Methods, devices, terminal equipment and storage media for flow field correction in coal-fired boilers

Abstract

This application relates to the field of industrial digital control technology, and in particular to a method, device, terminal equipment, and storage medium for correcting the flow field of a coal-fired boiler. The method includes: real-time acquisition of operating condition parameters and sensor measurement data of the coal-fired boiler; inputting the operating condition parameters and sensor measurement data into a trained reconstruction model to obtain a reconstruction result of the global flow field inside the boiler; evaluating the operating state of the coal-fired boiler based on the reconstruction result to obtain an evaluation result, and performing control optimization operations on the coal-fired boiler based on the evaluation result; and comparing the difference between the reconstruction result and the sensor observation data in real time. When the difference exceeds a preset threshold, the reconstruction model is adjusted to achieve accurate control of the coal-fired boiler.

Description

Technical Field

[0001] This application relates to the field of industrial digital control technology, and in particular to a method, apparatus, terminal equipment and storage medium for correcting the flow field of a coal-fired boiler. Background Technology

[0002] Under conditions of deep peak shaving and changes in coal quality, the flow field characteristics within the boiler furnace undergo significant alterations. The uniformity and stability of the gas-solid two-phase flow decrease, leading to uneven flow field distribution, intensified airflow scouring, and changes in nitrogen oxide generation characteristics. These issues severely restrict the stable and safe operation of the unit and the efficient and economical control of pollutants. Achieving rapid inference and precise optimization of the flow field state across all operating conditions and processes of a coal-fired boiler is a pressing technical challenge. Currently, there are two main approaches to address this problem. The most widely used method is Computational Fluid Dynamics (CFD), which constructs simulation models by solving the Navier-Stokes equations. However, the models developed using this method are relatively simple and primarily based on mechanistic theory, failing to align with engineering realities and easily resulting in distortions that deviate from actual unit operation. Another approach is based on machine learning methods such as artificial intelligence. However, purely data-driven fluid dynamics methods also have significant limitations: firstly, the scarcity of high-quality data often results in black-box models lacking clear physical meaning and mechanistic interpretability; secondly, poor generalization and robustness make it difficult to establish causal relationships in fluid motion. Summary of the Invention

[0003] In view of this, the embodiments of this application provide a method for correcting the flow field of a coal-fired boiler, which can effectively solve the problem of difficulty in establishing the causal relationship of fluid motion.

[0004] In a first aspect, embodiments of this application provide a method for correcting the flow field of a coal-fired boiler, including: Real-time acquisition of operating parameters and sensor measurement data of coal-fired boilers; The operating condition parameters and the sensor measurement data are input into the trained reconstruction model to obtain the reconstruction result of the global flow field inside the boiler. Based on the reconstruction results, the operating status of the coal-fired boiler is evaluated to obtain evaluation results, and the control optimization operation of the coal-fired boiler is performed based on the evaluation results. The reconstruction model is adjusted when the difference between the reconstruction result and the sensor observation data is compared in real time and the difference exceeds a preset threshold.

[0005] In some embodiments, the training method of the reconstructed model includes: Cold-state data, hot-state data, and online data of coal-fired boilers were collected as training data; The initial flow field is obtained by reconstructing the coal-fired boiler; A physical information neural network model incorporating residual learning is constructed as the base model for the reconstructed model, and the initial flow field is used as the initial flow field of the base model. The base model is trained based on physical constraints and observation assimilation mechanisms to obtain the reconstructed model.

[0006] In some embodiments, the collection of cold-state data, hot-state data, and online data from the coal-fired boiler as training data includes: During the operation of the coal-fired boiler, data was collected under cold conditions to obtain cold-state data. During the operation of the coal-fired boiler, data is collected at a preset height along the flame deflector at the boiler furnace outlet according to different load conditions to obtain thermal data; Obtain online historical data of the coal-fired boiler during operation within a preset time period.

