Boiler heating surface three-dimensional wall temperature monitoring and temperature field reconstruction method

Abstract

The invention discloses a boiler heating surface three-dimensional wall temperature monitoring and temperature field reconstruction method, and the method comprises the steps: constructing a full-dimensional wall temperature monitoring network through a fiber grating array sensing technology, and achieving the precise collection of the wall temperature of a boiler heating surface and the dynamic reconstruction of a three-dimensional temperature field in a boiler through the combination of full-size CFD simulation and a real-time monitoring data fusion algorithm. Firstly, a distributed optical fiber sensor is used for carrying out multi-measuring-point synchronous monitoring on key heating surfaces such as a water cooling wall, a high-pressure water cooling surface and a high-pressure water heating surface to obtain real-time wall temperature data; combustion condition parameters are corrected through primary air powder on-line monitoring data, and a full-size CFD simulation model is input; and finally, based on the heat flow distribution model and a big data analysis algorithm, visual reconstruction and safety early warning of the temperature field are completed. According to the invention, the problems of a heating surface temperature monitoring blind area and insufficient temperature field reconstruction precision under a deep peak regulation working condition are solved, accurate temperature monitoring in a full-load interval of 0-1800 DEG C is realized, and a technical support is provided for safe and economical operation of a unit.

Description

Technical Field

[0001] This invention relates to the field of deep peak shaving technology for thermal power units, and more specifically, to a method for three-dimensional wall temperature monitoring and temperature field reconstruction of boiler heating surfaces. Background Technology

[0002] Under the background of the construction of the new power system, thermal power units face the normalized operation requirements of deep peak shaving, and the load needs to change rapidly in the range of 20% to 100%, with a load change rate of 2% to 5% per minute. Under this operating condition, the boiler heating surface is prone to heat transfer deviation, heat transfer deterioration, and wall temperature overheating, which can lead to metal fatigue or even tube rupture accidents, seriously affecting the safe operation of the unit.

[0003] Existing monitoring technologies mostly rely on traditional thermocouple temperature measurement, which suffers from limited measuring points, large monitoring blind spots, and slow response speeds, making it difficult to cover temperature changes across the entire heating surface. Furthermore, temperature field reconstruction largely depends on pure simulation calculations, lacking dynamic correction based on real-time operational data and specific algorithmic support, resulting in insufficient reconstruction accuracy and failing to meet the safety monitoring requirements under deep peak-shaving conditions. In addition, existing technologies do not achieve linked analysis of wall temperature monitoring and combustion condition data; the adaptability between algorithms and technical features is poor, making it difficult to accurately reflect the impact of combustion status on the temperature distribution of the heating surface. Therefore, there is an urgent need to develop a technical method that integrates specialized algorithms and combines comprehensive monitoring with high-precision reconstruction to solve the safety monitoring challenges of boiler heating surfaces under deep peak-shaving conditions. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and to propose a method for three-dimensional wall temperature monitoring and temperature field reconstruction of boiler heating surfaces, comprising the following steps: S1. Construct a distributed optical fiber monitoring network: Fiber grating sensor arrays are deployed on the heating surfaces of the boiler water-cooled wall, high-pressure reheater, high-pressure reheater, screen reheater, and rear screen. Each sensor has multiple measuring points, and a single data acquisition instrument can achieve synchronous monitoring of 300+ measuring points. S2, Real-time acquisition of wall temperature data: The heated surface wall temperature data in the range of -250℃ to 960℃ is acquired through a fiber optic grating sensor. The sensor response time is ≤20ms, the measurement accuracy is ±0.5%, and the resolution is 0.01℃. S3, collect combustion condition related data: obtain the pulverized coal flow rate and concentration through the primary air-pulverized coal online monitoring device, and collect the NOx concentration and ammonia slip rate data at the SCR outlet through the ammonia injection optimization system to form a condition characteristic dataset; S4. Establish a full-size CFD simulation model: Based on the boiler structural parameters and combustion characteristics, construct a full-size CFD simulation model that includes furnace combustion and heat transfer, and set the calculation boundary conditions for the temperature measurement range of 0℃~1800℃. S5, Data Fusion and Model Correction: A weighted fuzzy control algorithm is used to fuse real-time wall temperature data, air-coal monitoring data, and ammonia injection optimization data. The heat exchange coefficient and combustion rate parameters of the CFD model are corrected through dynamic weight allocation (here, the weight of the wall temperature data is 0.4, the weight of the air-coal data is 0.3, and the weight of the ammonia injection data is 0.3). S6, Three-dimensional temperature field reconstruction: Based on the heat flow distribution model, combined with the corrected CFD simulation results and distributed monitoring data, the three-dimensional temperature field inside the furnace is reconstructed using a hierarchical radial basis function interpolation algorithm. S7, Safety Warning Judgment: Using the LSTM stress prediction algorithm, the stress state of the heated surface pipe section is calculated based on the reconstructed temperature field data, and compared with the dynamic safety threshold to trigger over-temperature and thermal deviation anomaly warnings.

