Ammonia spraying partition intelligent control method based on multi-objective optimization

Abstract

The invention discloses an ammonia spraying partition intelligent control method based on multi-objective optimization, and relates to the technical field of ammonia spraying partition intelligent control, and the method comprises the following steps: collecting ammonia spraying monitoring data including ammonia spraying data, flue gas data and ammonia spraying partition data, and carrying out the preprocessing of the collected data; calculating the amount of substance of nitrogen oxide by using the pretreated ammonia spraying data and flue gas data, and evaluating the relative content of ammonia nitrogen, the theoretical ammonia spraying amount and the ammonia gas escape amount by combining the pretreated ammonia spraying data; according to the method, a data acquisition technology, a weighted average technology, a weighted summation technology, a data calibration technology and a multiple linear regression algorithm are closely combined with a modern information technology, the data acquisition technology, the weighted average technology, the weighted summation technology, the data calibration technology and the multiple linear regression algorithm are combined with the modern information technology, and the method has the advantages of being high in accuracy and high in reliability. And the intelligent degree in the ammonia spraying partition intelligent control process based on multi-objective optimization is obviously enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of ammonia injection zone intelligent control, and in particular to an ammonia injection zone intelligent control method based on multi-objective optimization. Background Art

[0002] Under the current situation of increasingly stringent environmental protection requirements, the efficient and stable operation of the denitrification system is of vital importance. The traditional ammonia injection control method has gradually exposed many disadvantages when dealing with complex and changeable working conditions. On the one hand, the single-target control strategy may lead to a substantial increase in ammonia slip, which not only wastes ammonia resources, but also may cause secondary pollution and have adverse effects on subsequent equipment such as corrosion. On the other hand, due to the lack of precise control of the uneven distribution of nitrogen oxide concentrations in different areas of the flue, the unified ammonia injection method is difficult to meet the actual needs of each zone, resulting in unsatisfactory overall denitrification effect. In actual industrial processes, various ammonia injection monitoring parameters are in dynamic change, which further aggravates the complexity of ammonia injection control. With the rapid development of intelligent technology, realizing intelligent control of ammonia injection zones has become a key way to solve the problem. The ammonia injection zone intelligent control method based on multi-objective optimization can comprehensively consider multiple targets such as denitrification efficiency and ammonia slip, accurately adjust the ammonia injection amount of each zone, effectively improve the overall performance of the denitrification system, and provide a strong guarantee for the green and efficient operation of industrial production.

[0003] Although the existing technology has made great progress in the direction of intelligent control of ammonia injection partitions, there are still some problems that need to be optimized. The existing ammonia injection partition control technology is difficult to combine and analyze ammonia data, flue gas data, and ammonia injection partition data, and then accurately control the ammonia injection amount of the ammonia injection partition, resulting in poor nitrogen removal effect. Summary of the Invention

[0004] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for intelligent control of ammonia injection zones based on multi-objective optimization, comprising the following steps:

[0005] Step 1: Collect ammonia injection monitoring data including ammonia injection data, flue gas data, and ammonia injection partition data, and pre-process the collected data to provide a data basis for the calculation and evaluation processes in subsequent steps;

[0006] Step 2: Calculate the amount of nitrogen oxides using the pre-processed ammonia injection data and flue gas data. Combined with the pre-processed ammonia injection data, evaluate the relative ammonia nitrogen content, theoretical ammonia injection amount, and ammonia escape amount.

[0007] Step 3: Obtain the ammonia nitrogen reaction catalyst activity index by using the weighted summation method based on the pre-processed flue gas data, the ammonia injection partition data, and the relative ammonia nitrogen content;

[0008] Step 4: Based on the obtained ammonia nitrogen reaction catalyst activity index, an ammonia injection supply limitation model is constructed to obtain the maximum ammonia supply;

[0009] Step 5: Combine the ammonia escape rate and the maximum ammonia supply rate to construct an ammonia injection calibration model, and then obtain the ammonia injection calibration coefficient;

[0010] Step 6: Analyze the output results of the ammonia injection calibration model, calibrate the theoretical ammonia injection amount, and intelligently control the ammonia injection partitions. This solves the problem of difficulty in combining and analyzing ammonia data, flue gas data, and ammonia injection partition data to accurately control the ammonia injection amount of the ammonia injection partitions, resulting in poor nitrogen removal effect.

[0011] A further improvement of the technical solution of the present invention is that in step 1, the process of collecting ammonia injection monitoring data includes:

[0012] Deploy different types of data collection equipment and number each ammonia injection zone to collect ammonia injection data, flue gas data, and ammonia injection zone data for each ammonia injection zone, thereby achieving spatial synchronization of the collected data. The data collection equipment includes an ammonia sensor, an ammonia mass flow meter, a pitot tube anemometer, a flue gas flow meter, a Fourier transform infrared spectrometer, a temperature sensor, and a barometer.

[0013] The ammonia injection data includes the concentration and mass of the injected ammonia; the flue gas data includes the flue gas volume, flue gas flow, and the concentration and volume fraction of nitrous oxide, nitric oxide, and nitrogen dioxide in the flue gas; the ammonia injection partition data includes the temperature and air pressure of the ammonia injection partition; it should be explained that nitrogen oxides in the flue gas also include dinitrogen trioxide, dinitrogen tetroxide concentration, and dinitrogen pentoxide, but since dinitrogen trioxide, dinitrogen tetroxide concentration, and dinitrogen pentoxide are relatively unstable in the atmosphere, dinitrogen trioxide is easily decomposed into nitric oxide and nitrogen dioxide, dinitrogen tetroxide is easily decomposed into nitrogen dioxide, and dinitrogen pentoxide is easily decomposed into nitrogen dioxide and oxygen, so only nitrous oxide, nitric oxide, and nitrogen dioxide need to be analyzed, which provides a theoretical basis for reducing data complexity;

