Denitration monitoring system for SNCR (selective non-catalytic reduction) system of circulating fluidized bed boiler and use method

Abstract

The invention belongs to the technical field of flue gas denitration treatment of circulating fluidized bed boilers, and discloses a denitration monitoring system of an SNCR (selective non-catalytic reduction) system of a circulating fluidized bed boiler and a use method. The system comprises a separator and an air pre-heater, wherein a flue gas sampling device is connected between an inlet of the separator and an inlet of the air pre-heater; the flue gas sampling device is a pipeline with a sampling grid at an inlet; the inlet end of the flue gas sampling device is arranged at the upstream of the SNCR spray gun, a first monitoring system is arranged in a pipeline of the flue gas sampling device, and a second monitoring system is arranged at the inlet of the air pre-heater; the first monitoring system comprises a first oxygen monitoring device and a first NOX monitoring device; and the second monitoring system comprises a second oxygen monitoring device, a second NOX monitoring device and an NH3 monitoring device. The flue gas sampling grating is additionally arranged in front of the spray gun at the inlet of the separator, so that the problem of insufficient urea solution injection amount control precision caused by measurement data lag and large model error in the prior art is solved.

Description

Technical Field

[0001] The invention belongs to the technical field of circulating fluidized bed boiler flue gas denitration treatment, and in particular relates to a circulating fluidized bed boiler SNCR system denitration monitoring system and a use method thereof. Background Art

[0002] At present, the SNCR denitrification process is the most commonly used method for controlling nitrogen oxides in circulating fluidized bed boilers. Usually, several SNCR (selective non-catalytic reduction) spray guns are arranged at the flue duct at the inlet of the separator, so that the nitrogen oxides in the boiler flue gas react with ammonia in the absence of a catalyst to generate nitrogen and water that will not affect the environment. At present, the biggest problem in the operation of the SNCR system of circulating fluidized bed boilers is that the flue gas temperature at the inlet of the separator is high, and there is no suitable flue gas CEMS system to monitor the original NOx concentration value produced by the boiler; the second reason is that the NOx concentration value after denitrification of the SNCR system is monitored by the CEMS system (Continuous Emission Monitoring System) at the chimney inlet, and the data is seriously lagging; in addition, due to the absence of a catalyst, the denitrification efficiency of the SNCR system is closely related to the flue gas temperature at the inlet of the separator, the ammonia nitrogen molar ratio (NSR) of the SNCR, and the concentration of urea solution. Due to the lack of original NOx concentration value, serious lag of outlet NOx concentration value, separator inlet flue gas temperature, ammonia nitrogen molar ratio (NSR) and other comprehensive factors, the automatic control difficulty of the SNCR system is much greater than that of the SCR system of the pulverized coal furnace.

[0003] At present, the SNCR system of circulating fluidized bed boiler mostly uses PID control system to control the real-time flow of urea solution, that is, the error between the target value of NOx concentration at the chimney inlet and the real-time measured value is calculated by proportion, integration, differentiation, etc., and then the real-time flow of urea solution is controlled by controlling the opening of pneumatic control valve; then the real-time flow of dilution water is controlled by controlling the opening of pneumatic control valve according to the set urea solution concentration, so as to ensure that the urea solution concentration meets the operation requirements, and finally the NOx concentration value at the chimney inlet meets the emission standard. At present, the relatively more commonly used parameter determination method is engineering correction method, which requires repeated tests by engineering and technical personnel, and because the system has large inertia and time lag characteristics, it is too dependent on the experience and technical level of personnel. However, in the actual operation process, as the boiler load and working conditions change, the amount of urea solution is related to multiple factors such as flue gas volume and ammonia nitrogen molar ratio, so the PID control parameters should have adaptive correction function.

