A coal-fired boiler NO x Method and system for generating real-time concentration predictions

Abstract

The present application relates to the technical field of coal-fired boiler, and discloses a coal-fired boiler NOx generation concentration real-time prediction method and system, the method of which obtains multiple independent variable features of boiler NOx generation concentration, carries out correlation analysis on the independent variable features and boiler NOx measurement concentration, filters out main independent variable features, takes the ammonia injection moment of the denitration system of the boiler as the current moment, carries out time series analysis on each independent variable feature, determines the historical time of each independent variable feature relative to the boiler NOx measurement concentration, constructs multiple NOx concentration prediction data groups with time series, trains the multiple NOx concentration prediction data groups based on a linear model of least square method, constructs a NOx generation concentration prediction model, and carries out real-time prediction on the NOx generation concentration at the ammonia injection moment through the NOx generation concentration prediction model, so as to solve the hysteresis of ammonia injection and improve the accuracy of ammonia injection.

Description

Technical Field

[0001] This invention relates to the field of coal-fired boiler technology, and in particular to a method and system for real-time prediction of NOx generation concentration in coal-fired boilers. Background Technology

[0002] Currently, most coal-fired boilers adopt low-NOx combustion methods and are equipped with selective catalytic reduction (SCR) denitrification systems to achieve ultra-low NOx emissions. The SCR denitrification system is equipped with continuous emission monitoring systems at both the inlet and outlet, which can continuously monitor the NOx and O2 concentrations entering the system. This NOx concentration is the NOx generation concentration of the coal-fired boiler, and the ammonia injection control of the SCR denitrification system is also based on this NOx concentration.

[0003] However, currently, continuous emission monitoring systems all use the method of extracting flue gas into an analyzer and then measuring NOx concentration using infrared or chemiluminescence methods. This method has a significant lag relative to the ammonia injection time, and ammonia nitrogen cannot be accurately matched. As a result, it is impossible to truly achieve precise ammonia injection control of the denitrification system, which increases ammonia escape at the denitrification system outlet. This can easily cause a series of problems such as air preheater blockage and reduced denitrification efficiency, seriously affecting the safe operation of the boiler unit. Summary of the Invention

[0004] This invention provides a method and system for real-time prediction of NOx generation concentration in coal-fired boilers, which solves a series of problems caused by the lag in NOx generation concentration monitoring, such as inaccurate ammonia nitrogen matching control, which can easily lead to air preheater blockage, reduced denitrification efficiency, and other issues that seriously affect the safe operation of boiler units.

[0005] In view of this, the first aspect of the present invention provides a method for real-time prediction of NOx formation concentration in a coal-fired boiler, comprising the following steps:

[0006] Multiple independent variable features of NOx generation concentration in boilers are obtained, and a candidate pool of independent variable features is constructed.

[0007] Correlation analysis was performed on each independent variable feature in the candidate pool of independent variable features and the measured NOx concentration in the boiler to screen out the main independent variable features and construct the independent variable feature pool.

[0008] Using the ammonia injection time of the boiler's denitrification system as the current time, time series analysis is performed on each independent variable feature in the independent variable feature pool to determine the historical time of each independent variable feature relative to the measured NOx concentration of the boiler.

[0009] Multiple NOx concentration prediction data sets with time series are constructed by combining the independent variable feature pool with the historical time of the boiler NOx measurement concentration corresponding to each independent variable feature. The NOx concentration prediction data set contains several independent variable features and their corresponding boiler NOx generation concentrations.

[0010] A linear model based on the least squares method is trained on multiple NOx concentration prediction datasets to construct a NOx generation concentration prediction model. This model is then used to predict the NOx generation concentration at the ammonia injection time in real time.

[0011] Preferably, the candidate pool of independent variables includes the amount of coal fed into the furnace, the amount of secondary air, the nitrogen content of the coal fed into the furnace, the oxygen content of the boiler during operation, the unit power, the boiler load, and the furnace temperature.

[0012] Preferably, the method further includes:

[0013] The steps to obtain secondary air volume specifically include:

[0014] The resistance coefficient of the secondary air damper is calculated by the opening degree of the secondary air damper entering the furnace of the coal-fired boiler.

[0015] The air volume of the secondary air nozzle is calculated based on the area of ​​the secondary air nozzle, the differential pressure of the furnace air box, the secondary air density, and the resistance coefficient of the secondary air damper. The sum of the air volumes of each secondary air nozzle equals the total secondary air volume of the boiler.

