Method for correcting denitration equilibrium control in combination with ammonia escape x Method for correcting denitration equilibrium control in combination with ammonia escape

Abstract

This invention proposes a method based on the total discharge outlet NO x This invention relates to a denitrification equilibrium control method that combines ammonia slip correction with other methods, and pertains to the field of flue gas denitrification control in thermal power plants. Specifically, it provides a method based on NO from the total exhaust outlet. x A denitrification equilibrium control method that combines concentration and ammonia slip correction, which involves adjusting the total NO emission at the discharge outlet. x The hourly average concentration of NO is introduced into closed-loop control to dynamically calculate the NO concentration for the current hour. x NO emission surplus x As a baseline setting value, and combined with the NO values ​​at both entrances x The concentration difference is corrected by a baseline setpoint based on weighted load allocation. At the same time, fuzzy inference is used to constrain and correct the ammonia escape concentration on both sides. Finally, the opening of the ammonia injection valves on both sides is precisely adjusted by a model predictive controller, thereby minimizing the consumption of reducing agent while ensuring that environmental emissions meet the standards.

Description

Technical Field

[0001] This invention relates to the field of flue gas denitrification control technology in thermal power plants, specifically to NO control based on total exhaust outlet. x A denitrification equilibrium control method that combines correction with ammonia slip. Background Technology

[0002] Selective catalytic reduction (SCR) denitrification technology is currently the most advanced technology for removing nitrogen oxides (NOx) from flue gas in coal-fired power plants. x The mainstream technology for emission control. In a typical dual-flue parallel SCR denitrification system, the flue gas is divided into flue A and flue B, which are denitrified through their respective SCR denitrification reactors. The denitrified flue gas is then combined and treated by a desulfurization tower (FGD) before being finally discharged into the atmosphere through the chimney.

[0003] In existing technologies, denitrification control systems generally adopt a method of independent control of flue gas ducts A and B, that is, each flue gas duct is equipped with an independent NO3- duct. x The concentration reference setpoint is used, and the opening of the ammonia injection valve is adjusted through its respective closed-loop controller. However, this control method has the following technical problems:

[0004] In existing technologies, dual-flue parallel SCR denitrification systems employ independent control of side A and side B, with each flue controlling only the NO at its own outlet. x The deviation of the concentration from the benchmark setpoint was adjusted in a closed loop, but the total NO emission point was not included. x Concentration, export NO x Hourly average values ​​and ammonia slip concentrations on both sides, along with other multi-source correlated signals, are collaboratively incorporated into the automatic generation and dynamic adjustment mechanism of the baseline setpoint. This results in NO concentrations on both sides... x The determination of the baseline setpoint lacks a global optimal basis and can only rely on operators to manually adjust it based on experience. This makes it difficult for the denitrification system to minimize the consumption of reducing agent while meeting environmental emission assessment requirements, meaning it cannot automatically balance and optimize the balance between emission compliance and economic operation. Summary of the Invention

[0005] This invention addresses the problems existing in the prior art by providing a method based on the total discharge port NO x The denitrification equilibrium control method, which combines correction with ammonia slip, achieves the dual goals of environmental compliance and material conservation.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: Based on total discharge outlet NO x A denitrification equilibrium control method combined with ammonia slip correction, applied to the distributed control system of a coal-fired power generation unit with a dual-flue parallel selective catalytic reduction denitrification reactor, is characterized by comprising: Monitoring NO at both the inlet and outlet of the dual flue. x Concentration data, based on NO at the export level x NO concentration data were used to calculate x The concentration benchmark setting is then adjusted and feedforward compensation is applied to obtain the feedforward compensated NO. x Concentration baseline setting; NO based on entry x Concentration data and a preset weighted load allocation algorithm for feedforward compensation of NO x The concentration baseline setting is corrected to obtain the corrected NO at both inlets. x Concentration baseline setting; Based on a pre-set fuzzy rule base, fuzzy inference is performed using ammonia slip concentrations under different ammonia injection rates as constraints to obtain correction values. These correction values ​​are then used to adjust NO levels. x The concentration reference setpoint is calibrated to obtain the calibrated NO corresponding to both inlets. x Concentration baseline setting; The corrected NO corresponding to both inlets x The concentration baseline setting is the target, with NO at both outlets as the standard. x The concentration data serves as feedback. Based on the closed-loop calculation of the multivariate dynamic prediction model for the pre-set denitrification process, the optimal control increment of the opening of the dual flue gas ammonia injection valve is calculated, and ammonia injection denitrification is carried out.

[0007] In some embodiments, based on the export NO x NO concentration data were used to calculate x The process of setting the concentration benchmark includes: Monitor NO at both outlets of the dual flue x Concentration data, and obtain NO concentration data as required by environmental assessment standards. x Calculate the total NO concentration for the current hour based on the hourly average concentration limit and the total time of the entire hourly period. x Emissions budget; Obtain the running time within the current hour and the hourly average value at the current moment, and calculate the consumed NO. x Emissions; Based on the total NO for the current hourly segment x Emissions budget and NO consumed x Emissions, calculated to obtain the remaining available NO x Emissions; Based on the remaining available NO x Emissions, total time for the entire hourly period, and the current running time within this hourly period are used to calculate NO for the remaining time in the future. x Concentration baseline setting.

[0008] In some embodiments, feedforward compensation is performed to obtain the feedforward compensated NO. x The process of setting the concentration benchmark includes: Collect the current boiler load from the distributed control system and calculate the load change rate; Calculate the load factor based on the boiler load and the preset rated load; Based on the load feedforward gain coefficient, reference load coefficient, and load change rate feedforward gain coefficient set in the distributed control system; The feedforward compensation is calculated based on the load factor, load change rate, load feedforward gain factor, reference load factor, and load change rate feedforward gain factor. Add the feedforward compensation to NO x The initial NO concentration baseline setting was obtained. x Concentration baseline setting; For preliminary NO x By applying upper and lower limits to the concentration benchmark setpoint, feedforward-compensated NO is obtained. x Concentration baseline setting.

[0009] In some embodiments, for the initial NO x By applying upper and lower limits to the concentration benchmark setpoint, feedforward-compensated NO is obtained. x The steps for setting the concentration baseline include: Obtain the lower limit and upper limit of the baseline setting value; In response to the fact that the running time within this hourly period is less than the preset duration at the current moment, a preliminary NO will be initiated. x The NO concentration baseline setting is taken as the default value as the feedforward compensation value. x Concentration baseline setting; In response to the fact that the running time within this hourly period at the current moment is greater than the preset duration, a preliminary comparison of NO is made. x The minimum result is obtained by comparing the concentration baseline setpoint and the upper limit of the baseline setpoint. This minimum result is then compared to the lower limit of the baseline setpoint, and the maximum result is selected as the NO value for feedforward compensation. x Concentration baseline setting.

[0010] In some embodiments, based on the NO entry x Concentration data and a preset weighted load allocation algorithm for feedforward compensation of NO x The concentration baseline setting is corrected to obtain the corrected NO at both inlets. x The process of setting the concentration benchmark includes: Obtain the NO at both inlets of the dual flue separately. x Concentration data, and calculate the corresponding inlet load weighting factors on both sides; Based on the load weighting factors at both inlets and the NO of the feedforward compensationx The target denitrification efficiency on both sides is calculated using the concentration baseline setting and the preset load distribution adjustment coefficient, respectively. According to the NO at both inlets of the dual flue x The concentration data and the corresponding target denitrification efficiency are used to calculate the corresponding correction baseline settings on both sides; Calculate the average of the correction reference settings on both sides, and then calculate the average of the correction reference settings on both sides and the NO of the feedforward compensation. x The difference between the concentration baseline setpoints is used to obtain the corresponding correction values ​​on both sides. The arithmetic mean of these correction values ​​is equal to the NO value of the feedforward compensation. x Concentration baseline setting; The corresponding correction values ​​on both sides are superimposed onto the corresponding correction reference settings to obtain the correction NO corresponding to the two inlets of the dual flue. x Concentration baseline setting.

[0011] In some embodiments, the ammonia slip concentration under different ammonia injection rates is used as a constraint for fuzzy inference to correct NO. x The concentration reference setpoint is calibrated to obtain the calibrated NO corresponding to both inlets. x The process of setting the concentration benchmark includes: Build a fuzzy rule base; Based on the fuzzy rule base, fuzzy inference is performed on the ammonia escape concentration under different ammonia injection rates to obtain several fuzzy subsets; All fuzzy subsets are merged into a comprehensive fuzzy set, and the fuzziness is defuzzified using the centroid method. The centroid positions are calculated separately and used as correction values ​​for the corresponding reference settings on both sides. The correction values ​​corresponding to the reference settings on both sides are added to the correction NO values ​​corresponding to the two inlets of the dual flue. x From the concentration baseline setting, a preliminary correction value is obtained; Applying safety constraints to the initial correction values, we obtain the corrected NO values ​​corresponding to the two inlets of the dual flue. x Concentration baseline setting.

[0012] In some embodiments, the process of constructing a fuzzy rule base includes: Define fuzzy input variables and fuzzy output variables. The fuzzy input variables include a combination of linguistic variables for ammonia escape concentration and a combination of linguistic variables for the rate of change of ammonia escape concentration. The fuzzy output variables include a combination of linguistic variables for the baseline setpoint correction. A fuzzy rule base is established based on the relationship between fuzzy input variables and fuzzy output variables; The fuzzy rule base includes: When the ammonia escape concentration is low and the rate of change is zero or negative, output a language variable that increases the baseline setpoint correction. When the ammonia escape concentration is high and the rate of change is positive, output a language variable that reduces the baseline setting value for correction.

[0013] In some embodiments, the construction process of the multivariate dynamic prediction model for the denitrification process is as follows: The variable data of the historical denitrification process of both sides of the flue are obtained with a preset sampling period. The variable data of the historical denitrification process is subjected to outlier detection, filtering, smoothing and normalization to obtain the variable data of the historical denitrification process after preprocessing. The variable data of the historical denitrification process after preprocessing is divided into training set and validation set. The training set is transformed into Hankel matrix form, and singular value decomposition and least squares estimation are performed using a subspace-based state-space system identification method to obtain the initial identification model. The model obtained from the initial identification is validated using a validation set. The goodness of fit between the model output and the actual output is calculated. When the goodness of fit is greater than a preset threshold, the corresponding identification model is used as a multivariate dynamic prediction model for the denitrification process. The multivariate dynamic prediction model for the denitrification process is updated every preset update cycle using the latest historical variable data of the denitrification process.

