Combustor regulation and control system and method for reducing thermal shock of water cooling wall under variable load

Abstract

The invention provides a combustor regulation and control system and method for reducing thermal shock of a water cooling wall under variable loads. The method comprises the following steps that S101, boiler parameters, historical operation data of a boiler and boiler variables are obtained; s102, constructing a hearth combustion model based on the boiler parameters and verifying the hearth combustion model; s103, arranging different states of boiler variables into different operation combinations by adopting an orthogonal test method, and obtaining corresponding hearth water wall temperature distribution data; s104, inputting the operation combinations with different boiler variables and the corresponding hearth water-cooled wall temperature distribution data into a machine learning model for training; s105, temperature distribution data of the water cooling wall of the hearth under the load at the given moment in the future are predicted; s106, an operation combination of the boiler variables with the overtemperature data is screened out; and S107, the combustor combination enabling the heat shock borne by the water cooling wall to be minimum under the variable load is screened out. Thermal shock caused by load change to the water cooling wall pipeline is reduced, and safe and stable operation of the deep peak regulation boiler is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of boiler variable load control, and in particular to a burner control system and method for reducing thermal shock of a water-cooled wall under variable load. Background Art

[0002] With the large-scale grid connection of clean energy such as wind power and photovoltaic power, the main task of thermal power plants has gradually shifted to deep peak and frequency regulation to make up for the instability of new energy. However, the frequent changes in boiler load seriously affect the operating safety and service life of equipment and key components. As the main heating surface, the water-cooled wall is most affected by the change of boiler load. When the boiler operating conditions change, it is more likely to cause a pipe burst accident due to thermal shock, which will cause the boiler to shut down, resulting in huge economic losses and safety hazards. Summary of the invention

[0003] In view of this, an object of the present invention is to provide a burner control system and method for reducing thermal shock of a water-cooled wall under variable load, so as to solve or at least partially solve the above-mentioned problems existing in the prior art.

[0004] To achieve the above object, the first aspect of the present invention provides a burner control method for reducing thermal shock of a water-cooled wall under variable load, the method comprising the following steps: S101, obtaining boiler parameters, historical operation data of the boiler, and boiler variables; S102, constructing a furnace combustion model based on boiler parameters, and verifying the furnace combustion model based on historical operation data; S103, dividing the boiler variables into different states based on historical operation data, and using the orthogonal test method to arrange the different states of the boiler variables into different operation combinations, and obtaining furnace water-cooled wall temperature distribution data corresponding to different operation combinations through the furnace combustion model; S104, inputting different operation combinations of boiler variables and their corresponding furnace water-cooled wall temperature distribution data into a machine learning model for training; S105. Predicting furnace water wall temperature distribution data corresponding to different boiler variable operation combinations under load at a given moment in the future based on the trained machine learning model; S106, comparing the predicted furnace water wall temperature distribution data with the upper temperature limit of the corresponding area, screening out the operating combinations of boiler variables with over-temperature data, and discarding them; S107, comparing the furnace water-cooled wall temperature distribution data corresponding to the remaining operation combinations of the screened boiler variables with the furnace water-cooled wall temperature distribution data at the previous historical moment, and screening out the burner combination that minimizes the thermal shock to the water-cooled wall under variable load.

[0005] Furthermore, the boiler variables include boiler load, coal powder concentration, burner start-stop combination, burner angle, and burner powder feed rate.

[0006] Furthermore, the machine learning model includes an input layer, a hidden layer, and an output layer, and the hidden layer is provided with three gating mechanisms including a forget gate, an input gate, and an output gate.

[0007] Furthermore, the training process of the hidden layer includes: S41, delete some information through the forget gate, filter out valid data for storage, as shown below:

[0008] in, For the Gate of Forgetfulness, , for t The input variables of the moment model, Bias for the forget gate; S42, determining the input information ratio through the input gate, and screening the input information, as shown below:

[0009]

[0010]

[0011] in, is the input information ratio, is the input gate weight, i is the input gate, is the input gate bias, for t The candidate state of the memory unit at the moment, is the activation function, is the weight of the alternative state, c For the alternative state, is the alternative state bias, for t The state of the memory unit at the moment, for The state of the memory unit at the moment; S43, determine the output ratio through the output gate, and filter the output information to obtain the ideal output result, which is expressed as follows:

[0012]

[0013] in, is the output ratio, is the model output, Output gate weight, Bias for the output gate.