[0016] Preferably, the method further includes:

[0017] The air volume of the secondary air nozzle is calculated using the following formula based on the area of ​​the secondary air nozzle, the differential pressure of the furnace air box, the secondary air density, and the resistance coefficient of the secondary air damper:

[0018]

[0019] In the formula, Q2i is the air volume of the i-th secondary air nozzle; Ai is the area of ​​the i-th secondary air nozzle; ΔP is the differential pressure of the furnace air box; ρ is the secondary air density; and ξi is the resistance coefficient of the i-th secondary air damper.

[0020] Preferably, the method further includes:

[0021] Acquire the total moisture content, calorific value, and nitrogen content of several coal samples fed into the furnace. Use the total moisture content and calorific value as independent variables and the nitrogen content as the dependent variable to perform data fitting, and obtain fitting functions between the nitrogen content and the total moisture content and the calorific value, respectively.

[0022] Preferably, the step of performing correlation analysis between each independent variable feature in the candidate pool and the measured NOx concentration in the boiler, screening out the main independent variable features, and constructing the independent variable feature pool specifically includes:

[0023] Acquire historical data of each independent variable characteristic and historical data of NOx concentration measurement in the boiler during its historical operation.

[0024] The least squares method was used to perform correlation analysis between the historical data of each independent variable characteristic and the historical data of the measured NOx concentration in the boiler, and the first correlation index was obtained.

[0025] Independent variable features that have a first correlation index greater than a preset first correlation index threshold are selected and an independent variable feature pool is constructed.

[0026] Preferably, the step of performing time series analysis on each independent variable feature in the independent variable feature pool, using the ammonia injection time of the boiler's denitrification system as the current time, to determine the historical time of each independent variable feature relative to the measured NOx concentration of the boiler, specifically includes:

[0027] Using the ammonia injection time of the denitrification system as the current time, the least squares method is used to perform correlation analysis between the characteristics of the independent variable at the current time and the NOx generation concentration measured at the inlet of the denitrification system at multiple preset lag times, and the second correlation index is obtained.

[0028] The lag time corresponding to the highest second correlation index was selected as the historical time of each independent variable characteristic relative to the measured NOx concentration in the boiler.

[0029] Preferably, the method further includes:

[0030] The lag time of the coal feed rate and secondary air volume relative to the ammonia injection time is calculated based on the flow channel length from the preset measurement point of the coal feed rate and secondary air volume to the preset ammonia injection position and the flue gas velocity.

[0031] Preferably, the NOx generation concentration prediction model is as follows:

[0032]

[0033] In the formula, NOx is the NOx formation concentration, k0 is a constant term, ki, kj and kk are coefficients, P1i is the independent variable characteristic at the time t1 of the ammonia injection advance, n1 is the number of independent variable characteristics at the time t1 of the ammonia injection advance, i is the sequence number of the independent variable characteristic at the time t1 of the ammonia injection advance, P2j is the independent variable characteristic at the time of ammonia injection, n2 is the number of independent variable characteristics at the time of ammonia injection, j is the sequence number of the independent variable characteristic at the time of ammonia injection, P3k is the independent variable characteristic at the time t2 of the ammonia injection lag, n3 is the number of independent variable characteristics at the time t2 of the ammonia injection lag, and k is the sequence number of the independent variable characteristic at the time t2 of the ammonia injection lag.

[0034] Secondly, the present invention also provides a real-time prediction system for NOx formation concentration in coal-fired boilers, comprising:

[0035] The independent variable acquisition module is used to acquire multiple independent variable features of NOx generation concentration in boilers and construct an independent variable feature candidate pool.

[0036] The independent variable screening module is used to perform correlation analysis between each independent variable feature in the independent variable feature candidate pool and the measured NOx concentration in the boiler, screen out the main independent variable features, and construct the independent variable feature pool.

[0037] The lag analysis module is used to perform time series analysis on each independent variable feature in the independent variable feature pool, taking the ammonia injection time of the boiler's denitrification system as the current time, to determine the historical time of each independent variable feature relative to the measured NOx concentration of the boiler.

[0038] The time series construction module is used to construct multiple NOx concentration prediction data sets with time series by combining the independent variable feature pool with the historical time of the boiler NOx measurement concentration corresponding to each independent variable feature. The NOx concentration prediction data set contains several independent variable features and their corresponding boiler NOx generation concentrations.

