Boiler heating surface pipe temperature control method and system based on wall temperature monitoring
By screening key parameters through global and local sensitivity analysis, constructing a parametric finite element model and performing staged optimization, the problems of prediction error and parameter redundancy in boiler wall temperature control were solved, achieving precise boiler wall temperature control and preventing pipe overheating accidents.
Patent Information
- Application Number
- CN202511805529.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-03
- Publication Date
- 2026-02-27
AI Technical Summary
Existing boiler wall temperature control technologies suffer from problems such as large prediction errors, parameter redundancy, and inaccurate adjustment, making it difficult to effectively prevent pipe overheating accidents.
Key parameters were screened using global and local sensitivity analysis, a parameterized finite element model was constructed, and combined with experimental design and a phased optimization model, the boiler parameters were adjusted according to preset priorities by monitoring the wall temperature through sensors to achieve precise control.
Accurately identifying overheating risks reduces model development costs and time, avoids misjudgments or omissions, and enables precise control of boiler wall temperature, preventing pipe creep and pipe rupture accidents.
Smart Images
Figure CN121580744A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of boiler thermal control, more particularly to a boiler heating surface pipe temperature control method and system based on monitoring wall temperature. BACKGROUND
[0002] At present, in the process of boiler operation, the wall temperature of the heating surface pipe (such as the water wall, superheater, reheater pipe) is the core risk of causing pipe creep failure and pipe burst accidents, which directly threatens the safety and life of the boiler. The existing wall temperature control technology mainly uses single parameter adjustment (such as only relying on desuperheating water) or multi-parameter synchronous adjustment without priority, which mainly has the following problems: Traditional models directly apply single-stage prediction of boiler operating parameters to pipe wall temperature, without considering the mechanical transmission process of operating parameters, heating surface heat load and pipe wall temperature, ignoring the key role of heat load as an intermediate variable, resulting in large prediction error and difficulty in meeting engineering precision requirements; the wall temperature of the heating surface is affected by multiple system parameters such as combustion, steam, flue gas, etc. (such as fuel quantity, air supply quantity, water flow, flue gas temperature, etc.), the existing technology does not filter key parameters through sensitivity analysis, resulting in too many input dimensions, and the model training requires a large number of samples.
[0003] Therefore, how to realize accurate and efficient wall temperature control is a problem that needs to be solved by those skilled in the art. SUMMARY
[0004] Therefore, the present application provides a boiler heating surface pipe temperature control method and system based on monitoring wall temperature to solve the problems in the background art.
[0005] In order to achieve the above purpose, the present application adopts the following technical scheme: A boiler heating surface pipe temperature control method based on monitoring wall temperature, comprising: S1: collecting boiler data through a sensor, combining global sensitivity analysis and local sensitivity analysis, and filtering main parameters in the boiler data that significantly affect the heating surface heat load and pipe wall temperature; constructing a parameterized finite element model according to input parameters and output parameters in the main parameters; S2: obtaining training data by combining experimental design and parameterized finite element model; S3: constructing a stage optimization model through the training data and verifying the prediction accuracy; S4: extracting relevant parameters of the boiler in actual operation as input of the model, predicting and obtaining the heating surface load and pipe wall temperature, and when the pipe wall temperature exceeds the preset wall temperature threshold, adjusting the boiler parameters according to the preset adjustment priority to control the pipe wall temperature.
[0006] Preferably, the parameters in the boiler data in S1 that have the greatest impact on the output are selected as the main parameters; The input parameters include combustion system parameters: fuel quantity, air supply quantity, excess air coefficient; steam-water system parameters: feed water flow, desuperheating water flow, working medium temperature; flue gas system parameters: flue gas temperature, flue gas flow rate; pipe material characteristic parameters: pipe material thermal conductivity, wall thickness; The output parameters include: heating surface heat load, pipe wall temperature.
[0007] Preferably, the construction of the parameterized finite element model in S1 specifically includes: based on a finite element software, parameterized modeling is realized through a Python script interface: the screened main parameters are input, a thermal-structural coupling model of the heating surface pipe material is automatically generated, and the heating surface heat load and the pipe wall temperature under the corresponding working condition are output.