[0007] In some embodiments, reconstructing the initial flow field of the coal-fired boiler includes: Based on the training data, numerical simulations were performed on the furnace and flue area of ​​a coal-fired boiler to obtain three-dimensional flow field data corresponding to each operating condition. Based on the structural characteristics and flow direction of the boiler, multiple sub-regions are divided. Within each sub-region, the flow field is sliced ​​along the main flow direction or typical cross-section direction to construct a set of local flow field sample features. Feature extraction is performed on the flow field sample set of each sub-region to obtain low-order features. Then, the initial flow field is obtained by reconstructing the flow field using basis functions and the low-order features.

[0008] In some embodiments, the loss function of the reconstructed model includes partial differential equation residuals and boundary condition residuals; The expression for the loss function is: ; In the formula, L is the loss value. The residual of the partial differential equation is... The boundary condition residual, The weights for each loss.

[0009] In some embodiments, inputting the operating condition parameters and the sensor measurement data into the trained reconstruction model to obtain the reconstruction result of the global flow field inside the boiler includes: The reconstruction model outputs flow field residual data based on the input operating condition parameters and the sensor measurement data; The sum of the flow field residual data and the initial flow field of the reconstruction model is calculated to obtain the reconstruction result of the current global flow field inside the boiler.

[0010] In some embodiments, the control optimization operation based on the evaluation result includes: If the evaluation result is unstable, the operating parameters of the coal-fired boiler are dynamically corrected until the evaluation result is stable.

[0011] Secondly, this application also provides a flow field correction device for a coal-fired boiler, comprising: The data acquisition module is used to collect operating parameters and sensor measurement data of coal-fired boilers in real time. The reconstruction module is used to input the operating condition parameters and the sensor measurement data into the trained reconstruction model to obtain the reconstruction result of the global flow field inside the boiler. The optimization module is used to evaluate the operating status of the coal-fired boiler based on the reconstruction results, obtain evaluation results, and perform control optimization operations on the coal-fired boiler based on the evaluation results. An adjustment module is used to compare the reconstruction results and sensor observation data in real time. When the comparison difference exceeds a preset threshold, the reconstruction model is adjusted.

[0012] Thirdly, this application also provides a terminal device, which includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the described coal-fired boiler flow field correction method.

[0013] Fourthly, this application also provides a readable storage medium storing a computer program that, when executed on a processor, implements the aforementioned coal-fired boiler flow field correction method.

[0014] The embodiments of this application have the following beneficial effects: This embodiment analyzes and processes the real-time collected operating parameters and sensor measurement data of the coal-fired boiler through a reconstruction model to obtain reconstruction results. The reconstruction results are then used to control and optimize the operation of the coal-fired boiler. Furthermore, the reconstruction model can be adjusted online in real time to ensure that it keeps up with changes in hardware and environment, thereby improving work efficiency and safety. Combined with artificial intelligence technology, it effectively solves the problem of establishing causal relationships for fluid motion. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A schematic flowchart of a flow field correction method for a coal-fired boiler according to an embodiment of this application is shown; Figure 2 This paper illustrates a training process diagram of a reconstructed model according to an embodiment of the present application; Figure 3 The main structure diagram of the physical information neural network according to an embodiment of this application is shown; Figure 4 A schematic diagram of a flow field correction device for a coal-fired boiler according to an embodiment of this application is shown. Detailed Implementation

[0017] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0018] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0019] In the following text, the terms "comprising," "having," and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more combinations thereof. Furthermore, the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0020] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0021] The proposed solution analyzes and processes real-time collected operating parameters and sensor measurement data of a coal-fired boiler using a reconstruction model to obtain reconstruction results. These results are then used to control and optimize the operation of the coal-fired boiler. Furthermore, the reconstruction model can be adjusted online in real time to ensure it keeps pace with changes in hardware and the environment, thereby improving work efficiency and safety. Combined with artificial intelligence technology, this solution effectively addresses the challenge of establishing causal relationships for fluid motion.

[0022] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0023] The following examples illustrate the flow field correction method for this coal-fired boiler.

[0024] Figure 1 A flowchart illustrating a flow field correction method for a coal-fired boiler according to an embodiment of this application is shown. Exemplarily, the flow field correction method for a coal-fired boiler includes the following steps: Step S100: Real-time acquisition of operating condition parameters and sensor measurement data of coal-fired boiler.

