Coal-fired boiler real-time combustion fault identification method, system and equipment

Abstract

The invention discloses a coal-fired boiler real-time combustion fault identification method, system and equipment, and the method comprises the steps: carrying out the variable working condition CFD numerical simulation of the combustion state in a boiler, obtaining the temperature field data and speed field data in the boiler, carrying out POD decomposition, and extracting the POD modes and mode coefficient vectors of all simulated working conditions; acquiring actually-measured temperature data and actually-measured speed data of a site, and obtaining a temperature modal coefficient vector and a speed modal coefficient vector of a current working condition by combining the extracted POD mode and utilizing a gappy POD algorithm; the temperature modal coefficient vectors and the speed modal coefficient vectors of the current working condition and the simulated working conditions are traversed to compare included angles, the simulated working condition with the minimum included angle with the current working condition is obtained, and if the simulated working condition with the minimum included angle with the current working condition is an abnormal working condition, current combustion is abnormal. According to the invention, the combustion fault in the furnace can be identified in a short time.

Description

Technical Field

[0001] The invention relates to coal-fired boiler fault identification, and in particular to a coal-fired boiler real-time combustion fault identification method, system and equipment. Background Art

[0002] In order to ensure the smooth consumption of clean energy, deep peak regulation of thermal power units has become a normal practice, and the depth and duration of peak regulation are also increasing year by year, causing boilers to deviate from normal operating conditions. At the same time, in order to improve the economic efficiency of thermal power units, a large number of coal-fired power plants burn inferior coal, resulting in slagging of heating surfaces and blockage of nozzles, further causing problems such as unstable combustion and increased pollutant emissions.

[0003] Due to the high temperature and dust measurement environment in the furnace, it is difficult to accurately evaluate combustion faults in real time with limited on-site measurement methods. Most of the existing real-time combustion fault identification solutions use machine learning algorithms such as neural networks, and use a large amount of historical data to establish a prediction model to predict the current combustion state. There are two problems: on the one hand, the training process requires a large amount of combustion fault labeling data. The scale and quantity of the labeling data directly determine the quality of the prediction. However, in the actual operation of the power plant, the historical data lacks combustion fault labeling information; on the other hand, the power field operation data is mostly concentrated in medium and high loads, lacking low-load training samples, making it difficult to achieve low-load combustion state recognition and prediction. In addition, many researchers use CFD (Computational Fluid Dynamics) methods to study abnormal combustion states, but due to the time-consuming nature of CFD, it is more suitable for mechanism research rather than on-site prediction. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a method, system and equipment for real-time combustion fault identification of coal-fired boilers in view of the deficiencies in the prior art.

[0005] Technical solution: The real-time combustion fault identification method of a coal-fired boiler according to the present invention comprises:

[0006] (1) Perform CFD numerical simulation of the combustion state in the furnace under normal and abnormal conditions to obtain the temperature field data and velocity field data in the furnace and save them in matrix form;

[0007] (2) Perform POD decomposition on the temperature field data and velocity field data respectively, and extract the POD modes and modal coefficient vectors of all simulated working conditions;

[0008] (3) Obtain the measured temperature data and speed data on site, combine them with the extracted POD mode, and use the gappyPOD algorithm to obtain the temperature modal coefficient vector and speed modal coefficient vector of the current working condition;

[0009] (4) The temperature modal coefficient vector and the velocity modal coefficient vector of the current working condition and the simulated working condition are respectively traversed and compared in angle to obtain the simulated working condition with the smallest angle with the current working condition. If the simulated working condition with the smallest angle with the current working condition is an abnormal working condition, the current combustion is abnormal.

[0010] Furthermore, step (1) includes: establishing a three-dimensional physical model of the boiler and dividing it into grids; using computational fluid dynamics to couple chemical reactions to calculate the combustion dynamics field under different operating conditions; the models used in the simulation process: using the P-1 model for radiation heat transfer, using the realizable k-ε model for gas-solid flow, and using the DPM model for particle motion; after the model is set up, performing a large-scale variable operating condition numerical simulation to obtain the combustion dynamics field data in the furnace, including temperature field data and velocity field data.

[0011] Furthermore, step (1) further includes: converting the temperature field data and velocity field data in each working furnace into column vectors, and saving them in a matrix form in .csv or .mat format.

[0012] Furthermore, in step (1), the abnormal operating conditions include blockage of primary and secondary air and SOFA air nozzles, abnormal burner swing angle, abnormal coal powder particle size, and slagging on the heating surface.