[0008] Preferably, the obtaining of the training data in S2 specifically includes: determining the value range of each input parameter from the historical operation data of the target boiler, the design manual and the same type of engineering cases; using a Latin hypercube experimental design method to randomly sample in the parameter value range to generate a sample point set; The sample point set is substituted into the parameterized finite element model to calculate the corresponding heating surface heat load and pipe wall temperature to form an output response set. After normalization processing of the sample point set and the output response set, the training data set is obtained.
[0009] Preferably, the stage-wise optimization model in S3 specifically includes: ; Wherein, is a first-order regression model, is a regression coefficient; is a random process with a mean of 0 and a variance of The covariance is expressed as follows: ; Wherein, and are any two different points in the design space, is a correlation function related only to the spatial positions of the two points; ; Wherein, is a to-be-determined model parameter, is any one of a linear function, an exponential function, a Gaussian function, a spherical function and a cubic spline function; The training sample set and the response set The predicted response value of any new to-be-tested point can be obtained by substituting the known sample point response value The linear combination is performed to estimate: ; wherein, is the response value weight coefficient vector, and the prediction error is expressed as: .
[0010] Preferably, the verifying the prediction accuracy in S3 specifically comprises: dividing the stage-wise optimization model prediction process into two stages, a first-stage optimization model is used to predict the heat load of the heating surface, and a second-stage optimization model is used to predict the tube wall temperature; First-stage optimization model construction: selecting combustion system parameters, steam-water system basic parameters and flue gas system key parameters as inputs, taking the heat load of the heating surface as output, and using a Kriging model to establish a mapping relationship; the model hyperparameter θ1 is optimized through a particle swarm optimization algorithm, and the objective function is to minimize the heat load prediction mean square error; the model is trained using normalized training data, and a first-stage optimization model is obtained; Second-stage optimization model construction: taking the heat load of the heating surface output by the first-stage optimization model and the tube material characteristic parameters and working medium parameters as inputs, taking the tube wall temperature as output, and using the same type of Kriging model to establish a mapping relationship; the hyperparameter θ2 is also optimized through a particle swarm optimization algorithm, and the objective function is to minimize the wall temperature prediction mean square error; the model is trained using the same proportion of normalized sample data as the first stage, and a second-stage optimization model is obtained.
[0011] Preferably, the adjusting the boiler parameters according to the preset adjustment priority in S4 specifically comprises: the priorities are fuel oil system adjustment, steam-water system adjustment, and flue gas system adjustment in turn; When , priority adjustment is performed; wherein, is the wall temperature set value, is the over-temperature threshold value, is the predicted tube wall temperature; Fuel oil system adjustment: ; ; wherein, is the fuel adjustment coefficient, is the optimal heat load, is the optimal excess air coefficient, is the fuel adjustment amount, is the air supply adjustment amount, is the first-stage heat load of the heating surface, is the optimal air supply amount; When the over-temperature condition is exceeded for more than a preset observation period of the fuel oil, the over-temperature condition is still met, steam-water system adjustment is performed: ; ; wherein, is the steam-water regulation coefficient, is the feedwater regulation flow rate, is the desuperheating water regulation flow rate, is the pipe wall temperature set value, is the current feedwater flow rate, is the working medium specific heat capacity, is the steam flow rate, is the desuperheating cooling efficiency, is the desuperheating water temperature, is the heating surface inlet steam temperature; when the steam-water preset observation period is exceeded, still over-temperature, the flue gas system is regulated: ; wherein, is the flue gas regulation coefficient, is the flue gas recirculation rate regulation amount, is the optimal flue gas temperature, is the actual flue gas temperature.