[0039] The prediction model training module is used to train multiple NOx concentration prediction data sets based on a linear model using the least squares method, construct a NOx generation concentration prediction model, and use the NOx generation concentration prediction model to predict the NOx generation concentration at the ammonia injection time in real time.

[0040] As can be seen from the above technical solutions, the present invention has the following advantages:

[0041] This invention acquires multiple independent variable characteristics of NOx generation concentration in a boiler, performs correlation analysis between these characteristics and the measured NOx concentration, and identifies key independent variable characteristics to reduce computational load. Furthermore, using the ammonia injection time of the boiler's denitrification system as the current time, time series analysis is performed on each independent variable characteristic to determine its historical time relative to the measured NOx concentration. Multiple NOx concentration prediction data sets with time series are constructed. A linear model based on the least squares method is used to train these data sets, building a NOx generation concentration prediction model. This model then predicts the NOx generation concentration at the ammonia injection time in real time, thereby shortening the lag in ammonia injection timing, improving the accuracy of ammonia-nitrogen matching control at injection time, reducing the risk of air preheater blockage, reduced denitrification efficiency, and other problems, and ensuring the safe operation of the boiler unit. Attached Figure Description

[0042] Figure 1 A flowchart illustrating a real-time prediction method for NOx generation concentration in a coal-fired boiler, provided by an embodiment of the present invention;

[0043] Figure 2This is a schematic diagram of the time series results of the NOx concentration prediction data set provided in an embodiment of the present invention;

[0044] Figure 3 This is a schematic diagram of a real-time NOx generation concentration prediction system for a coal-fired boiler, provided in an embodiment of the present invention. Detailed Implementation

[0045] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] For easier understanding, please refer to Figure 1 The present invention provides a method for real-time prediction of NOx formation concentration in coal-fired boilers, comprising the following steps:

[0047] 101. Obtain multiple independent variable features of NOx generation concentration in boiler and construct a candidate pool of independent variable features.

[0048] Among them, the independent variable characteristic is the influencing factor of NOx formation concentration, which can be obtained through the NOx formation principle and boiler design.

[0049] The candidate pool of independent variable features includes the amount of coal fed into the furnace, the amount of secondary air, the nitrogen content of the coal fed into the furnace, the oxygen content of the boiler during operation, the unit power, the boiler load, and the furnace temperature.

[0050] It should be noted that coal-fired boilers generally only have the coal feed rate for each operating coal mill. When the boiler load changes, the coal feed rate is not equal to the amount of coal entering the furnace. If the coal feed rate is used to predict the NOx formation concentration, it will bring a large error and the effect will be poor. Therefore, the coal feed rate needs to be processed to obtain the amount of coal entering the furnace.

[0051] In this embodiment, the following methods are used to obtain the amount of coal to be fed into the furnace:

[0052] 1) Calculation method: Using the operating parameters of the coal mill, the amount of coal entering the boiler furnace is calculated. The specific calculation method adopts existing technology and will not be elaborated here.

[0053] 2) Delay method: The amount of coal fed into the furnace is delayed relative to the amount of coal fed into the furnace at a certain rate. The rate limit value is set according to different working conditions: when the coal feeder is just started, the rate limit is relatively large, for example, it is set at 0.1t / s; when the coal mill is running normally, another rate is set, for example, 0.5t / s; when the coal feeder is stopped, another rate is set, for example, 1t / s.

[0054] In this embodiment, it also includes:

[0055] The steps to obtain secondary air volume specifically include:

[0056] The resistance coefficient of the secondary air damper is calculated by measuring the opening degree of the secondary air damper entering the furnace of the coal-fired boiler.

[0057] In coal-fired boilers, the secondary air volume entering the furnace often lacks direct measurement points. The resistance characteristics of each secondary air damper generally conform to an exponential relationship, specifically:

[0058] ξi=ki×xi ni

[0059] In the formula: ξi is the resistance coefficient of the i-th secondary damper; xi is the opening degree of the i-th secondary damper; ki is a constant and ni is an exponent.

[0060] The air volume of the secondary air nozzle is calculated based on the area of ​​the secondary air nozzle, the differential pressure of the furnace air box, the secondary air density, and the resistance coefficient of the secondary air damper. The sum of the air volumes of each secondary air nozzle equals the total secondary air volume of the boiler.

