Hearth air distribution parameter optimization method and system for coal-fired power plant

Abstract

The invention relates to the technical field of heat energy engineering, and discloses a hearth air distribution parameter optimization method and system for a coal-fired power plant, and the method comprises the steps: obtaining power plant operation data of the coal-fired power plant under different working conditions, carrying out the preprocessing, obtaining a training set, inputting a preset initial hearth parameter optimization model through the training set, the method comprises the following steps: obtaining a plurality of optimization target values under each working condition, calculating a target loss value corresponding to each optimization target value, calculating a reference loss interval corresponding to each optimization target value by using power plant operation data and preset reference interval data, and when the target loss values under each working condition are in the associated reference loss interval, determining that the target loss values are in the reference loss interval. If yes, a target hearth parameter optimization model is generated, parameter optimization is conducted on the target operation data of the coal-fired power plant to be optimized through the target hearth parameter optimization model, the optimal hearth air distribution parameters are output, and the technical problem that existing parameter single target optimization cannot meet the multi-aspect requirements of the coal-fired power plant is solved.

Description

Technical Field

[0001] The present invention relates to the field of thermal energy engineering technology, and in particular to a method and system for optimizing furnace air distribution parameters of a coal-fired power plant. Background Art

[0002] In coal-fired power plants, furnace air distribution parameters directly affect combustion efficiency, pollutant emissions and equipment safety. Traditional furnace air distribution is often based on experience and has many problems. On the one hand, unreasonable air distribution will lead to incomplete combustion of fuel, increase coal consumption and reduce the economic efficiency of the power plant. For example, an improper ratio of primary air and secondary air may prevent coal powder from being fully mixed with oxygen. On the other hand, air distribution has a significant impact on the generation of pollutants such as nitrogen oxides (NOx). Improper air distribution will cause a large amount of NOx to be produced, aggravating environmental pollution. In addition, unreasonable air distribution may also cause problems such as furnace slagging and corrosion, affecting the service life and operational reliability of equipment. Therefore, the optimization of furnace air distribution parameters is very important.

[0003] At present, most of the existing parameter optimization methods set the optimization target as "minimizing" a specific indicator. However, the operation of coal-fired power plants is a complex system engineering. Single target setting cannot take into account all aspects of the needs. For example, when reducing NOx emissions, it may increase incomplete combustion losses, affect combustion efficiency and economy, and cannot meet the overall optimization effect of furnace air distribution in combustion efficiency, pollutant emissions, equipment safety and other aspects. Summary of the invention

[0004] The present invention provides a method and system for optimizing furnace air distribution parameters of a coal-fired power plant, which solves the technical problem that the existing single-objective optimization of parameters cannot meet the various requirements of coal-fired power plants.

[0005] A first aspect of the present invention provides a method for optimizing furnace air distribution parameters of a coal-fired power plant, comprising:

[0006] Obtaining and preprocessing the power plant operation data under different operating conditions of the coal-fired power plant to obtain a training set;

[0007] The training set is used to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under various working conditions;

[0008] Calculating the target loss value corresponding to each of the optimization target values;

[0009] Calculating the benchmark loss interval corresponding to each of the optimization target values ​​using the power plant operation data and preset benchmark interval data;

[0010] When the target loss values ​​under each operating condition are all within the associated reference loss range, a target furnace parameter optimization model is generated;

[0011] The target furnace parameter optimization model is used to optimize the target operating data of the coal-fired power plant to be optimized, and the optimal furnace air distribution parameters are output.

[0012] Optionally, it also includes:

[0013] When any of the target loss values ​​under any operating condition is not within the associated reference loss interval, all of the target loss values ​​under each operating condition are used to calculate the total target loss value;

[0014] Back-propagation adjustment of model parameters according to the total target loss value under each working condition;

[0015] Jump to the step of using the training set to input the preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under each operating condition, until the target loss values ​​under each operating condition are all in the associated benchmark loss range, then generate a target furnace parameter optimization model.

[0016] Optionally, the multiple optimization target values ​​under each operating condition are specifically the degassing water opening deviation value, denitrification inlet CO mean value, denitrification inlet CO deviation value, denitrification inlet NOx mean value, denitrification inlet NOx deviation value, exhaust gas temperature deviation value and denitrification inlet oxygen deviation value under each operating condition.

[0017] Optionally, the preset reference interval data includes first preset reference interval data and second preset reference interval data, and the using the power plant operation data and the preset reference interval data to calculate the reference loss interval corresponding to each of the optimization target values ​​includes:

[0018] When the operating condition is a steady-state operating condition, the power plant operating data and the first preset reference interval data are used to determine a first reference loss interval;

[0019] When the operating condition is a variable load operating condition, the power plant operating data and the second preset reference interval data are used to determine the second reference loss interval.

[0020] Optionally, the power plant operation data includes boiler main steam temperature, reheat steam temperature, boiler flue gas temperature, actual denitrification inlet oxygen content, main steam pressure, unit load and status data of each coal mill.

[0021] Optionally, the preprocessing includes data cleaning, normalization and smoothing.

[0022] A second aspect of the present invention provides a furnace air distribution parameter optimization system for a coal-fired power plant, comprising:

[0023] A preprocessing module is used to obtain and preprocess the power plant operation data of the coal-fired power plant under different operating conditions to obtain a training set;

[0024] A training module, used to use the training set to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under various working conditions;

[0025] A loss value calculation module, used to calculate the target loss value corresponding to each of the optimization target values;

[0026] A loss interval calculation module, used to calculate a reference loss interval corresponding to each of the optimization target values ​​using the power plant operation data and preset reference interval data;

[0027] A model generation module, used for generating a target furnace parameter optimization model when the target loss values ​​under various operating conditions are all within the associated reference loss interval;

[0028] The parameter optimization module is used to optimize the target operating data of the coal-fired power plant to be optimized through the target furnace parameter optimization model, and output the optimal furnace air distribution parameters.

[0029] A third aspect of the present invention provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for optimizing furnace air distribution parameters of a coal-fired power plant as described in any one of the above items.

[0030] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed, implements the method for optimizing furnace air distribution parameters of a coal-fired power plant as described in any one of the above items.

[0031] A fifth aspect of the present invention provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions, wherein when the program instructions are executed by a computer, the computer executes the method for optimizing furnace air distribution parameters of a coal-fired power plant as described in any one of the above items.

[0032] It can be seen from the above technical solutions that the present invention has the following advantages:

[0033] In the present invention, the power plant operation data of a coal-fired power plant under different operating conditions is obtained and preprocessed to obtain a training set. The training set is used to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under each operating condition, and the target loss value corresponding to each optimization target value is calculated. The power plant operation data and preset benchmark interval data are used to calculate the benchmark loss interval corresponding to each optimization target value. When the target loss values ​​under each operating condition are all in the associated benchmark loss interval, a target furnace parameter optimization model is generated. The target operation data of the coal-fired power plant to be optimized is optimized through the target furnace parameter optimization model, and the optimal furnace air distribution parameters are output.