[0014] Further, in step S105, the specific steps of predicting the furnace water wall temperature distribution data corresponding to different boiler variable operation combinations under load at a given time in the future are as follows: S51, reorganizing different states of boiler load, pulverized coal concentration and burner start / stop combination, burner angle, and burner pulverized coal amount at a given future time into a new operation combination; S52, inputting the new operation combination into the trained machine learning model; S53. Based on step S52, obtain the furnace water-cooled wall temperature distribution data corresponding to the new operation combination.

[0015] Furthermore, the specific steps of screening out the burner combination scheme that minimizes the thermal shock to the furnace water-cooled wall under variable load are as follows: S71, obtaining temperature distribution data of the furnace water-cooled wall at a previous historical moment; S72, comparing the furnace water wall temperature distribution data corresponding to the boiler variable operation combinations remaining after screening in step S106 with the furnace water wall temperature distribution data at the previous historical moment; S73, based on the comparison result of step S72, obtaining the temperature difference of the corresponding grid points of the furnace water-cooled wall at the previous and next moments; S74, calculating the average error of the temperature distribution data of the entire furnace water-cooled wall based on the temperature difference of the corresponding grid points of the furnace water-cooled wall at the previous and next moments, and selecting the furnace water-cooled wall temperature distribution data with the smallest average error as the optimal adjustment result; S75. The boiler variable operation combination corresponding to the furnace water-cooled wall temperature distribution data with the smallest average error is used as the burner combination that minimizes the thermal shock on the water-cooled wall under variable load.

[0016] A second aspect of the present invention provides a burner control system for reducing thermal shock of a water-cooled wall under variable load, comprising: Data acquisition module: used to obtain boiler parameters, boiler historical operation data and boiler variables; Modeling and simulation module: used to build a furnace combustion model according to boiler parameters, verify the furnace combustion model according to historical operation data, and divide boiler variables into different states based on historical operation data. Orthogonal test method is used to arrange different states of boiler variables to obtain different operation combinations, and furnace water-cooled wall temperature distribution data corresponding to different operation combinations are obtained through the furnace combustion model; Machine learning module: used to input different operating combinations of boiler variables and their corresponding furnace water-cooled wall temperature distribution data into the machine learning model for training, predict the furnace water-cooled wall temperature distribution data corresponding to different boiler variable operating combinations under load at a given moment in the future based on the trained machine learning model, compare the predicted furnace water-cooled wall temperature distribution data with the upper temperature limit of its corresponding area, screen out the operating combinations of boiler variables with over-temperature data, and discard their operating combinations, finally compare the furnace water-cooled wall temperature distribution data corresponding to the operating combinations of the remaining boiler variables after screening with the temperature distribution data of the furnace water-cooled wall at the previous historical moment, and screen out the burner combination that minimizes the thermal shock of the water-cooled wall under variable load.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention proposes a burner control system and method for reducing thermal shock of a water-cooled wall under variable load. The system uses a machine learning model to find the relationship between a boiler variable operation combination and the temperature distribution of a furnace water-cooled wall, predicts the temperature distribution of the furnace water-cooled wall, and screens the predicted results to optimize the combustion organization of the boiler under variable load, actively prevent overheating of the furnace water-cooled wall, and compares the furnace water-cooled wall temperature distribution data corresponding to the remaining boiler variable operation combinations with the temperature distribution data of the furnace water-cooled wall at the previous historical moment, respectively, to screen out a burner combination that minimizes the thermal shock of the water-cooled wall under variable load, thereby achieving smooth changes in the boiler load, reducing the thermal shock of the water-cooled wall pipeline caused by load changes, and extending the fatigue life of the heating surface, which is conducive to achieving safe and stable operation of deep peak-shaving boilers. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only preferred embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 A schematic diagram of a flow chart of a burner for reducing thermal shock of a water-cooled wall under variable load provided by an embodiment of the present invention; Figure 2 An example diagram of optimization of the burner start-stop combination provided by an embodiment of the present invention; Figure 3 An example diagram of an orthogonal test table provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The enumerated embodiments are only used to explain the present invention and are not used to limit the scope of the present invention.