[0061] The formula for calculating the air volume of the secondary air nozzle is as follows:

[0062]

[0063] In the formula: Q2i is the air volume of the i-th secondary air nozzle; Ai is the area of ​​the i-th secondary air nozzle; the area of ​​the secondary air nozzle can be obtained by referring to the boiler design data; ΔP is the furnace wind box differential pressure, measured value; ρ is the secondary air density, measured and calculated value.

[0064] In one example, the methods for obtaining the nitrogen content of the coal entering the furnace include:

[0065] 1) Direct measurement method

[0066] The nitrogen content of coal entering the furnace through the pulverized coal pipeline is directly measured using techniques such as laser-induced breakdown spectroscopy.

[0067] 2) Data fitting method

[0068] In one example, the total moisture content, calorific value, and nitrogen content of several coal samples fed into the furnace are obtained. The total moisture content and calorific value are used as independent variables, and the nitrogen content of the coal fed into the furnace is used as the dependent variable to perform data fitting, thereby obtaining fitting functions between the nitrogen content of the coal fed into the furnace and the total moisture content and the calorific value.

[0069] Specifically, several coal samples were taken for industrial and elemental analysis to obtain the total moisture (Mt), calorific value (Qnet), and nitrogen content (Nar) of the coal.

[0070] Using total moisture (Mt) and calorific value (Qnet) as independent variables and nitrogen content (Nar) of coal fed into the furnace as the dependent variable, the functional relationship between nitrogen content (Nar) of coal fed into the furnace and total moisture (Mt) and calorific value (Qnet) was fitted:

[0071] Nar = f1(Mt, Qnet)

[0072] In the formula, f1 represents a function;

[0073] The total moisture content (Mt) of the coal fed into the furnace is calculated through the heat balance of the coal mill, and the calorific value (Qnet) of the coal fed into the furnace is calculated through the heat balance of the unit. The specific calculation methods are not detailed here.

[0074] The nitrogen content of the coal fed into the furnace is obtained using a functional relationship.

[0075] 102. Correlation analysis was performed on each independent variable feature in the candidate pool with the measured NOx concentration in the boiler to screen out the main independent variable features and construct the independent variable feature pool.

[0076] Among them, the main independent variable features refer to the independent variable features whose correlation is greater than a certain threshold.

[0077] 103. Using the ammonia injection time of the boiler's denitrification system as the current time, perform time series analysis on each independent variable feature in the independent variable feature pool to determine the historical time of each independent variable feature relative to the measured NOx concentration of the boiler.

[0078] It should be noted that because infrared or chemiluminescence methods for measuring NOx formation concentration have a significant lag relative to the ammonia injection timing, ammonia-nitrogen levels cannot be precisely matched, thus hindering accurate ammonia injection control in the denitrification system and resulting in low accuracy in ammonia-nitrogen matching control. Therefore, this embodiment uses time-series analysis of the independent variable characteristics to determine the historical time of each independent variable characteristic relative to the boiler NOx measurement concentration. In other words, the independent variable characteristics directly affect the historical time of the boiler NOx measurement concentration, thereby shortening the lag in ammonia injection timing and improving the accuracy of ammonia-nitrogen matching control.

[0079] 104. Construct multiple NOx concentration prediction data sets with time series by combining the independent variable feature pool with the historical time of the boiler NOx measurement concentration corresponding to each independent variable feature. The NOx concentration prediction data set contains several independent variable features and their corresponding boiler NOx generation concentrations.

[0080] It is understandable that, since the historical time of each independent variable feature relative to the measured NOx concentration of the boiler is obtained as mentioned above, the ammonia injection time of the boiler's denitrification system can be used as the current time to form a NOx concentration prediction data set of several independent variable features and boiler NOx generation concentration. The independent variable feature in the NOx concentration prediction data set can be an independent variable feature or multiple independent variable features.

[0081] For example:

[0082] NOx concentration prediction data set 1 is [independent variable characteristics at time T-td, NOx generation concentration at time T], where td is the historical time of the boiler NOx measurement concentration corresponding to the independent variable characteristics, and time T is the ammonia injection time.

[0083] NOx concentration prediction data set 2 consists of [independent variable feature 1 at time T-t1, independent variable feature 2 at time T, independent variable feature 3 at time T+t2, and NOx measured concentration at time T+t3], where time T is the ammonia injection time, t1 is the advance time, and t2 and t3 are the lag times.