[0014] An ammonia sensor is used to collect the concentration of the injected ammonia; an ammonia mass flowmeter is used to measure the mass flow of the injected ammonia and record the ammonia mass flow collection time. The mass flow of ammonia is the mass of ammonia passing through a certain cross section per unit time. The purpose of collecting ammonia mass is achieved by multiplying the mass flow of the injected ammonia by the ammonia mass flow collection time; a pitot tube velocimeter is used to measure the dynamic pressure and static pressure of the flue gas, and the flue gas volume is collected by combining relevant calculations; a flue gas flowmeter is used to measure the flue gas flow; a Fourier transform infrared spectrometer is used to collect the concentrations of nitrous oxide, nitric oxide, and nitrogen dioxide in the flue gas; a Fourier transform infrared spectrometer is used to collect the volume fractions of nitrous oxide, nitric oxide, and nitrogen dioxide in the flue gas; a barometer is used to collect the air pressure of the ammonia injection zone;

[0015] Perform data cleaning and data normalization on the collected ammonia injection monitoring data, assign timestamps to the collected ammonia injection data, flue gas data, and ammonia injection partition data, and adjust the assigned timestamps to achieve synchronization of the collection time of the ammonia injection data, flue gas data, and ammonia injection partition data;

[0016] The preprocessed ammonia injection data, flue gas data and ammonia injection partition data are integrated to generate an ammonia injection monitoring data set, which is divided into a training set and a test set, with the ratio of the training set to the test set being 7:3.

[0017] A further improvement of the technical solution of the present invention is that in step 2, the calculation process of the amount of nitrogen oxides includes:

[0018] Calculate the amount of nitrous oxide, nitric oxide and nitrogen dioxide respectively by the volume fraction of nitrous oxide, nitric oxide and nitrogen dioxide in the flue gas and the flue gas volume;

[0019] The amount of nitrogen oxides is calculated by combining the number of nitrogen atoms in nitrous oxide, nitric oxide, and nitrogen dioxide. The calculation process includes:

[0020]

[0021]

[0022] in, is the amount of substance of nitrogen oxides, n NO and are the amounts of substances that are nitrous oxide, nitric oxide, and nitrogen dioxide, respectively, V NO and are the volume fractions of nitrous oxide, nitric oxide and nitrogen dioxide in the flue gas, V total is the flue gas volume, V mis the molar volume of gas under standard conditions, and V m Approximately equal to 22.4L / mol.

[0023] A further improvement of the technical solution of the present invention is that in step 2, the evaluation process of the relative content of ammonia nitrogen includes:

[0024] The amount of ammonia is calculated using the mass of the injected ammonia and the molar mass of the ammonia, where the molar mass of the ammonia is equal to 17.03 g / mol. Weights are assigned to the amount of nitrogen oxides and the amount of ammonia, respectively, and the assigned weights are summed to obtain the ammonia nitrogen weight sum.

[0025] Combined with the weighted average method, the relative content of ammonia nitrogen is evaluated. The product of the amount of nitrogen oxides and their corresponding weights and the product of the amount of ammonia and their corresponding weights are calculated to obtain the weighted values of nitrogen oxides and ammonia, respectively. The weighted values of nitrogen oxides and ammonia are added and then divided by the sum of the ammonia nitrogen weights to obtain the relative content of ammonia nitrogen. The relative content of ammonia nitrogen is then integrated into the ammonia injection monitoring data set.

[0026] A further improvement of the technical solution of the present invention is that in step 2, the evaluation process of the theoretical ammonia injection amount includes:

[0027] Ammonia injection data and flue gas data were extracted from the ammonia injection monitoring dataset. Using the training set data and a multivariate linear regression algorithm, the injected ammonia concentration, flue gas nitrous oxide concentration, flue gas nitric oxide concentration, and flue gas nitrogen dioxide concentration were used as inputs, and the theoretical ammonia injection amount was used as output. The linear relationship between the injected ammonia concentration, flue gas nitrous oxide concentration, flue gas nitric oxide concentration, flue gas nitrogen dioxide concentration, and the theoretical ammonia injection amount was learned to train a theoretical ammonia injection evaluation model.

[0028] Input the test set data into the theoretical ammonia injection evaluation model, adjust the intercept term and regression coefficient of the theoretical ammonia injection evaluation model, optimize the theoretical ammonia injection evaluation model, obtain the final theoretical ammonia injection evaluation model, and output the corresponding theoretical ammonia injection amount based on the injected ammonia concentration, nitrous oxide concentration in the flue gas, nitric oxide concentration in the flue gas, and nitrogen dioxide concentration in the flue gas;

[0029] Specifically, the expression of the theoretical ammonia injection evaluation model is:

[0030]

[0031] Where L is the theoretical ammonia injection amount, α1, α2, α3 and α4 are the regression coefficients of the injected ammonia concentration, the nitrous oxide concentration in the flue gas, the nitric oxide concentration in the flue gas and the nitrogen dioxide concentration in the flue gas, respectively; c1, c2, c3 and c4 are the regression coefficients of the injected ammonia concentration, the nitrous oxide concentration in the flue gas, the nitric oxide concentration in the flue gas and the nitrogen dioxide concentration in the flue gas, respectively; α0 and are the intercept term and error term of the theoretical ammonia injection evaluation model, respectively.

[0032] A further improvement of the technical solution of the present invention is that in step 2, the evaluation process of the ammonia escape amount includes:

[0033] Based on the impact of flue gas flow on ammonia escape, the flue gas flow is divided into low, medium, and high flue gas flow ranges. Ammonia escape scores are assigned to each divided flue gas flow range. Based on the flue gas flow range in which the current flue gas flow falls, the ammonia escape score corresponding to the flue gas flow is obtained.

