Method and system for predicting NOx concentration field of boiler based on multi-model integration weighting

Abstract

The invention relates to the field of boiler NOx concentration field online prediction, in particular to a method and system for predicting a boiler NOx concentration field based on multi-model integration weighting. The method comprises the steps of 1, constructing boiler operation typical working condition data; 2, boiler CFD numerical simulation data are obtained; 3, preprocessing the boiler numerical simulation data and the field operation data to obtain a training data set; 4, constructing a boiler NOx concentration field prediction initial model based on multi-model integration weighting; step 5, training the prediction initial model constructed in the step 4 by adopting the training data set obtained in the step 3 so as to obtain a prediction model; and 6, predicting the NOx concentration three-dimensional distribution of the target boiler by adopting the obtained trained prediction model. The method solves the problems that the boiler operation environment is severe, measurement equipment cannot perform holographic monitoring, and a single data driving model is difficult to comprehensively mine data deep features.

Description

Technical Field

[0001] This invention relates to the field of online prediction of NOx concentration fields in boilers, and in particular to a method and system for predicting NOx concentration fields in boilers based on multi-model integrated weighted prediction. Background Technology

[0002] Generating power plays a crucial role in deep peak shaving in new power systems. To meet the rapid adjustment demands of the power system, generator units need to operate under frequent load changes for extended periods. During this time, operators struggle to obtain accurate real-time combustion data (such as NOx concentration distribution) within the furnace, hindering real-time analysis of the boiler's combustion status. This significantly reduces the timeliness and efficiency of combustion process control, leading to decreased unit efficiency and increased pollutant emissions. However, because the boiler combustion process is a "black box," and the boiler environment is characterized by high temperatures, high noise levels, and high dust levels, there is currently a lack of methods for real-time, high-precision prediction of the three-dimensional NOx concentration distribution in boilers.

[0003] Currently, there are two main methods for measuring the three-dimensional distribution of NOx concentration in boilers: one is to use computational fluid dynamics (CFD) to numerically simulate the three-dimensional distribution of NOx concentration in boilers. However, this technique involves complex physical and chemical reaction processes such as turbulence, resulting in long computation times and slow convergence speeds, which cannot meet the needs of real-time boiler monitoring. The other method is to use a single data-driven neural network to predict the three-dimensional distribution of NOx concentration in boilers. However, the establishment of a single data-driven model ignores the deeper features of other data, resulting in insufficient model generalization ability and difficulty in obtaining high-precision prediction results that conform to the inherent characteristics of the system. Therefore, there is an urgent need to develop a method for real-time, high-precision monitoring of the three-dimensional distribution of NOx concentration in boilers. Summary of the Invention

[0004] This invention provides a method for predicting the NOx concentration field in boilers based on multi-model ensemble weighted prediction, which solves the problems of harsh boiler operating environment, inability of measurement equipment to perform holographic monitoring, and difficulty of fully mining deep features of data by a single data-driven model.

[0005] This invention provides a system for predicting boiler NOx concentration fields based on multi-model ensemble weighted prediction, thereby realizing a method for predicting boiler NOx concentration fields based on multi-model ensemble weighted prediction.

[0006] This invention is achieved through the following technical solution:

[0007] A method for predicting boiler NOx concentration fields based on multi-model ensemble weighted average, the method comprising the following steps:

[0008] Step 1: Obtain actual production data from the power plant and construct typical boiler operating condition data (D). iDCS ;

[0009] Step 2: Based on the typical boiler operating condition data D constructed in Step 1 iDCS Numerical simulations were performed on typical operating conditions to obtain boiler CFD numerical simulation data D. iCFD ;

[0010] Step 3: Transfer the boiler numerical simulation data D from Step 2 to... iCFD With field operation data D iDCS Preprocessing is performed to obtain the training dataset D. i ;

[0011] Step 4: Construct an initial model for predicting boiler NOx concentration field based on multi-model ensemble weighting, using BPNN neural network, ELM neural network and DNN neural network.

[0012] Step 5: Obtain the training dataset D using the methods described in Step 3. i The initial model for predicting the NOx concentration field in the boiler, constructed in step 4, is trained to obtain the NOx concentration field prediction model for the boiler.