[0084] like Figure 2 As shown, Figure 2 The time series results of the NOx concentration prediction data set are shown. When t1 = 6s, the independent variable characteristic 1 at time T-6s is the amount of coal fed into the furnace, the amount of secondary air, and the nitrogen content of the coal fed into the furnace; the independent variable characteristic 2 at time T is the operating oxygen content; when t2 = 20s, the independent variable characteristic 3 at time T+20s is the unit power; and when t3 = 150s, the NOx generation concentration measurement results at time T+150s are shown.

[0085] 105. A linear model based on the least squares method is trained on multiple NOx concentration prediction datasets to construct a NOx generation concentration prediction model. The NOx generation concentration at the ammonia injection time is then predicted in real time using the NOx generation concentration prediction model.

[0086] It is understandable that, through the aforementioned methods, a sufficient number of NOx concentration prediction data sets can be obtained. Based on the least squares linear model, multiple NOx concentration prediction data sets can be trained to construct a NOx generation concentration prediction model.

[0087] The NOx formation concentration prediction model is as follows:

[0088]

[0089] In the formula, NOx is the NOx formation concentration, k0 is a constant term, ki, kj and kk are coefficients, P1i is the independent variable characteristic at the time t1 of the ammonia injection advance, n1 is the number of independent variable characteristics at the time t1 of the ammonia injection advance, i is the sequence number of the independent variable characteristic at the time t1 of the ammonia injection advance, P2j is the independent variable characteristic at the time of ammonia injection, n2 is the number of independent variable characteristics at the time of ammonia injection, j is the sequence number of the independent variable characteristic at the time of ammonia injection, P3k is the independent variable characteristic at the time t2 of the ammonia injection lag, n3 is the number of independent variable characteristics at the time t2 of the ammonia injection lag, and k is the sequence number of the independent variable characteristic at the time t2 of the ammonia injection lag.

[0090] It should be noted that this invention obtains multiple independent variable features of the boiler NOx generation concentration, performs correlation analysis between these features and the measured NOx concentration in the boiler, and filters out the main independent variable features, thereby reducing the computational load. Furthermore, using the ammonia injection time of the boiler's denitrification system as the current time, time series analysis is performed on each independent variable feature to determine the historical time of each feature relative to the measured NOx concentration in the boiler. Multiple NOx concentration prediction data sets with time series are constructed. A linear model based on the least squares method is used to train these multiple NOx concentration prediction data sets, constructing a NOx generation concentration prediction model. This model then predicts the NOx generation concentration at the ammonia injection time in real time, thereby shortening the lag in ammonia injection timing, improving the accuracy of ammonia-nitrogen matching control, reducing the risk of air preheater blockage, reduced denitrification efficiency, and other related problems, and ensuring the safe operation of the boiler unit.

[0091] In one specific embodiment, step 102 specifically includes:

[0092] 1021. Obtain historical data of each independent variable characteristic and historical data of NOx concentration measurement of the boiler during its historical operation.

[0093] Historical data refers to data from the boiler's operation over the past few days, weeks, or months.

[0094] 1022. The least squares method was used to perform correlation analysis between the historical data of each independent variable characteristic and the historical data of the boiler NOx measurement concentration to obtain the first correlation index.

[0095] 1023. Select independent variable features whose first correlation index is greater than the preset first correlation index threshold, and construct an independent variable feature pool.

[0096] The preset threshold for the first correlation index can be 0.1, for example:

[0097] A correlation analysis was performed on the NOx formation concentration measured at the inlet of a denitrification system and the amount of coal (Ga) fed into mill A, yielding the fitting function: NOx=k11+k12*Ga

[0098] In the formula, k11 and K12 are constants;

[0099] The correlation index R2 of the fitting function is 0.32, indicating that the amount of coal Ga fed into the furnace from mill A is relatively correlated with the NOx generation concentration measured at the inlet of the denitrification system. Therefore, the amount of coal Ga fed into the furnace from mill A can be used as an independent variable feature for predicting the NOx generation concentration.

[0100] Similarly, if the correlation index R² of the correlation analysis between oxygen content (O2) and the NOx generation concentration measured at the inlet of the denitrification system is 0.75, it indicates that oxygen content (O2) is highly correlated with the NOx generation concentration measured at the inlet of the denitrification system, and oxygen content (O2) should be used as an independent variable feature for predicting NOx generation concentration.

[0101] If the correlation index R² of the secondary air volume Q2b in layer B and the NOx generation concentration measured at the inlet of the denitrification system is 0.043, it indicates that the correlation between the secondary air volume Q2b in layer B and the NOx generation concentration measured at the inlet of the denitrification system is very weak, and the secondary air volume Q2b in layer B can be excluded from the prediction of NOx generation concentration.