[0005] Preferably, the fiber Bragg grating sensor employs a femtosecond laser-written Bragg grating structure, with a density of no less than 1 grating / m in key areas. 2 Non-critical areas should have at least one per 3m 2 The total number of measurement points shall not be less than 400.

[0006] Preferably, the primary air-powder online monitoring device adopts a non-invasive full-section measurement structure, with 360-degree blind-spot-free measurement. Through the electrostatic detection principle combined with spatial filtering and noise reduction algorithm, it simultaneously collects flow rate and concentration data from 12 sets of devices, and the data update frequency is consistent with the wall temperature monitoring data. Preferably, the correction process of the full-size CFD simulation model includes adjusting the burner region calculation parameters based on the non-uniformity data of NOx concentration distribution at the SCR outlet using the particle swarm optimization (PSO) algorithm, and optimizing the furnace flue gas flow calculation model based on ammonia slip rate data.

[0007] Preferably, the layered RBF interpolation algorithm divides the furnace into combustion zone, superheated zone, and reheated zone calculation layers. Each layer constructs a radial basis function network based on data from at least three adjacent monitoring points, and interpolation calculation is achieved through a Gaussian kernel function (σ=0.8). The regional temperature error is controlled within 5%.

[0008] Preferably, in the safety warning judgment, the dynamic safety threshold is adjusted in real time through an adaptive PID algorithm, including warning of overheating of the heated surface wall, warning of uneven temperature field distribution, and warning of overstress in the pipe section. The threshold adjustment cycle is synchronized with the data acquisition cycle (50Hz).

[0009] Preferably, the fiber optic grating sensor integrates 8 measuring points per unit, with an adjacent measuring point spacing of 1.5m. The matching data acquisition instrument adopts a synchronous timing calibration algorithm, a sampling frequency of 50Hz, supports synchronous network monitoring of 3 units, and the data synchronization error is ≤10ms.

[0010] Preferably, the primary air-coal online monitoring device is installed in the coal mill outlet pulverized coal pipe. It achieves coordinated balancing of flow velocity and concentration through a dual-parameter coupled balancing algorithm. After balancing, the flow velocity deviation in each pulverized coal pipe is ≤5%, and the concentration measurement range is 4.2~9.8 kg / m³. 3 The data update frequency is 100ms.

[0011] Preferably, the CFD simulation model has a mesh size of ≥8 million, the combustion model adopts the EDC model, and the radiative heat transfer adopts the DO model. Based on the measured data of 20%, 30%, 50%, 75%, and 100% load, the heat exchange coefficient (correction range ±8%) and the combustion rate parameter (correction magnitude ±5.2%) are corrected by a multi-objective genetic optimization algorithm.

[0012] Preferably, the three-dimensional temperature field reconstruction results are visualized through an industrial control platform. The VTK visualization rendering algorithm is used to realize the real-time drawing of the temperature field cloud map, which supports zooming in on any area and backtracking of historical data within 3 months. The temperature data storage accuracy is 0.01℃.