[0014] In some embodiments, the corrected NO corresponding to both inlets is used. x The concentration baseline setting is the target, with NO at both outlets as the standard. x Using concentration data as feedback, the process of calculating the optimal control increment of the dual-flue ammonia injection valve opening, based on a pre-set multivariate dynamic prediction model for the denitrification process in a closed-loop manner, includes: Step 1: The pre-defined multivariate dynamic prediction model for the denitrification process uses the corrected NO at the two inlets of the dual flue. x The concentration benchmark setpoint is the target, with NO at both outlets of the dual flue gas ducts. x Concentration data is used for feedback; Step 2: Construct the optimization objective function and set constraints including control quantity constraints, control increment constraints, and output constraints. Transform the objective function and constraints into a quadratic programming problem and solve it using the interior point method to obtain the optimal control increment sequence. Step 3: Based on the preset rolling optimization strategy, the optimal control increment sequence is fed back for correction to obtain the corrected optimal control increment sequence. Step four: Iterate through steps one through three to construct and optimize the objective function until the NO outlets on both sides of the dual flue are reached. x Once the concentrations are equal to the threshold values, the first increment of the corrected optimal control increment sequence is taken as the optimal control increment for the corresponding ammonia injection valve opening in the dual flue.

[0015] In some embodiments, the process of performing feedback correction on the optimal control increment sequence according to a preset rolling optimization strategy to obtain the corrected optimal control increment sequence includes: Within the current time k, obtain the actual opening degree of the ammonia injection valve at the previous time k-1, apply the first increment of the optimal control increment sequence to the ammonia injection valve, calculate the output value of the actual opening degree of the ammonia injection valve at the current time k, and obtain the predicted output sequence at the current time k. Entering the next time step k+1, collect NO data from both outlets of the dual flues. x Concentration, obtain the predicted value of the current time k from the previous time k-1, and calculate the respective prediction error vector sequence based on the predicted value; By setting error correction coefficients and using the predicted error vector sequence and error correction coefficients to correct the output sequence predicted at the current time k, the corrected optimal control increment sequence is obtained.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention proposes a method based on the total discharge outlet NO x A denitrification equilibrium control method combined with ammonia slip correction is applied to the distributed control system of a coal-fired power generation unit equipped with a parallel selective catalytic reduction denitrification reactor with two flues. Its key feature is that it includes: monitoring the NO levels at the inlet and outlet of both flues. x Concentration data, based on NO at the export level x NO concentration data were used to calculate x The concentration benchmark setting is then adjusted and feedforward compensation is applied to obtain the feedforward compensated NO. x Concentration baseline setpoint; NO at the inlet x Concentration data and a preset weighted load allocation algorithm for feedforward compensation of NO x The concentration baseline setting is corrected to obtain the corrected NO at both inlets. x Concentration baseline setting; based on a preset fuzzy rule base, the ammonia escape concentration under different ammonia injection rates is used as a constraint condition for fuzzy inference to obtain a correction value, and the NO is corrected according to the correction value. x The concentration reference setpoint is calibrated to obtain the calibrated NO corresponding to both inlets. x Concentration reference setpoint; Corrected NO levels corresponding to both inlets x The concentration baseline setting is the target, with NO at both outlets as the standard. x The concentration data serves as feedback. Based on the closed-loop calculation of the multivariate dynamic prediction model for the pre-set denitrification process, the optimal control increment of the opening of the dual flue gas ammonia injection valve is calculated, and ammonia injection denitrification is carried out.

[0017] This invention provides a method based on the total discharge port NO xA denitrification equilibrium control method that combines concentration and ammonia slip correction, which involves adjusting the total NO emission at the discharge outlet. x The hourly average concentration is used to introduce closed-loop control, dynamically calculating the NO concentration for the current hourly period. x Emission margin, combined with NO at both inlets x The concentration difference is corrected by a baseline setpoint based on weighted load allocation. At the same time, fuzzy inference is used to constrain and correct the ammonia escape concentration on both sides. Finally, the opening of the ammonia injection valves on both sides is precisely adjusted by a model predictive controller, thereby minimizing the consumption of reducing agent while ensuring that environmental emissions meet the standards.

[0018] This invention uses the total discharge port NO x The hourly average concentration of NO was incorporated into the closed-loop control, and NO was dynamically calculated. x The system generates emission margins and baseline settings, enabling an automated, on-line operation strategy that minimizes reductant consumption while ensuring compliance with environmental emission standards.

[0019] Using an inlet concentration correction algorithm based on weighted load allocation, the NO concentration at both inlets is adjusted according to... x The concentration difference adaptively distributes the denitrification load on each side, achieving balanced utilization of the catalysts on both sides and extending the catalyst's service life.

[0020] By employing an ammonia escape constraint correction mechanism based on fuzzy reasoning, ammonia escape concentration information is used as the basis for adjusting the baseline setpoint, quickly determining whether the current ammonia injection rate is consistent with the inlet NO. x Concentration matching and dynamic adjustment of ammonia injection volume to ensure ammonia and NO concentration are matched. x While ensuring a full reaction, excessive ammonia injection is avoided, reducing ammonia waste and effectively preventing the waste of reducing agent and the risk of air preheater blockage caused by excessive ammonia injection.

[0021] Replacing traditional PID (proportional-integral-derivative) control with model predictive control can address industrial process control challenges such as multivariable coupling, process constraints, and time delays, improving control accuracy and response speed while reducing reliance on operator experience. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort.

[0023] Figure 1 The present invention provides a method based on the total discharge port NO x Flowchart of the denitrification equilibrium control method combined with ammonia slip correction.

[0024] Figure 2 This invention provides a method based on the total discharge port NO x Control diagram of the DCS control system for the denitrification equilibrium control method combined with ammonia slip correction.

[0025] Figure 3 A schematic diagram of the structure of an embodiment of the computer device provided by the present invention.

[0026] Figure 4 This is a schematic diagram of an embodiment of the computer-readable storage medium provided by the present invention.

[0027] Figure 5 The present invention provides a method based on the total discharge port NO x A flowchart of one embodiment of a denitrification equilibrium control method that combines correction with ammonia slip.

[0028] Figure 6 The present invention provides a method based on the total discharge port NO x Calculation of NO in the denitrification equilibrium control method with ammonia slip joint correction x Flowchart of concentration baseline setting values.

[0029] Figure 7 The present invention provides a method based on the total discharge port NO x A flowchart of inlet concentration correction based on weighted load allocation in a denitrification equilibrium control method that combines correction with ammonia slip.

[0030] Figure 8 The present invention provides a method based on the total discharge port NO x Fuzzy rule library graph in the denitrification equilibrium control method with ammonia slip joint correction.

[0031] Figure 9 The present invention provides a method based on the total discharge port NO x A flowchart of ammonia slip constraint correction based on fuzzy reasoning in the denitrification equilibrium control method with ammonia slip joint correction.

[0032] Figure 10 The present invention provides a method based on the total discharge port NO x Model predictive control flowchart in the denitrification equilibrium control method with ammonia slip joint correction. Detailed Implementation

[0033] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention. It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application.

[0034] It should be noted that all uses of "first" and "second" in the embodiments of the present invention are for the purpose of distinguishing two entities with the same name but different names or different parameters. It is clear that "first" and "second" are only for the convenience of description and should not be construed as limiting the embodiments of the present invention. Subsequent embodiments will not explain this in detail.

[0035] The existing control methods have the following problems: First, the total NO at the chimney outlet... x Concentration (i.e., the total NO emission at the desulfurization tower) x The concentration of NO is only used as a reference for the protection threshold and is not directly included in the NO concentration on both sides. x In the closed-loop control logic for concentration, when NO at both outlets... x When the concentrations are all near their respective baseline settings, the NO at the total discharge outlet x The concentration may deviate from the optimal operating point, resulting in insufficient environmental margin or waste of large quantities of materials.

[0036] Second, export NO x The hourly average concentration is an important indicator for environmental protection authorities, but the existing control system does not include this hourly average in the NO concentration data on both sides. x In the closed-loop control, operators need to manually adjust the NO on both sides based on the hourly average trend and absolute value. x The baseline setting is such that this operation not only consumes the energy of the operators, but is also limited by the experience level of the operators, making it difficult to achieve optimal operation.

[0037] Third, under the premise of meeting environmental emission limits, the export of NO x The closer the hourly average concentration is to the environmental assessment threshold, the more beneficial it is to save on reducing agent consumption, i.e., achieving operation close to the threshold. However, existing control systems lack automated strategies for operating close to the threshold, making it impossible to minimize material consumption while ensuring emissions meet standards.

[0038] Fourth, the ammonia slip concentration in both flues is a crucial indicator of whether the current ammonia injection rate is reasonable. A higher ammonia slip concentration indicates that the current ammonia injection rate is too high, resulting in a waste of reducing agent. However, the existing control system does not use the ammonia slip concentration as a basis for adjusting the baseline setpoint, which makes it impossible to effectively avoid material waste and the risk of air preheater blockage caused by excessive ammonia injection.

[0039] In summary, existing technologies lack a method that can comprehensively utilize NO from the total discharge outlet. x hourly average concentration of NO at both inlets x The method of coordinating and balancing the concentration difference and ammonia escape concentration information on both sides of the flue gas denitrification system is used to achieve the dual goals of environmental compliance and material conservation.

[0040] This invention proposes a method based on the total discharge outlet NO xFor a denitrification equilibrium control method that combines correction with ammonia slip, please refer to [link / reference]. Figure 1 , Figure 2 and Figure 5 It is applied in the distributed control system of a coal-fired power generation unit with a dual-flue parallel selective catalytic reduction denitrification reactor, including: Monitoring NO at both the inlet and outlet of the dual flue. x Concentration data, based on NO at the export level x NO concentration data were used to calculate x The concentration benchmark setting is then adjusted and feedforward compensation is applied to obtain the feedforward compensated NO. x Concentration baseline setting; NO based on entry x Concentration data and a preset weighted load allocation algorithm for feedforward compensation of NO x The concentration baseline setting is corrected to obtain the corrected NO at both inlets. x Concentration baseline setting; Based on a pre-set fuzzy rule base, fuzzy inference is performed using ammonia slip concentrations under different ammonia injection rates as constraints to obtain correction values. These correction values ​​are then used to adjust NO levels. x The concentration reference setpoint is calibrated to obtain the calibrated NO corresponding to both inlets. x Concentration baseline setting; The corrected NO corresponding to both inlets x The concentration baseline setting is the target, with NO at both outlets as the standard. x The concentration data serves as feedback. Based on the closed-loop calculation of the multivariate dynamic prediction model for the pre-set denitrification process, the optimal control increment of the opening of the dual flue gas ammonia injection valve is calculated, and ammonia injection denitrification is carried out.