[0012] A boiler heating surface pipe temperature control system based on monitoring wall temperature, characterized by comprising: a finite element model construction module, which acquires boiler data through a sensor, combines global sensitivity analysis and local sensitivity analysis, screens main parameters in the boiler data which have significant influence on heating surface heat load and pipe wall temperature, and constructs a parameterized finite element model according to input parameters and output parameters in the main parameters; a training data acquisition module, which acquires training data in combination with experimental design and the parameterized finite element model; a stage model construction module, which constructs a stage optimization model through the training data and verifies prediction accuracy; a regulation and control module, which extracts relevant parameters of the boiler in actual operation as input of the model, predicts and acquires heating surface load and pipe wall temperature, adjusts boiler parameters according to a preset adjustment priority when the pipe wall temperature exceeds a preset wall temperature threshold, and regulates and controls the pipe wall temperature.
[0013] Compared with the prior art, the boiler heating surface pipe temperature control method and system based on monitoring wall temperature provided by the technical solution disclosed in the present application can avoid misjudgment or missed judgment of over-temperature working conditions by using the stage type optimization model for operating parameters, heating surface heat load and pipe wall temperature in sequence, fitting the boiler heat transfer mechanism, combining global sensitivity analysis and local sensitivity analysis to screen key parameters, eliminating redundant interference, accurately identifying over-temperature risks, and using a progressive adjustment strategy of the combustion system, the steam-water system and the flue gas system to avoid side effects such as steam water carrying and combustion deterioration caused by excessive adjustment of a single parameter, thereby realizing accurate control of the wall temperature. BRIEF DESCRIPTION OF DRAWINGS
[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of the provided drawings.
[0015] Figure 1 The method step flowchart provided by the present application. DETAILED DESCRIPTION
[0016] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.
[0017] The embodiments of the present application disclose a boiler heating surface pipe temperature control method based on monitoring wall temperature, as shown in Figure 1 The method comprises the following steps: S1: Collecting boiler data by a sensor, combining global sensitivity analysis and local sensitivity analysis to screen the main parameters in the boiler data which have significant influence on the heating surface heat load and the pipe wall temperature; and constructing a parameterized finite element model according to the input parameters and the output parameters in the main parameters; S2: Obtaining training data by combining experimental design and the parameterized finite element model; S3: Constructing a stage type optimization model by using the training data and verifying the prediction accuracy; S4: Extract the relevant parameters of the actual running boiler as the input of the model, predict the heating surface load and pipe wall temperature, and when the pipe wall temperature exceeds the preset wall temperature threshold, adjust the boiler parameters according to the preset adjustment priority to control the pipe wall temperature.
[0018] In one specific embodiment, the parameters in the boiler data that have the greatest impact on the output are selected as the main parameters in S1; The input parameters include combustion system parameters: fuel quantity, air supply quantity, excess air coefficient; steam-water system parameters: feed water flow, desuperheating water flow, working medium temperature; flue gas system parameters: flue gas temperature, flue gas flow rate; pipe material characteristic parameters: pipe material thermal conductivity, wall thickness; The output parameters include: heating surface heat load, pipe wall temperature.
[0019] In one specific embodiment, the construction of the parameterized finite element model in S1 specifically includes: based on the finite element software, the parameterized modeling is realized through the Python script interface: the selected main parameters are input, the thermal-structural coupling model of the heating surface pipe is automatically generated, and the heating surface heat load and pipe wall temperature under the corresponding working condition are output.
[0020] In one specific embodiment, the training data in S2 specifically includes: determining the value range of each input parameter from the historical operation data of the target boiler, design manual and similar type engineering cases; using Latin hypercube experimental design method to randomly sample in the parameter value range to generate a sample point set; The sample point set is substituted into the parameterized finite element model to calculate the corresponding heating surface heat load and pipe wall temperature to form an output response set. After normalization processing of the sample point set and the output response set, the training data set is obtained.