[0013] Step 6: Using the trained boiler NOx concentration field prediction model obtained in Step 5, predict the three-dimensional distribution of NOx concentration in the target boiler.

[0014] Furthermore, step 1 specifically involves the following: operating parameters include total air volume TA, total coal volume TC, unit load UL, primary air temperature PT, secondary air temperature ST, primary air volume PA, secondary air volume SA, and burnout air volume SF; and the operating data are categorized according to the pulverizer's operating status.

[0015] Furthermore, step 2 specifically involves performing numerical simulations on typical operating conditions, setting boundary conditions based on the boiler's combustion characteristics, and determining the gas-phase turbulence model, gas-phase combustion model, radiative heat transfer model, and coke combustion model within the furnace.

[0016] The gas-phase turbulence model inside the furnace adopts the Realizable model; the kinetic / diffusion reaction rate model is used to simulate the coke combustion process; the radiative heat transfer between gas and solid is calculated using the DO model; M different sets of operating parameters are set, and m simulation operations are performed for each set of operating parameters to obtain the boiler CFD numerical simulation data D. iCFD .

[0017] Furthermore, step 3 specifically involves normalizing the basic modeling data to obtain the training dataset D. i .

[0018] Furthermore, step 4 specifically involves constructing a first initial prediction model, a second initial prediction model, and a third initial prediction model based on BPNN neural network, ELM neural network, and DNN neural network, respectively.

[0019] The training data was sequentially input into the first prediction initial model based on a BPNN neural network, the second prediction initial model based on an ELM neural network, and the third prediction initial model based on a DNN neural network. The prediction results obtained from the first, second, and third prediction initial models were inversely normalized and then integrated and weighted to obtain the output of the integrated and weighted boiler NOx concentration field prediction initial model:

[0020]

[0021] in, The output of the initial model for predicting the NOx concentration field in the boiler corresponding to the i-th training input data; The output of the first prediction initial model based on the BPNN neural network corresponding to the i-th training input data; The output of the second prediction initial model based on the ELM neural network corresponding to the i-th training input data; x1 is the output of the third prediction initial model based on the DNN neural network corresponding to the i-th training input data; x2 is the weight value of the first prediction initial model; x3 is the weight value of the third prediction initial model; the weights of the three initial prediction models are x1+x2+x3=1, and 0≤x1≤1,0≤x2≤1,0≤x3≤1.

[0022] Furthermore, step 5 specifically involves using the following formula as the objective function during the training process of the boiler NOx concentration field prediction model based on multi-model ensemble weighting:

[0023]

[0024] Where F is the objective function, and x1, x2, and x3 represent the prediction results of the BPNN, ELM, and DNN models, respectively, with x1, x2, and x3 being the weights of the prediction results for each model. The output of the initial model for predicting the NOx concentration field of the boiler corresponding to the i-th training input data.

[0025] Furthermore, step 6 specifically involves using the trained boiler NOx concentration field prediction model obtained in step 5 to verify the accuracy of the three-dimensional distribution of NOx concentration in the target boiler.

[0026] A system for predicting boiler NOx concentration fields based on multi-model ensemble weighted average, the system using the method described above for predicting boiler NOx concentration fields based on multi-model ensemble weighted average, the system comprising,

[0027] The module constructs typical boiler operating condition data (D) based on the acquired actual production data of the power plant. iDCS ;

[0028] The numerical simulation module performs numerical simulations on typical operating conditions to obtain boiler CFD numerical simulation data D. iCFD ;

[0029] The preprocessing module processes the boiler numerical simulation data D. iCFD With field operation data D iDCS Preprocessing is performed to obtain the training dataset D. i ;

[0030] The initial model building module, based on BPNN neural network, ELM neural network and DNN neural network, constructs an initial model for predicting NOx concentration field in boilers based on multi-model ensemble weighting;

[0031] The prediction model module obtains the training dataset D using step 3. i The initial model for predicting the NOx concentration field in the boiler, constructed in step 4, is trained to obtain the NOx concentration field prediction model for the boiler.

[0032] The trained boiler NOx concentration field prediction model is used to predict the three-dimensional distribution of NOx concentration in the target boiler.

[0033] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the method described above.

[0034] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described above.