[0034] According to the influence of the injected ammonia concentration on the ammonia escape amount, the injected ammonia concentration is divided into a low ammonia concentration range, a medium ammonia concentration range, and a high ammonia concentration range. Ammonia escape scores are assigned to each divided ammonia concentration range. Based on the ammonia concentration range in which the current injected ammonia concentration is located, the ammonia escape score corresponding to the injected ammonia concentration is obtained;

[0035] According to the influence of the relative ammonia nitrogen content on the ammonia escape amount, the relative ammonia nitrogen content is divided into a low ammonia nitrogen relative content range, a medium ammonia nitrogen relative content range and a high ammonia nitrogen relative content range, and an ammonia escape score is assigned to each divided relative ammonia nitrogen content range. According to the relative ammonia nitrogen content range in which the current relative ammonia nitrogen content is located, the ammonia escape score corresponding to the relative ammonia nitrogen content is obtained;

[0036] The ammonia escape amount is evaluated by simulating the power function relationship between the flue gas flow rate, the injected ammonia concentration, and the relative ammonia nitrogen content, thereby obtaining the ammonia escape amount. The ammonia escape amount evaluation process includes:

[0037]

[0038] Where T is the ammonia escape amount, t1, t2 and t3 are the ammonia escape fractions corresponding to the flue gas flow rate, injected ammonia concentration and relative ammonia nitrogen content, respectively, and a1, a2, a3 and k are constants.

[0039] A further improvement of the technical solution of the present invention is that in step 3, the process of obtaining the activity index of the ammonia nitrogen reaction catalyst includes:

[0040] According to the factors affecting the activity of the ammonia nitrogen reaction catalyst, weights are assigned to the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and ammonia nitrogen relative content respectively;

[0041] The weighted summation method is used to obtain the ammonia nitrogen reaction catalyst activity index through the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and relative ammonia nitrogen content and their corresponding weights. The ammonia nitrogen reaction catalyst activity index is integrated into the ammonia injection monitoring data set. The acquisition process includes:

[0042] H=w1T+w2P+w3V total +w4I+w5X

[0043] Among them, H is the activity index of the ammonia nitrogen reaction catalyst, w1, w2, w3, w4 and w5 are the weights of the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and ammonia nitrogen relative content, T, P, V total , I and X are the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and relative content of ammonia nitrogen, respectively.

[0044] A further improvement of the technical solution of the present invention is that in step 4, the process of constructing the ammonia injection supply limitation model and then obtaining the maximum ammonia supply includes:

[0045] Extract the ammonia injection partition data, ammonia nitrogen relative content and ammonia nitrogen reaction catalyst activity index from the ammonia injection monitoring data set;

[0046] Combining training data with a multivariate linear regression algorithm, the ammonia injection partition data, relative ammonia nitrogen content, and ammonia nitrogen reaction catalyst activity index are used as input, and the maximum ammonia supply rate is used as output. The linear relationship between the ammonia injection partition data, relative ammonia nitrogen content, ammonia nitrogen reaction catalyst activity index, and the maximum ammonia supply rate is learned to train the ammonia injection supply limitation model.

[0047] The test set data is input into the ammonia injection supply limitation model, and the intercept term and regression coefficient of the ammonia injection supply limitation model are adjusted to optimize the performance of the ammonia injection supply limitation model. The final ammonia injection supply limitation model is obtained, and the corresponding maximum ammonia supply rate is output based on the ammonia injection partition data, the relative ammonia nitrogen content, and the ammonia nitrogen reaction catalyst activity index. This maximum ammonia supply rate is then integrated into the ammonia injection monitoring data set.

[0048] The expression of the ammonia injection supply limitation model is as follows:

[0049] L max =β0+β1T+β2P+β3X+β4H+∈

[0050] Among them, L maxis the maximum ammonia supply, β1, β2, β3 and β4 are the regression coefficients of the ammonia injection partition temperature, ammonia injection partition pressure, ammonia nitrogen relative content and ammonia nitrogen reaction catalyst activity index, T, P, X and H are the ammonia injection partition temperature, ammonia injection partition pressure, ammonia nitrogen relative content and ammonia nitrogen reaction catalyst activity index, β0 and ∈ are the intercept term and error term of the ammonia injection supply limitation model, respectively.

[0051] A further improvement of the technical solution of the present invention is that in step 5, the process of constructing the ammonia injection calibration model and then obtaining the ammonia injection calibration coefficient includes:

[0052] Ammonia slip and maximum ammonia supply were extracted from the ammonia injection monitoring dataset. The training set data was then combined with a multivariate linear regression algorithm. The ammonia slip and maximum ammonia supply were used as inputs, and the ammonia injection calibration coefficient was used as output. The linear relationship between these factors was learned to train the ammonia injection calibration model.

[0053] Input the test set data into the ammonia injection calibration model, adjust the intercept term and regression coefficient of the ammonia injection calibration model, optimize the performance of the ammonia injection calibration model, deploy the ammonia injection calibration model into the system, and output the corresponding ammonia injection calibration coefficient based on the ammonia escape rate and the maximum ammonia supply rate;

[0054] The expression of the ammonia injection calibration model is:

[0055] R=γ0+γ1T+γ2L max +ε

[0056] Among them, R is the ammonia injection calibration coefficient, T and L max are the ammonia slip and the maximum ammonia supply, respectively; γ1 and γ2 are the regression coefficients of ammonia slip and the maximum ammonia supply, respectively; γ0 and ε are the intercept and error terms of the ammonia injection calibration model.