[0021] In one specific embodiment, the stage-wise optimization model in S3 specifically includes: ; Wherein, is a first-order regression model (providing global approximation), is a regression coefficient; is a random process with mean 0 and variance The covariance is expressed as follows: ; Wherein, and are any two different points in the design space, is a correlation function related only to the spatial positions of the two points; ; Wherein, is a to-be-determined model parameter, any one of linear function, exponential function, Gaussian function, spherical function and cubic spline function; training sample set and its response set The predicted response value of any new point to be tested can be estimated by linear combination of the known response values of the sample points : ; wherein, is the response weight coefficient vector, and the prediction error is expressed as: .
[0022] In a specific embodiment, the verification of the prediction accuracy in S3 specifically comprises: dividing the stage-wise optimization model prediction process into two stages, a first-stage optimization model is used to predict the heat load of the heating surface, and a second-stage optimization model is used to predict the pipe wall temperature; First-stage optimization model construction: select the combustion system parameters, steam-water system basic parameters and flue gas system key parameters as inputs, and the heat load of the heating surface as output, and use the Kriging model to establish the mapping relationship; the model hyperparameter θ1 is optimized through the particle swarm optimization algorithm, and the objective function is to minimize the heat load prediction mean square error; the model is trained using normalized training data, and the first-stage optimization model is obtained; Second-stage optimization model construction: the heat load of the heating surface output by the first-stage optimization model and the pipe material characteristic parameters and working medium parameters are taken as inputs, and the pipe wall temperature is taken as output, and the same type of Kriging model is used to establish the mapping relationship; similarly, the hyperparameter θ2 is optimized through the particle swarm optimization algorithm, and the objective function is to minimize the wall temperature prediction mean square error; the model is trained using the same proportion of normalized sample data as the first stage, and the second-stage optimization model is obtained.
[0023] The model hyperparameter optimization is realized through the particle velocity and position update formula, and specifically as follows: ; wherein, and are the velocity of the dth dimension of the ith particle in the k+1th iteration and the kth iteration, , is a learning factor; , is a random number between 0 and 1; ; wherein, and are the position of the dth dimension of the ith particle in the k+1th iteration and the kth iteration. Individual optimal position and global optimal position The update formula is as follows, where f is the fitness function and N is the population size: ; ; in, Let d be the optimal position of the i-th particle in the d-th dimension during a total of k iterations. During the algorithm's execution, the particle will gradually approach the global optimal solution through continuous searching and iteration.
[0024] In one specific embodiment, the boiler parameter adjustment in S4 according to the preset adjustment priority specifically includes the following priority order: fuel oil system adjustment, steam-water system adjustment, and flue gas system adjustment. when Priority adjustment is performed; among them, Set the wall temperature value. This exceeds the temperature threshold. To predict pipe wall temperature; Fuel system adjustment: ; ; in, This is the fuel adjustment coefficient. For optimal heat load, To achieve the optimal excess air coefficient, For fuel adjustment quantity, For air supply regulation, The heat load of the first stage heating surface, To achieve the optimal air supply volume; When the fuel consumption observation period has expired. Still overheating, adjust the soft drink system: ; ; in, This is the soda / water adjustment coefficient. To regulate water flow, To adjust the flow rate of the de-heating water, Set the pipe wall temperature value. This is the current water supply flow rate. The specific heat capacity of the working fluid, For steam flow rate, To reduce cooling efficiency, To reduce the temperature of the heated water, The inlet steam temperature of the heating surface; When the preset observation period for the soft drink is exceeded Still overheating, adjust the flue gas system: ; wherein, is the flue gas adjustment coefficient, is the flue gas recirculation rate adjustment amount, is the optimal flue gas temperature, is the actual flue gas temperature.
[0025] After the combustion system is adjusted: the combustion system responds the fastest, and the effect can be preliminarily reflected within 3-5s of observation.
[0026] After the steam-water system is adjusted: the working fluid flow and heat transfer need a longer time, and the wall temperature change is reflected within 10-30s of observation.
