Temperature field prediction model training method and temperature field reconstruction method

Abstract

The invention belongs to the field of data processing and deep learning, particularly relates to a temperature field prediction model training method and a temperature field reconstruction method, and aims to solve the problem of insufficient boiler furnace temperature measurement accuracy. The temperature field prediction model training method comprises the steps that a geometric model of a target boiler is obtained, grid division is conducted on the geometric model, and grids of the geometric model correspond to grid data used for positioning; acquiring temperature field data of the target boiler based on the fluid numerical simulation model of the target boiler; a target data set constructed by the grid data and the temperature field data is used as a training sample of the unit model, and a node temperature updating model and an edge feature updating model included in the unit model are respectively iterated for training; and synchronizing the model parameters of the trained unit model to the plurality of layers, and generating a temperature field prediction model composed of the plurality of layers. According to the invention, high-precision reconstruction of the internal temperature field of the target boiler can be realized through a small amount of actually measured temperature data.

Description

Technical Field

[0001] This application belongs to the field of data processing and deep learning, and specifically involves a temperature field prediction model training method and a temperature field reconstruction method. Background Art

[0002] Boilers are essential industrial equipment, and their operational efficiency and safety directly impact system energy efficiency, safety, and operational stability. During boiler operation, furnace temperature is a crucial parameter characterizing combustion status, fuel efficiency, and safe operation. It can be used to analyze core boiler characteristics such as coking, high-temperature corrosion, nitrogen oxide (NOx) generation, and heat transfer efficiency, ultimately optimizing boiler energy efficiency, environmental protection, and safety. Therefore, comprehensive, real-time, and accurate monitoring of boiler furnace temperature is crucial.

[0003] Currently, the commonly used methods for monitoring boiler furnace temperature mainly include thermocouple temperature measurement, infrared temperature measurement, acoustic wave temperature measurement, etc., or emerging methods such as furnace simulation and temperature field prediction through deep learning.

[0004] However, traditional boiler furnace temperature monitoring methods have insufficient measurement accuracy, incomplete monitoring, and are unable to perform real-time 3D monitoring. Existing numerical calculation and deep learning methods have low prediction accuracy and require a large amount of measured data as a basis. Summary of the Invention

[0005] In order to solve the above-mentioned problem in the prior art, namely, the problem of insufficient accuracy in boiler furnace temperature measurement, the first aspect of this application proposes a temperature field prediction model training method, comprising:

[0006] Obtaining a geometric model of the target boiler and performing mesh division on the geometric model, wherein the mesh of the geometric model corresponds to mesh data used for positioning;

[0007] Based on the fluid numerical simulation model of the target boiler, the temperature field data of the target boiler is obtained;

[0008] The target data set constructed by the grid data and the temperature field data is used as the training sample of the unit model, and the node temperature update model and the edge feature update model included in the unit model are trained iteratively respectively;

[0009] The model parameters of the trained unit model are synchronized to multiple layers to generate a temperature field prediction model composed of multiple layers, wherein the structure of the layers is the same as that of the unit model.

[0010] As a preferred embodiment, the node temperature update model and the edge feature update model included in the unit model are iteratively trained respectively, including:

[0011] Get the initial node model and initial edge model of the mesh, where

[0012] The input data of the initial node model include the initial node temperature and edge features. The output data of the initial node model is the updated node temperature. The edge features are the initial edge features or the updated edge features.

[0013] The input data of the initial edge model include the initial edge features and the node features corresponding to the nodes connected by the edge. The output data is the updated edge features, where the node features include the grid 3D coordinates corresponding to the node, the measured temperature, the initial temperature, the furnace pressure and the edge features.

[0014] The updated node temperature of the initial node model at the first moment is used as the initial node temperature at the second moment, and the initial node model is iterated to obtain a node temperature update model;

[0015] The updated edge features of the initial edge model at the first moment are used as the initial edge features at the second moment, and the initial edge model is iterated to obtain an updated edge feature model;

[0016] The node-edge critical matrix of the grid is obtained, and the node-edge critical matrix, the node temperature update model, and the edge feature update model are coupled to generate a unit model. The node-edge critical matrix is used to characterize the correspondence between the nodes and edges of the grid.

[0017] As a preferred implementation, when the edge feature update model converges, the target edge feature output by the edge feature update model is obtained as the edge feature in the input data of the node temperature update model.

[0018] As a preferred embodiment, obtaining the temperature field data of the target boiler based on the fluid numerical simulation model of the target boiler includes:

[0019] Based on the target boiler's fluid numerical simulation model, multiple boiler furnace combustion simulation data are obtained. The simulation data includes boiler geometry parameters, mesh data, and boiler furnace internal temperature field data. The simulation data is used to simulate the temperature field distribution under different furnace types and different operating conditions.

[0020] Perform data enhancement on the temperature field data inside the boiler furnace to obtain enhanced temperature field data;

[0021] The enhanced temperature field data is spliced with other data in the simulation data to generate the temperature field data of the target boiler.

[0022] As a preferred embodiment, based on the fluid numerical simulation model of the target boiler, the simulation data of the furnace combustion of multiple boilers is obtained, including:

[0023] Based on fluid numerical modeling and simulation, build the fluid, combustion and heat transfer model in any boiler furnace;

[0024] Perform model solution calculations on fluid, combustion and heat transfer models to obtain simulation calculation data sets as simulation data for combustion in any boiler furnace.