[0057] A further improvement of the technical solution of the present invention is that in step 6, the process of analyzing the output results of the ammonia injection calibration model, calibrating the theoretical ammonia injection amount, and intelligently controlling the ammonia injection zones includes:

[0058] According to the ammonia injection calibration coefficient output by the ammonia injection calibration model, the ammonia injection calibration coefficient is divided into three ammonia injection calibration intervals, and a calibration scheme is allocated to each ammonia injection calibration interval;

[0059] According to the calibration scheme corresponding to the ammonia injection calibration interval where the actual ammonia injection calibration coefficient is located, the calibrated ammonia injection amount is obtained, the theoretical ammonia injection amount is calibrated, and the ammonia injection amount of each ammonia injection zone is controlled to achieve intelligent control of the ammonia injection zone;

[0060] Among them, the calibration schemes corresponding to different ammonia injection calibration intervals are also different. The specific calibration schemes are as follows:

[0061] Set L0 as the calibrated ammonia injection amount. When the ammonia injection calibration coefficient is between 0 and 0.3, the calibrated ammonia injection amount is calculated according to the formula L0 = LR × L; when the ammonia injection calibration coefficient is between 0.3 and 0.7, the theoretical ammonia injection amount is not calibrated, that is, the theoretical ammonia injection amount is used as the calibrated ammonia injection amount; when the ammonia injection calibration coefficient is greater than 0.7, the calibrated ammonia injection amount is calculated according to the formula L0 = L + R × L.

[0062] The beneficial effects of the present invention are as follows: compared with the traditional multi-objective optimization ammonia injection partition intelligent control method, the data acquisition technology, weighted averaging technology, weighted summation technology, data calibration technology, and multivariate linear regression algorithm in the method of the present invention are closely integrated with modern information technology. By deploying multiple acquisition devices, ammonia injection data, flue gas data, and ammonia injection partition data are accurately captured, and key data such as the relative content of ammonia nitrogen, theoretical ammonia injection amount, ammonia escape amount, ammonia nitrogen reaction catalyst activity index, maximum ammonia supply amount, and ammonia injection calibration coefficient are obtained, thereby achieving accurate analysis and calculation of ammonia in the denitrification process and realizing accurate control of the ammonia injection amount of the ammonia injection partition. This solves the problem that traditional methods are difficult to combine and analyze multiple data to accurately control the ammonia injection amount, resulting in poor nitrogen removal effect. It ensures that the method of the present invention can refine the dynamic monitoring standard of an ammonia injection partition intelligent control method based on multi-objective optimization within a more precise range, so that the monitored data becomes a more accurate indicator under the same conditions. The development and application of this method significantly enhances the intelligence level of the ammonia injection partition intelligent control process based on multi-objective optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0064] Figure 1 This is a flow chart of an intelligent control method for ammonia injection partitioning based on multi-objective optimization of the present invention. DETAILED DESCRIPTION

[0065] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0066] like Figure 1 As shown, the present invention provides an intelligent control method for ammonia injection partitions based on multi-objective optimization, which consists of the following steps:

[0067] Step 1: Collect ammonia injection monitoring data including ammonia injection data, flue gas data, and ammonia injection partition data, and pre-process the collected data to provide a data basis for the calculation and evaluation processes in subsequent steps;

[0068] Step 2: Calculate the amount of nitrogen oxides using the pre-processed ammonia injection data and flue gas data. Combined with the pre-processed ammonia injection data, evaluate the relative ammonia nitrogen content, theoretical ammonia injection amount, and ammonia escape amount.

[0069] Step 3: Obtain the ammonia nitrogen reaction catalyst activity index by using the weighted summation method based on the pre-processed flue gas data, the ammonia injection partition data, and the relative ammonia nitrogen content;

[0070] Step 4: Based on the obtained ammonia nitrogen reaction catalyst activity index, an ammonia injection supply limitation model is constructed to obtain the maximum ammonia supply;

[0071] Step 5: Combine the ammonia escape rate and the maximum ammonia supply rate to construct an ammonia injection calibration model, and then obtain the ammonia injection calibration coefficient;

[0072] Step 6: Analyze the output results of the ammonia injection calibration model, calibrate the theoretical ammonia injection amount, and intelligently control the ammonia injection partitions. This solves the problem of difficulty in combining and analyzing ammonia data, flue gas data, and ammonia injection partition data to accurately control the ammonia injection amount of the ammonia injection partitions, resulting in poor nitrogen removal effect.

[0073] In step 1, the process of collecting ammonia injection monitoring data includes:

[0074] Deploy different types of data collection equipment and number each ammonia injection zone to collect ammonia injection data, flue gas data, and ammonia injection zone data for each ammonia injection zone, achieving spatial synchronization of the collected data. The data collection equipment includes ammonia sensors, ammonia mass flow meters, pitot tube anemometers, flue gas flow meters, Fourier transform infrared spectrometers, temperature sensors, and barometers.

[0075] Specifically, the ammonia injection data includes the concentration and mass of the injected ammonia; the flue gas data includes the flue gas volume, flue gas flow, and the concentration and volume fraction of nitrous oxide, nitric oxide, and nitrogen dioxide in the flue gas; the ammonia injection partition data includes the temperature and air pressure of the ammonia injection partition; among them, it should be explained that the nitrogen oxides in the flue gas also include dinitrogen trioxide, dinitrogen tetroxide concentration, and dinitrogen pentoxide, but since dinitrogen trioxide, dinitrogen tetroxide concentration, and dinitrogen pentoxide are relatively unstable in the atmosphere, dinitrogen trioxide is easily decomposed into nitric oxide and nitrogen dioxide, dinitrogen tetroxide is easily decomposed into nitrogen dioxide, and dinitrogen pentoxide is easily decomposed into nitrogen dioxide and oxygen, so only nitrous oxide, nitric oxide, and nitrogen dioxide need to be analyzed, which provides a theoretical basis for reducing data complexity;