[0035] The beneficial effects of this invention are:

[0036] This invention solves the problems of harsh boiler operating environment, inability of measurement equipment to conduct holographic monitoring, and difficulty of fully mining deep features of data using a single data-driven model.

[0037] This invention can better meet actual production needs and has broad application prospects. Attached Figure Description

[0038] Figure 1 This is a flowchart of the method of the present invention.

[0039] Figure 2 This is a comparison chart of the measured NOx distribution in the boiler furnace and the target value based on numerical simulation data, according to the present invention. Detailed Implementation

[0040] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of this application with unnecessary detail.

[0041] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0042] It should also be understood that the terminology used in this application specification is for the purpose of describing particular embodiments only and is not intended to limit the application. As used in this application specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0043] The following is in conjunction with the appendix to this application specification. Figure 1-2 The technical solutions in the embodiments of this application are clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0044] Many specific details are set forth in the following description in order to provide a full understanding of this application. However, this application may also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0045] Implementation Method 1

[0046] This embodiment provides a method for predicting boiler NOx concentration fields based on multi-model ensemble weighted prediction. The method includes the following steps:

[0047] Step 1: Obtain actual production data from the power plant and construct typical boiler operating condition data (D). iDCS ;

[0048] Step 2: Based on the typical boiler operating condition data D constructed in Step 1 iDCS Numerical simulations were performed on typical operating conditions to obtain boiler CFD numerical simulation data D. iCFD ;

[0049] Step 3: Transfer the boiler numerical simulation data D from Step 2 to... iCFD With field operation data D iDCS Preprocessing is performed to obtain the training dataset D. i ;

[0050] Step 4: Construct an initial model for predicting boiler NOx concentration field based on multi-model ensemble weighting, using BPNN neural network, ELM neural network and DNN neural network.

[0051] Step 5: Obtain the training dataset D using the methods described in Step 3. i The initial model for predicting the NOx concentration field in the boiler, constructed in step 4, is trained to obtain the NOx concentration field prediction model for the boiler.

[0052] Step 6: Using the trained boiler NOx concentration field prediction model obtained in Step 5, predict the three-dimensional distribution of NOx concentration in the target boiler.

[0053] Furthermore, step 1 specifically involves obtaining a typical operating condition dataset of the boiler from the power plant's DCS system. The operating parameters include total air volume TA, total coal volume TC, unit load UL, primary air temperature PT, secondary air temperature ST, primary air volume PA, secondary air volume SA, and burnout air volume SF. The operating condition data is then categorized according to the coal mill's operating status.

[0054] Furthermore, step 2 specifically involves performing numerical simulations on typical operating conditions, setting boundary conditions based on the boiler's combustion characteristics, and determining the gas-phase turbulence model, gas-phase combustion model, radiative heat transfer model, and coke combustion model within the furnace.

[0055] The gas-phase turbulence model inside the furnace adopts the Realizable model; the kinetic / diffusion reaction rate model is used to simulate the coke combustion process; the radiative heat transfer between gas and solid is calculated using the DO model; M different sets of operating parameters are set, and m simulation operations are performed for each set of operating parameters to obtain the boiler CFD numerical simulation data D. iCFD .

[0056] Furthermore, step 3 specifically involves fusing CFD numerical simulation data with DCS field operation data to construct basic modeling data. This process can be represented by the following formula. The basic modeling data is then normalized to obtain the training dataset D. i .

[0057] Furthermore, step 4 specifically involves constructing a first initial prediction model, a second initial prediction model, and a third initial prediction model based on BPNN neural network, ELM neural network, and DNN neural network, respectively.

[0058] The training data was sequentially input into the first prediction initial model based on a BPNN neural network, the second prediction initial model based on an ELM neural network, and the third prediction initial model based on a DNN neural network. The prediction results obtained from the first, second, and third prediction initial models were inversely normalized and then integrated and weighted to obtain the output of the integrated and weighted boiler NOx concentration field prediction initial model:

[0059]

[0060] in, The output of the initial model for predicting the NOx concentration field in the boiler corresponding to the i-th training input data; The output of the first prediction initial model based on the BPNN neural network corresponding to the i-th training input data; The output of the second prediction initial model based on the ELM neural network corresponding to the i-th training input data; x1 is the output of the third prediction initial model based on the DNN neural network corresponding to the i-th training input data; x2 is the weight value of the first prediction initial model; x3 is the weight value of the third prediction initial model; the weights of the three initial prediction models are x1+x2+x3=1, and 0≤x1≤1,0≤x2≤1,0≤x3≤1.