[0034] The present invention takes into account the optimization target values ​​of multiple different aspects, avoiding the situation that a single target setting method may cause the model to ignore other important factors during the optimization process, thereby reducing the overall optimization effect. At the same time, the present invention transforms each optimization target from the traditional "minimization" to "optimization within a certain range". Specifically, by setting the benchmark loss intervals of multiple optimization targets, rather than just pursuing the extreme optimal value of a certain indicator, this multi-target range setting method not only increases the amount of data corresponding to each optimization target, but also significantly improves the stability of model training and modeling accuracy. During the optimization process, the model can better capture the relationship between each target and generate an air distribution strategy that is effective for all optimization targets. Furthermore, the present invention can also effectively deal with data fluctuations and uncertainty problems that often occur in actual industrial applications. By finding the optimal solution within the multi-target range, the model can adapt to changes under different working conditions while ensuring overall performance, thereby providing a more robust air distribution solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0036] Figure 1 A flowchart of the steps of a method for optimizing furnace air distribution parameters of a coal-fired power plant provided in the first embodiment of the present invention;

[0037] Figure 2 A flowchart of the steps of a method for optimizing furnace air distribution parameters of a coal-fired power plant provided in the second embodiment of the present invention;

[0038] Figure 3 A structural block diagram of a furnace air distribution parameter optimization system for a coal-fired power plant provided in Embodiment 3 of the present invention;

[0039] Figure 4 This is a structural block diagram of a computer device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0040] The embodiment of the present invention provides a method and system for optimizing furnace air distribution parameters of a coal-fired power plant, which are used to solve the technical problem that the existing single-objective optimization of parameters cannot meet the various requirements of coal-fired power plants.

[0041] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] See also Figure 1 , Figure 1 A flowchart of the steps of a method for optimizing furnace air distribution parameters of a coal-fired power plant provided in Example 1 of the present invention.

[0043] The present invention provides a method for optimizing furnace air distribution parameters of a coal-fired power plant, comprising:

[0044] Step 101: Obtain power plant operation data of a coal-fired power plant under different operating conditions and perform preprocessing to obtain a training set.

[0045] Operating conditions refer to the comprehensive situation of various operating parameters and working conditions when the power plant is running. They are used to reflect the actual status of various equipment, systems and overall operation of the coal-fired power plant at a certain moment or within a certain period of time.

[0046] Power plant operation data refers to the collection of various related data involved in the operation of coal-fired power plants. It is used to reflect the status, performance and operation of each link of coal-fired power plants, including coal input, combustion process, energy conversion to power output and related auxiliary system operation.

[0047] It should be noted that the power plant operation data includes the real-time operation data of the power plant and the historical operation data of the power plant, which are specifically used to meet the model type selected by the subsequent preset initial furnace parameter optimization model.

[0048] Preprocessing refers to a collection of processing operations performed on the acquired power plant operation data to improve data quality and enhance the accuracy, efficiency and model performance of subsequent data analysis.

[0049] In an embodiment of the present invention, power plant operation data of a coal-fired power plant under different operating conditions is obtained, and the power plant operation data is preprocessed to obtain a training set.

[0050] Step 102: Use the training set to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under various working conditions.

[0051] The preset initial furnace parameter optimization model refers to a pre-set model for optimizing furnace-related parameters of a coal-fired power plant, wherein the preset initial furnace parameter optimization model is a long short-term memory network model or a linear regression model.

[0052] The optimization target value refers to the specific value used to measure the performance of the model or the quality of the results.

[0053] It is worth mentioning that the model of the present invention outputs multiple optimization target values ​​of the coal-fired power plant under different operating conditions, and the model simultaneously considers multiple different aspects of measurement criteria.

[0054] In an embodiment of the present invention, a training set is used to input a preset initial furnace parameter optimization model for training to obtain multiple optimization target values ​​under various working conditions.

[0055] Step 103: Calculate the target loss value corresponding to each optimization target value.

[0056] The target loss value refers to the value used to measure the difference between a model's predicted results and the actual results.

[0057] In the embodiment of the present invention, the target loss value corresponding to each optimization target value under each working condition is calculated.

[0058] Step 104: Calculate the benchmark loss interval corresponding to each optimization target value using the power plant operation data and the preset benchmark interval data.

[0059] The preset benchmark interval data refers to the upper and lower limits of the benchmark loss interval used to calculate each optimization target value. It is used to measure an acceptable loss range for each target optimization value. By comparing the loss value of each optimization target with its respective benchmark loss interval, the overall performance of the model under multi-objective optimization can be comprehensively evaluated.

[0060] In the embodiment of the present invention, the power plant operation data and the preset reference interval data are used to calculate the reference loss interval corresponding to each optimization target value under each working condition.

[0061] Step 105: When the target loss values ​​under various operating conditions are all within the associated reference loss range, a target furnace parameter optimization model is generated.

[0062] In the embodiment of the present invention, when each target loss value under each operating condition is within the associated reference loss interval, a target furnace parameter optimization model is generated.

[0063] Step 106: Optimize the target operating data of the coal-fired power plant to be optimized through the target furnace parameter optimization model, and output the optimal furnace air distribution parameters.

[0064] The target operating data refers to the power plant operating data corresponding to the coal-fired power plant to be optimized that requires parameter optimization.

[0065] In the embodiment of the present invention, the target operation data of the coal-fired power plant to be optimized is input into the trained target furnace parameter optimization model for parameter optimization, and the optimal furnace air distribution parameters are output, and the corresponding optimization target is indicated. Subsequent staff adjusts the air distribution mode of the boiler in real time according to the obtained optimal furnace air distribution parameters to ensure dynamic optimization of the combustion process, reduce NOx emissions, and improve the economy and environmental protection of boiler operation. The real-time optimization process ensures that the system can achieve rapid response and efficient optimization, ensuring optimal control of the combustion process.

[0066] In the present invention, the power plant operation data of a coal-fired power plant under different operating conditions is obtained and preprocessed to obtain a training set. The training set is used to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under each operating condition, and the target loss value corresponding to each optimization target value is calculated. The power plant operation data and preset benchmark interval data are used to calculate the benchmark loss interval corresponding to each optimization target value. When the target loss values ​​under each operating condition are all in the associated benchmark loss interval, a target furnace parameter optimization model is generated. The target operation data of the coal-fired power plant to be optimized is optimized through the target furnace parameter optimization model, and the optimal furnace air distribution parameters are output.

[0067] The present invention takes into account the optimization target values ​​of multiple different aspects, avoiding the situation that a single target setting method may cause the model to ignore other important factors during the optimization process, thereby reducing the overall optimization effect. At the same time, the present invention transforms each optimization target from the traditional "minimization" to "optimization within a certain range". Specifically, by setting the benchmark loss intervals of multiple optimization targets, rather than just pursuing the extreme optimal value of a certain indicator, this multi-target range setting method not only increases the amount of data corresponding to each optimization target, but also significantly improves the stability of model training and modeling accuracy. During the optimization process, the model can better capture the relationship between each target and generate an air distribution strategy that is effective for all optimization targets. Furthermore, the present invention can also effectively deal with data fluctuations and uncertainty problems that often occur in actual industrial applications. By finding the optimal solution within the multi-target range, the model can adapt to changes under different working conditions while ensuring overall performance, thereby providing a more robust air distribution solution.