[0076] An ammonia sensor is used to collect the concentration of the injected ammonia; an ammonia mass flowmeter is used to measure the mass flow of the injected ammonia and record the ammonia mass flow collection time. The mass flow of ammonia is the mass of ammonia passing through a certain cross section per unit time. The purpose of collecting ammonia mass is achieved by multiplying the mass flow of the injected ammonia by the ammonia mass flow collection time; a pitot tube velocimeter is used to measure the dynamic pressure and static pressure of the flue gas, and the flue gas volume is collected by combining relevant calculations; a flue gas flowmeter is used to measure the flue gas flow; a Fourier transform infrared spectrometer is used to collect the concentrations of nitrous oxide, nitric oxide, and nitrogen dioxide in the flue gas; a Fourier transform infrared spectrometer is used to collect the volume fractions of nitrous oxide, nitric oxide, and nitrogen dioxide in the flue gas; a barometer is used to collect the air pressure of the ammonia injection zone;

[0077] Perform data cleaning and data normalization on the collected ammonia injection monitoring data, assign timestamps to the collected ammonia injection data, flue gas data, and ammonia injection partition data, and adjust the assigned timestamps to achieve synchronization of the collection time of the ammonia injection data, flue gas data, and ammonia injection partition data;

[0078] The preprocessed ammonia injection data, flue gas data and ammonia injection partition data are integrated to generate an ammonia injection monitoring data set, which is divided into a training set and a test set, with the ratio of the training set to the test set being 7:3.

[0079] In step 2, the calculation process of the amount of nitrogen oxides includes:

[0080] Calculate the amount of nitrous oxide, nitric oxide and nitrogen dioxide respectively by the volume fraction of nitrous oxide, nitric oxide and nitrogen dioxide in the flue gas and the flue gas volume;

[0081] The amount of nitrogen oxides is calculated by combining the number of nitrogen atoms in nitrous oxide, nitric oxide, and nitrogen dioxide. The calculation process includes:

[0082]

[0083] in, is the amount of substance of nitrogen oxides, n NO and are the amounts of substances that are nitrous oxide, nitric oxide, and nitrogen dioxide, respectively, V NO and are the volume fractions of nitrous oxide, nitric oxide and nitrogen dioxide in the flue gas, V total is the flue gas volume, V m is the molar volume of gas under standard conditions, and V m Approximately equal to 22.4L / mol.

[0084] In step 2, the assessment process of the relative content of ammonia nitrogen includes:

[0085] The amount of ammonia is calculated using the mass of the injected ammonia and the molar mass of the ammonia, where the molar mass of the ammonia is equal to 17.03 g / mol. Weights are assigned to the amount of nitrogen oxides and the amount of ammonia, respectively, and the assigned weights are summed to obtain the ammonia nitrogen weight sum.

[0086] Combined with the weighted average method, the relative content of ammonia nitrogen is evaluated. The product of the amount of nitrogen oxides and their corresponding weights and the product of the amount of ammonia and their corresponding weights are calculated to obtain the weighted values of nitrogen oxides and ammonia, respectively. The weighted values of nitrogen oxides and ammonia are added and then divided by the sum of the ammonia nitrogen weights to obtain the relative content of ammonia nitrogen. The relative content of ammonia nitrogen is then integrated into the ammonia injection monitoring data set.

[0087] In step 2, the evaluation process of the theoretical ammonia injection amount includes:

[0088] Ammonia injection data and flue gas data were extracted from the ammonia injection monitoring dataset. Using the training set data and a multivariate linear regression algorithm, the injected ammonia concentration, flue gas nitrous oxide concentration, flue gas nitric oxide concentration, and flue gas nitrogen dioxide concentration were used as inputs, and the theoretical ammonia injection amount was used as output. The linear relationship between the injected ammonia concentration, flue gas nitrous oxide concentration, flue gas nitric oxide concentration, flue gas nitrogen dioxide concentration, and the theoretical ammonia injection amount was learned to train a theoretical ammonia injection evaluation model.

[0089] Input the test set data into the theoretical ammonia injection evaluation model, adjust the intercept term and regression coefficient of the theoretical ammonia injection evaluation model, optimize the theoretical ammonia injection evaluation model, obtain the final theoretical ammonia injection evaluation model, and output the corresponding theoretical ammonia injection amount based on the injected ammonia concentration, nitrous oxide concentration in the flue gas, nitric oxide concentration in the flue gas, and nitrogen dioxide concentration in the flue gas;

[0090] Specifically, the expression of the theoretical ammonia injection evaluation model is:

[0091]

[0092] Where L is the theoretical ammonia injection amount, α1, α2, α3 and α4 are the regression coefficients of the injected ammonia concentration, the nitrous oxide concentration in the flue gas, the nitric oxide concentration in the flue gas and the nitrogen dioxide concentration in the flue gas, respectively; c1, c2, c3 and c4 are the regression coefficients of the injected ammonia concentration, the nitrous oxide concentration in the flue gas, the nitric oxide concentration in the flue gas and the nitrogen dioxide concentration in the flue gas, respectively; α0 and are the intercept term and error term of the theoretical ammonia injection evaluation model, respectively.

[0093] In step 2, the ammonia escape assessment process includes:

[0094] Based on the impact of flue gas flow on ammonia escape, the flue gas flow is divided into low, medium, and high flue gas flow ranges. Ammonia escape scores are assigned to each divided flue gas flow range. Based on the flue gas flow range in which the current flue gas flow falls, the ammonia escape score corresponding to the flue gas flow is obtained.