[0061] Furthermore, step 5 specifically involves using the following formula as the objective function during the training process of the boiler NOx concentration field prediction model based on multi-model ensemble weighting:

[0062]

[0063] Where F is the objective function, and x1, x2, and x3 represent the prediction results of the BPNN, ELM, and DNN models, respectively, with x1, x2, and x3 being the weights of the prediction results for each model. The output of the initial model for predicting the NOx concentration field of the boiler corresponding to the i-th training input data.

[0064] Furthermore, step 6 specifically involves accuracy verification, which uses the trained boiler NOx concentration field prediction model obtained in step 5 to verify the accuracy of the three-dimensional distribution of NOx concentration in the target boiler. Figure 2 As can be seen from the results, this method has high model accuracy and can quickly predict the three-dimensional NOx concentration distribution in the boiler furnace. The method proposed in this invention better meets the actual production needs and has broad application prospects.

[0065] Implementation Method 2

[0066] This embodiment provides a system for predicting boiler NOx concentration fields based on multi-model ensemble weighted average. The system uses the method for predicting boiler NOx concentration fields based on multi-model ensemble weighted average as described in Embodiment 1. The system includes...

[0067] The module constructs typical boiler operating condition data (D) based on the acquired actual production data of the power plant. iDCS ;

[0068] The numerical simulation module performs numerical simulations on typical operating conditions to obtain boiler CFD numerical simulation data D. iCFD ;

[0069] The preprocessing module processes the boiler numerical simulation data D. iCFD With field operation data D iDCS Preprocessing is performed to obtain the training dataset D. i ;

[0070] The initial model building module, based on BPNN neural network, ELM neural network and DNN neural network, constructs an initial model for predicting NOx concentration field in boilers based on multi-model ensemble weighting;

[0071] The prediction model module obtains the training dataset D using step 3. i The initial model for predicting the NOx concentration field in the boiler, constructed in step 4, is trained to obtain the NOx concentration field prediction model for the boiler.

[0072] The trained boiler NOx concentration field prediction model is used to predict the three-dimensional distribution of NOx concentration in the target boiler.

[0073] Furthermore, the working process of the construction module is as follows: it obtains a typical operating condition dataset of the boiler during operation from the power plant DCS system. The operating parameters include total air volume TA, total coal volume TC, unit load UL, primary air temperature PT, secondary air temperature ST, primary air volume PA, secondary air volume SA, and burnout air volume SF; and classifies the operating condition data according to the coal mill's operating status.

[0074] Furthermore, the working process of the numerical simulation module specifically involves performing numerical simulations on typical operating conditions, setting boundary conditions based on the combustion characteristics of the boiler, and determining the gas-phase turbulence model, gas-phase combustion model, radiative heat transfer model, and coke combustion model within the furnace.

[0075] The gas-phase turbulence model inside the furnace adopts the Realizable model; the kinetic / diffusion reaction rate model is used to simulate the coke combustion process; the radiative heat transfer between gas and solid is calculated using the DO model; M different sets of operating parameters are set, and m simulation operations are performed for each set of operating parameters to obtain the boiler CFD numerical simulation data D. iCFD .

[0076] Furthermore, the preprocessing module specifically works by fusing CFD numerical simulation data with DCS field operation data to construct basic modeling data. This process can be represented by the following formula. The basic modeling data is then normalized to obtain the training dataset D. i .

[0077] Furthermore, the working process of the initial model construction module is specifically as follows: a first prediction initial model, a second prediction initial model, and a third prediction initial model are constructed based on BPNN neural network, ELM neural network, and DNN neural network, respectively.