[0068] See also Figure 2 , Figure 2 A flowchart of the steps of a method for optimizing furnace air distribution parameters of a coal-fired power plant provided in Embodiment 2 of the present invention.

[0069] The present invention provides a method for optimizing furnace air distribution parameters of a coal-fired power plant, comprising:

[0070] Step 201: Obtain power plant operation data of a coal-fired power plant under different operating conditions and perform preprocessing to obtain a training set.

[0071] Furthermore, the power plant operation data includes boiler main steam temperature, reheat steam temperature, boiler flue gas temperature, actual denitrification inlet oxygen content, main steam pressure, unit load and each coal mill status data.

[0072] In an embodiment of the present invention, the boiler main steam temperature, reheat steam temperature, boiler flue gas temperature, actual denitrification inlet oxygen content, main steam pressure, unit load and each coal mill status data are selected as input variables, wherein the unit load represents the current operating load of the boiler, and the coal mill operation status parameters (such as wA, wB, wC, wD, wE, wF) respectively represent the operation status of each coal mill (1 represents operation, 0 represents non-operation). These variables are important factors affecting furnace air distribution and NOx generation. Through real-time collection and input of these variables, the model can accurately capture the operation status of the boiler.

[0073] It should be noted that this application sets seven optimization targets to achieve comprehensive optimization of the combustion process. The optimization targets include the desuperheating water opening deviation value, the denitrification inlet CO mean value, the denitrification inlet CO deviation value, the denitrification inlet NOx mean value, the denitrification inlet NOx deviation value, the exhaust gas temperature deviation value and the denitrification inlet oxygen deviation value. Each optimization target corresponds to a set of optimal furnace air distribution data. Through these optimization targets, the model can comprehensively consider multiple factors to ensure the optimal combustion control under different operating conditions.

[0074] It is worth mentioning that this application sets constraints, including boiler efficiency ≥ 93.0, furnace outlet NOx concentration < 350 mg / m 3 , under this constraint, the model is solved using the input data.

[0075] Furthermore, preprocessing includes data cleaning, normalization and smoothing.

[0076] In the embodiment of the present invention, data preprocessing is a crucial link. The original data often contains various noises and outliers, which will directly affect the training effect and accuracy of the subsequent model. Therefore, the present application performs preprocessing steps such as cleaning and normalization on the collected power plant operation data.

[0077] Data cleaning is the first step in data preprocessing. In the training set of furnace air distribution data, you may encounter abnormal readings caused by sensor failure or invalid data caused by data transmission errors. Through data cleaning, you can remove invalid data, duplicate data and outliers in the original data.

[0078] Normalization is another important step in data preprocessing. Since the data features involved in furnace air distribution have different dimensions and value ranges, directly using these raw data for model training may lead to poor model performance. Normalization can scale the data features to the same value range, usually [0, 1] or [-1, 1], thereby eliminating the impact of dimensions and value ranges on model training. In the furnace air distribution data set, this application uses the minimum-maximum normalization method to scale the data features to the range of [0,1] to improve the stability and convergence speed of the model.

[0079] In addition to cleaning and normalization, this application performs other preprocessing on the collected data set, such as data smoothing, etc. By performing preprocessing steps such as cleaning and normalization on the power plant operation data, the quality and reliability of the data set are improved, providing strong support for subsequent model training and analysis.

[0080] Step 202: Use the training set to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under various working conditions.

[0081] Furthermore, the multiple optimization target values ​​under each operating condition are specifically the degassing water opening deviation value, denitrification inlet CO mean value, denitrification inlet CO deviation value, denitrification inlet NOx mean value, denitrification inlet NOx deviation value, exhaust gas temperature deviation value and denitrification inlet oxygen deviation value under each operating condition.

[0082] It is worth mentioning that the preset initial furnace parameter optimization model in this application is a long short-term memory network model or a linear regression model.

[0083] Specifically, the long short-term memory network (LSTM) is a powerful deep learning model that is particularly good at dealing with long-term dependencies in sequence data. By introducing a gating mechanism, the model performs well in dealing with furnace air distribution optimization. The LSTM structure consists of an input layer, a hidden layer (memory unit), and an output layer, which can achieve accurate optimization of the furnace air distribution method.

[0084] When the preset initial furnace parameter optimization model is a long short-term memory network model, the long short-term memory network model specifically includes:

[0085] First, determine the number of neurons in the LSTM layer. The choice of the number of neurons will affect the complexity and performance of the model. This application can test different numbers (such as 32, 64, 128, etc.) and determine the best number based on the performance of the model on the validation set. This application preferably constructs an LSTM layer with 64 neurons.

[0086] 1. Input layer: The input layer receives sequence data, including boiler main steam temperature, reheat steam temperature, boiler flue gas temperature, actual denitrification inlet oxygen content, main steam pressure, unit load, and each coal mill status data. These variables are key factors affecting furnace air distribution and NOx generation. By collecting and inputting these data in real time, the LSTM model can effectively capture the dynamic operating status of the boiler.

[0087] 2. Hidden layer (memory unit): The hidden layer is composed of LSTM units, each of which includes an input gate, a forget gate, an output gate, and a memory unit. These gating mechanisms enable LSTM to extract complex features of sequence data in a multi-layer structure and model long-term dependencies between input variables, thereby improving the model's predictive and generalization capabilities.

[0088] 3. Output layer: The output layer generates the furnace air distribution parameters required for actual control, which meets the optimization requirements. The specific output includes:

[0089] Air volume and damper opening of the front wall burnout air layer: used to adjust the air supply to the front wall area and optimize the combustion effect.

[0090] Air volume and damper opening of the secondary air layer on the front wall: ensure the reasonable supply of secondary air in the front wall area and reduce NOx generation.

[0091] Air volume and air door opening of the rear wall burnout air layer: used to optimize the combustion conditions in the rear wall area and improve combustion efficiency.

[0092] Air volume and damper opening of the rear wall secondary air layer: control the air flow in the rear wall area to further optimize combustion.

[0093] The output layer is constructed according to the desuperheating water opening deviation value, denitrification inlet CO mean value, denitrification inlet CO deviation value, denitrification inlet NOx mean value, denitrification inlet NOx deviation value, exhaust gas temperature deviation value and denitrification inlet oxygen deviation value that need to be output.

[0094] Since there are 7 output values, the number of neurons in the output layer is 7. A linear activation function can be used.

[0095] By calculating and outputting these parameters in real time based on the input data, the LSTM model can accurately drive the dynamic adjustment of the air in the furnace and achieve optimized control of the furnace air distribution.

[0096] In terms of optimization target setting, this application sets seven optimization targets to achieve comprehensive optimization of the combustion process. These optimization targets include the desuperheating water opening deviation value, the denitrification inlet CO mean value, the denitrification inlet CO deviation value, the denitrification inlet NOx mean value, the denitrification inlet NOx deviation value, the exhaust gas temperature deviation value, and the denitrification inlet oxygen deviation value. Each optimization target corresponds to a set of optimal furnace air distribution data. Through these optimization targets, the model can comprehensively consider multiple factors to ensure the optimal combustion control under different working conditions.