[0095] According to the influence of the injected ammonia concentration on the ammonia escape amount, the injected ammonia concentration is divided into a low ammonia concentration range, a medium ammonia concentration range, and a high ammonia concentration range. Ammonia escape scores are assigned to each divided ammonia concentration range. Based on the ammonia concentration range in which the current injected ammonia concentration is located, the ammonia escape score corresponding to the injected ammonia concentration is obtained;

[0096] According to the influence of the relative ammonia nitrogen content on the ammonia escape amount, the relative ammonia nitrogen content is divided into a low ammonia nitrogen relative content range, a medium ammonia nitrogen relative content range and a high ammonia nitrogen relative content range, and an ammonia escape score is assigned to each divided relative ammonia nitrogen content range. According to the relative ammonia nitrogen content range in which the current relative ammonia nitrogen content is located, the ammonia escape score corresponding to the relative ammonia nitrogen content is obtained;

[0097] The ammonia escape amount is evaluated by simulating the power function relationship between the flue gas flow rate, the injected ammonia concentration, and the relative ammonia nitrogen content, thereby obtaining the ammonia escape amount. The ammonia escape amount evaluation process includes:

[0098]

[0099] Where T is the ammonia escape amount, t1, t2 and t3 are the ammonia escape fractions corresponding to the flue gas flow rate, injected ammonia concentration and relative ammonia nitrogen content, respectively, and a1, a2, a3 and k are constants.

[0100] In step 3, the process of obtaining the activity index of the ammonia nitrogen reaction catalyst includes:

[0101] According to the factors affecting the activity of the ammonia nitrogen reaction catalyst, weights are assigned to the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and ammonia nitrogen relative content respectively;

[0102] The weighted summation method is used to obtain the ammonia nitrogen reaction catalyst activity index through the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and relative ammonia nitrogen content and their corresponding weights. The ammonia nitrogen reaction catalyst activity index is integrated into the ammonia injection monitoring data set. The acquisition process includes:

[0103] H=w1T+w2P+w3V total +w4I+w5X

[0104] Among them, H is the activity index of the ammonia nitrogen reaction catalyst, w1, w2, w3, w4 and w5 are the weights of the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and ammonia nitrogen relative content, T, P, V total , I and X are the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and relative content of ammonia nitrogen, respectively.

[0105] In step 4, the process of constructing an ammonia injection supply limitation model and obtaining the maximum ammonia supply includes:

[0106] Extract the ammonia injection partition data, ammonia nitrogen relative content and ammonia nitrogen reaction catalyst activity index from the ammonia injection monitoring data set;

[0107] Combining training data with a multivariate linear regression algorithm, the ammonia injection partition data, relative ammonia nitrogen content, and ammonia nitrogen reaction catalyst activity index are used as input, and the maximum ammonia supply rate is used as output. The linear relationship between the ammonia injection partition data, relative ammonia nitrogen content, ammonia nitrogen reaction catalyst activity index, and the maximum ammonia supply rate is learned to train the ammonia injection supply limitation model.

[0108] The test set data is input into the ammonia injection supply limitation model, and the intercept term and regression coefficient of the ammonia injection supply limitation model are adjusted to optimize the performance of the ammonia injection supply limitation model. The final ammonia injection supply limitation model is obtained, and the corresponding maximum ammonia supply rate is output based on the ammonia injection partition data, the relative ammonia nitrogen content, and the ammonia nitrogen reaction catalyst activity index. This maximum ammonia supply rate is then integrated into the ammonia injection monitoring data set.

[0109] The expression of the ammonia injection supply limitation model is as follows:

[0110] L max =β0+β1T+β2P+β3X+β4H+∈

[0111] Among them, L maxis the maximum ammonia supply, β1, β2, β3 and β4 are the regression coefficients of the ammonia injection partition temperature, ammonia injection partition pressure, ammonia nitrogen relative content and ammonia nitrogen reaction catalyst activity index, T, P, X and H are the ammonia injection partition temperature, ammonia injection partition pressure, ammonia nitrogen relative content and ammonia nitrogen reaction catalyst activity index, β0 and ∈ are the intercept term and error term of the ammonia injection supply limitation model, respectively.

[0112] In step 5, the process of constructing an ammonia injection calibration model and obtaining an ammonia injection calibration coefficient includes:

[0113] Ammonia slip and maximum ammonia supply were extracted from the ammonia injection monitoring dataset. The training set data was then combined with a multivariate linear regression algorithm. The ammonia slip and maximum ammonia supply were used as inputs, and the ammonia injection calibration coefficient was used as output. The linear relationship between these factors was learned to train the ammonia injection calibration model.

[0114] Input the test set data into the ammonia injection calibration model, adjust the intercept term and regression coefficient of the ammonia injection calibration model, optimize the performance of the ammonia injection calibration model, deploy the ammonia injection calibration model into the system, and output the corresponding ammonia injection calibration coefficient based on the ammonia escape rate and the maximum ammonia supply rate;

[0115] The expression of the ammonia injection calibration model is:

[0116] R=γ0+γ1T+γ2L max +ε

[0117] Among them, R is the ammonia injection calibration coefficient, T and L max are the ammonia slip and the maximum ammonia supply, respectively; γ1 and γ2 are the regression coefficients of ammonia slip and the maximum ammonia supply, respectively; γ0 and ε are the intercept and error terms of the ammonia injection calibration model.