[0041] A method based on total discharge outlet NO x A denitrification equilibrium control method combined with ammonia slip correction is applied to the DCS control system of a coal-fired power generation unit equipped with a parallel SCR denitrification reactor in a dual-fluehouse configuration. The method includes the following steps: Step S1: Data acquisition, real-time acquisition of process variable data of both flues at a preset sampling period; Step S2: Data preprocessing, including outlier detection, filtering, smoothing, and normalization of the collected process variable data; Step S3: Process model identification. Based on the preprocessed historical operating data, a dynamic prediction model for the denitrification process is established. Step S4: NO x Emission margin calculation based on total NOx emissions. x The concentration of NO is dynamically calculated based on the hourly average concentration and environmental assessment limit for the remaining time of the current hour. x Concentration baseline setting; Step S5: Inlet concentration correction based on weighted load allocation, according to the NO at both inlets x Concentration differences are addressed by allocating the baseline setting value to the respective corrected baseline setting values ​​on both sides. Step S6: Ammonia escape constraint correction based on fuzzy inference, using the ammonia escape concentrations on both sides to further correct the correction benchmark setting; Step S7: Model predictive control to correct NO x The concentration baseline setting is the target, with NO at both outlets. x The actual concentration value is used as feedback. The model predictive controller calculates the optimal control value for the opening of the ammonia injection valves on both sides to achieve closed-loop balanced control.

[0042] This embodiment uses a parallel SCR denitrification system with dual flues in a coal-fired power generating unit as the application scenario. The environmental assessment indicator is NO2 at the total emission outlet. x The hourly average concentration does not exceed 50 mg / m³.

[0043] like Figure 2 As shown, the dual-flue parallel SCR denitrification system involved in this invention includes: a flue on side A and a flue on side B. Each flue is equipped with an SCR denitrification reactor, an ammonia injection grid, an ammonia injection regulating valve, and an inlet NO. x Concentration detection device, outlet NO x Concentration detection devices and ammonia slip concentration detection devices are included. The denitrified flue gas from both sides merges and enters the desulfurization tower (FGD), where it is treated for desulfurization before being discharged through the chimney. A total NO emission outlet is installed at the chimney outlet. x Continuous Concentration Monitoring System (CEMS).

[0044] EntranceNO x Concentration detection device and outlet NO x Concentration detection devices are installed on the inlet and outlet flues of each SCR denitrification reactor. Ammonia slip concentration detection devices, employing tunable diode laser absorption spectroscopy (TDLAS) technology, are installed on the outlet flue of each SCR denitrification reactor to measure the ammonia slip concentration in real time. Total discharge NO x The Continuous Concentration Monitoring System (CEMS) is installed on the flue between the desulfurization tower outlet and the chimney inlet.

[0045] like Figure 5 As shown, the control method of the present invention is implemented in a DCS (Distributed Control System), and its overall control flow includes steps S1 to S7.

[0046] In step S1, the DCS collects the following process variable data in real time with a sampling period of 1 second: (1) Entrance on side A NO x concentration The unit is mg / m³; (2) Entrance on side B NO x concentration The unit is mg / m³; (3) Exit on side A NO x concentration The unit is mg / m³; (4) Exit on side B NO x concentration The unit is mg / m³; (5) Ammonia escape concentration on side A The unit is ppm; (6) Ammonia escape concentration on side B The unit is ppm; (7) Main discharge outlet NO x concentration That is, NO at the chimney outlet after passing through the desulfurization tower x Instantaneous concentration, in mg / m³; (8) Main discharge outlet NO x hourly average concentration The value is calculated by the CEMS system according to environmental monitoring standards, and the unit is mg / m³. (9) The running time of the current moment within this hourly segment The unit is minutes.

[0047] The aforementioned process variable data are input through the DCS's analog input module (AI module), converted by an A / D converter, and stored in the DCS's real-time database for subsequent steps. Signal transmission from each detection device uses a 4-20mA standard current signal or an RS-485 digital communication interface.

[0048] In step S2, the following preprocessing operations are performed sequentially on the process variable data collected in step S1: Step S2.1: Outlier Detection and Removal. For each process variable, outlier detection is performed using the median absolute deviation (MAD) method within a sliding window. Specifically, let the current sampling time be k, and take the most recent N sampled values ​​to form a sliding window, where N ranges from 30 to 120 (corresponding to historical data from 30 to 120 seconds). Calculate the median M and median absolute deviation (MAD) of the data within the sliding window. If the current sampled value... satisfy If the sampled value is found to be an outlier, it is replaced with the median M of the sliding window.

[0049] Where 1.4826 is the conversion constant between MAD and standard deviation under the standard normal distribution, and 3 is the threshold multiple for outlier detection.

[0050] Step S2.2: Low-pass filtering and smoothing. The data after outlier processing is smoothed using a first-order exponentially weighted moving average (EWMA) filter. The recursive calculation formula is as follows: ; in, This represents the filtered output value at the k-th sampling time. This represents the actual sampled value at the k-th sampling time after outlier handling. This represents the filtered output value at the (k-1)th sampling time. This represents the filter coefficient, where α ranges from 0.05 to 0.3. The smaller the value, the stronger the filtering effect, but the greater the signal response lag.

[0051] Step S2.3: Data Normalization. Perform Min-Max normalization on the filtered process variable data, mapping them to the [0,1] interval. The normalization calculation formula is: ; in, This represents the normalized value. These represent the lower and upper limits of the engineering range for the process variable, respectively. The normalized data is used in the model identification process in step S3.

[0052] This denitrification equilibrium control method is applied to the DCS control system of a coal-fired power generation unit equipped with a parallel SCR denitrification reactor in a dual-flue system. In the monitoring phase, relevant sensors are used to continuously monitor the NO levels at both outlets of the dual flues. x Concentration data, combined with these data, to calculate NO x Concentration baseline setting. This takes into account the impact of boiler load variations on NO. x Emissions have an impact, and boiler load parameters are introduced through the DCS control system to control NO. x Feedforward compensation is applied to the concentration baseline setpoint to obtain a feedforward compensation NO that more closely reflects actual operating conditions. x Concentration baseline setting.

[0053] Obtain the NO at both inlets of the dual flue. x Concentration data, using a preset weighted load allocation algorithm, combined with factors such as unit operating status and load demand, are used to adjust the NO concentration after feedforward compensation. xThe concentration baseline setting is corrected to obtain the corrected NO for each of the two inlets of the dual flue. x The concentration baseline setting ensures a more reasonable load distribution for the denitrification reactions on both sides.

[0054] Fuzzy inference was employed to determine the ammonia slip concentration under different ammonia injection rates. Since the ammonia slip concentration is complexly related to the ammonia injection rate and the reaction conditions within the reactor, fuzzy inference can comprehensively consider multiple fuzzy factors, thus improving the correction of NO₂ levels. x The concentration benchmark setting is calibrated to obtain a more accurate NO correction at both inlets of the dual flue. x Concentration baseline setting.

[0055] Based on a pre-defined multivariate dynamic prediction model for the denitrification process, to correct NO x The concentration benchmark setpoint is the target, with NO at both outlets of the dual flue gas ducts. x With concentration as feedback, the optimal control increment of the ammonia injection valve opening corresponding to the dual flue is obtained through closed-loop calculation, thereby achieving precise ammonia injection denitrification.

[0056] For example, when the boiler load of a certain unit is 500MW, NO2 is detected at both outlets of the dual flue. x The concentrations were 80 mg / m³ and 85 mg / m³, respectively. Feedforward compensation of NO was obtained through calculation and feedforward compensation. x The concentration baseline setting is 82 mg / m³. After correction using a weighted load allocation algorithm, the NO concentration at both inlets is corrected. x The initial concentration benchmarks were 80 mg / m³ and 84 mg / m³, which were corrected using fuzzy inference to become 79 mg / m³ and 83 mg / m³, respectively. The optimal control increment for the ammonia injection valve opening was then calculated using a model, achieving precise denitrification. This method improves denitrification efficiency, reduces ammonia slip, minimizes environmental impact, and enhances the economic efficiency and stability of the unit operation.

[0057] In some embodiments, please refer to Figure 6 NO based on exports x NO concentration data were used to calculate x The process of setting the concentration benchmark includes: Monitor NO at both outlets of the dual flue x Concentration data, and obtain NO concentration data as required by environmental assessment standards. x Calculate the total NO concentration for the current hour based on the hourly average concentration limit and the total time of the entire hourly period. x Emissions budget; Obtain the running time within the current hour and the hourly average value at the current moment, and calculate the consumed NO. x Emissions; Based on the total NO for the current hourly segment xEmissions budget and NO consumed x Emissions, calculated to obtain the remaining available NO x Emissions; Based on the remaining available NO x Emissions, total time for the entire hourly period, and the current running time within this hourly period are used to calculate NO for the remaining time in the future. x Concentration baseline setting.

[0058] like Figure 6 As shown, in step S4, based on the total discharge port NO x The concentration of NO is dynamically calculated based on the hourly average concentration and environmental assessment limit for the remaining time of the current hour. x Concentration baseline setting. The specific calculation process is as follows: Step S4.1: Calculate the total NO for the current hourly segment. x Emissions budget. Set NOx emission standards to meet environmental assessment requirements. x The hourly average concentration limit is In this embodiment =50mg / m³, the total time for the entire hour period is =60 minutes, then the NO of the current hour segment x Total Emissions Budget The calculation formula is: ; Step S4.2: Calculate the NO consumed x Emissions. Let the running time within this hourly period be . (Unit: minutes), the current hourly average reported by the CEMS system is... (Unit: mg / m³), then the NO consumed x Emissions The calculation formula is: ; Step S4.3: Calculate the remaining available NO x Emissions. Remaining available NO x Emissions The calculation formula is: ; Step S4.4: Calculate the baseline setting value. The remaining available NO... x The emissions are evenly distributed over the remaining time period to obtain the NOx emissions that should be controlled for the remaining time period. x Concentration benchmark setting value The calculation formula is as follows: ; For example, when =50mg / m³, current hourly average =35mg / m³, running time =At 15 minutes, the baseline setting value =(50×60-35×15) / (60-15)=(3000-525) / 45=55mg / m³. This calculation result indicates that in the remaining 45 minutes of the current hour, the exported NO... x The concentration can be controlled at 55mg / m³, while still ensuring that the hourly average does not exceed the environmental protection requirement of 50mg / m³, thereby achieving line operation and saving the consumption of reducing agent.