[0004] The Chinese patent publication number is CN115253636A, and the name is a patent application for a circulating fluidized bed boiler wide load denitration system, which includes a hot air generation system, a urea solution supply system, a urea pyrolysis furnace, a cyclone separator and a control system connected by pipelines; the hot air generated by the hot air generation system and the urea solution provided by the urea solution supply system enter the urea pyrolysis furnace and pyrolyze the urea solution to generate a reducing agent mixed gas containing ammonia, the reducing agent mixed gas is divided into multiple branches and mixed with the original flue gas from the furnace outlet respectively, and then enters the independent cyclone separators for denitration reaction, and finally discharged through the chimney; wherein the control system determines the supply amount of urea solution by collecting and testing the original flue gas components entering the cyclone separator, the flue gas components after denitration by the cyclone separator, and the flue gas components at the total outlet of the chimney. This patent application cannot solve the problem of insufficient control accuracy of urea solution injection caused by measurement data lag and large model errors. Summary of the invention

[0005] In order to overcome the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a denitrification monitoring system for a circulating fluidized bed boiler SNCR system and a method of use, by adding a flue gas sampling grid in front of the spray gun at the separator inlet, to solve the problem of insufficient control accuracy of urea solution injection amount caused by measurement data lag and large model error in the prior art.

[0006] To achieve the above object, the technical solution adopted by the present invention is: In the first aspect, the present invention provides a denitrification monitoring system for an SNCR system of a circulating fluidized bed boiler, comprising: a separator and an air preheater, wherein the outlet of the separator is connected to the inlet of the air preheater, and an SNCR spray gun is arranged at the inlet of the separator; a flue gas sampling device is connected between the inlet of the separator and the inlet of the air preheater; the flue gas sampling device is a pipe with a sampling grid at the inlet; the inlet end of the flue gas sampling device is arranged upstream of the SNCR spray gun, a first monitoring system is arranged in the pipe of the flue gas sampling device, and a second monitoring system is arranged at the inlet of the air preheater; the first monitoring system comprises a first oxygen monitoring device and a first NOx monitoring device; the second monitoring system comprises a second oxygen monitoring device, a second NOx monitoring device and an NH3 monitoring device.

[0007] Optionally, the pipeline of the smoke sampling device is connected to a compressed air pipeline.

[0008] Optionally, a superheater is provided between the outlet of the separator and the inlet of the air preheater.

[0009] Optionally, a reheater and an economizer are provided between the inlet of the superheater and the air preheater.

[0010] Optionally, the flue gas sampling device is arranged within a range of 1.5m to 2m upstream of the SNCR spray gun.

[0011] In a second aspect, the present invention provides a method for using the circulating fluidized bed boiler SNCR system denitration monitoring system, characterized in that it comprises the following steps: S1: setting an initial NSR design value according to the inlet NOx concentration value monitored by the first monitoring system; S2: According to the NO monitored by the first monitoring system X Urea consumption is calculated by concentration and flue gas flow rate; S3: Calculate the dilution water consumption according to the urea consumption; S4: Conducting field tests to determine the optimal urea solution concentration value under different loads and obtaining a fitting curve of the optimal urea solution concentration value; S5: Obtain historical operation data, and train the historical operation data to obtain a NOx generation prediction regression model and an NSR prediction regression model.

[0012] Optionally, a virtual SCR reactor model is established, a reaction loss function of the SNCR system is established, and an optimal solution is obtained by a differential evolution algorithm to obtain a SNCR denitration reaction model; The NOx concentration at the separator inlet is predicted by the NOx generation prediction regression model, and the urea solution consumption of the SNCR system is calculated and controlled by the NSR prediction regression model. The NOx concentration at the air preheater inlet is predicted by the virtual SCR reactor model of the SNCR system, and the dilution water volume is controlled by the fitting curve of the optimal urea solution concentration value.

[0013] Optionally, in step S2, the calculation formula for urea consumption is: ; Where: : Urea consumption, kg / h; : Flue gas flow rate at separator inlet at 6% O2 concentration, m3 / h; : Average NOX concentration at the separator inlet section at 6% O2 concentration, mg / m3; : NO2 molar mass, g / mol; NSR: NH3 / NOX molar ratio; : NH3 molar mass, g / mol.

[0014] Optionally, in step S3, the calculation formula for dilution water consumption is: ; ; Where: : Urea solution quality; : Concentration of urea solution prepared by the power plant; : The optimal urea solution mass concentration of the SNCR system determined by the experiment; : The amount of dilution water required by the SNCR system to dilute the urea solution to the optimal concentration.