[0118] In step 6, the output results of the ammonia injection calibration model are analyzed, the theoretical ammonia injection amount is calibrated, and the process of intelligently controlling the ammonia injection zones includes:

[0119] According to the ammonia injection calibration coefficient output by the ammonia injection calibration model, the ammonia injection calibration coefficient is divided into three ammonia injection calibration intervals, and a calibration scheme is allocated to each ammonia injection calibration interval;

[0120] According to the calibration scheme corresponding to the ammonia injection calibration interval where the actual ammonia injection calibration coefficient is located, the calibrated ammonia injection amount is obtained, the theoretical ammonia injection amount is calibrated, and the ammonia injection amount of each ammonia injection zone is controlled to achieve intelligent control of the ammonia injection zone;

[0121] Among them, the calibration schemes corresponding to different ammonia injection calibration intervals are also different. The specific calibration schemes are as follows:

[0122] Set L0 as the calibrated ammonia injection amount. When the ammonia injection calibration coefficient is between 0 and 0.3, the calibrated ammonia injection amount is calculated according to the formula L0 = LR × L; when the ammonia injection calibration coefficient is between 0.3 and 0.7, the theoretical ammonia injection amount is not calibrated, that is, the theoretical ammonia injection amount is used as the calibrated ammonia injection amount; when the ammonia injection calibration coefficient is greater than 0.7, the calibrated ammonia injection amount is calculated according to the formula L0 = L + R × L.

[0123] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A multi-objective optimization-based intelligent control method for ammonia injection zones, characterized by: The following steps are involved: Step 1: Collect ammonia injection monitoring data including ammonia injection data, flue gas data and ammonia injection partition data, and pre-process the collected data; Step 2: Calculate the amount of nitrogen oxides using the pre-processed ammonia injection data and flue gas data. Combined with the pre-processed ammonia injection data, evaluate the relative ammonia nitrogen content, theoretical ammonia injection amount, and ammonia escape amount. Step 3: Obtain the ammonia nitrogen reaction catalyst activity index by using the weighted summation method based on the pre-processed flue gas data, the ammonia injection partition data, and the relative ammonia nitrogen content; Step 4: Based on the obtained ammonia nitrogen reaction catalyst activity index, an ammonia injection supply limitation model is constructed to obtain the maximum ammonia supply; Step 5: Combine the ammonia escape rate and the maximum ammonia supply rate to construct an ammonia injection calibration model, and then obtain the ammonia injection calibration coefficient; Step 6: Analyze the output results of the ammonia injection calibration model, calibrate the theoretical ammonia injection amount, and perform intelligent control on the ammonia injection partitions.

2. The method for intelligent control of ammonia injection zones based on multi-objective optimization according to claim 1, characterized in that: In step 1, the process of collecting ammonia injection monitoring data includes: Deploy different types of data collection equipment and number each ammonia injection zone to collect ammonia injection data, flue gas data, and ammonia injection zone data for each ammonia injection zone, thereby achieving spatial synchronization of the collected data. The data collection equipment includes an ammonia sensor, an ammonia mass flow meter, a pitot tube anemometer, a flue gas flow meter, a Fourier transform infrared spectrometer, a temperature sensor, and a barometer. The ammonia injection data includes the concentration and mass of the injected ammonia; the flue gas data includes the flue gas volume, flue gas flow, the concentration and volume fraction of nitrous oxide, nitric oxide and nitrogen dioxide in the flue gas; the ammonia injection partition data includes the temperature and pressure of the ammonia injection partition; Perform data cleaning and data normalization on the collected ammonia injection monitoring data, assign timestamps to the collected ammonia injection data, flue gas data, and ammonia injection partition data, and adjust the assigned timestamps to achieve synchronization of the collection time of the ammonia injection data, flue gas data, and ammonia injection partition data; The preprocessed ammonia injection data, flue gas data and ammonia injection partition data are integrated to generate an ammonia injection monitoring dataset, which is then divided into a training set and a test set.

3. The method for intelligent control of ammonia injection zones based on multi-objective optimization according to claim 2, characterized in that: In step 2, the calculation process of the amount of nitrogen oxides includes: Calculate the amount of nitrous oxide, nitric oxide and nitrogen dioxide respectively by the volume fraction of nitrous oxide, nitric oxide and nitrogen dioxide in the flue gas and the flue gas volume; Combining the number of nitrogen atoms in nitrous oxide, nitric oxide, and nitrogen dioxide, calculate the amount of substance of nitrogen oxides.

4. The method for intelligent control of ammonia injection zones based on multi-objective optimization according to claim 3, characterized in that: In the step 2, the evaluation process of the relative content of ammonia nitrogen includes: The amount of ammonia is calculated using the mass of the injected ammonia and the molar mass of the ammonia, weights are assigned to the amount of nitrogen oxides and the amount of ammonia, and the assigned weights are summed to obtain the ammonia nitrogen weight sum; Combined with the weighted average method, the relative content of ammonia nitrogen is evaluated. The product of the amount of nitrogen oxides and their corresponding weights and the product of the amount of ammonia and their corresponding weights are calculated to obtain the weighted values of nitrogen oxides and ammonia, respectively. The weighted values of nitrogen oxides and ammonia are added and then divided by the sum of the ammonia nitrogen weights to obtain the relative content of ammonia nitrogen. The relative content of ammonia nitrogen is then integrated into the ammonia injection monitoring data set.

5. The method for intelligent control of ammonia injection zones based on multi-objective optimization according to claim 4, characterized in that: In step 2, the evaluation process of the theoretical ammonia injection amount includes: Ammonia injection data and flue gas data were extracted from the ammonia injection monitoring dataset. Using the training set data and a multivariate linear regression algorithm, the injected ammonia concentration, flue gas nitrous oxide concentration, flue gas nitric oxide concentration, and flue gas nitrogen dioxide concentration were used as inputs, and the theoretical ammonia injection amount was used as output. The linear relationship between the injected ammonia concentration, flue gas nitrous oxide concentration, flue gas nitric oxide concentration, flue gas nitrogen dioxide concentration, and the theoretical ammonia injection amount was learned to train a theoretical ammonia injection evaluation model. The test set data is input into the theoretical ammonia injection evaluation model, the intercept term and regression coefficient of the theoretical ammonia injection evaluation model are adjusted, the theoretical ammonia injection evaluation model is optimized, and the final theoretical ammonia injection evaluation model is obtained. The corresponding theoretical ammonia injection amount is output based on the injected ammonia concentration, the nitrous oxide concentration in the flue gas, the nitric oxide concentration in the flue gas, and the nitrogen dioxide concentration in the flue gas.