[0025] The method in this embodiment is applied to the operation of a coal-fired boiler, and therefore acquires the operating parameters and sensor measurement data of the coal-fired boiler in real time.

[0026] Operating data refers to system-level, parameterized, low-frequency time-series variables of the macroscopic operating status and boundary setting conditions of coal-fired boilers, such as load (MW), furnace negative pressure (Pa), total fuel quantity (t / h), and other related parameters. These parameters can generally be read directly from the control system.

[0027] Sensor measurement data refers to high-fidelity, spatially coordinated local physical quantity measurements acquired in real-time or near real-time at specific physical locations on the boiler. These measurements reflect the actual state of the flow field at a given location and require corresponding sensors to be installed on the coal-fired boiler to acquire them.

[0028] The operating parameters and sensor measurement data obtained above are all real-time data during the operation of the coal-fired boiler. These data can accurately and timely reflect the current flow field conditions inside the coal-fired boiler, thus supporting subsequent flow field reconstruction operations.

[0029] Step S200: Input the operating condition parameters and the sensor measurement data into the trained reconstruction model to obtain the reconstruction result of the global flow field inside the boiler.

[0030] This embodiment provides a reconstruction model. By inputting the operating condition parameters and sensor measurement data obtained in the aforementioned steps into the reconstruction model, the model will output the reconstruction results of the global flow field inside the boiler.

[0031] The model is a physical information neural network model that integrates residual learning. It internally sets an initial flow field. After acquiring operating parameters and sensor measurement data, the model processes the data and outputs low-order variable residuals or flow field residuals. These residuals represent the difference between the current actual flow field and the initial flow field within the model. By calculating the sum of the residuals and the initial flow field, the flow field reconstructed by the model based on these data can be obtained. Figure 3 The diagram shows the main structure of the physical information neural network according to an embodiment of this application. The main structure of the physical information neural network in this embodiment is similar. Figure 3 Regarding the structure, it should be noted that the specific parameters may be adjusted according to the circumstances in practical applications.

[0032] Step S300: Based on the reconstruction result, the operating status of the coal-fired boiler is evaluated to obtain the evaluation result, and the control optimization operation of the coal-fired boiler is performed based on the evaluation result.

[0033] Based on the reconstructed flow field, data on indicators such as flow uniformity, flame stability, nitride formation risk, heating surface safety, burnout performance, and airflow rigidity within the current coal-fired boiler can be directly obtained. These indicators can be directly acquired from the reconstruction results obtained in this embodiment.

[0034] Among them, flow uniformity is used to characterize whether the velocity at the furnace outlet section is uniform; flame stability is used to characterize whether the intensity of the reduction reaction in the furnace is too strong; nitride formation risk is used to characterize the risk of nitride explosion; heating surface safety is used to characterize the tube wall wear rate; burnout performance is used to determine the concentration of unburned carbon in the upper part of the furnace; airflow rigidity is used to characterize the performance of intracranial combustion by fitting the center streamline divergence angle of each burner nozzle, premature jet diffusion, short flame, excessive rigidity, and insufficient entrainment.

[0035] The above indicators can be used to assess the current operating status of the coal-fired boiler. For example, if any indicator deviates from the normal threshold range, the assessment result is considered unstable, triggering an alarm to provide a warning. Adjustments are then made accordingly until the assessment result becomes stable.

[0036] For example, if insufficient flow uniformity is detected, leading to nozzle deformation, in addition to issuing an alarm, the nozzle tilt angle can be adjusted, such as from 15 degrees to 13 degrees, to reduce the nozzle size and minimize flame damage. After adjusting the angle, continue monitoring to see if the flow uniformity reaches a stable threshold. If it does, stop adjusting; if not, further adjustments are needed until stability is achieved.

[0037] Step S400: Compare the reconstruction result and the sensor observation data in real time. When the comparison difference exceeds a preset threshold, adjust the reconstruction model.

[0038] In this embodiment, the reconstruction model will also be updated online. It can be understood that the reconstruction result of this embodiment is entirely based on the reconstruction operation of the reconstruction model. For this purpose, this embodiment will also acquire sensor observation data, which is the flow field sensor observation data obtained from inside the boiler through actual measurement.