[0015] Optionally, in step S5, the specific process of using a gated recurrent neural network to build a NOX generation prediction model is as follows: by setting the parameters of the activation function, learning rate, sample batch size, number of iterations and weight decay coefficient required to establish the NOX generation prediction model, a gated recurrent neural grid structure is used to build a NOX generation prediction model network structure; wherein: the expression of the activation function is , ; After training the NOX generation prediction model network structure through the training data set, the NOX generation prediction model is obtained. The expression of the NOX generation prediction model is: ; Where: is the dependent variable of the NOx generation prediction model and X1 is the normalized independent variable of the NOx generation prediction model.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention adds a flue gas sampling grid in front of the separator inlet spray gun, and allows a small proportion of flue gas to automatically flow into the air preheater inlet through the differential pressure between the separator and the air preheater inlet, and sets a first monitoring system on the pipeline; at the same time, the flue gas dissipates heat to the environment during flow, avoiding damage to the flue gas monitoring instrument by high temperature. By setting a second monitoring system at the air preheater inlet, the concentration of nitrogen oxides after denitrification is mastered. By selecting boiler operation data to predict the original NOX concentration at the separator inlet and the optimal value of NSR, the NOX content at the air preheater inlet is predicted in combination with the loss function of the virtual SCR reactor, and the urea solution consumption is adjusted after comparison with the actual measured value of the second monitoring system.

[0017] The problem of insufficient control accuracy of urea solution injection amount caused by measurement data lag and large model error in the prior art is thereby solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are for explanation purposes only and are not intended to limit the scope of the present invention in any way. In addition, the shapes and proportional dimensions of the components in the drawings are only for illustration purposes to help understand the present invention, and are not intended to specifically limit the shapes and proportional dimensions of the components of the present invention. In the drawings: Figure 1 This is a schematic diagram of NOX monitoring at the inlet and outlet of the SNCR system of the present invention; Figure 2 This is a schematic diagram of grid sampling of the inlet and outlet positions of the present invention; Figure 3 SNCR system diagram of Example 1 of the present invention; Among them, 1. boiler furnace; 2. cyclone separator; 3. flue gas sampling device; 4. air preheater; 21. SNCR spray gun; 31. compressed air pipeline; 32. first oxygen monitoring device; 33. first NOX monitoring device; 41. superheater; 42. reheater; 43. economizer; 44. second oxygen monitoring device; 45. second NOX monitoring device; 46. flow monitoring device; 47. NH3 monitoring device; 101. self-compressed air main pipe; 102. self-urea solution delivery pump; 103. self-dilution water pump. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0020] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] In the description of the embodiments of the present invention, it should be noted that if the terms "upper", "lower", "horizontal", "inner", etc. appear, the orientation or position relationship indicated is based on the orientation or position relationship shown in the accompanying drawings, or is the orientation or position relationship in which the product of the invention is usually placed when used. It is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0022] When an element is referred to as being "disposed on" another element, it may be directly on the other element or there may also be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may also be a central element. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only and are not intended to be the only embodiments. If the term "horizontal" appears, it does not mean that the component is required to be absolutely horizontal, but it can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", which does not mean that the structure must be completely horizontal, but it can be slightly tilted.

[0023] It should be noted that similar reference numerals and letters represent similar items in the following figures, so once an item is defined in one figure, it does not need to be further defined and explained in the subsequent figures. In the description of the present invention, it is understood that the terms "include" and "comprise" indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by those skilled in the art of the technical field of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, the singular forms of "a", "an" and "the" are intended to include plural forms unless the context clearly indicates otherwise.

[0025] The present invention is described in detail below with reference to the accompanying drawings.