[0013] Further, step (2) includes: taking the temperature field data as an example, firstly, the temperature field data of different working conditions are combined into a matrix S = [u (1) ,u (2) ,…,u (N) ],S∈M×N,u (1) ,u (2) ,…,u (N) is the temperature vector of N working conditions, M is the number of grids in the numerical simulation calculation process;

[0014] Let matrix R = S T S, calculate the eigenvalue of R, RΨ i =λ i Ψ i ,λ i is the i-th eigenvalue, Ψ i is the eigenvector corresponding to the i-th eigenvalue, then the i-th mode of the temperature field data matrix is for:

[0015]

[0016] Keep the first k modes and form the modal matrix with the first k modes

[0017] The modal coefficient matrix is: a = φ T S, each column of the modal coefficient matrix corresponds to the modal coefficient vector of each numerical simulation condition;

[0018] The calculation method of velocity modal coefficient vector is similar.

[0019] Further, step (3) includes: taking the temperature field as an example, first define the vector g R ∈M×1, where M is the number of grids in the numerical simulation calculation process, and the g at the temperature measurement point position in the furnace is set R The corresponding elements in are set to the measured temperature data, and the elements without measurement points are set to 0;

[0020] definition in is the i-th mode obtained by decomposing the temperature field data matrix, where m(x)=1 at the measurement point and m(x)=0 at other locations;

[0021] Define an overdetermined system of equations for the least squares fit:

[0022]

[0023] Among them, b i is the i-th modal coefficient under the current working condition; find the least square solution for the overdetermined equations, and form the vector b=(b 1 , b 2 , …, b k ), which is the modal coefficient vector of the current working condition;

[0024] The method for calculating the modal coefficient vector of the current working condition of the velocity field is similar.

[0025] Furthermore, step (3) also includes: preprocessing the measured temperature data and measured speed data obtained on site, including screening outliers, retaining significant digits, and performing basic addition, subtraction, multiplication, and division operations.

[0026] Furthermore, in step (4), the method for determining the angle γ between the current working condition and the simulated working condition is:

[0027] γ=w 1 α+w 2 β

[0028] Among them, α is the angle between the temperature modal coefficient vector of the current working condition and the simulated working condition, β is the angle between the speed modal coefficient vector of the current working condition and the simulated working condition, and w 1 is the temperature weight, w 2 is the speed weight;

[0029]

[0030]

[0031] Among them, b Tis the temperature modal coefficient vector of the current working condition, b v is the speed modal coefficient vector of the current working condition; a T is the temperature modal coefficient vector of the simulated working condition, a v is the velocity modal coefficient vector of the simulated working condition.

[0032] The coal-fired boiler real-time combustion fault identification system of the present invention comprises:

[0033] CFD numerical simulation module, used to perform variable operating condition CFD numerical simulation of the combustion state in the furnace, including normal operating condition and abnormal operating condition, to obtain the temperature field data and velocity field data in the furnace;

[0034] The sample database module is used to store the temperature field data and velocity field data obtained by the CFD numerical simulation module in matrix form;

[0035] The DCS data reading module is used to read the measured temperature data and measured speed data from the boiler DCS at a certain time interval, and pre-process them. The pre-processed results are passed to the POD feature extraction and coefficient prediction module;

[0036] The POD feature extraction and coefficient prediction module is used to perform POD decomposition on the temperature field data and velocity field data stored in the sample database module to obtain the POD modes and modal coefficient vectors of all simulated working conditions; based on the actual measured temperature data and actual measured velocity data on site, combined with the extracted POD modes, the gappy POD algorithm is used to obtain the temperature modal coefficient vector and velocity modal coefficient vector of the current working condition;

[0037] And, a combustion fault identification module is used to traverse and compare the angles of the temperature modal coefficient vector and the speed modal coefficient vector of the current operating condition and the simulated operating condition, respectively, to obtain the simulated operating condition with the smallest angle with the current operating condition. If the simulated operating condition with the smallest angle with the current operating condition is an abnormal condition, then the current combustion is abnormal.

[0038] The coal-fired boiler real-time combustion fault identification device of the present invention comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the coal-fired boiler real-time combustion fault identification method is implemented.

[0039] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: the present invention combines CFD numerical simulation results with on-site measurement data, and can identify combustion faults in the furnace in a short time, effectively solving the problem of difficulty in accurately and real-time evaluation of the combustion state due to limited on-site measurement means. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1is a flow chart of a method for identifying real-time combustion faults of a coal-fired boiler provided by an embodiment of the present invention;

[0041] Figure 2 It is a structural schematic diagram of a real-time combustion fault identification system for a coal-fired boiler provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The present invention will be further described below in conjunction with the accompanying drawings.