[0013] Compared with existing technologies, the beneficial effects of this invention are: This invention employs distributed fiber optic grating sensing technology to achieve full-area monitoring of the heated surface without blind spots, with more than 400 measuring points, thus solving the problem of insufficient coverage in traditional temperature measurement technologies. By fusing multi-source data and dynamically correcting the CFD model, the accuracy of temperature field reconstruction is improved, the regional temperature error is small, and the temperature distribution characteristics inside the furnace can be accurately reflected. It realizes the linkage analysis of wall temperature monitoring and combustion conditions. The reconstructed temperature field can respond in real time to changes in operating conditions such as air-coal parameters and ammonia injection, providing data support for combustion optimization. The sensor is easy to install, has a long service life, can adapt to the harsh operating environment of boilers with high temperature and high dust, has high operational reliability, and reduces maintenance costs. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the overall construction process of the three-dimensional wall temperature monitoring and temperature field reconstruction method for boiler heating surfaces proposed in this invention; Figure 2 This is a data table verifying the number of meshes in a 660MWW flame boiler CFD simulation model for a three-dimensional wall temperature monitoring and temperature field reconstruction method for boiler heating surfaces proposed in this invention. Figure 3 This is a table of specific experimental data for the weighted sensitivity analysis in the three-dimensional wall temperature monitoring and temperature field reconstruction method for boiler heating surfaces proposed in this invention. Detailed Implementation

[0015] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0016] Referring to the figure, this embodiment provides a method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface, including the following steps: S1. Construct a distributed optical fiber monitoring network: Fiber grating sensor arrays are deployed on the heating surfaces of the boiler water-cooled wall, high-pressure reheater, high-pressure reheater, screen reheater, and rear screen. Each sensor has multiple measurement points, and a single data acquisition instrument can achieve simultaneous monitoring of 300+ measurement points to ensure that there are no blind spots in the entire heating surface. S2, Real-time wall temperature data acquisition: The heated surface wall temperature data in the range of -250℃ to 960℃ is acquired through a fiber optic grating sensor. The sensor response time is ≤20ms, the measurement accuracy is ±0.5%, and the resolution is 0.01℃, which can accurately capture instantaneous changes in wall temperature. S3, collecting combustion condition related data: obtaining pulverized coal flow rate and concentration through the primary air-coal online monitoring device, and collecting SCR outlet NOx concentration and ammonia slip rate data through the ammonia injection optimization system to form a condition characteristic dataset; providing basic data for CFD model correction; S4. Establish a full-scale CFD simulation model: Based on the boiler structural parameters and combustion characteristics, construct a full-scale CFD simulation model that includes furnace combustion and heat transfer, and set the calculation boundary conditions for the temperature measurement range of 0℃ to 1800℃; the CFD simulation model has a mesh size of ≥8 million, the combustion model adopts the EDC model, and the radiative heat transfer adopts the DO model. Based on the measured data of 20%, 30%, 50%, 75%, and 100% load, the heat exchange coefficient (correction range ±8%) and combustion rate parameters (correction magnitude ±5.2%) are corrected through a multi-objective genetic optimization algorithm. S5, Data Fusion and Model Correction: A weighted fuzzy control algorithm is used to fuse real-time wall temperature data, air-coal monitoring data, and ammonia injection optimization data. The heat exchange coefficient and combustion rate parameters of the CFD model are corrected through dynamic weight allocation (here, the weight of the wall temperature data is 0.4, the weight of the air-coal data is 0.3, and the weight of the ammonia injection data is 0.3). S6, Three-dimensional temperature field reconstruction: Based on the heat flow distribution model, combined with the corrected CFD simulation results and distributed monitoring data, the three-dimensional temperature field inside the furnace is reconstructed using a hierarchical radial basis function interpolation algorithm. S7, Safety Warning Judgment: Using the LSTM stress prediction algorithm, the stress state of the heated surface pipe section is calculated based on the reconstructed temperature field data, and compared with the dynamic safety threshold to trigger over-temperature and thermal deviation anomaly warnings.