[0097] In the model training phase, the input variables are trained using the long short-term memory network (LSTM) model. This model can effectively capture the complex relationship between unit load, coal mill state parameters and NOx emissions. Through the training of historical data, the model generates the optimal air distribution data for the burnout air volume, secondary air volume and corresponding damper opening of each layer of the front wall and rear wall. The model parameters are continuously adjusted and optimized to ensure that the model can generate the optimal air distribution data under each optimization target.

[0098] During the real-time optimization process, the system collects unit load and coal mill operation status parameters in real time and inputs them into the trained model. Based on the real-time input data, the model outputs seven sets of optimal furnace air distribution data and indicates the corresponding optimization goals. By adjusting the air distribution mode of the boiler in real time, the dynamic optimization of the combustion process is ensured, NOx emissions are reduced, and the economy and environmental protection of boiler operation are improved. The real-time optimization process ensures that the system can achieve rapid response and efficient optimization, ensuring the best control of the combustion process.

[0099] When the preset initial furnace parameter optimization model is a linear regression model, the linear regression model specifically includes:

[0100] Firstly, the model expressions of desuperheating water opening deviation, denitrification inlet CO mean, denitrification inlet CO deviation, denitrification inlet NOx mean, denitrification inlet NOx deviation, exhaust gas temperature deviation and denitrification inlet oxygen deviation were constructed.

[0101] Assume the boiler main steam temperature is , the reheat steam temperature is , the unit load is , assuming there are m coal mills, their states are , the desuperheating water opening deviation value is .

[0102] The specific expression of the desuperheating water opening deviation value is:

[0103]

[0104] in, is the intercept term, is the regression coefficient, It is a random error term, which indicates the influence of other factors not included in the model on the deviation of the desuperheating water opening degree.

[0105] Assume the boiler flue gas temperature is , the unit load is , assuming there are m coal mills, their states are , the average value of denitrification inlet CO is .

[0106] The specific expression of the mean value of denitrification inlet CO is:

[0107]

[0108] in, is the intercept term, which means when hour The baseline value, is the regression coefficient, is a random error term, which indicates the impact of other factors not included in the model on the mean value of denitrification inlet CO.

[0109] Assume the boiler flue gas temperature is , the unit load is , assuming there are m coal mills, their states are , the denitrification inlet CO deviation value is .

[0110] The specific expression of the denitrification inlet CO deviation value is:

[0111]

[0112] in, is the intercept term, which means when hour The baseline value, is the regression coefficient, is a random error term, which indicates the impact of other factors not included in the model on the deviation of denitrification inlet CO.

[0113] Assume that the actual denitrification inlet oxygen quantity is , the main steam pressure is , assuming there are m coal mills, their states are , the average value of denitrification inlet NOx is .

[0114] The specific expression of the mean value of NOx at the denitrification inlet is:

[0115]

[0116] in, is the intercept term, which means when hour The baseline value, is the regression coefficient, It is a random error term, which indicates the influence of other factors not included in the model on the mean value of denitrification inlet NOx.

[0117] Assume the boiler flue gas temperature is , the actual denitrification inlet oxygen quantity is , the unit load is , assuming there are m coal mills, their states are , the denitrification inlet NOx deviation value is .

[0118] The specific expression of the denitrification inlet NOx deviation value is:

[0119]

[0120] in, is the intercept term, which means when hour The baseline value, is the regression coefficient, It is a random error term, which indicates the influence of other factors not included in the model on the deviation of denitrification inlet NOx.

[0121] Assume the boiler main steam temperature is , the boiler flue gas temperature is , the unit load is , assuming there are m coal mills, their states are , the exhaust gas temperature deviation is .

[0122] The specific expression of exhaust gas temperature deviation is:

[0123]

[0124] in, is the intercept term, which means when hour The baseline value, is the regression coefficient, It is a random error term, which indicates the influence of other factors not included in the model on the exhaust temperature deviation value.

[0125] Assume that the actual denitrification inlet oxygen quantity is , the unit load is , assuming there are m coal mills, their states are , the deviation value of denitrification inlet oxygen is .

[0126] The specific expression of the denitrification inlet oxygen deviation value is:

[0127]

[0128] in, is the intercept term, which means when hour The baseline value, is the regression coefficient, It is a random error term, which indicates the influence of other factors not included in the model on the deviation value of denitrification inlet oxygen content.

[0129] It should be noted that the regression coefficient is determined by the least squares method, and existing statistical software (such as the lm function in the R language) can also be used to perform linear regression analysis and obtain the regression coefficient.

[0130] In an embodiment of the present invention, a training set is used to input a preset initial furnace parameter optimization model for training to obtain multiple optimization target values ​​under various working conditions.

[0131] Step 203: Calculate the target loss value corresponding to each optimization target value.

[0132] The expression of the target loss value corresponding to each optimization target value is as follows:

[0133] The loss function of the desuperheating water opening deviation value is as follows:

[0134]

[0135] The loss function of the mean value of inlet CO for denitrification is:

[0136]

[0137] The loss function of the denitrification inlet CO deviation value is specifically:

[0138]

[0139] The loss function of the mean value of NOx at the denitrification inlet is:

[0140]

[0141] The loss function of the NOx deviation value of the denitrification inlet is specifically:

[0142]

[0143] The loss function of the exhaust temperature deviation value is specifically:

[0144]

[0145] The loss function of the denitrification inlet oxygen deviation value is as follows:

[0146]

[0147] In the formula, Represents the target loss value of a certain optimization target value, which is used to measure the model prediction value With reference value The error between reflects the degree of deviation from the optimization goal. It represents the benchmark value, which is used to indicate the expected result under given conditions. Represents the model prediction value, which is also the target optimization value.

[0148] For example, in the optimization of cooling water opening deviation, It is the actual opening deviation reference value under ideal conditions. It is the model prediction value, which means the model prediction output value generated by the model based on the current input parameters. The model tries to make near For example, in the optimization of cooling water opening deviation, is the opening deviation value predicted by the model, is the number of samples, which indicates how many data points are used to calculate this loss. In each loss function calculation, It is the total number of samples used to calculate the average error, which is used to ensure that the size of the loss value is proportional to the number of samples, and to avoid the impact of too large or too small sample size on the loss function.

[0149] Step 204: Calculate the benchmark loss interval corresponding to each optimization target value using the power plant operation data and the preset benchmark interval data.

[0150] Further, the preset reference interval data includes first preset reference interval data and second preset reference interval data, the power plant operation data also includes reference cooling water flow, reference denitration inlet CO concentration, reference denitration inlet NOx concentration, reference exhaust temperature, reference denitration inlet oxygen, the reference loss interval includes first reference loss interval and second reference loss interval, and step 204 may include the following steps:

[0151] It is worth mentioning that when the coal-fired power plant is in a steady-state operating condition, the first benchmark loss interval is selected for model training, and when the coal-fired power plant is in a variable load operating condition, the second benchmark loss interval is selected for model training.

[0152] S11. When the operating condition is a steady-state operating condition, the power plant operating data and the first preset reference interval data are used to determine a first reference loss interval.