[0039] This data is not the actual data during operation, but rather the flow field data of the boiler under various operating conditions obtained by testing the boiler under various constant conditions. These data are considered to be accurate data that have been calibrated.

[0040] When the reconstructed model obtains a set of operating condition parameters and sensor measurement data, it obtains a reconstruction result. If the reconstruction result deviates significantly from the originally recorded sensor observation data, it can be considered that the model reconstruction is not accurate enough. At this time, dynamic correction of the model will be performed to update the model and improve the reconstruction accuracy.

[0041] like Figure 2 As shown in the figure, this embodiment also provides a schematic diagram of the training process of the reconstructed model.

[0042] Step S500: Collect cold state data, hot state data, and online data of the coal-fired boiler as training data.

[0043] When the unit is shut down, on-site tests are conducted on the boiler under cold conditions using experimental instruments, and the data obtained are the cold-state data.

[0044] For example, inspect and measure the structural condition and installation angle of the burner and primary and secondary air nozzles, and measure the area of ​​the primary and secondary air nozzles. Each nozzle should not have severe deformation; otherwise, it will disrupt the flow of the outlet air. The burner tilt angle is measured using an angle gauge. Under the same tilt angle, the deviation between the actual tilt angles of each nozzle should be controlled within ±1.5º. Nozzles exceeding this deviation range should be adjusted.

[0045] Perform operational tests on the large airbox and secondary dampers to confirm that the scale on the damper shaft end matches the actual damper opening; confirm that the scale on the shaft end matches the local indication; confirm that the local indication matches the DCS command, and check the feedback signal. The damper opening can be measured locally using an angle gauge. The measurement results and the feedback value displayed on the DCS are then calibrated. Through inspection, calibration, and adjustment, ensure that the small damper opening displayed on the DCS is consistent with the actual opening. The deviation between the damper openings of burners on the same floor should be controlled within ±5%.

[0046] The burner nozzle velocity is measured, including the velocity of the primary air nozzle, secondary air nozzle, and SOFA air nozzle.

[0047] To ensure the accuracy of the sensor data, when measuring the wind speed at the boiler perimeter, an anemometer is used to measure four to five sets of wind speeds at different locations around the perimeter air nozzles, and the average value is taken as the final data.

[0048] The wind speed at the primary air, secondary air, and SOFA nozzles was measured using an anemometer at the primary and secondary air nozzles. Data was collected from 4-5 representative sensor points at the burner nozzle, and the average value was taken as the final result.

[0049] For wall-mounted wind speed measurement, when measuring the wall-mounted wind speed inside the furnace, primary and secondary air should be introduced according to the principle of equal momentum ratio, and the airflow motion state should be as close to the hot state as possible. The tangential and axial velocities near the furnace wall (0.1~0.2m away from the wall) should be measured at different elevations in the furnace. Generally, a measurement point should be selected every 0.5~1m. When measuring, pay attention to the direction of airflow.

[0050] For tracing the dynamic field inside the furnace, long ribbons are used to indicate the direction of airflow, while short ribbons are used to determine the extent of the light wind zone, recirculation zone, and vortex zone. A ribbon net is used to observe the overall airflow conditions at a specific cross-section.

[0051] Before the test, positioning coordinate lines should be drawn at different heights in the furnace. Generally, a cross line should be drawn at the height of each burner layer to facilitate recording the trajectory of the ribbon.

[0052] During the operation of the coal-fired boiler, data is collected at a preset height along the flame deflector at the boiler furnace outlet according to different load conditions to obtain thermal data.

[0053] The measurement steps for thermal data include measuring the NOx / SO2 / O2 concentration field. For example, at every 0.2-0.5m height, the flue gas component concentration field is tested across the entire cross section at the same height using a grid method. The flue gas component concentration is measured at each measuring point, and the arithmetic mean of the flue gas component concentrations at each grid point is calculated. The result is the actual flue gas component concentration value of that cross section, thereby obtaining the actual thermal data for each single point, single-layer cross section, and multi-layer cross section.