[0017] The specific implementation scheme for the above steps is as follows: S1, Construction of a distributed fiber optic monitoring network: At key heating surfaces such as boiler water-cooled walls, high-pressure reheaters, high-pressure reheaters, screen reheaters, and rear screens, at least one fiber optic monitoring network will be established per meter in key areas. 2 Non-critical areas should have at least one per 3m 2 A high-density fiber optic grating sensor array is deployed. The sensor uses a femtosecond laser-written Bragg grating structure, and a single sensor can integrate 8 measuring points with a spacing of 1.5m between adjacent measuring points. A single data acquisition instrument can achieve simultaneous monitoring of 300+ measuring points, ensuring that there are no blind spots in the monitoring of the entire heated surface.

[0018] The accompanying data acquisition instruments employ a synchronous timing calibration algorithm. By calibrating the phase of the sampling clocks of multiple data acquisition instruments, the data synchronization error is controlled to ≤10ms, ensuring the time consistency of data from different measuring points. The sensor installation utilizes a rapid installation process, enabling quick installation of individual measuring points, saving 90% of installation space, and avoiding interference with boiler operation.

[0019] S2. Multi-source data acquisition and preprocessing; S21 Wall Temperature Data Acquisition: Real-time acquisition of wall temperature data of the heated surface through fiber optic grating sensors. The sensor's temperature measurement range covers -250℃ to 960℃, with a response time of ≤20ms, measurement accuracy of ±0.5%, and resolution of 0.01℃, which can accurately capture instantaneous changes in wall temperature. S22 Air-Coal Data Acquisition: Twelve sets of primary air-coal online monitoring devices are installed in the coal mill outlet pulverized coal pipes, employing a non-invasive full-section measurement structure with 360-degree blind-spot-free measurement. The devices process electrostatic induction signals using a spatial filtering noise reduction algorithm, filtering out noise such as airflow disturbances and electromagnetic interference. They simultaneously acquire both coal powder flow rate and concentration parameters, with a data update frequency of 100ms. Subsequently, a coupling relationship model between flow rate and concentration is established using a dual-parameter coupling balancing algorithm to achieve coordinated balancing, ensuring that the flow rate deviation in each pulverized coal pipe is ≤5% after balancing. S23 Ammonia Injection Data Acquisition: The NOx concentration and ammonia slip rate at the SCR outlet are collected through the ammonia injection optimization system to form a complete set of operating condition characteristic data, providing basic data for CFD model correction.

[0020] S3. The establishment and correction of the full-size CFD simulation model is based on the actual structural parameters and combustion characteristics of the boiler. A full-size CFD simulation model including furnace combustion and heat transfer processes is constructed. The model has a mesh count of ≥8 million. The combustion model adopts the EDC model, and the radiative heat transfer adopts the DO model. The calculation boundary conditions are set for the temperature measurement range of 0℃ to 1800℃. Here, the mesh count is adapted and adjusted according to the specific boiler model and geometric dimensions. For the 660MWW flame boiler targeted in this invention, the 8 million mesh has passed geometric discretization, physical model adaptation, and independence verification, achieving an optimal balance between accuracy and efficiency. Specific parameters are detailed in [link to documentation]. Figure 2 ; Figure 2 When the number of grid cells increased from 5 million to 8 million, the calculation deviations for wall temperature and furnace center temperature decreased from 8.6% and 9.2% to 3.8% and 3.6%, respectively, showing a significant improvement in accuracy. However, when the number of grid cells increased from 8 million to 9 million, the deviations only decreased by 0.1 to 0.2 percentage points, indicating a diminishing marginal effect in accuracy improvement, but the computational efficiency decreased by 38% (the single-condition calculation time increased from 11.3 hours to 15.6 hours). Therefore, considering both accuracy and efficiency, 8 million grid cells were determined as the minimum number of grid cells, satisfying the technical requirement of "calculation deviation ≤ 4%" while ensuring computational efficiency in engineering applications.

[0021] The model correction process is achieved by using a combination of particle swarm optimization algorithm and multi-objective genetic optimization algorithm. Specifically, based on the non-uniformity of NOx concentration distribution at the SCR outlet, the calculation parameters such as fuel injection angle and wind speed distribution in the burner area are adjusted by the particle swarm optimization algorithm. With NOx concentration uniformity as the optimization objective, the parameters are iteratively updated until the requirements are met. In addition, based on the measured wall temperature data under five loads of 20%, 30%, 50%, 75%, and 100%, the heat exchange coefficient and combustion rate parameters are optimized simultaneously through a multi-objective genetic optimization algorithm. The heat exchange coefficient is corrected within a range of ±8%, and the combustion rate parameter is corrected within a range of ±5.2%, achieving high-precision adaptation of the model across the entire load range. In addition, the furnace flue gas flow calculation model was optimized based on ammonia slip rate data to further ensure that the temperature field calculation matches the actual operating conditions.