[0078] The training data was sequentially input into the first prediction initial model based on a BPNN neural network, the second prediction initial model based on an ELM neural network, and the third prediction initial model based on a DNN neural network. The prediction results obtained from the first, second, and third prediction initial models were inversely normalized and then integrated and weighted to obtain the output of the integrated and weighted boiler NOx concentration field prediction initial model:

[0079]

[0080] in, The output of the initial model for predicting the NOx concentration field in the boiler corresponding to the i-th training input data; The output of the first prediction initial model based on the BPNN neural network corresponding to the i-th training input data; The output of the second prediction initial model based on the ELM neural network corresponding to the i-th training input data; x1 is the output of the third prediction initial model based on the DNN neural network corresponding to the i-th training input data; x2 is the weight value of the first prediction initial model; x3 is the weight value of the third prediction initial model; the weights of the three initial prediction models are x1+x2+x3=1, and 0≤x1≤1,0≤x2≤1,0≤x3≤1.

[0081] Furthermore, the specific working process of the prediction model module is as follows: during the training process, the boiler NOx concentration field prediction model based on multi-model ensemble weighting uses the following formula as the objective function:

[0082]

[0083] Where F is the objective function, and x1, x2, and x3 represent the prediction results of the BPNN, ELM, and DNN models, respectively, with x1, x2, and x3 being the weights of the prediction results for each model. The output of the initial model for predicting the NOx concentration field of the boiler corresponding to the i-th training input data.

[0084] Furthermore, the verification process of the prediction model module specifically involves accuracy verification, which uses the trained boiler NOx concentration field prediction model obtained in step 5 to verify the accuracy of the three-dimensional distribution of NOx concentration in the target boiler. Figure 2 As can be seen from the results, this method has high model accuracy and can quickly predict the three-dimensional NOx concentration distribution in the boiler furnace. The method proposed in this invention better meets the actual production needs and has broad application prospects.

[0085] Implementation Method 3

[0086] This invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor. The memory stores software programs and modules, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory and processor are connected via a bus. Specifically, the processor implements any step in Embodiment 1 by running the computer program stored in the memory.

[0087] It should be understood that, in the embodiments of the present invention, the processor may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0088] Memory may include read-only memory, flash memory, and random access memory, and provides instructions and data to the processor. Some or all of the memory may also include non-volatile random access memory.

[0089] As can be seen from the above, the electronic device provided by the embodiments of the present invention can implement the method of predicting the NOx concentration field of a boiler based on multi-model ensemble weighted prediction as described in Embodiment 1 by running a computer program, thereby obtaining a new method for predicting the three-dimensional distribution of NOx concentration in a target boiler based on multi-model ensemble weighted prediction, which is used to predict the three-dimensional distribution of NOx concentration in a target boiler.

[0090] It should be understood that if the integrated modules / units described above are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods described above can also be implemented by a computer program instructing related hardware. This computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction.

[0091] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not 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.

[0092] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the above device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this invention. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0093] It should be noted that the methods and detailed examples provided in the above embodiments can be incorporated into the apparatus and devices provided in the embodiments for mutual reference, and will not be repeated here.

[0094] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in 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 implementations should not be considered beyond the scope of this invention.

[0095] In the embodiments provided by this invention, it should be understood that the disclosed apparatus / terminal devices and methods can be implemented in other ways. For example, the apparatus / device embodiments described above are merely illustrative. For instance, the division of modules or units described above is merely a logical functional division, and in actual implementation, it can be divided in other ways. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0096] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for predicting boiler NOx concentration fields based on multi-model ensemble weighted prediction, characterized in that, The method includes the following steps: Step 1: Obtain actual production data from the power plant and construct typical boiler operating condition data (D). iDCS ; Step 2: Based on the typical boiler operating condition data D constructed in Step 1 iDCS Numerical simulations were performed on typical operating conditions to obtain boiler CFD numerical simulation data D. iCFD ; Step 3: Transfer the boiler numerical simulation data D from Step 2 to... iCFD With field operation data D iDCS Preprocessing is performed to obtain the training dataset D. i ; Step 4: Construct an initial model for predicting boiler NOx concentration field based on multi-model ensemble weighting, using BPNN neural network, ELM neural network and DNN neural network. Step 5: Obtain the training dataset D using the methods described in Step 3. i The initial model for predicting the NOx concentration field in the boiler, constructed in step 4, is trained to obtain the NOx concentration field prediction model for the boiler. Step 6: Using the trained boiler NOx concentration field prediction model obtained in Step 5, predict the three-dimensional distribution of NOx concentration in the target boiler.