[0153] The first preset reference interval data includes the first preset denitration water opening deviation upper limit percentage, the first preset denitration water opening deviation lower limit percentage, the first preset denitration inlet CO mean upper limit percentage, the first preset denitration inlet CO mean upper limit percentage, the first preset denitration inlet CO deviation value upper limit percentage, the first preset denitration inlet CO deviation value lower limit percentage, the first preset denitration inlet NOx mean upper limit percentage, the first preset denitration inlet NOx mean lower limit percentage, the first preset denitration inlet NOx deviation value upper limit percentage, the first preset denitration inlet NOx deviation value lower limit percentage, the first preset exhaust temperature deviation value upper limit data, the first preset exhaust temperature deviation value lower limit data, the first preset denitration inlet oxygen deviation value upper limit percentage and the first preset denitration inlet oxygen deviation value lower limit percentage;

[0154] The first reference loss interval includes the first denitrification water opening deviation loss interval, the first denitrification inlet CO mean loss interval, the first denitrification inlet CO deviation loss interval, the first denitrification inlet NOx mean loss interval, the first denitrification inlet NOx deviation loss interval, the first exhaust gas temperature deviation loss interval and the first denitrification inlet oxygen deviation loss interval;

[0155] The reference cooling water flow rate is used to perform calculations with the first preset cooling water opening deviation upper limit percentage and the first preset cooling water opening deviation lower limit percentage respectively to form the upper limit value and the lower limit value of the first cooling water opening deviation value loss range;

[0156] The reference denitrification inlet CO concentration is used to calculate the first preset denitrification inlet CO mean upper limit percentage and the first preset denitrification inlet CO mean upper limit percentage respectively to form the upper limit value and the lower limit value of the first denitrification inlet CO mean loss range;

[0157] The reference denitration inlet CO concentration is used to perform calculations with the first preset denitration inlet CO deviation value upper limit percentage and the first preset denitration inlet CO deviation value lower limit percentage, respectively, to form the upper limit value and the lower limit value of the first denitration inlet CO deviation value loss range;

[0158] The reference denitration inlet NOx concentration is used to calculate the first preset denitration inlet NOx mean upper limit percentage and the first preset denitration inlet NOx mean lower limit percentage respectively, to form the upper limit value and the lower limit value of the first denitration inlet NOx mean loss range;

[0159] The reference denitration inlet NOx concentration is used to calculate the first preset denitration inlet NOx deviation value upper limit percentage and the first preset denitration inlet NOx deviation value lower limit percentage respectively, to form the upper limit value and the lower limit value of the first denitration inlet NOx deviation value loss range;

[0160] The reference exhaust temperature is used to calculate the upper limit data of the first preset exhaust temperature deviation value and the lower limit data of the first preset exhaust temperature deviation value to form the upper limit value and the lower limit value of the first exhaust temperature deviation value loss interval;

[0161] The reference denitrification inlet oxygen amount is used to perform calculations with the first preset denitrification inlet oxygen amount deviation value upper limit percentage and the first preset denitrification inlet oxygen amount deviation value lower limit percentage respectively to form the upper limit value and the lower limit value of the first denitrification inlet oxygen amount deviation value loss range.

[0162] For ease of understanding, taking the first cooling water opening deviation loss range as an example, if the benchmark cooling water flow rate designed for a coal-fired power plant boiler is 20t / h, and in steady-state operation conditions, the first preset cooling water opening deviation upper limit percentage and the first preset cooling water opening deviation lower limit percentage are ±3% and ±5% respectively, then calculate (20t / h multiplied by ±3% - 20t / h multiplied by ±5%), then the first cooling water opening deviation loss range is between ±0.6 - ±1t / h, therefore, the cooling water opening deviation should be controlled between ±0.6 - ±1t / h.

[0163] It is worth mentioning that, taking the first exhaust temperature deviation loss range as an example, if the benchmark exhaust temperature designed for a coal-fired power plant boiler is 130°C, and under steady-state operating conditions, the first preset exhaust temperature deviation upper limit data and the first preset exhaust temperature deviation lower limit data are ±3°C and ±8°C respectively, then they are calculated (130°C + ±3°C - 130°C + ±8°C), then the first exhaust temperature deviation loss range is between 133°C and 138°C or between 122°C and 127°C. Therefore, the exhaust temperature deviation should be controlled between 133°C and 138°C or between 122°C and 127°C.

[0164] In the embodiment of the present invention, when the operating condition is a steady-state operating condition, the details are as follows:

[0165] The first preset desuperheating water opening deviation upper limit percentage and the first preset desuperheating water opening deviation lower limit percentage are preferably ±3%-±5%;

[0166] The first preset denitration inlet CO average upper limit percentage and the first preset denitration inlet CO average upper limit percentage are preferably ±6%-±16%;

[0167] The first preset denitration inlet CO deviation value upper limit percentage and the first preset denitration inlet CO deviation value lower limit percentage are preferably ±10%-±30%;

[0168] The first preset denitration inlet NOx mean upper limit percentage and the first preset denitration inlet NOx mean lower limit percentage are preferably ±0.75%-±2.5%;

[0169] The first preset denitration inlet NOx deviation value upper limit percentage and the first preset denitration inlet NOx deviation value lower limit percentage are preferably ±0.75%-±2.5%;

[0170] The first preset exhaust temperature deviation value upper limit data and the first preset exhaust temperature deviation value lower limit data are preferably ±3°C-±8°C;

[0171] The first preset denitrification inlet oxygen amount deviation value upper limit percentage and the first preset denitrification inlet oxygen amount deviation value lower limit percentage are preferably ±0.3%-±1%.

[0172] S12. When the operating condition is a variable load operating condition, the power plant operating data and the second preset reference interval data are used to determine the second reference loss interval.

[0173] The second preset reference interval data includes the second preset denitration water opening deviation upper limit percentage, the second preset denitration water opening deviation lower limit percentage, the second preset denitration inlet CO mean upper limit percentage, the second preset denitration inlet CO mean upper limit percentage, the second preset denitration inlet CO deviation value upper limit percentage, the second preset denitration inlet CO deviation value lower limit percentage, the second preset denitration inlet NOx mean upper limit percentage, the second preset denitration inlet NOx mean lower limit percentage, the second preset denitration inlet NOx deviation value upper limit percentage, the second preset denitration inlet NOx deviation value lower limit percentage, the second preset exhaust temperature deviation value upper limit data, the second preset exhaust temperature deviation value lower limit data, the second preset denitration inlet oxygen deviation value upper limit percentage and the second preset denitration inlet oxygen deviation value lower limit percentage;

[0174] The second reference loss interval includes the second denitrification water opening deviation loss interval, the second denitrification inlet CO mean loss interval, the second denitrification inlet CO deviation loss interval, the second denitrification inlet NOx mean loss interval, the second denitrification inlet NOx deviation loss interval, the second exhaust gas temperature deviation loss interval and the second denitrification inlet oxygen deviation loss interval;

[0175] The reference cooling water flow rate is used to calculate with the second preset cooling water opening deviation upper limit percentage and the second preset cooling water opening deviation lower limit percentage respectively to form the upper limit value and the lower limit value of the second cooling water opening deviation value loss range;