[0022] S4. Data Fusion and 3D Temperature Field Reconstruction: A weighted fuzzy control algorithm is used for multi-source data fusion. Based on the influence weights of different data on the temperature field calculation, the fusion ratios of wall temperature data (weight 0.4), air-coal data (weight 0.3), and ammonia injection data (weight 0.3) are dynamically allocated. Fuzzy inference rules are used to correct the input parameters of the CFD model, eliminating calculation deviations caused by operating condition fluctuations. Here, the weights are determined from actual operating condition measurement data. Specific weight determination data is shown in the attached table in the instruction manual. The data in this table is based on measured data from Heze Power Plant Unit #3 (660MW anthracite boiler) under three typical load conditions of 30%, 50%, and 100%. Sensitivity tests were conducted using the controlled variable method to quantify the impact of various data fluctuations on the temperature field calculation deviation, thereby achieving weight allocation.

[0023] The three-dimensional temperature field reconstruction uses a hierarchical radial basis function (RBF) interpolation algorithm, the specific process of which is as follows: S41. Divide the furnace along the height direction into multiple calculation layers such as combustion zone, superheat zone, and reheat zone, and clarify the temperature characteristics and interpolation accuracy requirements of each layer; S42. Based on the measured data of at least 3 adjacent monitoring points, a radial basis function network is constructed for each layer. The Gaussian kernel function (σ=0.8) is selected as the basis function. The local interpolation characteristics of the kernel function are used to achieve accurate estimation of the temperature of the unmeasured points. S43. Based on the corrected CFD simulation results, the interpolation results are calibrated a second time to ensure that the regional temperature error is controlled within 5%, and finally the accurate reconstruction of the three-dimensional temperature field of the entire furnace is achieved.

[0024] S44. The reconstruction results are displayed through the industrial control platform. The VTK visualization rendering algorithm is used to convert the temperature field data into a real-time cloud map, which supports zooming in on any area and backtracking of historical data within 3 months. The temperature data storage accuracy is 0.01℃.

[0025] S5. Based on the reconstructed three-dimensional temperature field data, an LSTM stress prediction algorithm is used to construct a stress state calculation model for the heated surface pipe section: ① Using historical temperature data, material parameters, and operating parameters as inputs, the stress change trend of the pipe section in the future is predicted by leveraging the temporal memory characteristics of the LSTM network. ② The dynamic safety threshold is adjusted in real time through an adaptive PID algorithm. The wall temperature over-temperature threshold, temperature field unevenness threshold, and stress over-limit threshold are dynamically updated according to the current load, combustion status and other working conditions. The threshold adjustment cycle is synchronized with the data acquisition cycle (50Hz). ③ When the monitored data exceeds the corresponding threshold, the system automatically triggers an early warning, providing operators with precise operational guidance.

[0026] Working principle: The core working principle of this invention is to construct a temperature field system by monitoring the three-dimensional wall temperature of the boiler heating surface, thereby achieving full-dimensional monitoring and precise control of the boiler heating surface temperature under deep peak shaving conditions. The specific principle is as follows: First, using fiber Bragg grating sensing technology as the core, a comprehensive monitoring network covering key heated surfaces such as water-cooled walls, high-temperature cooling systems, and high-temperature reheating systems is constructed. The fiber Bragg grating sensor, through a femtosecond laser-etched Bragg grating structure, converts temperature changes into wavelength shifts in the reflected spectrum. Combined with a synchronous timing calibration algorithm for the data acquisition instrument, it achieves simultaneous monitoring of over 400 measuring points, accurately capturing instantaneous wall temperature changes within the range of -250℃ to 960℃.