[0025] As a preferred embodiment, data enhancement is performed on the temperature field data inside the boiler furnace to obtain enhanced temperature field data, including:

[0026] The intermediate iteration results of the simulation data in any iteration process are obtained as enhanced temperature field data.

[0027] As a preferred embodiment, data enhancement is performed on the temperature field data inside the boiler furnace to obtain enhanced temperature field data, further comprising:

[0028] Get the measured temperature at a random node position;

[0029] The temperatures at random node positions in the intermediate iteration results are replaced with the measured temperatures to obtain enhanced temperature field data.

[0030] In a second aspect of the present application, a temperature field reconstruction method is proposed, which is characterized by comprising:

[0031] Construct a measured data set based on the grid data and local measured temperature of the target boiler;

[0032] Inputting the measured data set into the temperature field prediction model to obtain the temperature field prediction result of the target boiler, wherein the temperature field prediction model is trained using the temperature field prediction model training method described above;

[0033] Based on the target measured temperature at the preset position and the temperature field prediction results, the temperature field of the target boiler is reconstructed.

[0034] As a preferred embodiment, based on the target measured temperature at the preset position and the temperature field prediction result, the temperature field of the target boiler is reconstructed, including:

[0035] Extracting the local predicted temperature at a preset location from the temperature field prediction results;

[0036] When the difference between the local predicted temperature and the target measured temperature is less than the deviation threshold, the temperature field prediction result is used as the temperature field reconstruction result of the target boiler;

[0037] When the difference between the local predicted temperature and the target measured temperature is greater than or equal to the deviation threshold, the temperature field prediction result is re-input into the temperature field prediction model for temperature field prediction until the difference between the local predicted temperature extracted from the temperature field prediction result and the target measured temperature is less than the deviation threshold. The output result of the temperature field prediction model is used as the temperature field reconstruction result of the target boiler.

[0038] As a preferred embodiment, the method further includes: obtaining measured data sets at multiple historical moments to form a historical data set;

[0039] Obtain forecast datasets based on historical datasets and time series forecasting models;

[0040] The predicted data set is input into the temperature field prediction model to obtain the predicted temperature field reconstruction result.

[0041] Beneficial effects of this application:

[0042] (1) First, the geometric model of the target boiler is obtained and the geometric model is meshed, wherein the mesh of the geometric model corresponds to the mesh data used for positioning. The three-dimensional shape of the target boiler is displayed by geometric modeling of the target boiler, and the collective model is cut by meshing to facilitate data processing in smaller units. Furthermore, based on the fluid numerical simulation model of the target boiler, the temperature field data of the target boiler is obtained. The simulation data of the target boiler is obtained as temperature field data through the simulation model, which can make up for the deficiency that the measured temperature data cannot be fully detected. The target data set constructed by the mesh data and the temperature field data is used as the training sample of the unit model. The node temperature update model and the edge feature update model included in the unit model are iteratively trained respectively. The unit model is trained by iteratively updating the node temperature and edge features, which enables the unit model to simulate the iterative simulation process of the temperature field and learn the simulation process of the temperature field of the target boiler. The model parameters of the trained unit model are synchronized to multiple layers to generate a temperature field prediction model composed of multiple layers. The distributed training method is used to reduce the data processing volume of the model training and improve the efficiency of the model training. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Other features, objects and advantages of the present application will become more apparent upon reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0044] Figure 1 This is a flow chart of a temperature field prediction model training method provided by the first embodiment of the present application;

[0045] Figure 2 This is a flowchart of the steps of a unit model training process provided in the first embodiment of the present application;

[0046] Figure 3 This is an example diagram of a unit model provided in the first embodiment of the present application;

[0047] Figure 4 This is a flow chart of a temperature field reconstruction method provided in the second embodiment of the present application;

[0048] Figure 5 This is an example diagram of a geometric model of a target boiler provided by an embodiment of the present application;

[0049] Figure 6 This is a system block diagram of the temperature field prediction model training system provided in the third embodiment of the present application;

[0050] Figure 7 is a system block diagram of a temperature field reconstruction system provided in a fourth embodiment of the present application;

[0051] Figure 8 It is a structural diagram of the boiler furnace temperature field reconstruction system provided in the fifth embodiment of the present application. DETAILED DESCRIPTION

[0052] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the relevant invention are shown in the accompanying drawings.

[0053] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0054] At present, common boiler furnace temperature monitoring technologies mainly include the following:

[0055] 1. Thermocouple temperature measurement technology:

[0056] A thermocouple is a traditional temperature sensor that is usually installed at a specific location inside a boiler furnace. It measures the electromotive force generated by the temperature difference to infer the temperature.

[0057] Advantages: simple measurement principle, low cost, suitable for medium and low temperature areas.

[0058] shortcoming:

[0059] (1) Limited high temperature resistance: Thermocouples are easily damaged in high temperature environments and have a short service life.

[0060] (2) Spatial limitations: Thermocouples can only measure the temperature at fixed points and cannot cover the temperature distribution inside the furnace.

[0061] (3) Slow response speed: When the temperature changes rapidly, the thermocouple may not be able to reflect the actual situation in time.

[0062] 2. Infrared temperature measurement technology:

[0063] Infrared temperature measurement is based on the principle that the infrared energy radiated by an object is related to its temperature, and the temperature is inferred by detecting the radiation intensity.

[0064] Advantages: Non-contact temperature measurement, suitable for high temperature environments, and able to monitor a wide range of areas.