[0102] In one specific embodiment, step 103 specifically includes:

[0103] 1031. Taking the ammonia injection time of the denitrification system as the current time, the least squares method is used to perform correlation analysis between the independent variable characteristics at the current time and the NOx generation concentration measured at the inlet of the denitrification system at multiple preset lag times, and the second correlation index is obtained.

[0104] 1032, the lag time corresponding to the highest second correlation index was selected as the historical time of each independent variable feature relative to the measured NOx concentration in the boiler.

[0105] For each independent variable feature, the NOx concentration measured at the inlet of the denitrification system has a lag time with each independent variable feature. Correlation analysis is performed on the NOx concentration measurements with multiple preset lag times and the independent variable features to determine the lag time corresponding to the highest correlation index, which is the historical time of each independent variable feature relative to the measured NOx concentration in the boiler.

[0106] For example: using the operating oxygen content (O2) as the independent variable, and taking the obtained operating oxygen content (O2) as the current time, the NOx concentration measured at the inlet of the denitrification system at lags of 50s, 100s, 150s, and 200s are used to form four sets of analytical data:

[0107] (1) O2 at time T corresponds to NOx at time T+50, O2 at time T+1 corresponds to NOx at time T+51, ...;

[0108] (2) O2 at time T corresponds to NOx at time T+100, and O2 at time T+1 corresponds to NOx at time T+101, ...;

[0109] (3) O2 at time T corresponds to NOx at time T+150, and O2 at time T+1 corresponds to NOx at time T+151, ...;

[0110] (4) O2 at time T corresponds to NOx at time T+200, and O2 at time T+1 corresponds to NOx at time T+201, ...;

[0111] Correlation analysis was performed on the above four sets of data, and the correlation indices were 0.85, 0.88, 0.90, and 0.89, respectively. This indicates that the current O2 is highly correlated with the NOx measurement at a time 150 seconds later, and the NOx measurement result lags the O2 measurement result by 150 seconds.

[0112] In another example, the lag time of NOx measurement concentration relative to the ammonia injection time was obtained experimentally as follows:

[0113] A standard NOx concentration, such as 300 mg / m³, is introduced into the flue gas extraction point of the continuous emission monitoring system at the inlet of the denitrification system. 3 A standard gas for NOx concentration was introduced, and timing began immediately. The continuous emission monitoring system recorded a value of 300 mg / m³. 3 The required time is the delay time between the measurement of NOx concentration and the ammonia injection time, for example, 180s in the test result.

[0114] In another example, it also includes:

[0115] The lag time of the coal feed rate and secondary air volume relative to the ammonia injection time is calculated based on the flow channel length from the preset measurement point of the coal feed rate and secondary air volume to the preset ammonia injection position and the flue gas velocity.

[0116] Specifically, the historical time of each predicted independent variable characteristic and NOx formation concentration (i.e., ammonia injection time) is obtained using calculation methods as follows:

[0117] The time required for the measured points of coal feed rate and secondary air volume to generate NOx after combustion in the furnace and then flow to the ammonia injection point is the time that the air and coal parameters are advanced relative to the ammonia injection time. This time is estimated by using the method of the total length of the flow channel and the flue gas velocity, where the total length of the flow channel is the distance from the measured points of coal feed rate and secondary air volume to the ammonia injection point.

[0118] For example, the amount of flue gas produced by combustion is 2500 t / h, the average flue gas temperature is 800℃, and the average flow area is 100 m². 2 If the total length of the flow channel is 120m, then the estimated flow time is:

[0119]

[0120] In the formula, The flue gas velocity is t; the amount of coal fed into the furnace and the amount of secondary air are approximately 6 seconds ahead of the ammonia injection time.

[0121] The above is a detailed description of an embodiment of a method for real-time prediction of NOx generation concentration in coal-fired boilers provided by the present invention. The following is a detailed description of an embodiment of a system for real-time prediction of NOx generation concentration in coal-fired boilers provided by the present invention.

[0122] For easier understanding, please refer to Figure 3 The present invention provides a real-time prediction system for NOx formation concentration in coal-fired boilers, comprising:

[0123] The independent variable acquisition module 100 is used to acquire multiple independent variable features of the NOx generation concentration in the boiler and construct an independent variable feature candidate pool.