[0027] Meanwhile, 12 sets of non-intrusive online coal dust monitoring devices were used to collect coal dust flow rate and concentration parameters. The spatial filtering noise reduction algorithm was used to filter out airflow disturbances and electromagnetic interference, and the flow rate and concentration were optimized in a coordinated manner through a dual-parameter coupling leveling algorithm. Simultaneously, NOx concentration and ammonia slip rate data at the SCR outlet were collected to form a multi-source dataset covering wall temperature, combustion conditions and environmental parameters, providing a foundation for subsequent analysis.

[0028] Then, based on the full-size boiler structural parameters, a CFD simulation model including an EDC combustion model and a DO radiation model was constructed, and the model accuracy was improved through multi-algorithm collaborative correction. First, a weighted fuzzy control algorithm was used to fuse multi-source data according to the dynamic weights of wall temperature data, air-coal data, and ammonia injection data to eliminate interference from operating condition fluctuations. Then, a particle swarm optimization algorithm was used to adjust the burner parameters to reduce the NOx concentration distribution non-uniformity to ≤27%. Finally, a multi-objective genetic optimization algorithm was used to correct the heat exchange coefficient and combustion rate to ensure that the deviation between the calculated and measured values ​​of the model is ≤4%, achieving accurate adaptation across the entire load range.

[0029] Furthermore, a hierarchical radial basis function (RBF) interpolation algorithm was employed, combined with the corrected CFD simulation results and distributed monitoring data, to reconstruct the three-dimensional temperature field of the furnace. First, the furnace was divided along its height into a combustion zone, a superheated zone, and a reheated zone. For each layer, a radial basis function network was constructed using a Gaussian kernel function (σ=0.8) based on data from at least three adjacent measuring points, accurately estimating the temperature at unmeasuring points. Then, the VTK visualization rendering algorithm was used to transform the reconstruction results into a real-time temperature field cloud map, supporting zoom-in viewing of any area and backtracking of 3 months of historical data, ensuring that the regional temperature error is ≤5%, thus achieving a visualized presentation of the temperature distribution within the furnace.

[0030] Finally, based on the reconstructed temperature field data, a pipe section stress calculation model is constructed using the LSTM stress prediction algorithm. Taking historical temperature and load data as input, the model predicts stress change trends 15 seconds in advance (prediction error ≤4.2%). Simultaneously, an adaptive PID algorithm is employed to dynamically adjust safety thresholds based on the current load (50Hz update frequency), forming three early warning mechanisms: wall temperature overheating, uneven temperature field, and stress exceeding limits. When monitored data exceeds the threshold, the system immediately triggers an early warning, guiding operators to adjust operating parameters such as air-coal ratio and ammonia injection rate. This forms a closed-loop control system of "monitoring-analysis-early warning-adjustment," preventing metal fatigue of the heated surfaces and pipe rupture accidents. The above embodiments employ distributed fiber optic grating sensing technology, achieving blind-spot-free monitoring of the entire heated surface area, with over 400 measuring points, thus solving the problem of insufficient coverage in traditional temperature measurement technologies. Through multi-source data fusion and dynamic correction of the CFD model, the accuracy of temperature field reconstruction is improved, with small regional temperature errors, accurately reflecting the temperature distribution characteristics within the furnace. Linked analysis of wall temperature monitoring and combustion conditions is achieved, and the reconstructed temperature field can respond in real-time to changes in operating conditions such as air-coal parameters and ammonia injection volume, providing data support for combustion optimization. The sensors are easy to install, have a long service life, can adapt to the harsh operating environment of high temperature and high dust in boilers, have high operational reliability, and reduce maintenance costs.