[0065] shortcoming:

[0066] (1) Susceptible to flue gas interference: Infrared sensors are easily affected by flue gas, dust and water vapor when the boiler is running, resulting in increased measurement errors.

[0067] (2) Complex calibration: Infrared sensors need to be calibrated regularly and have high requirements for monitoring angles and positions.

[0068] 3. Fiber optic temperature measurement technology:

[0069] Fiber optic temperature measurement uses the optical properties of optical fiber to infer temperature distribution by measuring the attenuation or scattering of transmitted light.

[0070] Advantages: High temperature resistance, anti-electromagnetic interference, suitable for distributed temperature monitoring in complex environments.

[0071] shortcoming:

[0072] (1) High cost: The manufacturing and maintenance costs of optical fiber temperature measurement equipment are high, making it unsuitable for large-scale promotion.

[0073] (2) Reliability issues: Optical fibers are easily damaged in high temperature, high pressure and corrosive gas environments.

[0074] 4. Sound temperature measurement technology:

[0075] Sound temperature measurement is an emerging technology that uses the principle that the propagation speed of sound waves in high-temperature gas changes with temperature to estimate the furnace temperature by measuring the propagation time of sound waves.

[0076] Advantages: It can realize non-contact temperature measurement, is suitable for high temperature and harsh environments, and can provide distributed temperature data.

[0077] shortcoming:

[0078] (1) Affected by background noise: combustion noise, mechanical vibration and steam noise in the boiler may interfere with the acoustic signal and affect the measurement accuracy.

[0079] (2) Complex signal processing: The propagation path of sound waves in the furnace is complex and may be affected by airflow disturbances. Complex algorithms are required for signal restoration and temperature calculation.

[0080] (3) High equipment calibration requirements: Acoustic temperature measurement equipment needs to be calibrated regularly to ensure measurement accuracy.

[0081] 5. Furnace simulation:

[0082] Furnace simulation is a fluid numerical calculation method that discretizes the Navier-Stokes equations according to the furnace geometry discrete grid. Finite element, finite volume or finite difference algorithms are used to calculate core field data such as temperature, pressure and flow rate at grid nodes or inside the grid to achieve the goal of combustion inside the reaction furnace.

[0083] shortcoming:

[0084] (1) The computational complexity is extremely high: for conventional furnaces, the number of grids is between 3 million and 10 million, and the computation time is in the order of hours, which makes it impossible to apply to real-time operation on site.

[0085] (2) Unable to adjust in real time: The fluid numerical calculation model is designed with a large number of empirical parameters, which cannot be adjusted immediately according to real-time data;

[0086] 6. Temperature field prediction based on deep learning:

[0087] Predicting the physical field based on deep learning is an emerging temperature field prediction calculation method. By accumulating massive data, establishing the relationship between boundary conditions and temperature fields based on deep learning models, and predicting the temperature field in real time, the goal of real-time prediction can be achieved.

[0088] shortcoming:

[0089] (1) Difficult to correct in real time: The current temperature field reconstruction algorithms based on deep learning methods are all static supervised learning methods, which establish the relationship between boundary parameters, auxiliary feature variables and target temperature fields, and do not effectively combine local measured temperature field data to improve the overall model prediction effect.

[0090] (2) There is no spatial correlation between temperature field data points: The current training process is a single-point prediction and evaluation of temperature field data, which is difficult to reflect the spatial relationship of temperature field measurement points and easily causes distortion of the spatial characteristics of the temperature field.

[0091] The present application provides a temperature field prediction model training method and a temperature field reconstruction method, which can achieve high-precision reconstruction of the internal temperature field of the target boiler through a small amount of measured temperature data, while taking into account the spatial characteristics and accuracy of the temperature field.

[0092] In order to more clearly explain the temperature field prediction model training method of this application, the following Figure 1 Each step in the embodiments of the present application is described in detail.

[0093] The temperature field prediction model training method of the first embodiment of the present application includes steps S110 to S140, each of which is described in detail as follows:

[0094] Step S110 , obtaining a geometric model of the target boiler and performing mesh division on the geometric model, wherein the meshes of the geometric model correspond to mesh data used for positioning.

[0095] Optionally, the geometric model is three-dimensional so as to fully reflect the three-dimensional distribution of the temperature field in the furnace of the target boiler.

[0096] As a possible implementation, a collective model is established by using fluid simulation software, for example, ANSYS Fluent is used to establish a geometric model of the target boiler.

[0097] It should be noted that the structures of boilers of different types may be different, and it is necessary to establish an adaptive geometric model for the target boiler for temperature field reconstruction.

[0098] Furthermore, the geometric model is meshed using a meshing tool that supports Adaptive Mesh Refinement (AMR), such as the HyperMesh tool.

[0099] It should be noted that AMR is a technology that dynamically adjusts the density of the computational grid, aiming to improve accuracy in critical areas by locally refining the grid while maintaining coarse granularity in non-critical areas to save computing resources.

[0100] The meshing tool is used to divide the geometric model into multiple meshes, where each mesh is a three-dimensional geometric block.

[0101] It should be noted that when using a meshing tool for meshing, the mesh type and density need to be determined based on the actual operating conditions of the target boiler to determine the optimal combination of computational complexity and accuracy, and balance computational complexity and accuracy.

[0102] Optionally, a structured grid or an unstructured grid may be used in the grid division process. For example, a structured grid may use a hexahedron structure, while an unstructured grid may use a tetrahedron structure.