[0176] The reference denitrification inlet CO concentration is calculated with the second preset denitrification inlet CO mean upper limit percentage and the second preset denitrification inlet CO mean upper limit percentage respectively to form the upper limit value and the lower limit value of the second denitrification inlet CO mean loss range;

[0177] The reference denitration inlet CO concentration is used to calculate the second preset denitration inlet CO deviation value upper limit percentage and the second preset denitration inlet CO deviation value lower limit percentage respectively, to form the upper limit value and the lower limit value of the second denitration inlet CO deviation value loss range;

[0178] The reference denitration inlet NOx concentration is used to calculate the second preset denitration inlet NOx mean upper limit percentage and the second preset denitration inlet NOx mean lower limit percentage respectively, to form the upper limit value and the lower limit value of the second denitration inlet NOx mean loss range;

[0179] The reference denitration inlet NOx concentration is used to calculate the second preset denitration inlet NOx deviation value upper limit percentage and the second preset denitration inlet NOx deviation value lower limit percentage respectively, to form the upper limit value and the lower limit value of the second denitration inlet NOx deviation value loss range;

[0180] The reference exhaust temperature is used to calculate with the second preset exhaust temperature deviation value upper limit data and the second preset exhaust temperature deviation value lower limit data respectively to form the upper limit value and the lower limit value of the second exhaust temperature deviation value loss interval;

[0181] The reference denitrification inlet oxygen amount is calculated with the second preset denitrification inlet oxygen amount deviation value upper limit percentage and the second preset denitrification inlet oxygen amount deviation value lower limit percentage respectively to form the upper limit value and the lower limit value of the second denitrification inlet oxygen amount deviation value loss range.

[0182] In the embodiment of the present invention, when the working condition is a variable load operating condition, the details are as follows:

[0183] The second preset desuperheating water opening deviation upper limit percentage and the second preset desuperheating water opening deviation lower limit percentage are preferably ±5%-±8%;

[0184] The second preset denitration inlet CO average upper limit percentage and the second preset denitration inlet CO average upper limit percentage are preferably ±16%-±30%;

[0185] The second preset denitration inlet CO deviation value upper limit percentage and the second preset denitration inlet CO deviation value lower limit percentage are preferably ±30%-±60%;

[0186] The second preset denitration inlet NOx mean upper limit percentage and the second preset denitration inlet NOx mean lower limit percentage are preferably ±2%-±5%;

[0187] The second preset denitration inlet NOx deviation value upper limit percentage and the second preset denitration inlet NOx deviation value lower limit percentage are preferably ±2%-±5%;

[0188] The second preset exhaust temperature deviation value upper limit data and the second preset exhaust temperature deviation value lower limit data are preferably ±8°C-±15°C;

[0189] The second preset denitrification inlet oxygen amount deviation value upper limit percentage and the second preset denitrification inlet oxygen amount deviation value lower limit percentage are preferably ±3%-±3%.

[0190] Step 205: When the target loss values ​​under each operating condition are all within the associated reference loss range, a target furnace parameter optimization model is generated.

[0191] In an embodiment of the present invention, when the target loss values ​​under various operating conditions are all within the associated reference loss interval, it indicates that the model training is completed and a target furnace parameter optimization model is generated.

[0192] Step 206: Optimize the target operating data of the coal-fired power plant to be optimized through the target furnace parameter optimization model, and output the optimal furnace air distribution parameters.

[0193] In an embodiment of the present invention, the target operation data of the coal-fired power plant to be optimized is optimized by using a target furnace parameter optimization model, and the optimal furnace air distribution parameters are output.

[0194] It is worth mentioning that the attemperating water opening deviation aims to minimize the attemperating water opening deviation. By taking the boiler main steam temperature, reheat steam temperature, unit load and each pulverizer status as input, the model can predict the required attemperating water opening, thereby reducing the deviation between actual and expected and ensuring steam temperature stability.

[0195] The mean CO value at the denitrification inlet is optimized by modeling the boiler flue gas temperature, unit load and pulverizer status, so as to reduce the mean CO value as much as possible, thereby improving combustion efficiency and reducing pollution emissions.

[0196] The denitrification inlet CO deviation minimizes the fluctuation of CO emissions by optimizing the boiler flue gas temperature, unit load and pulverizer operation status to improve the balance and stability of combustion.

[0197] Similarly, the mean value of NOx at the denitrification inlet is optimized by combining the actual denitrification inlet oxygen content, main steam pressure and the status of each coal mill to reduce the NOx generation concentration, which not only helps to reduce pollutant emissions but also reduces the burden on the denitrification system.

[0198] The denitrification inlet NOx deviation is minimized through real-time measurement and prediction of boiler flue gas temperature, actual denitrification inlet oxygen content, unit load and pulverizer status, thereby improving the balance and effect of denitrification.

[0199] The exhaust temperature deviation minimizes the fluctuation of exhaust temperature by optimizing the main steam temperature, flue gas temperature, unit load and pulverizer status to improve the overall thermal efficiency of the boiler and ensure the stability of the combustion process.

[0200] In addition, the deviation of the denitrification inlet oxygen amount is reduced by optimizing the actual denitrification inlet oxygen amount, unit load and pulverizer status, thereby reducing the energy loss caused by the excess air coefficient and achieving more efficient combustion.

[0201] In summary, this application defines seven optimization target values, takes the boiler operating status, unit load and the operating status of each coal mill as input parameters, closely associates them with the optimization objectives, and performs real-time prediction and optimization of the combustion process based on the model.

[0202] Step 207: When any target loss value under any working condition is not in the associated reference loss interval, all target loss values ​​under each working condition are used to calculate the total target loss value.

[0203] In the embodiment of the present invention, when any target loss value under any working condition is not in the associated reference loss interval, all target loss values ​​under each working condition are used to calculate the total target loss value, and the expression of the total target loss value is as follows:

[0204]

[0205] In the formula, represents the total target loss value under the current working conditions, represents the weight of a certain optimization target value under the current working conditions, represents the target loss value of a certain optimization target value under the current working conditions, Represents an optimization target value.

[0206] Step 208: Back-propagate and adjust the model parameters according to the total target loss value under each working condition.

[0207] In an embodiment of the present invention, model parameters under the current working condition, such as loss weight, are dynamically adjusted according to the total target loss value under each working condition.

[0208] Step 209, jump to the step of using the training set to input the preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under various operating conditions, until the target loss values ​​under various operating conditions are all in the associated benchmark loss range, then generate the target furnace parameter optimization model.

[0209] In an embodiment of the present invention, the process jumps to the step of using a training set to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under various operating conditions, until the target loss values ​​under various operating conditions are all in the associated benchmark loss range, and then a target furnace parameter optimization model is generated.

[0210] The following are application examples provided by the present invention:

[0211] In the furnace air distribution optimization experiment, the target furnace parameter optimization model was used to obtain 7 sets of optimal furnace air distribution parameters, and the 7 sets of optimal furnace air distribution parameters were used for experiments, and the combustion efficiency and pollutant emission concentration under various modes were recorded, as shown in Table 1. These data provide an important basis for evaluating the performance of the target furnace parameter optimization model in furnace air distribution optimization.