[0059] Obtain NO from the environmental assessment documents for the area where the unit is located. x The hourly average concentration limit is set, and the total duration of the current hourly period is defined. Based on this, the total NO concentration for the current hourly period is calculated. x Emissions budget. Total NOx x The emissions budget is the allowable NO emissions from the unit for the current hour. x The total upper limit. Retrieve the duration already run within this hourly timeframe at the current moment, and the NO values ​​recorded from the start to the current moment. x Hourly average concentration. The amount of NO consumed at the current moment is calculated by multiplying the hourly average by the running time. x Emissions. This reflects the emission allowances already used by the unit in the current phase, expressed as the total NOx for the current hourly period. x Emissions budget minus NO consumed x The emissions will give the remaining usable NO. x Emissions. Remaining available emissions reflect the amount of NO that the unit can still emit during the remaining time of the current hour. x Total amount. Based on the remaining available NO x Based on emissions, the total time of the entire hourly period, and the running time already completed within this hourly period at the current moment, the NO levels for the remaining time in the future can be calculated. x Concentration baseline setting.

[0060] For example, if the total duration of a certain unit's current hourly segment is 1 hour, the total NO... x The emission budget is 100kg, and it has been running for 30 minutes. The current NO is... x The hourly average concentration is 50 mg / m³, and the consumed emissions are 25 kg. Therefore, the remaining available emissions are 75 kg. What is the NO emission level in the next 30 minutes? x The concentration baseline setpoint can be calculated to guide denitrification control. This method can accurately control NO. x Emissions will be reduced to meet environmental assessment requirements while optimizing unit operation and lowering denitrification costs.

[0061] In some embodiments, feedforward compensation is performed to obtain the feedforward compensated NO. x The process of setting the concentration benchmark includes: Collect the current boiler load from the distributed control system and calculate the load change rate; Calculate the load factor based on the boiler load and the preset rated load; Based on the load feedforward gain coefficient, reference load coefficient, and load change rate feedforward gain coefficient set in the distributed control system; The feedforward compensation is calculated based on the load factor, load change rate, load feedforward gain factor, reference load factor, and load change rate feedforward gain factor. Add the feedforward compensation to NO x The initial NO concentration baseline setting was obtained. x Concentration baseline setting; For preliminary NO x By applying upper and lower limits to the concentration benchmark setpoint, feedforward-compensated NO is obtained. x Concentration baseline setting.

[0062] Step S4.5: Feedforward compensation correction based on boiler load The baseline setpoints calculated in steps S4.1 to S4.4 are based solely on the emitted NO. x Rolling extrapolation from historical NO concentration data is a feedback-based margin calculation strategy that does not consider the impact of boiler operating conditions on future NO levels. x The impact of generation trends. In actual operation, changes in boiler load will significantly alter the inlet NO. x The generation characteristics of the system mean that without feedforward compensation, the calculation of the reference setpoint will lag behind changes in operating conditions, causing the control system to experience NO during load fluctuations. x The problem of excessive emissions or waste of reducing agents.

[0063] Specifically, during boiler operation, load changes affect NO x The mechanism of NO generation is as follows: When the boiler is operating at low load, the combustion intensity in the furnace weakens, the flame temperature decreases, and the combustion stability declines. To maintain combustion stability and safety, operators or the automatic combustion control system will increase the air supply to provide sufficient combustion support, leading to an increase in the oxygen concentration in the flue gas. In an oxygen-rich combustion environment, thermal NO... x and fuel-type NO x The formation rate of both increases: on the one hand, higher oxygen concentration promotes the oxidation reaction of N2 and O2 in high-temperature regions, increasing the production rate of thermally generated NO. x On the one hand, sufficient oxygen content allows for more complete oxidation of nitrogen-containing compounds in pulverized coal, increasing the production of fuel-type NO. xThe release rate. Therefore, the NO release at the SCR reactor inlet under low load conditions. x Concentrations are often relatively high. If the baseline setting is still extrapolated based on the linear trend of the current hourly average, it will underestimate future NO levels. x Emission intensity leads to higher baseline settings, which in turn creates the risk of exceeding hourly average limits.

[0064] Conversely, when a boiler gradually increases its load from low to high, the furnace temperature rises, combustion tends to stabilize, and operators will correspondingly reduce the excess air coefficient, resulting in a decrease in oxygen concentration and NO content in the flue gas. x The generation rate then decreases. If the baseline setting value is still the conservative value used in the low-load stage, it will result in the baseline setting value being too low, causing excessive ammonia injection and waste of reducing agent.

[0065] To address the aforementioned issues, this step introduces a feedforward compensation correction mechanism based on boiler load. The current boiler load signal is acquired in real-time from the DCS system. (Unit: MW), and calculate the load change rate. : ; in, For the calculation time window of the load change rate, it is recommended to take [time window value]. =60~120 seconds.

[0066] Define the feedforward compensation correction amount ΔC_ff, and calculate it as follows: First, determine the load factor based on the ratio of boiler load P(t) to rated load P_rated: ; Then, based on the load factor γ and the load change rate... Calculate the feedforward compensation. When the load factor is low (i.e., low load condition), the expected inlet NO is... x The concentration will increase, so the baseline setting should be lowered to allow for a larger denitrification margin; when the load factor is high (i.e., high load condition), NO x As generation stabilizes, the baseline setting can be appropriately relaxed. Feedforward compensation amount. The calculation formula is: ; in, This is the load feedforward gain coefficient, representing the degree of influence of unit load deviation on the correction amount of the reference setpoint. It is recommended to take [value missing]. =5~15mg / m³ (the specific value should be determined based on the characteristics of the unit during the commissioning phase); As a reference load factor, it is usually taken as =0.75 (corresponds to 75% of rated load, which is the typical dividing point for medium load conditions); This is the load change rate feedforward gain coefficient, representing the degree of influence of the load change rate on the correction amount of the reference setpoint. It is recommended to take [value missing]. =0.5~2.0(mg / m³) / (MW / min), its function is to provide an advanced correction signal when the load changes rapidly; The load change rate is calculated using formula (10).

[0067] The physical meaning of formula (12) is: when the actual load is lower than the reference load factor ( )hour, When the value is negative, the baseline setpoint is lowered, causing the denitrification system to increase the ammonia injection rate in advance to cope with the increased inlet NO under low load conditions. x Concentration; when the actual load is higher than the reference load factor ( )hour, If the value is positive, the baseline setting is increased to reduce unnecessary ammonia injection; load change rate item It provides the function of differential feedforward when the load drops rapidly ( When the load increases rapidly, the baseline setting value is further lowered. The baseline setting value is further increased to enable the compensation signal to predict in advance.

[0068] feedforward compensation The final reference setpoint is obtained by superimposing it onto the constraint output of step S4.4. That is, NO of feedforward compensation x Concentration baseline setting: ; For example: Let the rated load be... Current load Then the load factor =360 / 600=0.6. Take... =0.75, =10mg / m³, then the steady-state component of the load feedforward compensation is ×(0.6-0.75)=10×(-0.15)=-1.5mg / m³. Assume the current load change rate... =-2MW / min (load is being reduced). =1.0, then the rate of change component is 1.0 × (-2) = -2.0 mg / m³. Total feedforward compensation =-1.5+(-2.0)=-3.5mg / m³. If the result calculated in step S4.4 is... =47.5mg / m³, then =47.5 + (-3.5) = 44.0 mg / m³. This correction enables the denitrification system to pre-reduce NO under low-load and reduced-load conditions. x The baseline setting is the inlet NO that will be raised soon. x Sufficient control margin has been reserved for the concentration.

[0069] Boiler load via DCS control system for NO x Feedforward compensation is performed on the concentration baseline setpoint to obtain the feedforward compensated NO. x During the concentration benchmark setting process, the boiler load data at the current moment is accurately acquired from the DCS control system. This reflects the current operating power status of the unit. Simultaneously, the rate of change of the boiler load is calculated. By comparing the boiler load at the current moment with that at the previous moment, the load change per unit time is determined. Based on the preset rated load, the load factor γ is calculated. The load factor γ is the ratio of the current boiler load to the rated load, representing the degree of the current boiler load relative to the rated load. Then, the preset load feedforward gain coefficient, reference load coefficient, and load change rate feedforward gain coefficient are obtained from the DCS control system. These coefficients are empirical values ​​derived from extensive experiments and data analysis; they respectively reflect the influence of load, reference load coefficient, and load change rate on NO. x The extent of the impact generated.

[0070] Based on the previously calculated load factor, load change rate, and various gain coefficients, the feedforward compensation is calculated using a specific formula. This feedforward compensation comprehensively considers the current state and dynamic changes of the boiler load on NO. x The impact of generation.

[0071] The calculated feedforward compensation amount is added to NO. x At the concentration baseline setting, preliminary NO was obtained. x Concentration baseline setting. However, to ensure the rationality and safety of the baseline setting, it is also necessary to conduct preliminary NO concentration baseline setting. x Upper and lower limits are imposed on the concentration benchmark setting to prevent the benchmark setting from exceeding the actual operating capacity of the unit.

[0072] In some embodiments, for the initial NO x By applying upper and lower limits to the concentration benchmark setpoint, feedforward-compensated NO is obtained. x The steps for setting the concentration baseline include: Obtain the lower limit and upper limit of the baseline setting value; In response to the fact that the running time within this hourly period is less than the preset duration at the current moment, a preliminary NO will be initiated. x The NO concentration baseline setting is taken as the default value as the feedforward compensation value. x Concentration baseline setting; In response to the fact that the running time within this hourly period at the current moment is greater than the preset duration, a preliminary comparison of NO is made. x The minimum result is obtained by comparing the concentration baseline setpoint and the upper limit of the baseline setpoint. This minimum result is then compared to the lower limit of the baseline setpoint, and the maximum result is selected as the NO value for feedforward compensation. x Concentration baseline setting.

