is matched based on at least one of the following: the overall pipe wall temperature of the heated surface, the working fluid temperature, the hot spot location, and the overheating probability. Based on the aforementioned over-temperature level, generate over-temperature graded early warning information to issue over-temperature alerts; And / or, match the corresponding operation control recommendations according to the over-temperature level, and prompt the user with the operation control recommendations.

4. A device for predicting the global temperature of a power plant boiler heating surface based on multi-source coupling, characterized in that, include: The acquisition module is used to acquire at least one of the following from the target power plant boiler: geometry, heating surface arrangement, burner arrangement, steam and water pipeline network, coal quality, operating parameters of the distributed control system, wall temperature measurement points, soot blowing records, and maintenance and scaling information. A module is established to create a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, key location measurement point data, and a calibration model based on at least one of the above. A multi-source coupling interface is used to exchange non-uniform heat load on the heating surface, working fluid flow rate in the pipe, working fluid temperature, heat transfer boundary parameters, and measurement point calibration deviations between the models and measured data. This allows for the acquisition of multi-source coupling calculation results after meeting the multi-source coupling convergence conditions. These conditions include: the relative deviation of the total heat absorption of the heating surface between two adjacent iterations is less than or equal to a preset percentage; the deviation between the working fluid at the outlet of the key pipe screen and the temperature of the key pipe wall is less than or equal to a preset threshold; the calibration residual at the on-site measurement points is less than a preset difference; and the location of the local high heat load area reaches a preset condition where significant drift no longer occurs. Specifically, based on at least one of the above, a three-dimensional combustion heat transfer model on the flue gas side, a thermal-hydraulic model on the steam-water side, key location measurement point data, and field operation data calibration model are established. A multi-source coupling interface is used to exchange non-uniform heat load on the heating surface, working fluid flow rate in the pipe, working fluid temperature, heat transfer boundary parameters, and measurement point calibration deviations between the models and measured data to obtain multi-source coupling calculation results. This includes: using the heat load mapping module of the multi-source coupling interface to map the radiative heat flow, convective heat flow, and total heat load on the three-dimensional flue gas side simulation mesh or surface element to the steam-water side heat transfer boundary according to the heating surface tube screen, pipe segment, or single pipe number; using the flow rate and temperature feedback module of the multi-source coupling interface to feed back the pipe mass flow rate, inlet enthalpy, outlet enthalpy, pressure loss, and inner wall heat transfer coefficient calculated on the one-dimensional steam-water side to the three-dimensional combustion heat transfer model on the flue gas side; and using the measured deviation calibration module of the multi-source coupling interface to correct the model boundary and prediction results based on the wall temperature measurement points, the operating parameters of the distributed control system, and the coal quality. The prediction module is used to construct a hybrid sample library using the multi-source coupling calculation results and field measured data to extract at least one flue gas side feature, at least one steam-water side feature, at least one operation side feature and at least one historical temperature feature. It trains a global temperature prediction model for the heating surface of a power plant boiler, which includes a spatial feature extraction module, a temporal feature modeling module and a hotspot attention module, to output the global tube wall temperature, working fluid temperature, hotspot location and over-temperature probability of the heating surface.

5. The device for predicting the global temperature of a power plant boiler heating surface based on multi-source coupling according to claim 4, characterized in that, The prediction module includes: The output unit is used to deploy the global temperature prediction model of the heating surface of the power plant boiler as a real-time proxy model, connect it to the online monitoring system of the target power plant boiler, and output the global tube wall temperature of the heating surface, the working fluid temperature, the hot spot location and the over-temperature probability.

6. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and capable of running on the processor, wherein the processor executes the program to implement the method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling as described in any one of claims 1-3.

7. A non-volatile computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling as described in any one of claims 1-3.

8. A computer program product, comprising a computer program, characterized in that, The computer program is executed to implement the method for predicting the global temperature of the heating surface of a power plant boiler based on multi-source coupling as described in any one of claims 1-3.

Images