2. The method according to claim 1, characterized in that, Step 1 specifically involves configuring operating parameters including total air volume TA, total coal volume TC, unit load UL, primary air temperature PT, secondary air temperature ST, primary air volume PA, secondary air volume SA, and burnout air volume SF; and classifying the operating data according to the coal mill's operating status.

3. The method according to claim 1, characterized in that, Step 2 specifically involves performing numerical simulations on typical operating conditions, setting boundary conditions based on the boiler's combustion characteristics, and determining the gas-phase turbulence model, gas-phase combustion model, radiative heat transfer model, and coke combustion model within the furnace. The gas-phase turbulence model inside the furnace adopts the Realizable model; the kinetic / diffusion reaction rate model is used to simulate the coke combustion process; the radiative heat transfer between gas and solid is calculated using the DO model; M different sets of operating parameters are set, and m simulation operations are performed for each set of operating parameters to obtain the boiler CFD numerical simulation data D. iCFD .

4. The method according to claim 1, characterized in that, Step 3 specifically involves normalizing the basic modeling data to obtain the training dataset D. i .

5. The method according to claim 4, characterized in that, Step 4 specifically involves constructing a first prediction initial model, a second prediction initial model, and a third prediction initial model based on BPNN neural network, ELM neural network, and DNN neural network, respectively. The training data was sequentially input into the first prediction initial model based on a BPNN neural network, the second prediction initial model based on an ELM neural network, and the third prediction initial model based on a DNN neural network. The prediction results obtained from the first, second, and third prediction initial models were inversely normalized and then integrated and weighted to obtain the output of the integrated and weighted boiler NOx concentration field prediction initial model: in, The output of the initial model for predicting the NOx concentration field in the boiler corresponding to the i-th training input data; The output of the first prediction initial model based on the BPNN neural network corresponding to the i-th training input data; The output of the second prediction initial model based on the ELM neural network corresponding to the i-th training input data; x1 is the output of the third initial prediction model based on the DNN neural network corresponding to the i-th training input data; x2 is the weight value of the first initial prediction model; x3 is the weight value of the second initial prediction model; the weights of the three initial prediction models are x1+x2+x3=1, and 0≤x1≤1,0≤x2≤1,0≤x3≤1.

6. The method according to claim 5, characterized in that, Step 5 specifically involves using the following formula as the objective function during the training process of the boiler NOx concentration field prediction model based on multi-model ensemble weighting: Where F is the objective function, and x1, x2, and x3 represent the prediction results of the BPNN, ELM, and DNN models, respectively, with x1, x2, and x3 being the weights of the prediction results for each model. The output of the initial model for predicting the NOx concentration field of the boiler corresponding to the i-th training input data.

7. The method according to claim 2, characterized in that, Step 6 specifically involves using the trained boiler NOx concentration field prediction model obtained in step 5 to verify the accuracy of the three-dimensional distribution of NOx concentration in the target boiler.

8. A system for predicting boiler NOx concentration fields based on multi-model ensemble weighted prediction, characterized in that, The system uses the method for predicting boiler NOx concentration fields based on multi-model ensemble weighted prediction as described in any one of claims 1-7, and the system comprises, The module constructs typical boiler operating condition data (D) based on the acquired actual production data of the power plant. iDCS ; The numerical simulation module performs numerical simulations on typical operating conditions to obtain boiler CFD numerical simulation data D. iCFD ; The preprocessing module processes the boiler numerical simulation data D iCFD With field operation data D iDCS Preprocessing is performed to obtain the training dataset D. i ; The initial model building module, based on BPNN neural network, ELM neural network and DNN neural network, constructs an initial model for predicting NOx concentration field in boilers based on multi-model ensemble weighting; The prediction model module obtains the training dataset D using step 3. i The initial model for predicting the NOx concentration field in the boiler, constructed in step 4, is trained to obtain the NOx concentration field prediction model for the boiler. The trained boiler NOx concentration field prediction model is used to predict the three-dimensional distribution of NOx concentration in the target boiler.

9. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1-7.

Images