[0073] Step S4.6: Baseline Setpoint Constraint Processing. To ensure system operational safety, the calculated baseline setpoints are constrained. Apply upper and lower limit constraints: ; in, This represents the lower limit of the baseline setting (taken as 20 mg / m³), used to prevent excessive ammonia injection due to an excessively low baseline setting; the safety margin factor is set as... (Recommended) =0.95), then the upper limit of the benchmark setting value is Furthermore, when When the duration is less than the preset duration, due to the hourly average The statistical sample is insufficient, at this time Take the default value The value is 45 mg / m³.

[0074] In the preliminary NO x Apply upper and lower limit constraints to the concentration benchmark setpoint to obtain feedforward compensated NO x During the process of setting the concentration benchmark value, the lower and upper limits of the benchmark setpoint are obtained from the system's preset parameters. The running time within the current hour is determined. If the running time within the current hour is less than the preset duration, the monitoring data and calculation results may have significant errors because the unit's operating state is not yet stable. To ensure the NO feedforward compensation... x The accuracy and reliability of the concentration benchmark setting will initially affect NO x The NO concentration baseline setting is taken as the default value as the feedforward compensation value. x Concentration baseline setting. This default value is set based on the unit's historical operating data and experience, and can provide a relatively reasonable NO concentration during the initial startup phase of the unit. x Concentration control target.

[0075] If the current running time within this hourly period is greater than the preset duration, it indicates that the unit has entered a relatively stable operating state. At this point, the initial NO... x The concentration baseline setting and the upper limit of the baseline setting are compared, and the lower of the two is taken. This is to prevent initial NO... xThe concentration benchmark setpoint exceeds the upper limit that the unit's denitrification system can handle, to avoid incomplete denitrification due to an excessively high benchmark setpoint, thus preventing NO from reaching its maximum capacity. x Emissions exceeded the standard. Then, the minimum result obtained was compared with the lower limit of the baseline setpoint, and the maximum result was selected as the NO for feedforward compensation. x Concentration baseline setting. Ensure the baseline setting does not fall below the minimum NO concentration required for normal unit operation. x Control the level to ensure the denitrification system can operate stably.

[0076] For example, the lower limit of the reference setpoint is 30 mg / m³, and the upper limit of the reference setpoint is 80 mg / m³, for preliminary NO x The concentration baseline setting is 85 mg / m³. Since the running time has exceeded the preset duration by 5 minutes, the minimum value (80 mg / m³) between 80 mg / m³ and 85 mg / m³, and then the maximum value (80 mg / m³) between 80 mg / m³ and 30 mg / m³, also 80 mg / m³, is used as the NO feedforward compensation. x Concentration baseline setting. This method ensures NO concentration. x The concentration benchmark setting is within a reasonable range, which improves the stability and reliability of the denitrification system.

[0077] In some embodiments, please refer to Figure 7 NO based on the entry point x Concentration data and a preset weighted load allocation algorithm for feedforward compensation of NO x The concentration baseline setting is corrected to obtain the corrected NO at both inlets. x The process of setting the concentration benchmark includes: Obtain the NO at both inlets of the dual flue separately. x Concentration data, and calculate the corresponding inlet load weighting factors on both sides; Based on the load weighting factors at both inlets and the NO of the feedforward compensation x The target denitrification efficiency on both sides is calculated using the concentration baseline setting and the preset load distribution adjustment coefficient, respectively. According to the NO at both inlets of the dual flue x The concentration data and the corresponding target denitrification efficiency are used to calculate the corresponding correction baseline settings on both sides; Calculate the average of the correction reference settings on both sides, and then calculate the average of the correction reference settings on both sides and the NO of the feedforward compensation. x The difference between the concentration baseline setpoints is used to obtain the corresponding correction values ​​on both sides. The arithmetic mean of these correction values ​​is equal to the NO value of the feedforward compensation. x Concentration baseline setting; The corresponding correction values ​​on both sides are superimposed onto the corresponding correction reference settings to obtain the correction NO corresponding to the two inlets of the dual flue. xConcentration baseline setting.

[0078] In step S5, based on the NO at both inlets x The difference in concentration will be used to calculate the baseline setpoint in step S4. The settings are allocated to the corrected reference values ​​for both side A and side B. The principle of this allocation strategy is: Input NO... x The side with higher NO concentration bears a higher denitrification load (i.e., reduces the NO at the outlet of that side). x (Base setting value), Input NO x The baseline setting value is appropriately increased on the side with lower concentration to achieve balanced utilization of denitrification efficiency on both sides and avoid overuse of catalyst on one side.

[0079] In actual operation, the NO at the flue inlets on both sides A and B x The concentration often varies, mainly due to the combined effects of multiple factors: uneven powder distribution leads to differences in local combustion intensity and temperature, resulting in varying NO concentrations. x Production deviation; asymmetrical burner operation alters the local temperature field and excess air coefficient; uneven secondary air distribution causes differences in oxygen distribution, thus affecting NO. x Generation; asymmetry in the flue gas flow field and flue distribution within the furnace causes the amount of flue gas entering the SCRs on both sides to differ from the amount of NO. x Different total amounts; under the condition of blending multiple coal types, the differences in coal quality on each side are further amplified. x The resulting imbalance. The above factors are time-varying and coupled, causing the NO at both inlets to be uneven. x Concentration exhibits dynamic fluctuations, thus necessitating the establishment of a distribution and control mechanism capable of sensing differences in real time and adaptively adjusting.

[0080] Step S5.1: Calculate NO at both inlets x Normalized weighting factor for concentration. Define the inlet load weighting factor for side A. and the inlet load weighting factor on side B The calculation formula is as follows: ; in, and These represent the NO values ​​at inlets A and B after preprocessing in step S2, respectively. x Concentration value (actual engineering value after filtering but not normalization). When When the concentration approaches zero (i.e., the concentrations at both inlets are extremely low), =0.5.

[0081] Step S5.2: Calculate the target denitrification efficiency on both sides. Based on the inlet load weighting factor and the baseline setpoint, calculate the target denitrification efficiency on side A and side B respectively. and : ; in, This represents the load distribution adjustment coefficient, used to control the degree of differentiation between the baseline setpoints on both sides. The value range is from 0 to 1. When When =0, the baseline settings on both sides are not affected by the difference in inlet concentration; when When =1, the degree of differentiation is the greatest. The default value is... =0.5.

[0082] Step S5.3: Calculate the corrected baseline settings on both sides. Based on the target denitrification efficiency, calculate the corrected baseline settings on side A respectively. and B-side correction reference setting value : ; Step S5.4: Consistency Verification of Corrected Reference Settings. To ensure that the average of the corrected reference settings on both sides is consistent with the reference setting, a consistency verification and adjustment are performed: C_set,B = C_set,B + ​​ΔC; ; After this consistency check, the arithmetic mean of the corrected reference settings on both sides equals the reference setting value. This ensures that the total discharge outlet NO x The hourly average concentration does not deviate from the target value due to the allocation of the baseline setpoint.

[0083] NO at the two inlets of the dual flues was obtained respectively. x Concentration data, and feedforward compensation of NO based on a preset weighted load allocation algorithm. x The concentration baseline setting is corrected to obtain the corrected NO corresponding to the two inlets of the dual flue. x During the process of setting the concentration benchmark value, the NO at both inlets of the dual flue is obtained. x Concentration data reflect the NO concentration at both inlets. x The initial concentration conditions are determined. The inlet load weighting factors on both sides are calculated. These factors comprehensively consider the operating parameters such as flow rate, temperature, and pressure of the flues on both sides, as well as the current load distribution of the unit, reflecting the load proportion undertaken by the flues on both sides in the overall denitrification process.

[0084] Based on the calculated inlet load weighting factors on both sides and the NO of the feedforward compensation xThe target denitrification efficiency on both sides is calculated using the concentration baseline setpoint and the preset load distribution adjustment coefficient. The load distribution adjustment coefficient, derived through extensive experiments and simulation optimization, is used to adjust the denitrification efficiency of the flues on both sides to adapt to different operating conditions and denitrification requirements.

[0085] With NO at both entrances x Based on the concentration data and the corresponding target denitrification efficiency, the corrected baseline setpoints for both sides can be calculated separately. Combining the inlet concentration and target efficiency, the theoretically achievable NO₂ levels for both flues can be derived. x Concentration baseline setpoint. Calculate the average of the two corrected baseline setpoints and compare it with the NO concentration of the feedforward compensation. x The difference between the concentration baseline setpoints is used to obtain the corresponding correction values ​​on both sides. Here, it is required that the arithmetic mean of the corresponding correction values ​​on both sides equals the NO of the feedforward compensation. x The concentration baseline setpoint is to ensure the balance and accuracy of overall denitrification control. The corresponding correction values ​​on both sides are superimposed onto the corresponding correction baseline setpoints to obtain the corrected NO at both inlets of the dual flue. x Concentration baseline setpoint, i.e., the corrected NO at both inlets of the dual flue. x Concentration baseline setting.

[0086] In some embodiments, please refer to Figure 9 Using the ammonia slip concentration under different ammonia injection rates as constraints, fuzzy reasoning was performed to correct the NO... x The concentration reference setpoint is calibrated to obtain the calibrated NO corresponding to both inlets. x The process of setting the concentration benchmark includes: Build a fuzzy rule base; Based on the fuzzy rule base, fuzzy inference is performed on the ammonia escape concentration under different ammonia injection rates to obtain several fuzzy subsets; All fuzzy subsets are merged into a comprehensive fuzzy set, and the fuzziness is defuzzified using the centroid method. The centroid positions are calculated separately and used as correction values ​​for the corresponding reference settings on both sides. The correction values ​​corresponding to the reference settings on both sides are added to the correction NO values ​​corresponding to the two inlets of the dual flue. x From the concentration baseline setting, a preliminary correction value is obtained; Applying safety constraints to the initial correction values, we obtain the corrected NO values ​​corresponding to the two inlets of the dual flue. x Concentration baseline setting.

[0087] Step S6.3: Fuzzy Inference and Defuzzification. A Mamdani-type fuzzy inference engine is used to infer the fuzzy rule base from Step S6.2. The centroid method (COG) is employed for defuzzification to calculate the precise baseline setting correction δ_A. Similarly, the same fuzzy inference process is performed on side B to obtain the baseline setting correction δ_B for side B.

[0088] Step S6.4: Calculate the final correction NO x Concentration baseline setting value. The correction value obtained from fuzzy inference is superimposed on the corrected baseline setting value from step S5 to obtain the final corrected NO. x Concentration baseline setting: ; Step S6.5: Safety constraints on the final reference setting. For the final calibration NO... x Safety constraints are imposed on the concentration baseline setpoint to ensure it remains within a reasonable range: ; in, Indicates export NO x The absolute lower limit of the concentration benchmark setting (taken as 10 mg / m³). Indicates export NO x The absolute upper limit of the concentration benchmark setting (taken as 48 mg / m³).