[0103] Optionally, the grid data may be grid three-dimensional coordinates that can be used to locate nodes within the grid.

[0104] Step S120 : acquiring temperature field data of the target boiler based on the fluid numerical simulation model of the target boiler.

[0105] Optionally, based on the fluid numerical simulation model of the target boiler, simulation data of combustion in multiple boiler furnaces are obtained: based on fluid numerical modeling and simulation, fluid, combustion and heat transfer models in any boiler furnace are constructed; model solution calculations are performed on the fluid, combustion and heat transfer models to obtain a simulation calculation data set as simulation data for combustion in any boiler furnace, wherein the simulation data includes boiler furnace geometric parameters, mesh data, and boiler furnace internal temperature field data, and the simulation data is used to simulate the temperature field distribution under different furnace types and different operating conditions.

[0106] In the embodiments of the present application, the physical model includes but is not limited to a thermodynamic property model, a turbulence model (such as k-ε, k-omega and related extended models, etc.), a combustion chemical reaction kinetics model (based on the chemical composition of the fuel) and a radiation heat transfer model (P-1 or DOM model, etc.), a combustion model, etc.

[0107] In the embodiments of the present application, boundary conditions are set and processed, such as primary air volume or flow rate, secondary air volume or flow rate, fuel feed rate or flow rate, steam-water side heat exchange boundary, etc. The boundary condition setting needs to be expanded based on the measured data.

[0108] As an example, the measured data is detection data of a temperature, pressure and flow sensor acquired by a distributed control system.

[0109] Furthermore, the simulation data is enhanced to expand the simulation data set.

[0110] Optionally, obtain simulation data of combustion in multiple boiler furnaces; perform data enhancement on the temperature field data inside the boiler furnace to obtain enhanced temperature field data; and splice the enhanced temperature field data with other data in the simulation data to generate temperature field data of the target boiler.

[0111] As an example, an intermediate iteration result of the simulation data in any iteration process is obtained as enhanced temperature field data.

[0112] As another example, the measured temperature at a random node position is obtained; the temperature at the random node position in the intermediate iteration result is replaced with the measured temperature to obtain enhanced temperature field data.

[0113] As another example, the initial temperature field is randomly initialized. The random initialization strategy includes but is not limited to: uniform initialization of random values, random initialization with a given random distribution, etc., to improve the generalization and iteration capability of the model's random initialization.

[0114] As another example, the initialization strategy of the intermediate iteration results of the initial temperature field: multiple groups of simulation iteration intermediate calculation results are used to initialize the temperature field to approximately simulate the combustion initialization effects at different time nodes.

[0115] As another example, random sampling of the converged result space iteratively uses random noise to simulate the measured temperature field, avoiding the input missing problem caused by the lack of measured temperature field data; at the same time, large sample space sampling is used to increase the model's generalization ability for different temperature sensor installation locations.

[0116] Step S130 : Using the target data set constructed by the grid data and the temperature field data as a training sample of the unit model, the node temperature update model and the edge feature update model included in the unit model are trained by iteration respectively.

[0117] Optional, combined Figure 2 Flowchart of the steps to train the unit model:

[0118] Step S131: Obtain the initial node model and initial edge model of the grid.

[0119] Among them, the input data of the initial node model includes the initial node temperature and edge features, the output data of the initial node model is the updated node temperature, and the edge features are the initial edge features or the updated edge features; the input data of the initial edge model includes the initial edge features and the node features corresponding to the nodes connected by the edges, and the output data is the updated edge features.

[0120] Optionally, the input data of the initial node model is a 2nd-order tensor, with the tensor dimensions being the number of nodes N v and the number of node features N e , that is, the tensor is an N v ×N e The node characteristics are the initial node temperatures corresponding to the grid nodes and various temperature-related parameters during the boiler furnace combustion process, including at least the initial node temperatures and edge characteristics, which include but are not limited to the three-dimensional coordinates of the nodes, whether the temperature is measured, the measured temperature value, the initial node temperature, the furnace pressure, the furnace components, whether it is a boundary, the boundary type, the boundary amount, the edge characteristics and other variables in the embodiment of the present application.

[0121] Among them, the initial node temperature can be a randomly initialized temperature, a uniform temperature value determined based on experience, or a predicted temperature value at the previous moment; the edge feature is an edge implicit space feature; the boundary type is a Diriclet boundary type (represented by 0 in the node feature) and a Neumann boundary type (represented by 1 in the node feature); the boundary quantity is the quantitative data of the operating parameters, such as fuel velocity, heat transfer density, fuel composition, etc.; the edge feature is a combination of the edge features of each edge related to the node.

[0122] In an embodiment of the present application, the model structure of the initial node model is a graph neural network (GNN), and the matrix composed of the node features of each node in the grid is input into the initial node model as input data. Batch normalization and fully connected layer processing are performed through the initial node model to output the updated node temperature, wherein the updated node temperature is a prediction of the initial node temperature, and the loss function is trained by the difference between the updated node temperature and the true label, wherein the loss function can be mean square error loss, mean absolute error loss, etc.

[0123] Optionally, the input of the edge model is edge features and node combination features. The edge features include but are not limited to edge vectors, edge lengths, adjacent node feature vectors, and the ratio of edge length to boiler feature length. The node combination feature is a combination of node features corresponding to the nodes connected by the edge where the edge feature is located.