[0054] Flue gas velocity field measurements were conducted at 0.2-0.5m heights across the entire cross-section at the same height using high-temperature Pitot tubes and a grid method. Flue gas velocity was measured at each measuring point, and the arithmetic mean of the flue gas velocities at each grid point was calculated. The result represents the actual flue gas velocity value at that cross-section, thus obtaining the actual thermal data for each single point, single-layer cross-section, and multi-layer cross-section.

[0055] Flue gas temperature field measurements were conducted at 0.2-0.5m heights across the entire cross-section at the same height using a fast-response temperature probe thermocouple and a grid method. Flue gas temperature was measured at each measuring point, and the arithmetic mean of the flue gas temperatures at each grid point was calculated. The result represents the actual flue gas temperature value at that cross-section, thus obtaining the actual thermal data for each single point, single-layer cross-section, and multi-layer cross-section.

[0056] Online historical data collection refers to selecting specific data from the boiler within a predetermined period, such as 6 months, excluding downtime. Data is collected at intervals of 30 seconds to 5 minutes, depending on the actual situation of the unit. The types of data collected include, but are not limited to, boiler load, furnace negative pressure, reheat steam pressure, reheat steam temperature, feedwater flow rate, feedwater temperature, and total fuel consumption. There are many types of data collected here, which will not be listed one by one.

[0057] The abundant data obtained through the above methods ensures broad and accurate data coverage, thus providing a solid foundation for subsequent model training.

[0058] Step S600: Reconstruct the coal-fired boiler to obtain the initial flow field.

[0059] After acquiring the data, the initial flow field is obtained by reconstructing based on the coal-fired boiler. This initial flow field is used as the initial flow field in the reconstructed model that is trained next. It can be considered that the initial flow field acts as an origin, so that the model has an original benchmark for comparison.

[0060] As mentioned earlier, the reconstruction model output in this embodiment is the residual. The sum of the residual and the initial flow field is the actual flow field of the current coal-fired boiler. Therefore, the reconstruction of the initial flow field needs to be close enough to the real situation.

[0061] This embodiment employs CFD (Computational Fluid Dynamics) methods to numerically simulate the furnace and flue region of a coal-fired boiler, obtaining three-dimensional flow field data corresponding to various operating conditions. Based on the boiler's structural characteristics and flow direction, the overall computational domain is divided into multiple sub-regions, each corresponding to functional sections such as the combustion zone, recirculation zone, or flue region. Within each sub-region, the flow field is sliced ​​along the main flow direction or typical cross-sectional direction to construct a local flow field sample set. The flow field samples from each sub-region are then subjected to a POD (Orthogonal Decomposition) model for feature extraction, mapping the original high-dimensional flow field into low-order variable representations.

[0062] Using the existing basis functions and modal coefficients, the initial flow field results are reconstructed using Equation (1), and this initial flow field estimate serves as the baseline input for the subsequent residual learning model.

[0063] (1) In the formula, To reconstruct the flow field, that is, the initial flow field, The time average of the target variable. These are the basis functions of POD. For time coefficient, The number of eigenvalues.

[0064] Step S700: Construct a physical information neural network model that integrates residual learning as the base model of the reconstructed model, and use the initial flow field as the initial flow field of the base model.

[0065] Specifically, a physical information neural network model integrating residual learning is constructed, representing the real flow field as the sum of the initial flow field of the base model and the residual correction: (2) Where q0 is the initial flow field output by the low-fidelity base model, and Δq is the flow field correction to be learned.

[0066] In this embodiment, a PINNs (Physical Information Neural Network) model can be built based on the Nvidia Modulus framework. The main steps include creating the geometry, creating network nodes, setting constraints (including governing equations and boundary conditions), adding validators and monitors, and training. This PINN takes operating parameters, sensor observation data, and spatial coordinates as input, and low-order variable residuals or flow field residuals as output. By learning only the residual corrections, the learning difficulty of the network is effectively reduced, improving the model's convergence speed and generalization ability.

[0067] Step S800: Based on physical constraints and observation assimilation mechanisms, the base model is trained to obtain the reconstructed model.

[0068] Physical constraints and observation assimilation mechanisms are introduced during neural network training. The physical constraints include at least the conservation of mass, momentum, energy, and boundary conditions.