[0043] like Figure 1 As shown, the embodiment of the present invention provides a method for real-time combustion fault identification of a coal-fired boiler, which specifically includes the following steps:

[0044] (1) Perform CFD numerical simulation of the combustion state in the furnace under normal and abnormal conditions to obtain the temperature field data and velocity field data in the furnace and save them in matrix form;

[0045] A three-dimensional physical model of the boiler is established and meshed; CFD simulation conditions are designed according to the actual on-site operation conditions, including normal conditions and abnormal conditions. Abnormal conditions include blockage of primary and secondary air and SOFA air (burnout air) nozzles, abnormal burner swing angle, abnormal coal powder particle size, and slagging on the heating surface; then, computational fluid dynamics is coupled with chemical reactions to calculate the combustion dynamic field under different conditions; the models used in the simulation process are as follows: the P-1 model is used for radiation heat transfer, the k-ε model is used for gas-solid flow, and the DPM model is used for particle motion; after the model is set up, a large-scale variable condition numerical simulation is carried out to obtain the combustion dynamic field data in the furnace, including temperature field data and velocity field data;

[0046] Variable operating conditions include variable load, variable coal distribution, variable air distribution and other conditions.

[0047] The temperature field data and velocity field data in each working furnace are converted into column vectors and saved in matrix form in .csv or .mat format.

[0048] (2) Perform POD decomposition on the temperature field data and velocity field data respectively, and extract the POD modes and modal coefficient vectors of all simulated working conditions (including normal working conditions and abnormal working conditions);

[0049] Step (2) includes: taking the temperature field data as an example, firstly, the temperature field data of different working conditions are combined into a matrix S = [u (1) ,u (2) ,…,u (N) ],S∈M×N,u (1) ,u (2) ,…,u (N) is the temperature vector of N working conditions, M is the number of grids in the numerical simulation calculation process;

[0050] Let matrix R = S T S, calculate the eigenvalue of R, RΨ i =λ i Ψ i ,λ i is the i-th eigenvalue, Ψ i is the eigenvector corresponding to the i-th eigenvalue, then the i-th mode of the temperature field data matrix is for:

[0051]

[0052] In order to reduce the amount of calculation, the first k modes are retained and the first k modes are combined into a modal matrix

[0053] The modal coefficient matrix is: a = φ T S, each column of the modal coefficient matrix corresponds to the modal coefficient vector of each numerical simulation condition;

[0054] The calculation method of velocity modal coefficient vector is similar.

[0055] (3) Obtain the measured temperature data and speed data on site, combine them with the extracted POD mode, and use the gappyPOD algorithm to obtain the temperature modal coefficient vector and speed modal coefficient vector of the current working condition;

[0056] The gappy POD algorithm is a data reconstruction method based on proper orthogonal decomposition (POD).

[0057] Step (3) includes: preprocessing the measured temperature data and measured speed data obtained on site, including screening outliers, retaining significant digits, and basic addition, subtraction, multiplication, and division operations.

[0058] Taking the temperature field as an example, first define the vector g R ∈M×1, where M is the number of grids in the numerical simulation calculation process, and the g at the temperature measurement point position in the furnace is set R The corresponding elements in are set to the measured temperature data, and the elements without measurement points are set to 0;

[0059] definition in is the i-th mode obtained by decomposing the temperature field data matrix, where m(x)=1 at the measurement point and m(x)=0 at other locations;

[0060] Define an overdetermined system of equations for the least squares fit:

[0061]

[0062] Among them, b i is the i-th modal coefficient under the current working condition; find the least square solution for the overdetermined equations, and form the vector b=(b 1 , b 2 , …, b k ), which is the modal coefficient vector of the current working condition;

[0063] The method for calculating the modal coefficient vector of the current working condition of the velocity field is similar.

[0064] (4) The temperature modal coefficient vector and the velocity modal coefficient vector of the current working condition and the simulated working condition are respectively traversed and compared in angle to obtain the simulated working condition with the smallest angle with the current working condition. If the simulated working condition with the smallest angle with the current working condition is an abnormal working condition, the current combustion is abnormal.

[0065] The present invention considers that the simulated operating condition with the smallest angle is closest to the current operating condition: if the current operating condition is close to the normal operating condition in the simulated operating condition, the current combustion state is considered normal; if the current operating condition is close to the abnormal operating condition in the simulated operating condition, the current combustion is considered abnormal. Then, the characteristics characterizing the abnormal operating condition (such as blockage of primary and secondary air and SOFA air nozzles, abnormal burner swing angle, abnormal coal powder particle size, slagging on the heating surface, etc.) can be further determined as the cause of the current combustion failure.