[0212] Table 1 shows the test results

[0213]

[0214] The experimental results are shown in Table 1. Experiment No. 1 represents the air distribution scheme composed of traditional air distribution parameters, with a combustion efficiency of 85.3%, a CO emission concentration of 79 mg / m³, and a NOx emission concentration of 241 mg / m³. This set of data reflects the combustion state of the furnace and the corresponding pollutant emission level without optimization.

[0215] The target furnace parameter optimization model proposed in this application optimizes the air distribution method and obtains seven different optimization schemes, corresponding to experimental numbers 2 to 8 respectively.

[0216] Under the optimized air distribution scheme 1 of the model in this paper, the combustion efficiency is increased to 88.1%, and the emission concentrations of CO and NOx are reduced to 68mg / m³ and 221mg / m³ respectively. Compared with the traditional air distribution method, this optimization scheme significantly improves the combustion efficiency and reduces the emission of pollutants.

[0217] The solution of optimizing air distribution 2 increased the combustion efficiency to 89.7%, while reducing the emission concentrations of CO and NOx to 71mg / m³ and 214mg / m³, respectively. This result shows that by adjusting the parameters of the LSTM model, this paper can further optimize the air distribution method to obtain better combustion effects and lower pollutant emissions.

[0218] In the scheme of optimizing air distribution 4, this paper achieved more significant results. The combustion efficiency reached 93.5%, the CO emission concentration was reduced to 59mg / m³, and the NOx emission concentration was also reduced to 174mg / m³. This set of data fully demonstrates the potential and advantages of the model proposed in this paper in optimizing furnace air distribution.

[0219] By optimizing the furnace air distribution method through the model of this application, the combustion efficiency can be significantly improved and the emission concentration of pollutants can be reduced. The experimental data fully proves the effectiveness and feasibility of this method, providing strong support for the energy efficiency improvement and environmental protection of industrial furnaces.

[0220] In summary, in order to achieve real-time optimization of the furnace air distribution method, this application proposes a flexible optimization target setting method. Traditional optimization methods usually set the optimization target as "minimizing" certain specific indicators, such as pollutant emissions or fuel consumption generated during the combustion process. However, this single target setting method may cause the model to ignore other important factors during the optimization process, thereby reducing the overall optimization effect.

[0221] To overcome this limitation, this application innovatively transforms each optimization objective from "minimization" to "optimization within a certain range". Specifically, this method makes full use of more historical data for modeling and optimization by setting a reasonable range of multiple objectives, rather than just pursuing the extreme optimal value of a certain indicator. This multi-objective range setting method not only increases the amount of data corresponding to each optimization objective, but also significantly improves the stability of model training and modeling accuracy. During the optimization process, the model can better capture the relationship between various objectives and generate wind distribution strategies that are effective for all optimization objectives.

[0222] In addition, the method of this application can also effectively deal with data fluctuations and uncertainty problems that often occur in actual industrial applications. By finding the optimal solution within a multi-objective range, the model can adapt to changes under different working conditions while ensuring overall performance, thereby providing a more robust air distribution solution. Experimental results show that this flexible optimization target setting method not only improves combustion efficiency, but also performs well in pollutant emission control, demonstrating its potential and advantages in optimizing industrial furnace air distribution.

[0223] In the present invention, the power plant operation data of a coal-fired power plant under different operating conditions is obtained and preprocessed to obtain a training set. The training set is used to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under each operating condition, and the target loss value corresponding to each optimization target value is calculated. The power plant operation data and preset benchmark interval data are used to calculate the benchmark loss interval corresponding to each optimization target value. When the target loss values ​​under each operating condition are all in the associated benchmark loss interval, a target furnace parameter optimization model is generated. The target operation data of the coal-fired power plant to be optimized is optimized through the target furnace parameter optimization model, and the optimal furnace air distribution parameters are output.

[0224] The present invention takes into account the optimization target values ​​of multiple different aspects, avoiding the situation that a single target setting method may cause the model to ignore other important factors during the optimization process, thereby reducing the overall optimization effect. At the same time, the present invention transforms each optimization target from the traditional "minimization" to "optimization within a certain range". Specifically, by setting the benchmark loss intervals of multiple optimization targets, rather than just pursuing the extreme optimal value of a certain indicator, this multi-target range setting method not only increases the amount of data corresponding to each optimization target, but also significantly improves the stability of model training and modeling accuracy. During the optimization process, the model can better capture the relationship between each target and generate an air distribution strategy that is effective for all optimization targets. Furthermore, the present invention can also effectively deal with data fluctuations and uncertainty problems that often occur in actual industrial applications. By finding the optimal solution within the multi-target range, the model can adapt to changes under different working conditions while ensuring overall performance, thereby providing a more robust air distribution solution.

[0225] See also Figure 3 , Figure 3 This is a structural block diagram of a furnace air distribution parameter optimization system for a coal-fired power plant provided in Example 3 of the present invention.

[0226] The present invention provides a furnace air distribution parameter optimization system for a coal-fired power plant, comprising:

[0227] A preprocessing module 301 is used to obtain and preprocess the power plant operation data of the coal-fired power plant under different operating conditions to obtain a training set;

[0228] A training module 302 is used to input a preset initial furnace parameter optimization model using a training set to obtain multiple optimization target values ​​under various working conditions;

[0229] The loss value calculation module 303 is used to calculate the target loss value corresponding to each optimization target value;

[0230] The loss interval calculation module 304 is used to calculate the reference loss interval corresponding to each optimization target value by using the power plant operation data and the preset reference interval data;

[0231] The model generation module 305 is used to generate a target furnace parameter optimization model when the target loss values ​​under each operating condition are all within the associated reference loss interval;

[0232] The parameter optimization module 306 is used to optimize the target operating data of the coal-fired power plant to be optimized through the target furnace parameter optimization model, and output the optimal furnace air distribution parameters.

[0233] Furthermore, it also includes:

[0234] A total target loss value module is used to calculate the total target loss value using all target loss values ​​under each working condition when any target loss value under any working condition is not within the associated reference loss interval;

[0235] A model parameter adjustment module, used to adjust the model parameters by back-propagation according to the total target loss value under each working condition;

[0236] The jump module is used to jump to the step of using the training set to input the preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under various working conditions, until the target loss values ​​under various working conditions are all in the associated benchmark loss range, then the target furnace parameter optimization model is generated.

[0237] Furthermore, the power plant operation data includes boiler main steam temperature, reheat steam temperature, boiler flue gas temperature, actual denitrification inlet oxygen content, main steam pressure, unit load and each coal mill status data.

[0238] Furthermore, the multiple optimization target values ​​under each operating condition are specifically the degassing water opening deviation value, denitrification inlet CO mean value, denitrification inlet CO deviation value, denitrification inlet NOx mean value, denitrification inlet NOx deviation value, exhaust gas temperature deviation value and denitrification inlet oxygen deviation value under each operating condition.