[0124] In an embodiment of the present application, the model structure of the edge model is also a graph neural network (GNN). The edge features and node combination features are merged as input data and input into the edge model. After processing through the fully connected layer, the updated edge features are output. The updated edge features are predictions of the initial edge features. The loss function is trained by the difference between the updated edge features and the true labels. The loss function can be mean square error loss, mean absolute error loss, etc.

[0125] It should be noted that the real label is the simulation data obtained by performing furnace combustion simulation on the target boiler.

[0126] Step S132 : Using the updated node temperature of the initial node model at the first moment as the initial node temperature at the second moment, iterating the initial node model to obtain an updated node temperature model.

[0127] Optionally, the updated node temperature of each node of the grid at the first moment replaces the corresponding initial node temperature as the initial node temperature at the second moment, and a new matrix is formed with other node features as the input data of the initial node model. The loss function of the initial node model is continued to be trained, and the initial node model is iterated to obtain the node temperature update model.

[0128] Step S133 : Using the updated edge features of the initial edge model at the first moment as the initial edge features at the second moment, iterating the initial edge model to obtain an updated edge feature model.

[0129] Optionally, the updated edge features of each node of the grid at the first moment replace the corresponding initial edge features as the initial edge features at the second moment, and are merged with the corresponding node combination features to form new input data. The loss function of the initial edge model is continued to be trained, and the initial edge model is iterated to obtain the edge feature update model.

[0130] Step S134 , obtaining the node-edge critical matrix of the grid, and coupling the node-edge critical matrix, the node temperature update model, and the edge feature update model to generate a unit model, wherein the node-edge critical matrix is used to characterize the correspondence between nodes and edges.

[0131] Optionally, obtain the corresponding relationship between each node and each edge in the grid. If edge k is not connected to node j, the corresponding relationship coefficient M kj =0, otherwise M kj is a non-zero learnable relationship coefficient, and the node-edge critical matrix is composed of the relationship coefficients in the grid

[0132] Furthermore, the node-edge critical matrix, node temperature update model, and edge feature update model are coupled to generate a unit model, which can be expressed by the following formula:

[0133] in, represents the i-th unit model, x i represents the input data of the i-th unit model, represents the initial node model, represents a composite function, represents the number of iterations of the initial node model, Represents iteration The node temperature update model after the times, represents the initial edge model, x' i represents the input data of the initial edge model, represents the number of iterations of the initial edge model, Represents iteration The edge features after this time update the model.

[0134] As an example, Figure 3 In the example unit model shown, the initial edge model can be iteratively updated. Once the edge feature update model converges, the target edge feature output by the edge feature update model is obtained and used as the edge feature in the input data of the node temperature update model. The output of the updated edge feature update model is used as a parameter in the node feature of the node temperature update model, and the node temperature update model is iteratively updated.

[0135] Step S140 , synchronizing the model parameters of the trained unit model to multiple layers, generating a temperature field prediction model composed of multiple layers, wherein the structure of the layers is the same as that of the unit model.

[0136] Optionally, the result parameters of the trained unit model are copied to multiple layers with the same structure as the unit model to generate a temperature field prediction model composed of multiple layers, which can further iterate the unit model.

[0137] The temperature field prediction model structure is designed as a multi-layer iterative structure with the same structure in each layer. By copying the unit model to different layers, the iteration capability of the unit model can be improved and the model convergence efficiency can be improved.

[0138] As a possible implementation method, the model composed of multiple layers can be continuously trained until convergence to obtain a temperature field prediction model.

[0139] In an embodiment of the present application, a geometric model of a target boiler is first obtained and the geometric model is meshed, wherein the mesh of the geometric model corresponds to mesh data for positioning. The three-dimensional shape of the target boiler is displayed by geometric modeling of the target boiler, and the set model is cut by meshing to facilitate data processing in smaller units. Furthermore, the temperature field data of the target boiler is obtained based on the fluid numerical simulation model of the target boiler. The simulation data of the target boiler is obtained as temperature field data through the simulation model, which can make up for the deficiency that the measured temperature data cannot be fully detected. The target data set constructed by the mesh data and the temperature field data is used as a training sample of the unit model. The node temperature update model and the edge feature update model included in the unit model are iteratively trained respectively. The training of the unit model is realized by iteratively updating the node temperature and edge features, which enables the unit model to simulate the iterative simulation process of the temperature field and learn the simulation process of the temperature field of the target boiler. The model parameters of the trained unit model are synchronized to multiple layers to generate a temperature field prediction model composed of multiple layers. The distributed training method is used to reduce the data processing volume of the model training and improve the efficiency of the model training.

[0140] See also Figure 4 The second embodiment of the present application provides a temperature field reconstruction method based on the above-mentioned temperature field prediction model training method, and the specific steps include:

[0141] Step S410: constructing a measured data set based on the grid data and the local measured temperature of the target boiler.

[0142] Optionally, a local measured temperature is obtained at a sampling point selected for the target boiler. For example, the local measured temperature can be obtained through a sensor, and the local measured temperature is correspondingly spliced with the grid data to generate a measured data set.

[0143] As a possible implementation, the measured data set may also include node three-dimensional coordinates, whether the temperature is measured, the measured temperature value, initial node temperature, furnace pressure, furnace composition, whether it is a boundary, boundary type, boundary quantity, edge characteristics and other variables.

[0144] Step S420: Input the measured data set into the temperature field prediction model to obtain the temperature field prediction result of the target boiler.