[0124] The independent variable screening module 200 is used to perform correlation analysis between each independent variable feature in the independent variable feature candidate pool and the measured NOx concentration in the boiler, screen out the main independent variable features, and construct the independent variable feature pool.

[0125] The lag analysis module 300 is used to perform time series analysis on each independent variable feature in the independent variable feature pool, taking the ammonia injection time of the boiler's denitrification system as the current time, and to determine the historical time of each independent variable feature relative to the measured NOx concentration of the boiler.

[0126] The time series construction module 400 is used to construct multiple NOx concentration prediction data sets with time series by combining the independent variable feature pool with the historical time of the boiler NOx measurement concentration corresponding to each independent variable feature. The NOx concentration prediction data set contains several independent variable features and their corresponding boiler NOx generation concentrations.

[0127] The prediction model training module 500 is used to train a linear model based on the least squares method on multiple NOx concentration prediction datasets to build a NOx generation concentration prediction model. The NOx generation concentration prediction model is used to predict the NOx generation concentration at the ammonia injection time in real time.

[0128] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0129] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0130] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0131] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0132] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for real-time prediction of NOx formation concentration in coal-fired boilers, characterized in that, Includes the following steps: Multiple independent variable features of NOx generation concentration in boilers are obtained, and a candidate pool of independent variable features is constructed. Correlation analysis was performed on each independent variable feature in the candidate pool of independent variable features and the measured NOx concentration in the boiler to screen out the main independent variable features and construct the independent variable feature pool. Using the ammonia injection time of the boiler's denitrification system as the current time, time series analysis is performed on each independent variable feature in the independent variable feature pool to determine the historical time of each independent variable feature relative to the measured NOx concentration of the boiler, including: Using the ammonia injection time of the denitrification system as the current time, the least squares method is used to perform correlation analysis between the characteristics of the independent variable at the current time and the NOx generation concentration measured at the inlet of the denitrification system at multiple preset lag times, and the second correlation index is obtained. The lag time corresponding to the highest second correlation index was selected to be the historical time of each independent variable characteristic relative to the measured NOx concentration in the boiler. Multiple NOx concentration prediction data sets with time series are constructed by combining the independent variable feature pool with the historical time of the boiler NOx measurement concentration corresponding to each independent variable feature. The NOx concentration prediction data set contains several independent variable features and their corresponding boiler NOx generation concentrations. A linear model based on the least squares method is trained on multiple NOx concentration prediction datasets to construct a NOx generation concentration prediction model. This model is then used to predict the NOx generation concentration at the ammonia injection time in real time. The NOx generation concentration prediction model is as follows: In the formula, NOx is the NOx formation concentration, k0 is a constant term, ki, kj and kk are coefficients, P1i is the independent variable characteristic at the time t1 of the ammonia injection advance, n1 is the number of independent variable characteristics at the time t1 of the ammonia injection advance, i is the sequence number of the independent variable characteristic at the time t1 of the ammonia injection advance, P2j is the independent variable characteristic at the time of ammonia injection, n2 is the number of independent variable characteristics at the time of ammonia injection, j is the sequence number of the independent variable characteristic at the time of ammonia injection, P3k is the independent variable characteristic at the time t2 of the ammonia injection lag, n3 is the number of independent variable characteristics at the time t2 of the ammonia injection lag, and k is the sequence number of the independent variable characteristic at the time t2 of the ammonia injection lag.

2. The method for real-time prediction of NOx formation concentration in coal-fired boilers according to claim 1, characterized in that, The candidate pool of independent variables includes the amount of coal fed into the furnace, the amount of secondary air, the nitrogen content of the coal fed into the furnace, the oxygen content of the boiler during operation, the unit power, the boiler load, and the furnace temperature.

3. The method for real-time prediction of NOx formation concentration in coal-fired boilers according to claim 2, characterized in that, Also includes: The steps to obtain secondary air volume specifically include: The resistance coefficient of the secondary air damper is calculated by the opening degree of the secondary air damper entering the furnace of the coal-fired boiler. The air volume of the secondary air nozzle is calculated based on the area of ​​the secondary air nozzle, the differential pressure of the furnace air box, the secondary air density, and the resistance coefficient of the secondary air damper. The sum of the air volumes of each secondary air nozzle equals the total secondary air volume of the boiler.