[0239] Further, the preset reference interval data includes first preset reference interval data and second preset reference interval data, and the loss interval calculation module 304 includes:

[0240] A first reference loss interval submodule is used to determine a first reference loss interval by using power plant operation data and a first preset reference interval data when the operating condition is a steady-state operating condition;

[0241] The second reference loss interval submodule is used to determine the second reference loss interval by using the power plant operation data and the second preset reference interval data when the operating condition is a variable load operating condition.

[0242] Furthermore, preprocessing includes data cleaning, normalization and smoothing.

[0243] In the present invention, the power plant operation data of a coal-fired power plant under different operating conditions is obtained and preprocessed to obtain a training set. The training set is used to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under each operating condition, and the target loss value corresponding to each optimization target value is calculated. The power plant operation data and preset benchmark interval data are used to calculate the benchmark loss interval corresponding to each optimization target value. When the target loss values ​​under each operating condition are all in the associated benchmark loss interval, a target furnace parameter optimization model is generated. The target operation data of the coal-fired power plant to be optimized is optimized through the target furnace parameter optimization model, and the optimal furnace air distribution parameters are output.

[0244] The present invention takes into account the optimization target values ​​of multiple different aspects, avoiding the situation that a single target setting method may cause the model to ignore other important factors during the optimization process, thereby reducing the overall optimization effect. At the same time, the present invention transforms each optimization target from the traditional "minimization" to "optimization within a certain range". Specifically, by setting the benchmark loss intervals of multiple optimization targets, rather than just pursuing the extreme optimal value of a certain indicator, this multi-target range setting method not only increases the amount of data corresponding to each optimization target, but also significantly improves the stability of model training and modeling accuracy. During the optimization process, the model can better capture the relationship between each target and generate an air distribution strategy that is effective for all optimization targets. Furthermore, the present invention can also effectively deal with data fluctuations and uncertainty problems that often occur in actual industrial applications. By finding the optimal solution within the multi-target range, the model can adapt to changes under different working conditions while ensuring overall performance, thereby providing a more robust air distribution solution.

[0245] See also Figure 4 , Figure 4 This is a structural block diagram of a computer device provided in Embodiment 4 of the present invention.

[0246] An electronic device according to an embodiment of the present invention comprises: a memory 401 and a processor 402, wherein the memory 402 stores a computer program; when the computer program is executed by the processor 402, the processor 402 executes a method for optimizing furnace air distribution parameters of a coal-fired power plant as described in any of the above embodiments.

[0247] The memory 401 may be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk or a ROM. The memory 401 has a storage space 403 for a program code 413 for executing any method step in the above method. For example, the storage space 403 for the program code may include individual program codes 413 for implementing the various steps in the above method, respectively. These program codes may be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards or floppy disks. The program code may be compressed, for example, in an appropriate form. When these codes are run by a computing and processing device, the computing and processing device performs the various steps in the method described above. These program codes may be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards or floppy disks. The program code may be compressed, for example, in an appropriate form. When these codes are executed by a computing and processing device, they cause the computing and processing device to execute the various steps in the above-described method for optimizing furnace air distribution parameters of a coal-fired power plant.

[0248] Embodiment 5 of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for optimizing furnace air distribution parameters of a coal-fired power plant as in any of the above embodiments is implemented.

[0249] Embodiment 6 of the present invention further provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions, wherein when the program instructions are executed by a computer, the computer executes a method for optimizing furnace air distribution parameters of a coal-fired power plant as described in any of the above embodiments.

[0250] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0251] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0252] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0253] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0254] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0255] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing furnace air distribution parameters of a coal-fired power plant, characterized in that: include: Obtaining and preprocessing the power plant operation data under different operating conditions of the coal-fired power plant to obtain a training set; The training set is used to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under various working conditions; Calculating the target loss value corresponding to each of the optimization target values; Calculating the benchmark loss interval corresponding to each of the optimization target values ​​using the power plant operation data and preset benchmark interval data; When the target loss values ​​under each operating condition are all within the associated reference loss range, a target furnace parameter optimization model is generated; The target furnace parameter optimization model is used to optimize the target operating data of the coal-fired power plant to be optimized, and the optimal furnace air distribution parameters are output.

2. The method for optimizing furnace air distribution parameters of a coal-fired power plant according to claim 1, characterized in that: Also includes: When any of the target loss values ​​under any operating condition is not within the associated reference loss interval, all of the target loss values ​​under each operating condition are used to calculate the total target loss value; Back-propagation adjustment of model parameters according to the total target loss value under each working condition; Jump to the step of using the training set to input the preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under each operating condition, until the target loss values ​​under each operating condition are all in the associated benchmark loss range, then generate a target furnace parameter optimization model.

3. The method for optimizing furnace air distribution parameters of a coal-fired power plant according to claim 1, characterized in that: The multiple optimization target values ​​under each operating condition are specifically the degassing water opening deviation value, denitrification inlet CO mean value, denitrification inlet CO deviation value, denitrification inlet NOx mean value, denitrification inlet NOx deviation value, exhaust gas temperature deviation value and denitrification inlet oxygen deviation value under each operating condition.

4. The method for optimizing furnace air distribution parameters of a coal-fired power plant according to claim 3, characterized in that: The preset reference interval data includes first preset reference interval data and second preset reference interval data, and the calculation of the reference loss interval corresponding to each of the optimization target values ​​using the power plant operation data and the preset reference interval data includes: When the operating condition is a steady-state operating condition, the power plant operating data and the first preset reference interval data are used to determine a first reference loss interval; When the operating condition is a variable load operating condition, the power plant operating data and the second preset reference interval data are used to determine the second reference loss interval.

5. The method for optimizing furnace air distribution parameters of a coal-fired power plant according to any one of claims 1 to 4, characterized in that: The power plant operation data include boiler main steam temperature, reheat steam temperature, boiler flue gas temperature, actual denitrification inlet oxygen content, main steam pressure, unit load and status data of each coal mill.

6. The method for optimizing furnace air distribution parameters of a coal-fired power plant according to any one of claims 1 to 4, characterized in that: The preprocessing includes data cleaning, normalization and smoothing.

7. A furnace air distribution parameter optimization system for a coal-fired power plant, characterized in that: include: A preprocessing module is used to obtain and preprocess the power plant operation data of the coal-fired power plant under different operating conditions to obtain a training set; A training module, used to use the training set to input a preset initial furnace parameter optimization model to obtain multiple optimization target values ​​under various working conditions; A loss value calculation module, used to calculate the target loss value corresponding to each of the optimization target values; A loss interval calculation module, used to calculate a reference loss interval corresponding to each of the optimization target values ​​using the power plant operation data and preset reference interval data; A model generation module, used for generating a target furnace parameter optimization model when the target loss values ​​under various operating conditions are all within the associated reference loss interval; The parameter optimization module is used to optimize the target operating data of the coal-fired power plant to be optimized through the target furnace parameter optimization model, and output the optimal furnace air distribution parameters.

8. An electronic device, characterized in that: It comprises a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of the method for optimizing furnace air distribution parameters of a coal-fired power plant as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, the method for optimizing furnace air distribution parameters of a coal-fired power plant as described in any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The computer program product includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions, wherein when the program instructions are executed by a computer, the computer executes the method for optimizing furnace air distribution parameters of a coal-fired power plant as described in any one of claims 1-6.

Images