[0066] The method for determining the angle γ between the current working condition and the simulated working condition is:

[0067] γ=w 1 α+w 2 β

[0068] Among them, α is the angle between the temperature modal coefficient vector of the current working condition and the simulated working condition, β is the angle between the speed modal coefficient vector of the current working condition and the simulated working condition; w 1 is the temperature weight, w 2 is the speed weight, which is set manually;

[0069]

[0070] Among them, b T is the temperature modal coefficient vector of the current working condition, b v is the speed modal coefficient vector of the current working condition; a T is the temperature modal coefficient vector of the simulated working condition, a v is the velocity modal coefficient vector of the simulated working condition.

[0071] By repeating the above steps (3) and (4), real-time combustion fault identification of coal-fired boilers can be achieved.

[0072] like Figure 2As shown, the embodiment of the present invention further provides a real-time combustion fault identification system for a coal-fired boiler, which is used to implement the real-time combustion fault identification method for a coal-fired boiler according to the embodiment of the present invention, comprising:

[0073] CFD numerical simulation module 1, used for performing variable working condition CFD numerical simulation of the combustion state in the furnace including normal working condition and abnormal working condition, and obtaining temperature field data and velocity field data in the furnace;

[0074] The sample database module 2 is used to store the temperature field data and velocity field data obtained by the CFD numerical simulation module 1 in a matrix form;

[0075] The DCS data reading module 3 is used to read the measured temperature data and measured speed data at the site from the boiler DCS at a certain time interval, and preprocess them, including abnormal value screening, retaining the number of significant digits and basic addition, subtraction, multiplication and division operations; the preprocessed results are transmitted to the POD feature extraction and coefficient prediction module 4;

[0076] The POD feature extraction and coefficient prediction module 4 is used to receive data from the sample database module 2 and the DCS data reading module 3 respectively, firstly perform POD decomposition on the temperature field data and the velocity field data stored in the sample database module respectively, and obtain the POD mode and modal coefficient vector of all simulated working conditions (including normal working conditions and abnormal working conditions); then, based on the actual measured temperature data and the actual measured velocity data (i.e., the data read by the DCS data reading module 3) on site, combined with the extracted POD mode, the gappy POD algorithm is used to obtain the temperature modal coefficient vector and the velocity modal coefficient vector of the current working condition;

[0077] And, the combustion fault identification module 5 is used to receive the CFD simulation working condition and the current working condition modal coefficient vector from the POD feature extraction and coefficient prediction module, and traverse and compare the angles of the temperature modal coefficient vector and the speed modal coefficient vector of the current working condition and the simulated working condition, and obtain the simulated working condition with the smallest angle with the current working condition. If the simulated working condition with the smallest angle with the current working condition is an abnormal working condition, then the current combustion is abnormal.

[0078] An embodiment of the present invention also provides a coal-fired boiler real-time combustion fault identification device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the coal-fired boiler real-time combustion fault identification method described in the embodiment of the present invention is implemented.

Claims

1. A method for real-time combustion fault identification of a coal-fired boiler, characterized in that: include: (1) Perform CFD numerical simulation of the combustion state in the furnace under normal and abnormal conditions to obtain the temperature field data and velocity field data in the furnace and save them in matrix form; (2) Perform POD decomposition on the temperature field data and velocity field data respectively, and extract the POD modes and modal coefficient vectors of all simulated working conditions; (3) Obtain the measured temperature data and speed data on site, combine them with the extracted POD mode, and use the gappy POD algorithm to obtain the temperature modal coefficient vector and speed modal coefficient vector of the current working condition; (4) The temperature modal coefficient vector and the velocity modal coefficient vector of the current working condition and the simulated working condition are respectively traversed and compared in angle to obtain the simulated working condition with the smallest angle with the current working condition. If the simulated working condition with the smallest angle with the current working condition is an abnormal working condition, the current combustion is abnormal.

2. The method for real-time combustion fault identification of a coal-fired boiler according to claim 1, characterized in that: Step (1) includes: establishing a three-dimensional physical model of the boiler and dividing it into grids; using computational fluid dynamics to couple chemical reactions and calculate the combustion dynamics field under different working conditions; the models used in the simulation process: using the P-1 model for radiation heat transfer, using the realizable k-ε model for gas-solid flow, and using the DPM model for particle motion; after the model is set up, performing a large-scale variable working condition numerical simulation to obtain the combustion dynamics field data in the furnace, including temperature field data and velocity field data.

3. The method for real-time combustion fault identification of a coal-fired boiler according to claim 2, characterized in that: Step (1) also includes: converting the temperature field data and velocity field data in each working furnace into column vectors, saving them in matrix form in .csv or .mat format.