[0026] A denitration monitoring system for an SNCR system of a circulating fluidized bed boiler, characterized in that it comprises: a separator 2 and an air preheater 4, wherein the outlet of the separator 2 is connected to the inlet of the air preheater 4, and an SNCR spray gun 21 is arranged at the inlet of the separator 2; a flue gas sampling device 3 is connected between the inlet of the separator 2 and the inlet of the air preheater 4; the flue gas sampling device 3 is a pipeline with a sampling grid at the inlet; the inlet end of the flue gas sampling device 3 is arranged upstream of the SNCR spray gun 21, a first monitoring system is arranged in the pipeline of the flue gas sampling device 3, and a second monitoring system is arranged at the inlet of the air preheater 4; the first monitoring system comprises a first oxygen monitoring device 32 and a first NOx monitoring device 33; the second monitoring system comprises a second oxygen monitoring device 44, a second NOx monitoring device 45 and an NH3 monitoring device 47.

[0027] The method for using the denitration monitoring system of the circulating fluidized bed boiler SNCR system comprises the following steps: S1: setting an initial NSR design value according to the inlet NOx concentration value monitored by the first monitoring system; S2: According to the NO monitored by the first monitoring system X Urea consumption is calculated by concentration and flue gas flow rate; S3: Calculate the dilution water consumption according to the urea consumption; S4: Conducting field tests to determine the optimal urea solution concentration value under different loads and obtaining a fitting curve of the optimal urea solution concentration value; S5: Obtain historical operation data, and train the historical operation data to obtain a NOx generation prediction regression model and an NSR prediction regression model.

[0028] The present invention adds a flue gas sampling grid in front of the separator inlet spray gun, and allows a small proportion of flue gas to automatically flow into the air preheater 4 inlet through the differential pressure between the cyclone separator 2 and the air preheater 4 inlet, and sets a first monitoring system on the pipeline; at the same time, the flue gas dissipates heat to the environment during flow, avoiding damage to the flue gas monitoring instrument by high temperature. By setting a second monitoring system at the air preheater 4 inlet, the concentration of nitrogen oxides after denitrification is mastered. By selecting boiler operation data to predict the original NOX concentration at the separator inlet and the optimal value of NSR, the NOX content at the air preheater inlet is predicted in combination with the loss function of the virtual SCR reactor, and the urea solution consumption is adjusted after comparison with the actual measured value of the second monitoring system. The problem of insufficient control accuracy of the urea solution injection amount caused by the lag of measurement data and large model error is solved.

[0029] Example 1 like Figure 3 As shown, in this embodiment, the SNCR system has a self-compressed air main pipe 101, a self-urea solution delivery pump 102 and a self-dilution water pump 103, wherein the self-compressed air main pipe 101 is used for atomizing air. The self-compressed air main pipe 101 is connected to the compressed air pipeline 31.

[0030] A circulating fluidized bed boiler SNCR system denitration monitoring system of this embodiment includes: a separator 2 and an air preheater 4, the outlet of the separator 2 is connected to the inlet of the air preheater 4, and an SNCR spray gun 21 is provided at the inlet of the separator 2. The separator 2 is connected to the boiler furnace 1 upstream, and is connected to the superheater 41, the reheater 42, the economizer 43 and the air preheater 4 in sequence downstream.

[0031] In order to prevent the injected urea solution from affecting the test of the original NOx concentration value, a flue gas sampling device 3 is connected between the inlet of the separator 2 and the inlet of the air preheater 4; the flue gas sampling device 3 is a pipe with a sampling grid at the inlet; the inlet end of the flue gas sampling device 3 is arranged upstream of the SNCR spray gun 21. The flue gas is led to the inlet of the air preheater 4 by the pressure difference between the inlet of the separator 2 and the inlet of the air preheater 4.

[0032] A first monitoring system is arranged in the pipeline of the flue gas sampling device 3, and a second monitoring system is arranged at the inlet of the air preheater 4, which are used to monitor the original NOx concentration value produced by the boiler.

[0033] The first monitoring system includes a first oxygen monitoring device 32 and a first NOx monitoring device 33 connected to the flue gas sampling device 3 through a pipeline. The second monitoring system includes a second oxygen monitoring device 44, a second NOx monitoring device 45, a flow monitoring device 46 and an NH3 monitoring device 47 connected to the inlet of the air preheater 4 through a pipeline.

[0034] Optionally, the compressed air pipeline 31 is connected to the power plant's miscellaneous compressed air.