[0027] A boiler heating surface pipe temperature control system based on monitoring wall temperature, comprising: a finite element model construction module, which collects boiler data through a sensor, combines global sensitivity analysis and local sensitivity analysis, screens main parameters in the boiler data which have significant influence on heating surface heat load and pipe wall temperature, and constructs a parameterized finite element model according to input parameters and output parameters in the main parameters; a training data acquisition module which acquires training data in combination with experimental design and the parameterized finite element model; a stage model construction module which constructs a stage optimization model through the training data and verifies the prediction accuracy; a regulation and control module which extracts relevant parameters of the boiler in actual operation as inputs of the model, predicts and acquires heating surface load and pipe wall temperature, adjusts the boiler parameters according to a preset adjustment priority when the pipe wall temperature exceeds a preset wall temperature threshold, and regulates and controls the pipe wall temperature.
[0028] First-stage optimization model (heat load prediction): Input: key basic parameters of combustion / steam / water / flue gas; Output: heating surface heat load; Training: the model is trained with 160-240 groups of normalized training sets, and the accuracy is verified with 40-60 groups of verification sets.
[0029] Second-stage optimization model (wall temperature prediction): Input: heat load predicted in the first stage + pipe material characteristic parameters + working fluid parameters (no additional sampling is required, the heat load is directly transmitted from the first stage, and the pipe material / working fluid parameters are extracted from the basic parameters); Output: pipe wall temperature; Training: the model is trained with the same training set as the first stage (after parameter transmission), and the accuracy is verified.
[0030] The various embodiments described in this specification are implemented in a progressive manner, each embodiment focusing on the differences from other embodiments, and the same or similar parts between embodiments can be mutually referred to. For the apparatus disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.
[0031] The above description of disclosed embodiments enables one of ordinary skill in the art to make or use the application. Various modifications to these embodiments will be readily apparent to those of ordinary skill in the art, and the generic principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Therefore, the application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for controlling the temperature of a boiler heating surface pipe based on monitoring the wall temperature, characterized by, The application relates to a boiler parameter adjustment method based on a stage-wise optimization model. The method comprises the following steps: S1: collecting boiler data through a sensor, screening main parameters which have a significant influence on heating surface heat load and pipe wall temperature in the boiler data by combining global sensitivity analysis and local sensitivity analysis, and constructing a parameterized finite element model according to input parameters and output parameters in the main parameters; S2: obtaining training data by combining a test design and the parameterized finite element model; S3: constructing a stage-wise optimization model through the training data and verifying prediction accuracy; 2. The method of claim 1, wherein the temperature of the heating surface of the boiler is controlled based on the wall temperature. S4: extracting relevant parameters of an actual running boiler as inputs of the model, predicting and obtaining heating surface load and pipe wall temperature, and adjusting boiler parameters according to a preset adjustment priority when the pipe wall temperature exceeds a preset wall temperature threshold value to control the pipe wall temperature. In S1, parameters which have the greatest influence on outputs in the boiler data are selected as the main parameters. Input parameters include combustion system parameters (fuel quantity, air supply quantity and excess air coefficient), steam-water system parameters (feed water flow, desuperheating water flow and working medium temperature), flue gas system parameters (flue gas temperature and flue gas flow rate) and pipe material characteristic parameters (pipe material thermal conductivity and wall thickness). Output parameters include heating surface heat load and pipe wall temperature. In S1, the parameterized finite element model is constructed by inputting the screened main parameters, automatically generating a heat-structure coupling model of a heating surface pipe and outputting heating surface heat load and pipe wall temperature under corresponding working conditions based on a finite element software through a Python script interface.
3. The method of claim 2, wherein the temperature of the heating surface tube of the boiler is controlled based on the wall temperature. In S2, the training data are obtained by determining the value range of each input parameter from historical running data of a target boiler, a design manual and a same type engineering case, randomly sampling sample point sets in the parameter value range by using a Latin hypercube test design method, substituting the sample point sets into the parameterized finite element model to calculate corresponding heating surface heat load and pipe wall temperature to form an output response set, and normalizing the sample point set and the output response set to obtain a training data set.