[0031] The above specific embodiments further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the technical scope disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0032] Furthermore, it should be understood in the description of this invention that the terms indicating orientation or positional relationship are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0033] Furthermore, in this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

Claims

1. A method for three-dimensional wall temperature monitoring and temperature field reconstruction of boiler heating surfaces, characterized in that, Includes the following steps: S1. Construct a distributed optical fiber monitoring network: Fiber grating sensor arrays are deployed on the heating surfaces of the boiler water-cooled wall, high-pressure reheater, high-pressure reheater, screen reheater, and rear screen. Each sensor has multiple measurement points, and a single data acquisition instrument can achieve synchronous monitoring of multiple measurement points. S2, Real-time acquisition of wall temperature data: The wall temperature data of the heated surface in the range of -250℃ to 960℃ is acquired through a fiber optic grating sensor. The fiber optic grating sensor can achieve a response time of ≤20ms, a measurement accuracy of ±0.5%, and a resolution of 0.01℃. S3, collect combustion condition related data: obtain the pulverized coal flow rate and concentration through the primary air-pulverized coal online monitoring device, and collect the NOx concentration and ammonia slip rate data at the SCR outlet through the ammonia injection optimization system to form a condition characteristic dataset; S4. Establish a full-size CFD simulation model: Based on the boiler structural parameters and combustion characteristics, construct a full-size CFD simulation model that includes furnace combustion and heat transfer, and set the calculation boundary conditions for the temperature measurement range. S5, Data Fusion and Model Correction: Real-time wall temperature data, air-coal monitoring data and ammonia injection optimization data are input into the CFD simulation model, and the model parameters are corrected through fuzzy control algorithm; S6, Three-dimensional temperature field reconstruction: Based on the heat flow distribution model, combined with the corrected CFD simulation results and distributed monitoring data, the three-dimensional temperature field inside the furnace is reconstructed through a hierarchical interpolation algorithm. S7, Safety Warning Judgment: Calculate the stress state of the heated surface pipe section based on the reconstructed temperature field data, compare it with the safety threshold, and trigger an alarm for over-temperature and thermal deviation.

2. The method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface according to claim 1, characterized in that: The fiber Bragg grating sensor employs a femtosecond laser-written Bragg grating structure.

3. The method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface according to claim 1, characterized in that: The primary air-powder online monitoring device adopts a non-invasive full-section measurement structure. It uses electrostatic detection principle combined with spatial filtering technology to simultaneously collect flow velocity and concentration data. The data update frequency is consistent with the wall temperature monitoring data.

4. The method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface according to claim 1, characterized in that: The correction process of the full-scale CFD simulation model includes a. adjusting the burner region calculation parameters based on the non-uniformity of NOx concentration distribution at the SCR outlet, and b. optimizing the furnace flue gas flow calculation model based on ammonia slip rate data.

5. The method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface according to claim 1, characterized in that: The layered interpolation algorithm divides the furnace into combustion zone, superheat zone, and reheat zone calculation layers. Each layer is based on interpolation of data from at least three adjacent monitoring points, and the regional temperature error is controlled within 5%.

6. The method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface according to claim 1, characterized in that: The safety early warning judgment includes early warning of excessive wall temperature of the heated surface, early warning of uneven temperature field distribution, and early warning of excessive stress in the pipe section. The early warning threshold is dynamically adjusted according to the material characteristics of the heated surface and the operating conditions.

7. The method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface according to claim 2, characterized in that: The fiber optic grating sensor integrates 8 measuring points per unit, with a spacing of 1.5m between adjacent measuring points. The matching data acquisition instrument has a sampling frequency of 50Hz and supports synchronous network monitoring of 3 units.

8. The method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface according to claim 3, characterized in that: The primary air-coal online monitoring device is installed in the coal mill outlet pulverized coal pipe. Through the air-coal and leveling device, the deviation of coal powder flow rate and concentration in each pulverized coal pipe at the coal mill outlet is adjusted online to achieve uniform combustion in the boiler and improve the boiler's stable combustion capability under low load.

9. The method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface according to claim 4, characterized in that: The CFD simulation model has a mesh size of ≥8 million, the combustion model adopts the EDC model, and the radiative heat transfer adopts the DO model. The heat exchange coefficient and combustion rate parameters are corrected based on measured data of 20%, 30%, 50%, 75%, and 100% load.

10. The method for three-dimensional wall temperature monitoring and temperature field reconstruction of a boiler heating surface according to claim 1, characterized in that: The three-dimensional temperature field reconstruction results are visualized through an industrial control platform, supporting zooming in on any area and backtracking of historical data within 3 months, with a temperature data storage accuracy of 0.01℃.

Images