[0145] Optionally, a measured data set consisting of the temperatures actually measured on the target boiler is input into the temperature field prediction model, so as to obtain a more accurate temperature condition of the target boiler at the current moment.

[0146] It can be understood that the temperature field prediction result is a temperature prediction of the entire geometric model of the target boiler, and can describe the temperature distribution inside the target boiler.

[0147] Step S430 : reconstructing the temperature field of the target boiler based on the target measured temperature at the preset position and the temperature field prediction result.

[0148] Optionally, a target measured temperature at a preset position is obtained through a sensor, and the target measured temperature is used to verify and correct the temperature field prediction result to obtain a temperature field reconstruction result.

[0149] In an embodiment of the present application, a local predicted temperature of a preset position is extracted from the temperature field prediction result; when the difference between the local predicted temperature and the target measured temperature is less than a deviation threshold, the temperature field prediction result is used as the temperature field reconstruction result of the target boiler; when the difference between the local predicted temperature and the target measured temperature is greater than or equal to the deviation threshold, the temperature field prediction result is re-input into the temperature field prediction model for temperature field prediction, until the difference between the local predicted temperature extracted from the temperature field prediction result and the target measured temperature is less than the deviation threshold, and the output result of the temperature field prediction model is used as the temperature field reconstruction result of the target boiler.

[0150] By comparing the local prediction results with the target measured temperature, local sampling and detection of the temperature field prediction results can be performed. When the difference between the local predicted temperature and the target measured temperature is less than the deviation threshold, the temperature field prediction result is more accurate than the measured temperature and can be used as the temperature field reconstruction result of the entire target boiler. When the difference between the local predicted temperature and the target measured temperature is greater than or equal to the deviation threshold, the temperature field prediction result is not ideal. At this time, the temperature field prediction result is used as the initial temperature field, the temperature field prediction model is re-called, a new temperature field prediction result is obtained, and the comparison with the target measured temperature is continued until the difference between the local predicted temperature extracted from the temperature field prediction result and the target measured temperature is less than the deviation threshold. It is determined that the accuracy of the temperature field prediction result is within the error range, and the output result of the temperature field prediction model is used as the temperature field reconstruction result of the target boiler.

[0151] In an embodiment of the present application, the temperature field prediction model can be called multiple times, so the initial temperature field can be the local measured temperature of the target boiler, or it can be the temperature field prediction result when the difference between the local predicted temperature and the target measured temperature is greater than or equal to the deviation threshold.

[0152] Furthermore, the temperature field prediction model can also establish a time series prediction model for the model input parameters to predict the dynamic change trend of future model data parameters, predict the dynamic changes of the target boiler temperature field in the future time period, identify possible combustion anomalies or local high temperature risks, and issue early warnings or plan prevention strategies in advance.

[0153] As an example, measured data sets at multiple historical moments are obtained to form a historical data set; a predicted data set is obtained based on the historical data set and a time series prediction model; the predicted data set is input into a temperature field prediction model to obtain a predicted temperature field reconstruction result.

[0154] As an example, a 300MW subcritical pressure, single-intermediate reheat, single-furnace, natural circulation, balanced ventilation, solid slag removal, drum-type, four-corner tangential coal-fired boiler is used as the target boiler. The boiler is pre-installed with a real-time furnace temperature measurement system based on high-temperature infrared transmitters and receivers. 4x4=16 points of temperature data are measured on the 60m and 75m sections of the furnace.

[0155] First, the numerical simulation of the boiler furnace is carried out using ANSYS Fluent 2025 and OpenFOAM V12. The target boiler is geometrically modeled and meshed: the boiler furnace model is established using ANSYS Fluent, using a structured grid with a total grid size of 3 million and an edge size of approximately 9 million. Figure 5 shown.

[0156] The physical model is set as: turbulence model: RNG k-ε model; combustion model: non-premixed combustion model (EDM); radiation heat transfer model: P-1 radiation model; chemical reaction mechanism: based on the chemical reaction mechanism of coal powder combustion.

[0157] The boundary conditions are set as follows: primary air speed: 10-15m / s; secondary air speed: 6-20m / s; fuel flow: 4000kcal pulverized coal, 100-200t / h; water-cooled wall temperature: 200-300 degrees Celsius; superheater / reheater wall temperature: 400-560 degrees Celsius; solution and calculation: SIMPLE algorithm and first-order format are used for pressure-velocity coupling solution; time step size is 0.001s, calculation steps are 20,000, and a single case can converge in about 6 hours to obtain the calculation results of the target boiler simulation data.

[0158] For the above boundary conditions, a single sampling rate of 5 was used, resulting in 3125 simulation cases. Each case stored 100 steps of intermediate and final convergence data. Without measured temperature data, the simulation was performed at the sensor installation location using the convergence data with noise added.

[0159] The temperature field prediction model is designed as follows: the number of iterations of the unit model is 5 times, the number of iterations of the node model and the edge model are 3 times respectively, the number of grid nodes is 3 million, and the number of edges is about 9 million.

[0160] The target boiler's temperature field dataset was fed into the temperature field prediction model. The CFD simulation took six hours. The ARMGNet:8xA100 inference server took 3 to 5 seconds to infer. The prediction accuracy was <3%.

[0161] In practical applications, there are the following beneficial effects:

[0162] Improved dynamic response capabilities: achieve temperature prediction within seconds and quickly identify abnormal combustion areas; enhanced safety: avoid furnace overheating and increase equipment life; energy-saving optimization: adjust the air-coal ratio and reduce coal consumption by 1.5%.