[0021] This embodiment provides a burner control method for reducing thermal shock of a water-cooled wall under variable load, characterized in that the method comprises the following steps: like Figure 1 As shown, a flow diagram of a burner for reducing thermal shock of a water-cooled wall under variable load is provided, including: S101, obtaining boiler parameters, historical operation data of the boiler and boiler variables, specifically including: The boiler variables include: boiler load, coal powder concentration, burner start-stop combination, burner angle, and burner powder feeding amount.

[0022] S102: construct a furnace combustion model based on boiler parameters, and verify the furnace combustion model based on historical operation data.

[0023] S103, dividing the boiler variables into different states based on historical operation data, and using the orthogonal test method to arrange the different states of the boiler variables into different operation combinations, and obtaining the furnace water-cooled wall temperature distribution data corresponding to the different operation combinations through the furnace combustion model, specifically including: The working states of the burners do not affect each other, and different start-stop combinations can be performed. The start-stop combinations of the burners can be changed according to the characteristics of different boilers. Taking the opposed pulverized coal boiler with three burners in front and three in the back as an example, six burners are arranged in each row. During normal operation, five rows of burners are enabled for flame hedging. Before and after variable load cutting, the center of the furnace flame is easily offset. Figure 2 This is the optimized schematic diagram of the burner start-stop combination. Figure 2 The loads corresponding to (a), (b), (c), and (d) in the figure are decreasing continuously, but the flame deviation can be minimized by optimizing the burner start-stop combination and other parameters. The burner working state is encoded as 0,1, where 0 means the burner stops working and 1 means the burner starts running, forming a matrix input into the combustion model. Figure 2 The optimized schematic diagram of the burner start-stop combination shown is only an example, and other combinations can be made according to the site conditions; Taking the boiler variable divided into four levels as an example, an orthogonal test table is designed, such as Figure 3 As shown in the figure, the boiler load, pulverized coal concentration, burner start-stop combination, burner angle, and burner pulverized coal amount are marked as factors A, B, C, D, and E in turn. Figure 2 The combinations (a) to (d) shown are marked as level 1, 2, 3, and 4 of factor C, respectively, and 16 operating combinations are obtained. By inputting them into the furnace combustion model, 16 furnace water-cooled wall temperature distribution data under the corresponding operating combinations can be obtained. Figure 3 The orthogonal test table shown is only an example. In actual operation, different numbers of levels can be set according to the variation range of historical data to obtain richer and more comprehensive operation combinations.

[0024] S104, inputting different operation combinations of boiler variables and their corresponding furnace water wall temperature distribution data into a machine learning model for training, specifically including: With boiler load, pulverized coal concentration, burner start-stop combination, burner angle, and burner pulverized coal quantity as input, and furnace water wall temperature distribution data as output, the machine learning model can grasp the mapping relationship between input and output. The machine learning model may select a long short-term memory network model (Long Short-Time Momery, LSTM). The LSTM model is mainly composed of an input layer, a hidden layer, and an output layer. To solve the gradient vanishing problem, the hidden layer is provided with three gating mechanisms, namely a forget gate, an input gate, and an output gate, which is suitable for solving the long-term dependency problem. The training process of the hidden layer includes: S41, deleting part of the information through the forget gate, screening out different operation combinations of boiler variables and their corresponding valid data of furnace water-cooled wall temperature distribution data for storage, as shown below:

[0025] in, For the Gate of Forgetfulness, , for t The input variables of the moment model, Bias for the forget gate; S42, determining the input information ratio through the input gate, and screening the input information, as shown below:

[0026]

[0027]

[0028] in, is the input information ratio, is the input gate weight, i is the input gate, is the input gate bias, for t The candidate state of the memory unit at that moment, is the activation function, is the weight of the alternative state, c For the alternative state, is the alternative state bias, for t The state of the memory unit at the moment, for The state of the memory unit at the moment; S43, determine the output ratio through the output gate, and filter the output information to obtain the ideal output result, which is expressed as follows:

[0029]

[0030] in, is the output ratio, is the model output, Output gate weight, Bias for the output gate.

[0031] S105, predicting the furnace water-cooled wall temperature distribution data corresponding to different boiler variable operation combinations under load at a given moment in the future based on the trained machine learning model, specifically comprising the following steps: S51, reorganizing different states of boiler load, pulverized coal concentration and burner start / stop combination, burner angle, and burner pulverized coal amount at a given future time into a new operation combination; S52, inputting the new operation combination into the trained machine learning model; S53. Based on step S52, obtain the furnace water-cooled wall temperature distribution data corresponding to the new operation combination.