[0035] Optionally, the flue gas sampling device 3 is a grid-method flue gas sampling device, which merges the flue gas into a DN25 pipe through the flue gas sampling grid and introduces it into the inlet of the air preheater 4. Since the flue gas temperature at the sampling point is as high as 850~950℃, the steel pipe used for sampling by the flue gas sampling device 3 needs to be made of high-temperature resistant stainless steel. Optionally, the pipe of the flue gas sampling device 3 is made of 310S stainless steel. In order to ensure that the flue gas temperature reaching the inlet of the air preheater 4 is reduced to less than 400℃, the DN25 flue gas stainless steel pipe does not need to be insulated and dissipates heat naturally.

[0036] Preferably, the flue gas sampling device 3 is arranged within a range of 1.5 m to 2 m upstream of the SNCR spray gun 21 .

[0037] Optionally, the pipeline of the flue gas sampling device 3 is connected to a compressed air pipeline 31. In order to prevent dust in the flue gas from clogging the pipeline and the sampling port of the flue gas sampling grid during long-term operation, a compressed air pipeline 31 is provided on the pipeline for backblowing, and the backblowing frequency is determined according to the dust concentration of the flue gas.

[0038] Optionally, the second monitoring system is a CEMS (Continuous Emission Monitoring System), which is used to monitor the flue gas composition after denitrification, to guide the combustion control of the power plant, and to adjust the operating oxygen content.

[0039] Example 2 Since the SCR system has a catalytic effect, under the premise that the catalyst activity meets the performance requirements, the NSR is directly calculated from the denitrification efficiency and ammonia slip, and the urea consumption of the SCR system can be directly calculated through the flue gas volume, NSR, and inlet NOx concentration.

[0040] During actual operation, the specific NSR is related to factors such as boiler load (flue gas volume), flue gas temperature at the inlet of separator 2, concentration of diluted urea solution, and original NOx concentration value. Once the specific NSR value is obtained, the consumption of urea solution is directly calculated through the boiler load (flue gas volume) and the original NOx concentration value.

[0041] The amount of urea granules is calculated as follows:

[0042] Where: : Urea consumption, kg / h; :6%O 2 Flue gas flow rate at separator 2 inlet at concentration, m 3 / h; :6%O 2 Separator 2 inlet cross section NO X Average concentration, mg / m 3 ; : NO 2 Molar mass, g / mol; NSR :NH 3 / NO X Molar ratio; :NH 3 Molar mass, g / mol.

[0043] The urea consumption is calculated by the above formula. The power plant first prepares the urea granules into 40%~50% urea solution for storage.

[0044]

[0045] Where: : That is, the urea solution of a certain concentration prepared and stored by the power plant that needs to be consumed, with a mass concentration of 40%~50%, which is also the mass of the urea solution measured by the urea solution electromagnetic flowmeter in the SNCR system; : Concentration of urea solution prepared by the power plant.

[0046] The SNCR system uses dilution water to dilute the 40%~50% urea solution to the required concentration and controls the concentration to 8%~12%. The specific amount needs to be confirmed by on-site tests. The consumption of dilution water is:

[0047] Where: : The optimal urea solution mass concentration of the SNCR system determined by experiment.

[0048] : The amount of dilution water required by the SNCR system to dilute the urea solution to the optimal concentration, which is also the amount of dilution water measured by the turbine flowmeter set up in the SNCR system.

[0049] The effect of the diluted urea solution concentration on the NSR value was determined through field tests.

[0050] Taking the stable load as an example, the power plant operators stabilize the unit operating load, try to ensure the stability of coal quality, stable parameters such as operating oxygen content, control the amount of 50% concentration of urea solution unchanged, adjust the dilution water flow rate to adjust the concentration of the diluted urea solution, calculate the denitrification efficiency based on the NOX concentration of the first monitoring system and the second monitoring system added this time, and find the concentration of the diluted urea solution corresponding to the highest denitrification efficiency. The concentration of the diluted urea solution is between 5% and 15%.

[0051] By adjusting the dilution water flow rate under certain stable load, operating oxygen content, coal quality, urea solution flow rate and other conditions, urea solutions of different concentrations are obtained, the changes in the NOx concentration value at the inlet of the air preheater 4 are observed, and the optimal urea solution concentration corresponding to the load is selected.