4. The method of claim 1, wherein the temperature of the heating surface of the boiler is monitored by measuring the temperature of the wall of the boiler. The stage-wise optimization model in S3 specifically comprises: In S3, the prediction process of the stage-wise optimization model is divided into two stages, a first stage optimization model is used to predict heating surface heat load, and a second stage optimization model is used to predict pipe wall temperature.
5. The method of claim 1, wherein the temperature of the heating surface of the boiler is monitored by measuring the temperature of the wall of the boiler. The first stage optimization model is constructed by selecting combustion system parameters, steam-water system basic parameters and flue gas system key parameters as inputs, taking heating surface heat load as output, establishing a mapping relationship by using a Kriging model, optimizing a model hyperparameter theta1 by using a particle swarm optimization algorithm, taking a minimum heat load prediction mean square error as a target function, training the model by using normalized training data, and obtaining the first stage optimization model. ； wherein, is a first order regression model, is a regression coefficient; is a random process with mean 0 and variance The covariance is expressed as follows: ； wherein and are any two different points in the design space, is a correlation function that depends only on the spatial positions of the two points. ； wherein, is a pending model parameter, is any one of a linear function, an exponential function, a Gaussian function, a spherical function, and a cubic spline function; training sample set and its response set , the predicted response value of any new point to be tested can be estimated by linear combination of the known response values of the sample points ； wherein, The predicted error is expressed as: 。 6. The method of claim 1, wherein the temperature of the heating surface of the boiler is monitored by a temperature sensor. The second stage optimization model is constructed by taking the heating surface heat load output by the first stage optimization model and pipe material characteristic parameters and working medium parameters as inputs, taking pipe wall temperature as output, establishing a mapping relationship by using a same type Kriging model, optimizing a hyperparameter theta2 by using a particle swarm optimization algorithm, taking a minimum wall temperature prediction mean square error as a target function, training the model by using normalized sample data in a same proportion as that in the first stage, and obtaining the second stage optimization model. 7. The method of claim 1, wherein the temperature of the heating surface of the boiler is monitored by a temperature sensor. The boiler parameter adjustment according to the preset adjustment priority in the S4 specifically includes: the adjustment priorities are fuel oil system adjustment, steam-water system adjustment and flue gas system adjustment in sequence; When a priority adjustment is made; wherein, is a wall temperature setpoint, is an over-temperature threshold, is a predicted pipe wall temperature; The fuel oil system adjustment includes: ； ； wherein, is a fuel adjustment factor, is an optimal heat load, is an optimal excess air factor, is a fuel adjustment amount, is an air supply adjustment amount, is a first stage heating surface heat load, is an optimal air supply amount; When the fuel preset observation period is exceeded, Still over temperature, adjust the steam-water system: ； ； wherein, is a steam-water regulation coefficient, is a feed-water regulation flow rate, is a desuperheating water regulation flow rate, is a pipe material wall temperature set value, is a current feed-water flow rate, is a working medium specific heat capacity, is a steam flow rate, is a desuperheating cooling efficiency, is a desuperheating water temperature, is a heating surface inlet steam temperature; When the superheated steam exceeds the preset observation period, Still over temperature, adjust the flue gas system: ； wherein, is a flue gas adjustment factor, is a flue gas recirculation rate adjustment, is an optimal flue gas temperature, is an actual flue gas temperature.
8. A boiler heating surface tube material temperature control system based on monitoring wall temperature, characterized by, The finite element model construction module collects boiler data through a sensor, screens main parameters which have significant influence on the heating surface heat load and the pipe wall temperature in the boiler data by combining global sensitivity analysis and local sensitivity analysis, and constructs a parameterized finite element model according to input parameters and output parameters in the main parameters; The training data acquisition module acquires training data by combining experimental design and the parameterized finite element model; The stage model construction module constructs a stage optimization model by the training data and verifies the prediction accuracy; The regulation and control module extracts related parameters of the boiler in actual operation as inputs of the model, predicts and acquires the heating surface load and the pipe wall temperature, adjusts the boiler parameters according to the preset adjustment priority when the pipe wall temperature exceeds a preset wall temperature threshold, and regulates and controls the pipe wall temperature.