[0089] Fuzzy inference was performed on the ammonia slip concentration under different ammonia injection rates to correct the NO at the two inlets of the dual flue. x The concentration baseline setting is used to obtain the corrected NO at both inlets of the dual flue. x In the process of setting the concentration benchmark, a fuzzy rule base was constructed based on a large amount of experimental data, practical operational experience, and expert knowledge. It covers the complex relationship between different ammonia injection rates and ammonia slip concentration, clarifying the possible range of ammonia slip concentration under various injection rates and the corresponding fuzzy descriptions, such as low, medium, and high.

[0090] like Figure 8 As shown, when the ammonia slip concentration is low (L): If the ammonia escape concentration change rate is negative, the output will be a slight increase in NO. x Reference setting value; If the ammonia escape concentration change rate is zero, the output will be a significant increase in NO. x Reference setting value; If the rate of change in ammonia escape concentration is positive, the output will be a slight increase in NO. x Baseline setting value.

[0091] When the ammonia escape concentration is medium (M): If the rate of change of ammonia escape concentration is negative, the output is the constant baseline setpoint correction value; If the rate of change of ammonia escape concentration is zero, the output is the constant baseline setpoint correction value; If the change rate of ammonia escape concentration is positive, the output will be a slight reduction in the baseline setting value as a correction amount.

[0092] When the ammonia escape concentration is high (H): If the rate of change of ammonia escape concentration is negative, the output will be a slight reduction in the baseline setpoint correction amount. If the ammonia escape concentration change rate is zero, the output will be a significant reduction in the baseline setpoint correction amount. If the change rate of ammonia escape concentration is positive, the output will be a significant reduction in the baseline setpoint correction amount.

[0093] Based on the constructed fuzzy rule base, fuzzy inference is performed on the ammonia escape concentration actually monitored under different ammonia injection rates. In this process, based on the specific numerical value of the ammonia escape concentration, it is categorized into the corresponding fuzzy sets in the fuzzy rule base, thus obtaining several fuzzy subsets. These fuzzy subsets reflect the fuzzy state of the ammonia escape concentration under different ammonia injection rates. All fuzzy subsets are merged into a comprehensive fuzzy set. To obtain a clear numerical value that can be used for practical correction from the fuzzy set, the centroid defuzzification method is used for defuzzification. The centroid defuzzification method is a commonly used defuzzification method to convert fuzzy output into a clear numerical value.

[0094] By calculating the centroid position of the comprehensive fuzzy set, the corresponding baseline setting correction values ​​on both sides are obtained. The centroid method can comprehensively consider the influence of all elements in the fuzzy set, making the obtained correction values ​​more reasonable and accurate.

[0095] The calculated correction values ​​for the baseline settings on both sides are added to the correction NO values ​​corresponding to the two inlets of the dual flue. x A preliminary correction value is obtained from the concentration benchmark setpoint. However, this preliminary correction value may exceed the safe operating range of the unit, therefore, safety constraints need to be imposed. These safety constraints are determined based on factors such as the unit's operating characteristics, environmental requirements, and equipment capacity, ensuring that the corrected NO... x The concentration benchmark setting meets environmental protection standards without damaging the unit equipment.

[0096] For example, assuming the left-side baseline setting correction is +3 mg / m³ and the right-side correction is -2 mg / m³ after fuzzy inference, the left-side NO correction... x The concentration baseline setting is 65 mg / m³, and the right side is 75 mg / m³. After superposition, the initial correction value is 68 mg / m³ on the left and 73 mg / m³ on the right. After safety constraints, the final corrected NO value is obtained.x Concentration baseline setpoint. This method can accurately correct the baseline setpoint, effectively control ammonia escape, and improve the safety and environmental friendliness of the denitrification system.

[0097] In some embodiments, please refer to Figure 8 The process of building a fuzzy rule base includes: Define fuzzy input variables and fuzzy output variables. The fuzzy input variables include a combination of linguistic variables for ammonia escape concentration and a combination of linguistic variables for the rate of change of ammonia escape concentration. The fuzzy output variables include a combination of linguistic variables for the baseline setpoint correction. A fuzzy rule base is established based on the relationship between fuzzy input variables and fuzzy output variables; The fuzzy rule base includes: When the ammonia escape concentration is low and the rate of change is zero or negative, output a language variable that increases the baseline setpoint correction. When the ammonia escape concentration is high and the rate of change is positive, output a language variable that reduces the baseline setting value for correction.

[0098] In step S6, the ammonia slip concentration information from both sides is used to further correct the modified baseline setpoint obtained in step S5. The ammonia slip concentration reflects whether the current ammonia injection rate is reasonable; an excessively high ammonia slip concentration indicates that the ammonia injection rate is too high, posing a risk of reducing agent waste and air preheater blockage. This step uses a fuzzy inference engine to map the ammonia slip concentration to a correction value for the baseline setpoint, achieving adaptive ammonia slip constraint control.

[0099] Step S6.1: Define fuzzy input and output variables. Taking side A as an example (the processing on side B is exactly the same), define the following fuzzy variables: Fuzzy input variable 1: Ammonia escape concentration (NH4+) on side A 3,A (Values ​​after preprocessing in step S2), with a universe of discourse of [0,10] ppm. Three linguistic variables and their corresponding membership functions are defined: Low (L), Medium (M), and High (H). Low uses a descending semi-trapezoidal membership function with parameters [0,0,1,3]; Medium uses a triangular membership function with parameters [1,3,5]; and High uses an ascending semi-trapezoidal membership function with parameters [3,5,10,10] (these parameters must match the actual ammonia escape range).

[0100] Fuzzy input variable 2: Change rate of ammonia escape concentration on side A ΔNH 3,A It is defined as the difference between the ammonia escape concentration at the current sampling time and the previous sampling time divided by the sampling period, i.e. The domain is [-2,2]ppm / s. Three linguistic variables are defined: negative (N), zero (Z), and positive (P), using descending semi-trapezoidal, triangular, and ascending semi-trapezoidal membership functions, respectively.

[0101] Fuzzy output variable: Baseline setting value correction coefficient δ A The domain of discourse is [-10, 10] mg / m³. Define five linguistic variables: significantly increased (PB), slightly increased (PS), unchanged (ZE), slightly decreased (NS), and significantly decreased (NB).

[0102] Step S6.2: Establish the fuzzy rule base. The fuzzy rule base is used to draw FIG.9 fuzzy rule tables, including, for example... Figure 8 The rules.

[0103] The above fuzzy rule means that when the ammonia slip concentration is low and the rate of change is zero or negative, it indicates that the current ammonia injection rate is too low, and the NO injection rate can be appropriately increased. x The baseline setting should be used to reduce the denitrification load and conserve reducing agent. When the ammonia slip concentration is high and the rate of change is positive, it indicates that the ammonia injection rate is significantly too high, and the NO injection rate should be drastically reduced. x The baseline setting value is used to increase the denitrification load in order to reduce ammonia slip.

[0104] In some embodiments, please refer to Figure 10 The construction process of the multivariate dynamic prediction model for the denitrification process is as follows: The variable data of the historical denitrification process of both sides of the flue are obtained with a preset sampling period. The variable data of the historical denitrification process is subjected to outlier detection, filtering, smoothing and normalization to obtain the variable data of the historical denitrification process after preprocessing. The variable data of the historical denitrification process after preprocessing is divided into training set and validation set. The training set is transformed into Hankel matrix form, and singular value decomposition and least squares estimation are performed using a subspace-based state-space system identification method to obtain the initial identification model. The model obtained from the initial identification is validated using a validation set. The goodness of fit between the model output and the actual output is calculated. When the goodness of fit is greater than a preset threshold, the corresponding identification model is used as a multivariate dynamic prediction model for the denitrification process. The multivariate dynamic prediction model for the denitrification process is updated every preset update cycle using the latest historical variable data of the denitrification process.

[0105] In step S3, based on the historical operating data preprocessed in step S2, a multivariate dynamic prediction model for the denitrification process is established using the subspace identification method. This model is used in the model predictive controller in step S7.

[0106] Step S3.1: Model Structure Definition. The denitrification process is modeled as a multiple-input multiple-output (MIMO) discrete state-space model, whose state-space expression is: ; in, This represents the state vector at the k-th sampling time, with a dimension of n×1; This represents the control input vector at the k-th sampling time. ,in These represent the opening degrees of the ammonia injection valves on side A and side B, respectively. This represents the output vector at the k-th sampling time. , representing the NO values ​​at outlets A and B respectively. x Concentration; A, B, C, and D represent the system matrix, input matrix, output matrix, and direct transmission matrix, respectively.

[0107] Step S3.2: Subspace Identification. Collect a sufficiently long period of historical operating data under steady-state and dynamic disturbance conditions, with a data length of no less than 2000 sampling points. Organize the input and output data into Hankel matrix form. Using the N4SID (subspace-based state-space system identification method) algorithm, through singular value decomposition (SVD) and least squares estimation of the Hankel matrix, identify the system matrices A, B, C, and D of the state-space model and the state-space dimension n. The state-space dimension n ranges from 2 to 8.

[0108] Step S3.3: Model Validation. The identified model is validated using an independent validation dataset. The goodness-of-fit (FIT) between the model output and the actual output is calculated. The formula for FIT is: ; in, This represents the actual output data sequence. This represents the model's predicted output data sequence. represents the mean of the actual output data, and ||·|| represents the Euclidean norm of the vector. When the FIT value is greater than 70%, the model accuracy is considered to meet the control requirements. If the FIT value does not meet the requirements, return to step S3.2, adjust the data window length or the state space dimension n, and then re-identify.

[0109] Step S3.4: Online Model Update. During the operation of the control system, every preset model update cycle T_update (ranging from 24 to 72 hours), steps S3.2 and S3.3 are re-executed using the latest historical operating data to update the process model and ensure that the model can adapt to the effects of changes in operating conditions and catalyst aging.

[0110] Transforming the training set into Hankel matrix form allows for the capture of the system's dynamic characteristics. The N4SID algorithm is then used to perform singular value decomposition and least squares estimation on the Hankel matrix. Singular value decomposition extracts the main features from the data, while least squares estimation optimizes the model parameters, resulting in an initial identification model.