[0052] Field tests were conducted to determine the optimum urea solution concentration values ​​at 50%, 60%, 70%, 80%, 90% and 100% boiler loads.

[0053] When the load is stable, the coal quality is stable, and the operating oxygen content is stable, the basic temperature of the original NOX concentration produced can be considered. The optimal urea solution concentration value refers to the 50% concentration urea solution flow rate when the denitrification efficiency is the highest and the NOX concentration of the second monitoring system is the lowest. The relationship curve between the boiler load and the optimal urea solution concentration is fitted based on the optimal urea solution concentration value determined by the above 6 boiler loads and input into the denitrification DCS system.

[0054] Obtain the historical operation data of the boiler, determine the dependent variable y and independent variable x of the NOx generation prediction model, and obtain the training data set; Among them, boiler load, coal flow, primary air volume, secondary air volume, boiler operating oxygen content, furnace temperature and dry desulfurization limestone amount in the furnace are used as independent variables of the NOX generation prediction model; the NOX concentration at the separator inlet is used as the dependent variable of the NOX generation prediction model.

[0055] Among them: the independent variables need to be normalized before the NOX generation prediction model can be trained; The normalization formula is as follows:

[0056] Where: is the minimum value of the independent variable X, is the maximum value of the independent variable X, Model the data after normalizing the independent variables.

[0057] The gated recurrent neural network model is used to build the network structure of the NOx generation prediction model, and the NOx generation prediction regression model is obtained after training the network structure of the NOx generation prediction model with the training data set.

[0058] Preferably, the specific process of using a gated recurrent neural network to build a NOX generation prediction model is as follows: By setting the parameters of the activation function, learning rate, sample batch size, number of iterations and weight decay coefficient required to establish the NOX generation prediction model, a gated recurrent neural grid structure is used to build the NOX generation prediction model network structure; Among them: The expression of the activation function is , ; Preferably, the NOx generation prediction model is obtained by training the NOx generation prediction model network structure with the training data set. The NOx generation prediction model is expressed as:

[0059] Where: is the dependent variable of the NOX generation prediction model and X 1 It is the normalized independent variable of the NOX generation prediction model.

[0060] The boiler historical operation data include: boiler load, coal flow rate, primary air volume, secondary air volume, exhaust gas oxygen content, furnace temperature, dry desulfurization limestone volume in the furnace, and NOx concentration at the separator 2 inlet.

[0061] Obtain the historical operation data of the boiler, determine the dependent variables and independent variables of the NSR prediction model, and obtain the training data set; use the gated recurrent neural network to build the NSR prediction model network structure, and train the NSR prediction model network structure with the training data set to obtain the NSR prediction regression model.

[0062] Among them, boiler load, coal flow, primary air volume, secondary air volume, boiler operating oxygen content, furnace temperature, dry desulfurization limestone volume in the furnace, and separator inlet NOX concentration are used as independent variables of the NSR prediction model; the quality of urea solution of a certain concentration prepared by the power plant required for this project is used as the dependent variable of the NSR prediction model.

[0063] Among them: the independent variables need to be normalized before the NSR prediction model can be trained; The normalization formula is as follows: ; Where: is the minimum value of the independent variable X, is the maximum value of the independent variable X, Model the data after normalizing the independent variables.

[0064] The gated recurrent neural network model is used to build the NSR prediction model network structure, and the NSR prediction regression model is obtained by training the NSR prediction model network structure with a training data set.

[0065] Preferably, the specific process of using a gated recurrent neural network to build an NSR prediction model is as follows: By setting the parameters of the activation function, learning rate, sample batch size, number of iterations and weight decay coefficient required to establish the NSR prediction model, the gated recurrent neural grid structure is used to build the NSR prediction model network structure; Among them: The expression of the activation function is , ; Preferably, the NSR prediction model is obtained by training the NSR prediction model network structure through the training data set. The expression of the NSR prediction model is:

[0066] Where: is the dependent variable of the NSR prediction model and X 2 It is the normalized independent variable of the NSR prediction model.