[0163] The embodiment of the present application first constructs a measured data set based on the grid data and local measured temperatures of the target boiler; inputs the measured data set into the temperature field prediction model to obtain the temperature field prediction result of the target boiler; based on the target measured temperature at the preset position and the temperature field prediction result, the temperature field of the target boiler is reconstructed, and the temperature field prediction model can realize the autoregression of the unit model data, and then correct the data, and obtain more accurate temperature data while reflecting the spatial relationship to complete the correction of the initial temperature field and obtain a relatively accurate temperature field; finally, the temperature field prediction result is verified by a small amount of local measured temperatures, and high-precision reconstruction of the internal temperature field of the target boiler can be achieved through a small amount of measured temperature data, providing theoretical support for combustion optimization, equipment safety management and improvement of operation efficiency.

[0164] Although the various steps in the above embodiment are described in the above-mentioned order, those skilled in the art will understand that in order to achieve the effect of this embodiment, different steps do not have to be executed in such an order. They can be executed simultaneously (in parallel) or in a reverse order. These simple changes are within the scope of protection of this application.

[0165] See also Figure 6 The temperature field prediction model training system of the third embodiment of the present application includes: a geometric model acquisition module 610, a temperature field data acquisition module 620, a unit model training module 630 and a temperature field prediction model generation module 640.

[0166] A geometric model acquisition module 610 is used to acquire a geometric model of a target boiler and perform mesh division on the geometric model, wherein the mesh of the geometric model corresponds to mesh data used for positioning;

[0167] The temperature field data acquisition module 620 is used to acquire the temperature field data of the target boiler based on the fluid numerical simulation model of the target boiler;

[0168] The unit model training module 630 is configured to use the target data set constructed from the grid data and the temperature field data as a training sample for the unit model, and to iteratively train the node temperature update model and the edge feature update model included in the unit model.

[0169] The temperature field prediction model generation module 640 is used to synchronize the model parameters of the trained unit model to multiple layers to generate a temperature field prediction model composed of multiple layers, wherein the structure of the layers is the same as that of the unit model.

[0170] See also Figure 7 The temperature field reconstruction system of the fourth embodiment of the present application includes: a measured data set construction module 710, a temperature field prediction result acquisition module 720 and a temperature field reconstruction module 730.

[0171] A measured data set construction module 710 constructs a measured data set based on the grid data and the local measured temperature of the target boiler;

[0172] The temperature field prediction result acquisition module 720 inputs the measured data set into the temperature field prediction model to obtain the temperature field prediction result of the target boiler, wherein the temperature field prediction model is trained using the temperature field prediction model training method described above;

[0173] The temperature field reconstruction module 730 reconstructs the temperature field of the target boiler based on the target measured temperature at the preset position and the temperature field prediction result.

[0174] See also Figure 8 The temperature field reconstruction system of the fifth embodiment of the present application includes a sensor, a communication card, a data server and an inference server.

[0175] In the embodiment of the present application, the sensor is used to measure the furnace temperature of the target boiler.

[0176] Among them, the furnace temperature sensor can directly or indirectly measure the furnace internal or wall temperature based on thermocouples, infrared, sound waves, etc.

[0177] The communication card is used to transmit the temperature data measured by the sensor to the data server.

[0178] The communication card is a communication card of a distributed control system (DCS).

[0179] The data server is used to receive temperature data; the data server is also used to construct a measured data set; the data server is also used to transmit the measured data set to the inference server; the data server is also used to receive the temperature field prediction results from the inference server, and reconstruct the temperature field of the target boiler based on the temperature field prediction results and the target measured temperature.

[0180] The inference server stores a temperature field prediction model, which is used to receive a measured data set from the data server, execute the temperature field prediction model based on the measured data set to obtain a temperature field prediction result, and return the temperature field prediction result to the data server.

[0181] Furthermore, the data server can also obtain temperature data alarm data of key areas based on the temperature field reconstruction results and transmit it through the communication card; the data server can also store the continuous data of the target boiler in the local database.

[0182] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the system described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0183] It should be noted that the boiler furnace temperature field reconstruction system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present application can be decomposed or combined. For example, the modules in the above embodiment can be combined into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present application are only for distinguishing the modules or steps and are not considered to be improper limitations on the present application.

[0184] An electronic device according to a sixth embodiment of the present application includes:

[0185] at least one processor; and

[0186] a memory communicatively connected to at least one of the processors; wherein,

[0187] The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the above-mentioned boiler furnace temperature field reconstruction method.

[0188] A seventh embodiment of the present application is a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the above-mentioned boiler furnace temperature field reconstruction method.

[0189] An eighth embodiment of the present application provides a computer program product. When the computer program product is run on an electronic device, the electronic device is enabled to execute the above-mentioned boiler furnace temperature field reconstruction method.

[0190] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes and related instructions of the electronic device and computer-readable storage medium described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0191] Those skilled in the art should be able to appreciate that, in conjunction with the modules and method steps of each example described in the embodiments disclosed herein, it is possible to implement them with electronic hardware, computer software, or a combination of the two, and the programs corresponding to the software modules and method steps can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the art. In order to clearly illustrate the interchangeability of electronic hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in electronic hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0192] Computer program code for performing the operations of the present application may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0193] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0194] The terms "first", "second", etc. are used to distinguish similar objects, rather than to describe or indicate a particular order or sequence.