[0032] S106. Compare the predicted furnace water-cooled wall temperature distribution data with the upper temperature limit of its corresponding area, screen out boiler variable operation combinations with over-temperature data, and discard them.

[0033] S107, comparing the furnace water wall temperature distribution data corresponding to the remaining boiler variable operation combinations with the furnace water wall temperature distribution data at the previous historical moment, and selecting a burner combination that minimizes the thermal shock to the furnace water wall under variable load, specifically comprising the following steps: S71, obtaining temperature distribution data of the furnace water-cooled wall at a previous historical moment; S72, comparing the furnace water wall temperature distribution data corresponding to the boiler variable operation combinations remaining after screening in step S106 with the furnace water wall temperature distribution data at the previous historical moment; S73, based on the comparison result of step S72, obtaining the temperature difference of the corresponding grid points of the furnace water-cooled wall at the previous and next moments; S74, calculating the average error of the temperature distribution data of the entire furnace water-cooled wall based on the temperature difference of the corresponding grid points of the furnace water-cooled wall at the previous and next moments, and selecting the furnace water-cooled wall temperature distribution data with the smallest average error as the optimal adjustment result; S75. The boiler variable operation combination corresponding to the furnace water-cooled wall temperature distribution data with the smallest average error is used as the burner combination that minimizes the thermal shock on the water-cooled wall under variable load.

[0034] Another embodiment of the present invention provides a burner control system for reducing thermal shock of a water-cooled wall under variable load, comprising: Data acquisition module: used to obtain boiler parameters, boiler historical operation data and boiler variables; Modeling and simulation module: used to build a furnace combustion model according to boiler parameters, verify the furnace combustion model according to historical operation data, and divide boiler variables into different states based on historical operation data. Orthogonal test method is used to arrange different states of boiler variables to obtain different operation combinations, and furnace water-cooled wall temperature distribution data corresponding to different operation combinations are obtained through the furnace combustion model; Machine learning module: used to input different operating combinations of boiler variables and their corresponding furnace water-cooled wall temperature distribution data into the machine learning model for training, predict the furnace water-cooled wall temperature distribution data corresponding to different boiler variable operating combinations under the load at a given moment in the future based on the trained machine learning model, compare the predicted furnace water-cooled wall temperature distribution data with the upper temperature limit of the corresponding area, screen out the operating combinations of boiler variables with over-temperature data, and discard the operating combinations, and finally compare the furnace water-cooled wall temperature distribution data corresponding to the remaining operating combinations of boiler variables after screening with the temperature distribution data of the furnace water-cooled wall at the previous historical moment, and screen out the burner combination that minimizes the thermal shock of the water-cooled wall under variable load; Data storage module: used to store the temperature distribution data of the furnace water-cooled wall at the previous historical moment. The machine learning module can access the data stored in this module at any time to compare with the prediction results. The optimal result selected by the machine learning module is also stored in the data storage module for comparison with the next machine learning model prediction.

[0035] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A burner control method for reducing thermal shock of water-cooled wall under variable load, characterized in that: The method comprises the following steps: S101, obtaining boiler parameters, historical operation data of the boiler, and boiler variables; S102, constructing a furnace combustion model based on boiler parameters, and verifying the furnace combustion model based on historical operation data; S103, dividing the boiler variables into different states based on historical operation data, and using the orthogonal test method to arrange the different states of the boiler variables into different operation combinations, and obtaining furnace water-cooled wall temperature distribution data corresponding to different operation combinations through the furnace combustion model; S104, inputting different operation combinations of boiler variables and their corresponding furnace water-cooled wall temperature distribution data into a machine learning model for training; S105. Predicting furnace water wall temperature distribution data corresponding to different boiler variable operation combinations under load at a given moment in the future based on the trained machine learning model; S106, comparing the predicted furnace water wall temperature distribution data with the upper temperature limit of the corresponding area, screening out the operating combinations of boiler variables with over-temperature data, and discarding them; S107, comparing the furnace water-cooled wall temperature distribution data corresponding to the remaining operation combinations of the screened boiler variables with the furnace water-cooled wall temperature distribution data at the previous historical moment, and screening out the burner combination that minimizes the thermal shock to the water-cooled wall under variable load.