[0067] The reaction loss function of the SNCR system is established through the independent variables and dependent variables of the virtual SCR reactor model of the SNCR system. The differential evolution algorithm is used to solve the optimal solution of the loss function, thereby establishing the SNCR denitrification reaction model.

[0068] The specific process of establishing the virtual SCR reactor model of the SNCR system is as follows: The loss function is established through the independent variables and dependent variables of the virtual SCR reactor denitrification reaction model, and the expression is:

[0069] in: is a random term that follows a standard normal distribution, For variables The reaction coefficient, For variables The reaction coefficient, The loss function is calculated as: ; Where: e is represented by boiler load s, separator inlet temperature T, ammonia nitrogen molar ratio The denitrification reaction efficiency determined by SR, , , are boiler load s, reactor inlet flue gas temperature T, ammonia nitrogen molar ratio Contribution of SR to denitrification reaction efficiency; The differential evolution algorithm is used to solve the parameters in the loss function. The parameters of the optimization results can be used to establish the denitrification reaction model of the virtual SCR reactor. The expression of the SCR denitrification reaction model is: ; Where: It is the dependent variable of the SCR denitrification reaction model.

[0070] Preferably, the specific process of predicting nitrogen oxides at the separator inlet by using the NOx generation prediction regression model is as follows: Obtain the data of the current operating status of the boiler, including boiler load, coal flow, primary air volume, secondary air volume, boiler operating oxygen content, furnace temperature, and dry desulfurization limestone volume in the furnace, and normalize them to form the prediction independent variable X 1 ; The NOX generation prediction regression model is used to predict the NOX concentration at the separator inlet, and its expression is:

[0071] Preferably, the specific process of the SCR denitration reaction model predicting the nitrogen oxides at the reactor outlet is: Based on the current boiler load s, the separator inlet flue gas temperature T, and the separator inlet NOX molar concentration , using the current urea capacity flow , calculate the molar concentration of NH3 , ammonia nitrogen molar ratio SR; then use the virtual SCR reactor to predict the NOX molar concentration at the reactor outlet , whose expression is:

[0072] Preferably, the specific process of controlling the ammonia injection flow rate of the reactor is: NOX control target through reactor outlet , control tolerance deviation , if the predicted value If the following conditions are met, the current urea capacity flow rate The recommended value is: ; If not satisfied, adjust and update the urea capacity flow rate with a fixed step size Repeat the process of using the virtual SCR reactor reaction model to predict the nitrogen oxides at the air preheater inlet and the NOX generation prediction regression model to predict the nitrogen oxides at the reactor inlet until the conditions are met.

[0073] According to the determined urea capacity flow The dilution water flow rate is controlled in combination with the optimal urea solution concentration after dilution selected above.

[0074] According to the current operating state of the virtual SCR reactor of the SNCR system, the NOx concentration at the inlet of the separator 2 is predicted by the NOx generation prediction regression model of S4.

[0075] The urea solution consumption of the SNCR system is calculated and controlled by using the NSR prediction regression model 4; the virtual SCR reactor model of the SNCR system is used to predict the NOx concentration at the inlet of the air preheater 4. The dilution water volume is controlled by the curve fitting of the optimal urea solution concentration corresponding to different loads determined by the above field test.

[0076] Unless otherwise specified, the device elements involved in the above embodiments are all conventional device elements, and the structural settings, working modes or control modes involved are all conventional settings, working modes or control modes in the art unless otherwise specified.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Other modifications or equivalent substitutions made to the technical solution of the present invention by ordinary technicians in the field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. A circulating fluidized bed boiler SNCR system denitrification monitoring system, characterized in that: include: A separator (2) and an air preheater (4), wherein the outlet of the separator (2) is communicated with the inlet of the air preheater (4), and an SNCR spray gun (21) is arranged at the inlet of the separator (2); a flue gas sampling device (3) is connected between the inlet of the separator (2) and the inlet of the air preheater (4); the flue gas sampling device (3) is a pipeline having a sampling grid at the inlet; the inlet end of the flue gas sampling device (3) is arranged upstream of the SNCR spray gun (21), a first monitoring system is arranged in the pipeline of the flue gas sampling device (3), and a second monitoring system is arranged at the inlet of the air preheater (4); the first monitoring system comprises a first oxygen monitoring device (32) and a first NOx monitoring device (33); the second monitoring system comprises a second oxygen monitoring device (44), a second NOx monitoring device (45) and an NH3 monitoring device (47).