[0111] For example, with a preset sampling period of 1 minute, 1000 sets of historical data were acquired. After processing, the data were divided into a training set of 800 sets and a validation set of 200 sets. The initial model fit was 0.8, and the preset threshold was 0.75, indicating that the model was qualified. The model constructed by this method can accurately predict the denitrification process, providing a basis for optimized control and improving denitrification efficiency and stability.

[0112] In some embodiments, please refer to Figure 10 The corrected NO corresponding to both inlets x The concentration baseline setting is the target, with NO at both outlets as the standard. x Using concentration data as feedback, the process of calculating the optimal control increment of the dual-flue ammonia injection valve opening, based on a pre-set multivariate dynamic prediction model for the denitrification process in a closed-loop manner, includes: Step 1: The pre-defined multivariate dynamic prediction model for the denitrification process uses the corrected NO at the two inlets of the dual flue. x The concentration benchmark setpoint is the target, with NO at both outlets of the dual flue gas ducts. x Concentration data is used for feedback; Step 2: Construct the optimization objective function and set constraints including control quantity constraints, control increment constraints, and output constraints. Transform the objective function and constraints into a quadratic programming problem and solve it using the interior point method to obtain the optimal control increment sequence. Step 3: Based on the preset rolling optimization strategy, the optimal control increment sequence is fed back for correction to obtain the corrected optimal control increment sequence. Step four: Iterate through steps one through three to construct and optimize the objective function until the NO outlets on both sides of the dual flue are reached. x Once the concentrations are equal to the threshold values, the first increment of the corrected optimal control increment sequence is taken as the optimal control increment for the corresponding ammonia injection valve opening in the dual flue.

[0113] like Figure 10 As shown, in step S7, the final corrected NO obtained in step S6 is used. x Concentration benchmark setting value The target value is the NO at both ends of the outlet. x The actual measured concentration is the feedback value. The optimal control increment of the ammonia injection valve opening on both sides is calculated by the model predictive controller (MPC) to achieve closed-loop balanced control.

[0114] Step S7.1: Define the prediction model. The state-space model identified in step S3 is used as the prediction model for MPC. At the start time k of each control cycle, based on the current state estimate... Given (k) and the future control input sequence, predict the output sequence at P sampling times in the future. Here, P represents the prediction time domain, and M represents the control time domain, satisfying M ≤ P. The value of P ranges from 10 to 30, and the value of M ranges from 3 to 10.

[0115] Step S7.2: Construct the optimization objective function. The MPC controller solves the following quadratic programming (QP) optimization problem in each control cycle: ; in, This indicates the prediction at time k. The output vector at time (i.e., the output NO at both sides A and B) x (Predicted concentration) express The reference setpoint vector at time, ; This represents the control increment vector at step j. , i.e., the increment of the ammonia injection valve opening on side A and side B; Q represents the output error weight matrix, which is a positive semi-definite diagonal matrix; R represents the control increment weight matrix, which is a positive definite diagonal matrix. The values ​​of the diagonal elements of Q and R directly affect the trade-off between tracking speed and stability in control performance. The diagonal elements of Q range from 1 to 10, and the diagonal elements of R range from 0.1 to 5.

[0116] Step S7.3: Set constraints. The optimization problem must satisfy the following constraints: (1) Control quantity constraints: ,in and These represent the lower and upper limits of the ammonia injection valve opening (ranging from 0% to 100%), respectively. (2) Control increment constraints: ,in These represent the lower and upper limits of the change in the opening degree of the ammonia injection valve within each control cycle, respectively, with values ​​ranging from ±2% to ±5%, to prevent the valve from operating too violently. (3) Output constraints: To ensure the predicted export NO x The concentration is within the safe range.

[0117] Step S7.4: Solve the optimization problem. In each control cycle, the objective function of step S7.2 and the constraints of step S7.3 are transformed into a standard quadratic programming problem, which is solved using the Active Set Method or the Interior Point Method to obtain the optimal control increment sequence. .

[0118] Step S7.5: Apply the rolling optimization strategy. Following the rolling optimization principle of model predictive control, only the first element of the optimal control increment sequence is selected. Apply it to the controlled object, that is: ; in, , These represent the actual opening output values ​​of the ammonia injection valves on sides A and B, respectively. These opening output values ​​are transmitted to the actuators of the ammonia injection regulating valves on each side via the DCS's analog output module (AO module) as a 4-20mA standard current signal.

[0119] Step S7.6: Feedback Correction. In the next control cycle, by collecting the new output NO... x The actual concentration measurement value is used to correct the model prediction output, eliminating biases caused by model mismatch and unknown disturbances. ; in, This represents the prediction error vector at time k. This represents the actual output measurement value at time k. Indicates in The predicted value at time k. This represents the error correction coefficient for step i. The value of is in the range of 0 to 1 and decreases as i increases (using ). In the form of, The value range is from 0.6 to 0.9.

[0120] Step S7.7: Execute cyclically. Return to step S1 to begin data acquisition for the next control cycle, and repeat steps S1 to S7 to form a continuous closed-loop balanced control.

[0121] For example, assuming NO correction x The concentration baseline setpoints are 65 mg / m³ on the left and 75 mg / m³ on the right. The optimal control increment sequence was calculated. After feedback correction and cyclic optimization, the first increment is 2% on the left and 1% on the right. This increment precisely controls the ammonia injection valve opening. This method can accurately control the ammonia injection rate, improve denitrification efficiency, and reduce the risk of ammonia escape.

[0122] In some embodiments, please refer to Figure 10 The process of obtaining the corrected optimal control increment sequence by performing feedback correction on the optimal control increment sequence according to the preset rolling optimization strategy includes: Within the current time k, obtain the actual opening degree of the ammonia injection valve at the previous time k-1, apply the first increment of the optimal control increment sequence to the ammonia injection valve, calculate the output value of the actual opening degree of the ammonia injection valve at the current time k, and obtain the predicted output sequence at the current time k. Entering the next time step k+1, collect NO data from both outlets of the dual flues. x Concentration, obtain the predicted value of the current time k from the previous time k-1, and calculate the respective prediction error vector sequence based on the predicted value; By setting error correction coefficients and using the predicted error vector sequence and error correction coefficients to correct the output sequence predicted at the current time k, the corrected optimal control increment sequence is obtained.

[0123] During the feedback correction of the optimal control increment sequence according to the preset rolling optimization strategy to obtain the corrected optimal control increment sequence, at the current time k, the system acquires the actual opening degree of the ammonia injection valve at the previous time k-1, which is a true reflection of the system's previous operating state. Next, the first increment of the optimal control increment sequence is applied to the ammonia injection valve, and based on the relationship between the valve opening and parameters such as the ammonia injection rate, the actual opening degree output value of the ammonia injection valve at the current time k is calculated. Simultaneously, using the preset multivariate dynamic prediction model for the denitrification process, combined with the currently known variable information, the predicted output sequence for the current time k is obtained. This sequence contains predicted values ​​for parameters related to the denitrification effect over a future period.

[0124] When entering the next time step k+1, the system will collect NO from both outlets of the dual flues. x Actual concentration values. These actual values ​​reflect the true results of the denitrification process at the current moment. Simultaneously, the predicted value of the current moment k at the previous moment k-1 is obtained. By comparing the actual values ​​with the predicted values, their respective prediction error vector sequences are calculated. The prediction error vector sequence reflects the degree of deviation between the model prediction and the actual situation.

[0125] An error correction coefficient is set, which is determined comprehensively based on factors such as system characteristics, historical data, and actual operating experience. This coefficient determines the strength of the correction for the prediction error. The output sequence predicted at the current time k is fed back and corrected using the prediction error vector sequence and the error correction coefficient. This adjusts the predicted value to be closer to the actual situation, thereby obtaining the corrected optimal control increment sequence.

[0126] For example, at time k-1, the actual opening degree of the ammonia injection valve is 30%, the first increment of the optimal control increment sequence is 5%, and after application, the NO in the predicted output sequence at time k is... x The concentration was 50 mg / m³. Actual NO was collected at time k+1. xWith a concentration of 55 mg / m³ and a prediction error of 5 mg / m³, if the error correction coefficient is 0.8, the optimal control increment sequence will be adjusted accordingly after correction, making the control more precise. This method can adjust the control strategy in real time according to the actual situation, improving the accuracy and stability of the denitrification system control.

[0127] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 3 As shown, an embodiment of the present invention also provides a computer device 30, which includes a processor 310 and a memory 320. The memory 320 stores a computer program 321 that can be run on the processor. When the processor 310 executes the program, it performs the steps of the method described above.

[0128] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 4 As shown, embodiments of the present invention also provide a computer-readable storage medium 40, which stores a computer program 410 that, when executed by a processor, performs the methods described above.

[0129] Embodiments of the present invention may also include a corresponding computer device. The computer device includes a memory, at least one processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes any of the methods described above when executing the program.

[0130] The memory, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as program instructions / modules in the embodiments of this application. The processor executes various functional applications and data processing of the device by running the non-volatile software programs, instructions, and modules stored in the memory, thereby implementing the above-described method.

[0131] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the device. Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the local module via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0132] Finally, it should be noted that those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The storage medium for the program can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. The above computer program embodiments can achieve the same or similar effects as any of the corresponding foregoing method embodiments.

[0133] Those skilled in the art will also understand that the various exemplary logic blocks, modules, circuits, and algorithm steps described in conjunction with the disclosure herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the functionality of various illustrative components, blocks, modules, circuits, and steps has been generally described. Whether this functionality is implemented as software or as hardware depends on the specific application and the design constraints imposed on the system as a whole. Those skilled in the art can implement the functionality in various ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the embodiments disclosed herein.

[0134] The above are exemplary embodiments disclosed in this invention. However, it should be noted that various changes and modifications can be made without departing from the scope of the embodiments of this invention as defined by the claims. The functions, steps, and / or actions of the methods according to the disclosed embodiments described herein do not need to be performed in any particular order. The sequence numbers of the disclosed embodiments of this invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. Furthermore, although the elements disclosed in the embodiments of this invention may be described or claimed individually, they may be understood as multiple unless explicitly limited to a singular number.

[0135] It should be understood that, as used herein, the singular form "one" is intended to include the plural form as well, unless the context clearly supports an exception. It should also be understood that, as used herein, "and / or" refers to any and all possible combinations of one or more of the associated listed items.

[0136] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples. Within the framework of the invention, technical features of the above embodiments or different embodiments can be combined, and many other variations of different aspects of the invention exist, which are not provided in the details for the sake of brevity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the protection scope of the invention.