4. The method for real-time prediction of NOx formation concentration in coal-fired boilers according to claim 3, characterized in that, Also includes: The air volume of the secondary air nozzle is calculated using the following formula based on the area of ​​the secondary air nozzle, the differential pressure of the furnace air box, the secondary air density, and the resistance coefficient of the secondary air damper: In the formula, Q2i is the air volume of the i-th secondary air nozzle; Ai is the area of ​​the i-th secondary air nozzle; The differential pressure of the furnace air box; Secondary wind density; i is the resistance coefficient of the i-th secondary damper.

5. The method for real-time prediction of NOx formation concentration in coal-fired boilers according to claim 2, characterized in that, Also includes: Acquire the total moisture content, calorific value, and nitrogen content of several coal samples fed into the furnace. Use the total moisture content and calorific value as independent variables and the nitrogen content as the dependent variable to perform data fitting, and obtain fitting functions between the nitrogen content and the total moisture content and the calorific value, respectively.

6. The method for real-time prediction of NOx formation concentration in coal-fired boilers according to claim 2, characterized in that, The steps of performing correlation analysis between each independent variable feature in the candidate pool and the measured NOx concentration in the boiler, screening out the main independent variable features, and constructing the independent variable feature pool specifically include: Acquire historical data of each independent variable characteristic and historical data of NOx concentration measurement in the boiler during its historical operation. The least squares method was used to perform correlation analysis between the historical data of each independent variable characteristic and the historical data of the measured NOx concentration in the boiler, and the first correlation index was obtained. Independent variable features that have a first correlation index greater than a preset first correlation index threshold are selected and an independent variable feature pool is constructed.

7. The method for real-time prediction of NOx formation concentration in coal-fired boilers according to claim 2, characterized in that, Also includes: The lag time of the coal feed rate and secondary air volume relative to the ammonia injection time is calculated based on the flow channel length from the preset measurement point of the coal feed rate and secondary air volume to the preset ammonia injection position and the flue gas velocity.

8. A real-time prediction system for NOx generation concentration in a coal-fired boiler, characterized in that, include: The independent variable acquisition module is used to acquire multiple independent variable features of NOx generation concentration in boilers and construct an independent variable feature candidate pool. The independent variable screening module is used to perform correlation analysis between each independent variable feature in the independent variable feature candidate pool and the measured NOx concentration in the boiler, screen out the main independent variable features, and construct the independent variable feature pool. The lag analysis module is used to perform time series analysis on each independent variable feature in the independent variable feature pool, taking the ammonia injection time of the boiler's denitrification system as the current time, to determine the historical time of each independent variable feature relative to the measured NOx concentration of the boiler. Using the ammonia injection time of the boiler's denitrification system as the current time, time series analysis is performed on each independent variable feature in the independent variable feature pool to determine the historical time of each independent variable feature relative to the measured NOx concentration of the boiler, including: Using the ammonia injection time of the denitrification system as the current time, the least squares method is used to perform correlation analysis between the characteristics of the independent variable at the current time and the NOx generation concentration measured at the inlet of the denitrification system at multiple preset lag times, and the second correlation index is obtained. The lag time corresponding to the highest second correlation index was selected to be the historical time of each independent variable characteristic relative to the measured NOx concentration in the boiler. The time series construction module is used to construct multiple NOx concentration prediction data sets with time series by combining the independent variable feature pool with the historical time of the boiler NOx measurement concentration corresponding to each independent variable feature. The NOx concentration prediction data set contains several independent variable features and their corresponding boiler NOx generation concentrations. The prediction model training module is used to train a linear model based on the least squares method on multiple NOx concentration prediction data sets to construct a NOx generation concentration prediction model. This model is then used to predict the NOx generation concentration at the ammonia injection time in real time. The NOx generation concentration prediction model is as follows: In the formula, NOx is the NOx formation concentration, k0 is a constant term, ki, kj and kk are coefficients, P1i is the independent variable characteristic at the time t1 of the ammonia injection advance, n1 is the number of independent variable characteristics at the time t1 of the ammonia injection advance, i is the sequence number of the independent variable characteristic at the time t1 of the ammonia injection advance, P2j is the independent variable characteristic at the time of ammonia injection, n2 is the number of independent variable characteristics at the time of ammonia injection, j is the sequence number of the independent variable characteristic at the time of ammonia injection, P3k is the independent variable characteristic at the time t2 of the ammonia injection lag, n3 is the number of independent variable characteristics at the time t2 of the ammonia injection lag, and k is the sequence number of the independent variable characteristic at the time t2 of the ammonia injection lag.

Images