2. A burner control method for reducing thermal shock of water-cooled wall under variable load according to claim 1, characterized in that: The boiler variables include boiler load, coal powder concentration, burner start-stop combination, burner angle, and burner powder feeding amount.

3. The burner control method for reducing thermal shock of water-cooled wall under variable load according to claim 1 is characterized in that: The machine learning model includes an input layer, a hidden layer, and an output layer. The hidden layer is provided with three gating mechanisms including a forget gate, an input gate, and an output gate.

4. A burner control method for reducing thermal shock of water-cooled wall under variable load according to claim 3, characterized in that: The training process of the hidden layer includes: S41, delete some information through the forget gate, filter out valid data for storage, as shown below: in, For the Gate of Forgetfulness, , for t The input variables of the moment model, Bias for the forget gate; S42, determining the input information ratio through the input gate, and screening the input information, as shown below: in, is the input information ratio, is the input gate weight, i is the input gate, is the input gate bias, for t The candidate state of the memory unit at the moment, is the activation function, is the weight of the alternative state, c For the alternative state, is the alternative state bias, for t The state of the memory unit at the moment, for The state of the memory unit at the moment; S43, determine the output ratio through the output gate, and filter the output information to obtain the ideal output result, which is expressed as follows: in, is the output ratio, is the model output, Output gate weight, Bias for the output gate.

5. The burner control method for reducing thermal shock of water-cooled wall under variable load according to claim 1, characterized in that: In step S105, the specific steps of predicting the furnace water wall temperature distribution data corresponding to different boiler variable operation combinations under load at a given time in the future are as follows: S51, reorganizing different states of boiler load, pulverized coal concentration and burner start / stop combination, burner angle, and burner pulverized coal amount at a given future time into a new operation combination; S52, inputting the new operation combination into the trained machine learning model; S53. Based on step S52, obtain the furnace water-cooled wall temperature distribution data corresponding to the new operation combination.

6. A burner control method for reducing thermal shock of water-cooled wall under variable load according to claim 1, characterized in that: The specific steps of screening out the burner combination scheme that minimizes the thermal shock to the furnace water-cooled wall under variable load are as follows: S71, obtaining temperature distribution data of the furnace water-cooled wall at a previous historical moment; S72, comparing the furnace water wall temperature distribution data corresponding to the boiler variable operation combinations remaining after screening in step S106 with the furnace water wall temperature distribution data at the previous historical moment; S73, based on the comparison result of step S72, obtaining the temperature difference of the corresponding grid points of the furnace water-cooled wall at the previous and next moments; S74, calculating the average error of the temperature distribution data of the entire furnace water-cooled wall based on the temperature difference of the corresponding grid points of the furnace water-cooled wall at the previous and next moments, and selecting the furnace water-cooled wall temperature distribution data with the smallest average error as the optimal adjustment result; S75. The boiler variable operation combination corresponding to the furnace water-cooled wall temperature distribution data with the smallest average error is used as the burner combination that minimizes the thermal shock on the water-cooled wall under variable load.

7. A burner control system for reducing thermal shock of water-cooled wall under variable load, characterized in that: include: Data acquisition module: used to obtain boiler parameters, boiler historical operation data and boiler variables; Modeling and simulation module: used to build a furnace combustion model according to boiler parameters, verify the furnace combustion model according to historical operation data, and divide boiler variables into different states based on historical operation data. Orthogonal test method is used to arrange different states of boiler variables to obtain different operation combinations, and furnace water-cooled wall temperature distribution data corresponding to different operation combinations are obtained through the furnace combustion model; Machine learning module: used to input different operating combinations of boiler variables and their corresponding furnace water-cooled wall temperature distribution data into the machine learning model for training, predict the furnace water-cooled wall temperature distribution data corresponding to different boiler variable operating combinations under load at a given moment in the future based on the trained machine learning model, compare the predicted furnace water-cooled wall temperature distribution data with the upper temperature limit of its corresponding area, screen out the operating combinations of boiler variables with over-temperature data, and discard their operating combinations, finally compare the furnace water-cooled wall temperature distribution data corresponding to the operating combinations of the remaining boiler variables after screening with the temperature distribution data of the furnace water-cooled wall at the previous historical moment, and screen out the burner combination that minimizes the thermal shock of the water-cooled wall under variable load.

Images