Claims

1. Based on the total discharge outlet NO x A denitrification equilibrium control method combined with ammonia slip correction is applied to the distributed control system of a coal-fired power generation unit equipped with a parallel selective catalytic reduction denitrification reactor with dual flues. Its characteristics are... include: Monitoring NO at both the inlet and outlet of the dual flue. x Concentration data, based on NO at the export level x NO concentration data were used to calculate x The concentration benchmark setting is then adjusted and feedforward compensation is applied to obtain the feedforward compensated NO. x Concentration baseline setting; NO based on entry x Concentration data and a preset weighted load allocation algorithm for feedforward compensation of NO x The concentration baseline setting is corrected to obtain the corrected NO at both inlets. x Concentration baseline setting; Based on a pre-set fuzzy rule base, fuzzy inference is performed using ammonia slip concentrations under different ammonia injection rates as constraints to obtain correction values. These correction values ​​are then used to adjust NO levels. x The concentration reference setpoint is calibrated to obtain the calibrated NO corresponding to both inlets. x Concentration baseline setting; The corrected NO corresponding to both inlets x The concentration baseline setting is the target, with NO at both outlets as the standard. x The concentration data serves as feedback. Based on the closed-loop calculation of the multivariate dynamic prediction model for the pre-set denitrification process, the optimal control increment of the opening of the dual flue gas ammonia injection valve is calculated, and ammonia injection denitrification is carried out.

2. The method based on total discharge outlet NO according to claim 1 x The denitrification equilibrium control method combined with ammonia slip correction is characterized by, NO based on exports x NO concentration data were used to calculate x The process of setting the concentration benchmark includes: Monitor NO at both outlets of the dual flue x Concentration data, and obtain NO concentration data as required by environmental assessment standards. x Calculate the total NO concentration for the current hour based on the hourly average concentration limit and the total time of the entire hourly period. x Emissions budget; Obtain the running time within the current hour and the hourly average value at the current moment, and calculate the consumed NO. x Emissions; Based on the total NO for the current hourly segment x Emissions budget and NO consumed x Emissions, calculated to obtain the remaining available NO x Emissions; Based on the remaining available NO x Emissions, total time for the entire hourly period, and the current running time within this hourly period are used to calculate NO for the remaining time in the future. x Concentration baseline setting.

3. The method based on total discharge outlet NO according to claim 2 x The denitrification equilibrium control method combined with ammonia slip correction is characterized by, And perform feedforward compensation to obtain the NO of the feedforward compensation. x The process of setting the concentration benchmark includes: Collect the current boiler load from the distributed control system and calculate the load change rate; Calculate the load factor based on the boiler load and the preset rated load; Based on the load feedforward gain coefficient, reference load coefficient, and load change rate feedforward gain coefficient set in the distributed control system; The feedforward compensation is calculated based on the load factor, load change rate, load feedforward gain factor, reference load factor, and load change rate feedforward gain factor. Add the feedforward compensation to NO x The initial NO concentration baseline setting was obtained. x Concentration baseline setting; For preliminary NO x By applying upper and lower limits to the concentration benchmark setpoint, feedforward-compensated NO is obtained. x Concentration baseline setting.

4. The method based on total discharge outlet NO according to claim 3 x The denitrification equilibrium control method combined with ammonia slip correction is characterized by, For preliminary NO x By applying upper and lower limits to the concentration benchmark setpoint, feedforward-compensated NO is obtained. x The steps for setting the concentration baseline include: Obtain the lower limit and upper limit of the baseline setting value; In response to the fact that the running time within this hourly period is less than the preset duration at the current moment, a preliminary NO will be initiated. x The NO concentration baseline setting is taken as the default value as the feedforward compensation value. x Concentration baseline setting; In response to the fact that the running time within this hourly period at the current moment is greater than the preset duration, a preliminary comparison of NO is made. x The minimum result is obtained by comparing the concentration baseline setpoint and the upper limit of the baseline setpoint. This minimum result is then compared to the lower limit of the baseline setpoint, and the maximum result is selected as the NO value for feedforward compensation. x Concentration baseline setting.

5. The method based on total discharge outlet NO according to claim 1 x The denitrification equilibrium control method combined with ammonia slip correction is characterized by, NO based on entry x Concentration data and a preset weighted load allocation algorithm for feedforward compensation of NO x The concentration baseline setting is corrected to obtain the corrected NO at both inlets. x The process of setting the concentration benchmark includes: Obtain the NO at both inlets of the dual flue separately. x Concentration data, and calculate the corresponding inlet load weighting factors on both sides; Based on the load weighting factors at both inlets and the NO of the feedforward compensation x The target denitrification efficiency on both sides is calculated using the concentration baseline setting and the preset load distribution adjustment coefficient, respectively. According to the NO at both inlets of the dual flue x The concentration data and the corresponding target denitrification efficiency are used to calculate the corresponding correction baseline settings on both sides; Calculate the average of the correction reference settings on both sides, and then calculate the average of the correction reference settings on both sides and the NO of the feedforward compensation. x The difference between the concentration baseline setpoints is used to obtain the corresponding correction values ​​on both sides. The arithmetic mean of these correction values ​​is equal to the NO value of the feedforward compensation. x Concentration baseline setting; The corresponding correction values ​​on both sides are superimposed onto the corresponding correction reference settings to obtain the correction NO corresponding to the two inlets of the dual flue. x Concentration baseline setting.

6. The method based on total discharge outlet NO according to claim 5 x The denitrification equilibrium control method combined with ammonia slip correction is characterized by, Using ammonia slip concentrations under different ammonia injection rates as constraints, fuzzy reasoning was performed to correct NO. x The concentration reference setpoint is calibrated to obtain the calibrated NO corresponding to both inlets. x The process of setting the concentration benchmark includes: Build a fuzzy rule base; Based on the fuzzy rule base, fuzzy inference is performed on the ammonia escape concentration under different ammonia injection rates to obtain several fuzzy subsets; All fuzzy subsets are merged into a comprehensive fuzzy set, and the fuzziness is defuzzified using the centroid method. The centroid positions are calculated separately and used as correction values ​​for the corresponding reference settings on both sides. The correction values ​​corresponding to the reference settings on both sides are added to the correction NO values ​​corresponding to the two inlets of the dual flue. x From the concentration baseline setting, a preliminary correction value is obtained; Applying safety constraints to the initial correction values, we obtain the corrected NO values ​​corresponding to the two inlets of the dual flue. x Concentration baseline setting.

7. The method based on total discharge outlet NO according to claim 6 x The denitrification equilibrium control method combined with ammonia slip correction is characterized by, The process of building a fuzzy rule base includes: Define fuzzy input variables and fuzzy output variables. The fuzzy input variables include a combination of linguistic variables for ammonia escape concentration and a combination of linguistic variables for the rate of change of ammonia escape concentration. The fuzzy output variables include a combination of linguistic variables for the baseline setpoint correction. A fuzzy rule base is established based on the relationship between fuzzy input variables and fuzzy output variables; The fuzzy rule base includes: When the ammonia escape concentration is low and the rate of change is zero or negative, output a language variable that increases the baseline setpoint correction. When the ammonia escape concentration is high and the rate of change is positive, output a language variable that reduces the baseline setting value for correction.

8. The method based on total discharge outlet NO according to claim 1 x The denitrification equilibrium control method combined with ammonia slip correction is characterized by, The construction process of the multivariate dynamic prediction model for the denitrification process is as follows: The variable data of the historical denitrification process of both sides of the flue are obtained with a preset sampling period. The variable data of the historical denitrification process is subjected to outlier detection, filtering, smoothing and normalization to obtain the variable data of the historical denitrification process after preprocessing. The variable data of the historical denitrification process after preprocessing is divided into training set and validation set. The training set is transformed into Hankel matrix form, and singular value decomposition and least squares estimation are performed using a subspace-based state-space system identification method to obtain the initial identification model. The model obtained from the initial identification is validated using a validation set. The goodness of fit between the model output and the actual output is calculated. When the goodness of fit is greater than a preset threshold, the corresponding identification model is used as a multivariate dynamic prediction model for the denitrification process. The multivariate dynamic prediction model for the denitrification process is updated every preset update cycle using the latest historical variable data of the denitrification process.

9. The method based on total discharge outlet NO according to claim 8 x The denitrification equilibrium control method combined with ammonia slip correction is characterized by, The corrected NO corresponding to both inlets x The concentration baseline setting is the target, with NO at both outlets as the standard. x Using concentration data as feedback, the process of calculating the optimal control increment of the dual-flue ammonia injection valve opening, based on a pre-set multivariate dynamic prediction model for the denitrification process in a closed-loop manner, includes: Step 1: The pre-defined multivariate dynamic prediction model for the denitrification process uses the corrected NO at the two inlets of the dual flue. x The concentration benchmark setpoint is the target, with NO at both outlets of the dual flue gas ducts. x Concentration data is used for feedback; Step 2: Construct the optimization objective function and set constraints including control quantity constraints, control increment constraints, and output constraints. Transform the objective function and constraints into a quadratic programming problem and solve it using the interior point method to obtain the optimal control increment sequence. Step 3: Based on the preset rolling optimization strategy, the optimal control increment sequence is fed back for correction to obtain the corrected optimal control increment sequence. Step four: Iterate through steps one through three to construct and optimize the objective function until the NO outlets on both sides of the dual flue are reached. x Once the concentrations are equal to the threshold values, the first increment of the corrected optimal control increment sequence is taken as the optimal control increment for the corresponding ammonia injection valve opening in the dual flue.

10. The method based on total discharge outlet NO according to claim 9 x The denitrification equilibrium control method combined with ammonia slip correction is characterized by, The process of obtaining the corrected optimal control increment sequence by performing feedback correction on the optimal control increment sequence according to the preset rolling optimization strategy includes: Within the current time k, obtain the actual opening degree of the ammonia injection valve at the previous time k-1, apply the first increment of the optimal control increment sequence to the ammonia injection valve, calculate the output value of the actual opening degree of the ammonia injection valve at the current time k, and obtain the predicted output sequence at the current time k. Entering the next time step k+1, collect NO data from both outlets of the dual flues. x Concentration, obtain the predicted value of the current time k from the previous time k-1, and calculate the respective prediction error vector sequence based on the predicted value; By setting error correction coefficients and using the predicted error vector sequence and error correction coefficients to correct the output sequence predicted at the current time k, the corrected optimal control increment sequence is obtained.

Images