2. The circulating fluidized bed boiler SNCR system denitration monitoring system according to claim 1 is characterized in that: The pipeline of the smoke sampling device (3) is connected to a compressed air pipeline (31).

3. The circulating fluidized bed boiler SNCR system denitration monitoring system according to claim 1, characterized in that: A superheater (41) is provided between the outlet of the separator (2) and the inlet of the air preheater (4).

4. The denitration monitoring system for the SNCR system of a circulating fluidized bed boiler according to claim 3 is characterized in that: A reheater (42) and an economizer (43) are provided between the superheater (41) and the inlet of the air preheater (4).

5. The denitration monitoring system for the SNCR system of a circulating fluidized bed boiler according to claim 1, characterized in that: The flue gas sampling device (3) is arranged within a range of 1.5 m to 2 m upstream of the SNCR spray gun (21).

6. The method for using the denitration monitoring system of the circulating fluidized bed boiler SNCR system according to any one of claims 1 to 5, characterized in that: The following steps are involved: S1: setting an initial NSR design value according to the inlet NOx concentration value monitored by the first monitoring system; S2: According to the NO monitored by the first monitoring system X Urea consumption is calculated by concentration and flue gas flow rate; S3: Calculate the dilution water consumption according to the urea consumption; S4: Conducting field tests to determine the optimal urea solution concentration value under different loads and obtaining a fitting curve of the optimal urea solution concentration value; S5: Obtain historical operation data, and train the historical operation data to obtain a NOx generation prediction regression model and an NSR prediction regression model.

7. The method for using the denitration monitoring system of the SNCR system of a circulating fluidized bed boiler according to claim 6 is characterized in that: A virtual SCR reactor model is established, a reaction loss function of the SNCR system is established, and the optimal solution is obtained through the differential evolution algorithm to obtain the SNCR denitrification reaction model; The NOx concentration at the inlet of the separator (2) is predicted by the NOx generation prediction regression model, and the urea solution consumption of the SNCR system is calculated and controlled by the NSR prediction regression model (4). The NOx concentration at the inlet of the air preheater (4) is predicted by the virtual SCR reactor model of the SNCR system, and the dilution water volume is controlled by the fitting curve of the optimal urea solution concentration value.

8. The method for using the denitration monitoring system of the SNCR system of a circulating fluidized bed boiler according to claim 6 is characterized in that: In step S2, the calculation formula for urea consumption is: ; Where: : Urea consumption, kg / h; : Flue gas flow rate at the separator (2) inlet at 6% O2 concentration, m3 / h; : Average NOX concentration at the inlet section of separator (2) at 6% O2 concentration, mg / m3; : NO2 molar mass, g / mol; NSR: NH3 / NOX molar ratio; : NH3 molar mass, g / mol.

9. The method for using the denitration monitoring system of the SNCR system of a circulating fluidized bed boiler according to claim 6 is characterized in that: In step S3, the calculation formula for dilution water consumption is: ; ; Where: : Urea solution quality; : Concentration of urea solution prepared by the power plant; : The optimal urea solution mass concentration of the SNCR system determined by the experiment; : The amount of dilution water required by the SNCR system to dilute the urea solution to the optimal concentration.

10. The method for using the denitration monitoring system of the SNCR system of a circulating fluidized bed boiler according to claim 6, characterized in that: In step S5, the specific process of using the gated recurrent neural network to build the NOX generation prediction model is as follows: by setting the parameters of the activation function, learning rate, sample batch size, number of iterations and weight decay coefficient required to establish the NOX generation prediction model, the gated recurrent neural grid structure is used to build the NOX generation prediction model network structure; wherein: the expression of the activation function is , ; After training the NOX generation prediction model network structure through the training data set, the NOX generation prediction model is obtained. The expression of the NOX generation prediction model is: ; Where: is the dependent variable of the NOx generation prediction model and X1 is the normalized independent variable of the NOx generation prediction model.

Images