[0195] The term "comprise" or any other similar term is intended to cover non-exclusive inclusion such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0196] Thus far, the technical solutions of the present application have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is readily understood by those skilled in the art that the scope of protection of the present application is obviously not limited to these specific embodiments. Without departing from the principles of the present application, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present application.

Claims

1. A temperature field prediction model training method, characterized in that: include: Acquire a geometric model of a target boiler and perform mesh division on the geometric model, wherein the mesh of the geometric model corresponds to mesh data for positioning; Acquiring temperature field data of the target boiler based on a fluid numerical simulation model of the target boiler; Using the target data set constructed by the grid data and the temperature field data as a training sample of the unit model, and iterating and training the node temperature update model and the edge feature update model included in the unit model respectively; The model parameters of the trained unit model are synchronized to multiple layers to generate a temperature field prediction model consisting of multiple layers, wherein the structure of the layers is the same as that of the unit model.

2. The temperature field prediction model training method according to claim 1, characterized in that: The step of iteratively training the node temperature update model and the edge feature update model included in the unit model comprises: Obtain the initial node model and initial edge model of the grid, where: The input data of the initial node model includes the initial node temperature and the edge feature, the output data of the initial node model is the updated node temperature, and the edge feature is the initial edge feature or the updated edge feature. The input data of the initial edge model includes the initial edge features and the node features corresponding to the nodes connected by the edges, and the output data is the updated edge features, wherein the node features include the grid three-dimensional coordinates corresponding to the nodes, the measured temperature, the initial temperature, the furnace pressure and the edge features; Using the updated node temperature of the initial node model at the first moment as the initial node temperature at the second moment, iterating the initial node model to obtain the node temperature update model; Using the updated edge features of the initial edge model at the first moment as the initial edge features at the second moment, iterating the initial edge model to obtain the edge feature update model; Obtain a node-edge critical matrix of the grid, and couple the node-edge critical matrix, the node temperature update model, and the edge feature update model to generate the unit model, wherein the node-edge critical matrix is used to characterize the correspondence between the nodes and edges of the grid.

3. The temperature field prediction model training method according to claim 2, characterized in that: When the edge feature updating model converges, the target edge feature output by the edge feature updating model is obtained as the edge feature in the input data of the node temperature updating model.

4. The temperature field prediction model training method according to claim 1, characterized in that: The acquiring temperature field data of the target boiler based on the fluid numerical simulation model of the target boiler includes: Based on the fluid numerical simulation model of the target boiler, obtaining simulation data of multiple boiler furnace combustion, wherein the simulation data includes boiler furnace geometry parameters, mesh data, and boiler furnace internal temperature field data, and the simulation data is used to simulate the temperature field distribution under different furnace types and different operating conditions; Performing data enhancement on the temperature field data inside the boiler furnace to obtain enhanced temperature field data; The enhanced temperature field data is spliced with other data of the simulation data to generate the temperature field data of the target boiler.

5. The temperature field prediction model training method according to claim 4, characterized in that: The step of obtaining simulation data of furnace combustion of multiple boilers based on the fluid numerical simulation model of the target boiler includes: Based on fluid numerical modeling and simulation, build the fluid, combustion and heat transfer model in any boiler furnace; Perform model solution calculations on fluid, combustion and heat transfer models to obtain simulation calculation data sets as simulation data for combustion in any boiler furnace.

6. The temperature field prediction model training method according to claim 4, characterized in that: The step of enhancing the temperature field data inside the boiler furnace to obtain enhanced temperature field data includes: An intermediate iteration result of the simulation data in any iteration process is obtained as the enhanced temperature field data.

7. The temperature field prediction model training method according to claim 6, characterized in that: The step of enhancing the temperature field data inside the boiler furnace to obtain enhanced temperature field data further includes: Get the measured temperature at a random node position; The temperature at the random node position in the intermediate iteration result is replaced by the measured temperature to obtain the enhanced temperature field data.

8. A temperature field reconstruction method based on the temperature field prediction model training method according to any one of claims 1 to 7, characterized in that: include: Construct a measured data set based on the grid data and local measured temperature of the target boiler; Inputting the measured data set into a temperature field prediction model to obtain a temperature field prediction result of the target boiler; Based on the target measured temperature at the preset position and the temperature field prediction result, the temperature field of the target boiler is reconstructed.

9. The temperature field reconstruction method according to claim 8, characterized in that: The temperature field reconstruction of the target boiler based on the target measured temperature at the preset position and the temperature field prediction result includes: Extracting the local predicted temperature of the preset position from the temperature field prediction result; When the difference between the local predicted temperature and the target measured temperature is less than a deviation threshold, using the temperature field prediction result as the temperature field reconstruction result of the target boiler; When the difference between the local predicted temperature and the target measured temperature is greater than or equal to the deviation threshold, the temperature field prediction result is re-input into the temperature field prediction model for temperature field prediction until the difference between the local predicted temperature extracted from the temperature field prediction result and the target measured temperature is less than the deviation threshold, and the output result of the temperature field prediction model is used as the temperature field reconstruction result of the target boiler.

10. The temperature field reconstruction method according to claim 8, characterized in that: The method further comprises: Obtain measured data sets at multiple historical moments to form a historical data set; Acquire a forecast data set based on the historical data set and the time series forecasting model; The predicted data set is input into the temperature field prediction model to obtain a predicted temperature field reconstruction result.

Images