[0069] In the formula, t represents time, ux, uy, and uz represent the velocity components along the x, y, and z directions, respectively, ρ represents density; p is pressure, τxx – τzz are stress components in different directions, ρfx, ρfy, and ρfz are volume forces caused by gravity; h is the total enthalpy of the fluid, T is temperature, and Sh is the enthalpy source term, which includes heat energy generated by radiation, chemical reactions, and interphase heat transfer. , ,and Let represent the time term, convection term, and diffusion term of the i-th chemical substance, respectively. Ri represents the mass change of the i-th chemical substance due to the chemical reaction.

[0070] The deviation between sensor observations and model reconstruction results is used as an observation assimilation constraint to guide the model output to be consistent with the actual operating state.

[0071] like Figure 3 As shown, the loss function for PINNs generally consists of the residuals of the PDE and the residuals of the boundary conditions, where... Figure 3 The weights of each loss have been omitted.

[0072] (7) Here, L represents the weight of each loss, and L is the total loss value. The residual of the partial differential equation is... The boundary condition residual is given.

[0073] For the loss function described above, it is of the MSE form: (8) In the formula, N is the total number of samples. Let be the theoretical value for the i-th sample. This represents the actual value of the i-th sample.

[0074] The method in this embodiment collects operating condition parameters and sensor measurement data in real time during the actual operation of the boiler and inputs them into a trained model to quickly output the reconstruction results of the global flow field inside the boiler. When the deviation between the sensor observation data and the reconstruction results exceeds a preset threshold, the model is dynamically corrected through incremental updates or local parameter adjustments to achieve continuous updates of the coal-fired boiler model. The reconstructed flow field information is then used for boiler operating status assessment, combustion diagnosis, and optimized control.

[0075] The beneficial effects of this invention are as follows: This invention provides a method for reconstructing the flow field and creating a digital twin of a coal-fired boiler based on a physical information neural network. By integrating multi-condition numerical simulation data, flow field partitioning and order reduction representation, and residual learning mechanisms, it achieves rapid reconstruction of the entire internal flow field of a coal-fired boiler while satisfying physical constraints such as the conservation of mass, momentum, and energy. This method can accurately reflect changes in the internal flow and combustion state of the boiler, relying only on limited operating parameters and a small amount of sensor measurement data. It also supports online updating and dynamic correction of the model, thus providing reliable digital support for boiler operation status monitoring, combustion optimization, and safe and stable operation.

[0076] Figure 4 A schematic diagram of a flow field correction device for a coal-fired boiler according to an embodiment of this application is shown. Exemplarily, the flow field correction device for a coal-fired boiler includes: The data acquisition module 10 is used to collect the operating condition parameters and sensor measurement data of the coal-fired boiler in real time. The reconstruction module 20 is used to input the operating condition parameters and the sensor measurement data into the trained reconstruction model to obtain the reconstruction result of the global flow field inside the boiler. The optimization module 30 is used to evaluate the operating status of the coal-fired boiler based on the reconstruction result, obtain the evaluation result, and perform control optimization operations on the coal-fired boiler based on the evaluation result. The adjustment module 40 is used to compare the reconstruction result and the sensor observation data in real time. When the comparison difference exceeds a preset threshold, the reconstruction model is adjusted.

[0077] It is understood that the device in this embodiment corresponds to the coal-fired boiler flow field correction method in the above embodiment, and the options in the above embodiment are also applicable to this embodiment, so they will not be described again here.

[0078] This application also provides a terminal device, exemplary of which includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to enable the terminal device to perform the functions of the various modules in the above-described coal-fired boiler flow field correction method or the above-described coal-fired boiler flow field correction device.

[0079] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including at least one of a Central Processing Unit (CPU), Graphics Processing Unit (GPU), Network Processor (NP), Digital Signal Processor (DSP), Application-Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application.

[0080] The memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory is used to store computer programs, and the processor can execute the computer programs accordingly after receiving execution instructions.

[0081] This application also provides a readable storage medium for storing the computer program used in the aforementioned terminal device.