6. The method for intelligent control of ammonia injection zones based on multi-objective optimization according to claim 5, characterized in that: In step 2, the evaluation process of the ammonia escape amount includes: Based on the impact of flue gas flow on ammonia escape, the flue gas flow is divided into low, medium, and high flue gas flow ranges. Ammonia escape scores are assigned to each divided flue gas flow range. Based on the flue gas flow range in which the current flue gas flow falls, the ammonia escape score corresponding to the flue gas flow is obtained. According to the influence of the injected ammonia concentration on the ammonia escape amount, the injected ammonia concentration is divided into a low ammonia concentration range, a medium ammonia concentration range, and a high ammonia concentration range. Ammonia escape scores are assigned to each divided ammonia concentration range. Based on the ammonia concentration range in which the current injected ammonia concentration is located, the ammonia escape score corresponding to the injected ammonia concentration is obtained; According to the influence of the relative ammonia nitrogen content on the ammonia escape amount, the relative ammonia nitrogen content is divided into a low ammonia nitrogen relative content range, a medium ammonia nitrogen relative content range and a high ammonia nitrogen relative content range, and an ammonia escape score is assigned to each divided relative ammonia nitrogen content range. According to the relative ammonia nitrogen content range in which the current relative ammonia nitrogen content is located, the ammonia escape score corresponding to the relative ammonia nitrogen content is obtained; The ammonia escape amount is evaluated by simulating the power function relationship between the flue gas flow rate, the injected ammonia concentration and the relative content of ammonia nitrogen, and then the ammonia escape amount is obtained.

7. The method for intelligent control of ammonia injection zones based on multi-objective optimization according to claim 6, characterized in that: In step 3, the process of obtaining the activity index of the ammonia nitrogen reaction catalyst includes: According to the factors affecting the activity of the ammonia nitrogen reaction catalyst, weights are assigned to the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and ammonia nitrogen relative content respectively; The weighted summation method is used to obtain the ammonia-nitrogen reaction catalyst activity index through the ammonia injection zone temperature, ammonia injection zone pressure, flue gas volume, flue gas flow rate and relative ammonia nitrogen content and their corresponding weights, and the ammonia-nitrogen reaction catalyst activity index is integrated into the ammonia injection monitoring data set.

8. The method for intelligent control of ammonia injection zones based on multi-objective optimization according to claim 7, characterized in that: In step 4, the process of constructing an ammonia injection supply restriction model and obtaining the maximum ammonia supply includes: Extract the ammonia injection partition data, ammonia nitrogen relative content and ammonia nitrogen reaction catalyst activity index from the ammonia injection monitoring data set; Combining training data with a multivariate linear regression algorithm, the ammonia injection partition data, relative ammonia nitrogen content, and ammonia nitrogen reaction catalyst activity index are used as input, and the maximum ammonia supply rate is used as output. The linear relationship between the ammonia injection partition data, relative ammonia nitrogen content, ammonia nitrogen reaction catalyst activity index, and the maximum ammonia supply rate is learned to train the ammonia injection supply limitation model. The test set data is input into the ammonia injection supply limitation model, the intercept term and regression coefficient of the ammonia injection supply limitation model are adjusted, the performance of the ammonia injection supply limitation model is optimized, and the final ammonia injection supply limitation model is obtained. Combined with the ammonia injection partition data, the relative content of ammonia nitrogen and the activity index of the ammonia nitrogen reaction catalyst, the corresponding maximum ammonia supply is output and integrated into the ammonia injection monitoring data set.

9. The method for intelligent control of ammonia injection zones based on multi-objective optimization according to claim 8, characterized in that: In step 5, the process of constructing an ammonia injection calibration model and then obtaining an ammonia injection calibration coefficient includes: Ammonia slip and maximum ammonia supply were extracted from the ammonia injection monitoring dataset. The training set data was then combined with a multivariate linear regression algorithm. The ammonia slip and maximum ammonia supply were used as inputs, and the ammonia injection calibration coefficient was used as output. The linear relationship between these factors was learned to train the ammonia injection calibration model. The test set data is input into the ammonia injection calibration model, the intercept term and regression coefficient of the ammonia injection calibration model are adjusted, the performance of the ammonia injection calibration model is optimized, and the ammonia injection calibration model is deployed into the system. Combined with the ammonia escape amount and the maximum ammonia supply amount, the corresponding ammonia injection calibration coefficient is output.

10. The method for intelligent control of ammonia injection zones based on multi-objective optimization according to claim 9, characterized in that: In step 6, the process of analyzing the output results of the ammonia injection calibration model, calibrating the theoretical ammonia injection amount, and intelligently controlling the ammonia injection zones includes: According to the ammonia injection calibration coefficient output by the ammonia injection calibration model, the ammonia injection calibration coefficient is divided into three ammonia injection calibration intervals, and a calibration scheme is allocated to each ammonia injection calibration interval; According to the calibration scheme corresponding to the ammonia injection calibration interval where the actual ammonia injection calibration coefficient is located, the calibrated ammonia injection amount is obtained, the theoretical ammonia injection amount is calibrated, and the ammonia injection amount of each ammonia injection zone is controlled to realize intelligent control of the ammonia injection zone.

Images