4. The method for real-time combustion fault identification of a coal-fired boiler according to claim 1, characterized in that: In step (1), the abnormal operating conditions include blockage of primary and secondary air and SOFA air nozzles, abnormal burner swing angle, abnormal coal powder particle size, and slagging on the heating surface.

5. The method for real-time combustion fault identification of a coal-fired boiler according to claim 1, characterized in that: Step (2) includes: taking the temperature field data as an example, firstly, the temperature field data of different working conditions are combined into a matrix S = [u (1) ,u (2) ,…,u (N) ],S∈M×N,u (1) ,u (2) ,…,u (N) is the temperature vector of N working conditions, M is the number of grids in the numerical simulation calculation process; Let matrix R = S T S, calculate the eigenvalue of R, RΨ i =λ i Ψ i ,λ i is the i-th eigenvalue, Ψ i is the eigenvector corresponding to the i-th eigenvalue, then the i-th mode of the temperature field data matrix is for: Keep the first k modes and form the modal matrix with the first k modes The modal coefficient matrix is: a = φ T S, each column of the modal coefficient matrix corresponds to the modal coefficient vector of each numerical simulation condition; The calculation method of velocity modal coefficient vector is similar.

6. The method for real-time combustion fault identification of a coal-fired boiler according to claim 1, characterized in that: Step (3) includes: Taking the temperature field as an example, first define the vector g R ∈M×1, where M is the number of grids in the numerical simulation calculation process, and the g at the temperature measurement point position in the furnace is set R The corresponding elements in are set to the measured temperature data, and the elements without measurement points are set to 0; definition in is the i-th mode obtained by decomposing the temperature field data matrix, where m(x)=1 at the measurement point and m(x)=0 at other locations; Define an overdetermined system of equations for the least squares fit: Among them, b i is the i-th modal coefficient under the current working condition; find the least square solution for the overdetermined equations, and form the k coefficients into a vector b=(b1, b2, …, b k ), which is the modal coefficient vector of the current working condition; The method for calculating the modal coefficient vector of the current working condition of the velocity field is similar.

7. The method for real-time combustion fault identification of a coal-fired boiler according to claim 6, characterized in that: Step (3) also includes: preprocessing the measured temperature data and measured speed data obtained on site, including screening outliers, retaining significant digits, and basic addition, subtraction, multiplication, and division operations.

8. The method for real-time combustion fault identification of a coal-fired boiler according to claim 1, characterized in that: In step (4), the method for determining the angle γ between the current working condition and the simulated working condition is: γ=w1α+w2β Among them, α is the angle between the temperature modal coefficient vector of the current working condition and the simulated working condition, β is the angle between the speed modal coefficient vector of the current working condition and the simulated working condition, w1 is the temperature weight, and w2 is the speed weight; Among them, b T is the temperature modal coefficient vector of the current working condition, b v is the speed modal coefficient vector of the current working condition; a T is the temperature modal coefficient vector of the simulated working condition, a v is the velocity modal coefficient vector of the simulated working condition.

9. A real-time combustion fault identification system for a coal-fired boiler, characterized in that: include: CFD numerical simulation module, used to perform variable operating condition CFD numerical simulation of the combustion state in the furnace, including normal operating condition and abnormal operating condition, to obtain the temperature field data and velocity field data in the furnace; The sample database module is used to store the temperature field data and velocity field data obtained by the CFD numerical simulation module in matrix form; The DCS data reading module is used to read the measured temperature data and measured speed data from the boiler DCS at a certain time interval, and pre-process them. The pre-processed results are passed to the POD feature extraction and coefficient prediction module; The POD feature extraction and coefficient prediction module is used to perform POD decomposition on the temperature field data and velocity field data stored in the sample database module to obtain the POD modes and modal coefficient vectors of all simulated working conditions; based on the actual measured temperature data and actual measured velocity data on site, combined with the extracted POD modes, the gappy POD algorithm is used to obtain the temperature modal coefficient vector and velocity modal coefficient vector of the current working condition; And, a combustion fault identification module is used to traverse and compare the angles of the temperature modal coefficient vector and the speed modal coefficient vector of the current operating condition and the simulated operating condition, respectively, to obtain the simulated operating condition with the smallest angle with the current operating condition. If the simulated operating condition with the smallest angle with the current operating condition is an abnormal condition, then the current combustion is abnormal.

10. A coal-fired boiler real-time combustion fault identification device, comprising a memory and a processor, wherein a computer program is stored in the memory, characterized in that: When the computer program is executed by a processor, the method for real-time combustion fault identification of a coal-fired boiler according to any one of claims 1 to 8 is implemented.

Images