[0082] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that, in alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0083] In addition, the functional modules or units in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0084] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0085] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method of coal-fired boiler flow field modification, characterized by, include: Real-time acquisition of operating parameters and sensor measurement data of coal-fired boilers; The operating condition parameters and the sensor measurement data are input into the trained reconstruction model to obtain the reconstruction result of the global flow field inside the boiler. Based on the reconstruction results, the operating status of the coal-fired boiler is evaluated to obtain evaluation results, and the control optimization operation of the coal-fired boiler is performed based on the evaluation results. The reconstruction model is adjusted when the difference between the reconstruction result and the sensor observation data is compared in real time and the difference exceeds a preset threshold.

2. The coal-fired boiler flow field modification method of claim 1, wherein, The training method for the reconstructed model includes: Cold-state data, hot-state data, and online data of coal-fired boilers were collected as training data; The initial flow field is obtained by reconstructing the coal-fired boiler; A physical information neural network integrating residual learning is constructed as the base model of the reconstructed model, and the initial flow field is used as the initial flow field of the base model; The base model is trained based on physical constraints and observation assimilation mechanisms to obtain the reconstructed model.

3. The method for correcting the flow field of a coal-fired boiler according to claim 2, characterized in that, The collected cold-state data, hot-state data, and online data of the coal-fired boiler are used as training data, including: During the operation of the coal-fired boiler, data was collected under cold conditions to obtain cold-state data. During the operation of the coal-fired boiler, data is collected at a preset height along the flame deflector at the boiler furnace outlet according to different load conditions to obtain thermal data; Obtain online historical data of the coal-fired boiler during operation within a preset time period.

4. The method for correcting the flow field of a coal-fired boiler according to claim 2, characterized in that, The process of reconstructing the initial flow field of the coal-fired boiler includes: Based on the training data, numerical simulations were performed on the furnace and flue area of ​​a coal-fired boiler to obtain three-dimensional flow field data corresponding to each operating condition. Based on the structural characteristics and flow direction of the boiler, multiple sub-regions are divided. Within each sub-region, the flow field is sliced ​​along the main flow direction or typical cross-section direction to construct a set of local flow field sample features. Feature extraction is performed on the flow field sample set of each sub-region to obtain low-order features. Then, the initial flow field is obtained by reconstructing the flow field using basis functions and the low-order features.

5. The method for correcting the flow field of a coal-fired boiler according to claim 2, characterized in that, The loss function of the reconstructed model includes the partial differential equation residuals and the boundary condition residuals; The expression for the loss function is: ； In the formula, L is the loss value. The residual of the partial differential equation is... The boundary condition residual, The weights for each loss.

6. The method for correcting the flow field of a coal-fired boiler according to claim 1, characterized in that, The step of inputting the operating condition parameters and the sensor measurement data into the trained reconstruction model to obtain the reconstruction result of the global flow field inside the boiler includes: The reconstruction model outputs flow field residual data based on the input operating condition parameters and sensor measurement data. The sum of the flow field residual data and the initial flow field of the reconstruction model is calculated to obtain the reconstruction result of the current global flow field inside the boiler.

7. The method for correcting the flow field of a coal-fired boiler according to claim 1, characterized in that, The control optimization operation based on the evaluation results includes: If the evaluation result is unstable, the operating parameters of the coal-fired boiler are dynamically corrected until the evaluation result is stable.

8. A flow field correction device for a coal-fired boiler, characterized in that, include: The data acquisition module is used to collect operating parameters and sensor measurement data of coal-fired boilers in real time. The reconstruction module is used to input the operating condition parameters and the sensor measurement data into the trained reconstruction model to obtain the reconstruction result of the global flow field inside the boiler. The optimization module is used to evaluate the operating status of the coal-fired boiler based on the reconstruction results, obtain evaluation results, and perform control optimization operations on the coal-fired boiler based on the evaluation results. An adjustment module is used to compare the reconstruction results and sensor observation data in real time. When the comparison difference exceeds a preset threshold, the reconstruction model is adjusted.

9. A terminal device, characterized in that, The terminal device includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the coal-fired boiler flow field correction method according to any one of claims 1-7.

10. A readable storage medium, characterized in that, It stores a computer program that, when executed on a processor, implements the coal-fired boiler flow field correction method according to any one of claims 1-7.

Images