Flue gas denitration control system and control method

Abstract

The application discloses a flue gas denitration control system and a control method. According to the application, the data processing module can correct the measurement lag problem of the denitration inlet NOx concentration through correlation analysis, improve the data accuracy, and provide reliable data basis for intelligent control; the ammonia injection total amount control module realizes low-delay closed-loop control in cooperation with the double targets of the denitration outlet NOx concentration and the chimney outlet NOx concentration, and constructs a feedforward unit, a predictive feedback unit and a weight adjustment unit; the weight adjustment unit reasonably configures the weights of the feedforward unit and the predictive feedback unit, can accurately calculate the ammonia injection total amount, improves the flue gas emission reliability, and is suitable for different working conditions.

Description

Technical Field

[0001] This invention relates to the field of flue gas treatment technology, specifically to a flue gas denitrification control system and control method. Background Technology

[0002] The existing flue gas denitrification control technologies mainly include the following methods:

[0003] The first method relies on single-point sensors to collect NOx concentrations at the SCR inlet and outlet, supplemented by manual input of offline parameters such as coal quality and catalyst. This method simply uses the measured data for calculations, without considering factors such as measurement lag and abnormal data, resulting in relatively low control accuracy.

[0004] The second method is the total ammonia injection control method, which mainly uses PID feedback control to adjust the total ammonia injection through proportional-integral-derivative adjustment. This control method does not consider denitrification and NOx factors at the chimney outlet, resulting in insufficient control precision and timeliness, and making it difficult to adapt to changes in operating conditions.

[0005] As can be seen from the above description, each of the existing flue gas denitrification control technologies has different defects, and the control accuracy is not high.

[0006] Therefore, how to improve the control accuracy of flue gas denitrification control methods is a technical problem that those skilled in the art have been trying to solve. Summary of the Invention

[0007] The purpose of this application is to provide a flue gas denitrification control system that improves the accuracy of flue gas denitrification control. Another purpose of this application is to provide a control method based on the above-described flue gas denitrification control system.

[0008] This application provides a flue gas denitrification control system, including:

[0009] The data processing module, including the correction unit, is capable of calculating the dynamic time warping distance between the time series of coal consumption and the NOx concentration at the denitrification inlet, and between the NOx concentration at the denitrification inlet and the NOx concentration at the outlet, based on a dynamic time warping correlation analysis algorithm, thereby correcting the time lag problem in the measurement of the NOx concentration at the denitrification inlet.

[0010] The ammonia injection total control module includes a nonlinear mapping unit, a feedforward unit, a predictive feedback unit, and a weight adjustment unit. The nonlinear mapping unit is used to establish a nonlinear mapping model between the NOx concentration at the denitrification outlet and the NOx concentration at the chimney outlet using historical data, so as to generate a control value for the NOx concentration at the denitrification outlet. The feedforward unit can calculate the first total ammonia injection based at least on the corrected NOx concentration at the denitrification inlet, the unit load, and the start-up and shutdown status of the coal mill. The predictive feedback unit can estimate the predicted NOx concentration at the denitrification outlet in the future time period based on historical data, and then calculate the second total ammonia injection, wherein the historical data includes the unit load, the NOx concentration at the denitrification inlet, and the NOx concentration at the denitrification outlet.

[0011] The weighting adjustment unit adjusts the weights of the first total ammonia injection and the second total ammonia injection to calculate the total ammonia injection, so that the flue gas meets the emission standards.

[0012] In this embodiment, the data processing module can correct the lag problem in the measurement of NOx concentration at the denitrification inlet through correlation analysis, thereby improving data accuracy and providing a reliable data foundation for intelligent control. The ammonia injection total control module achieves low-delay closed-loop control by coordinating the NOx concentration at the denitrification outlet and the NOx concentration at the chimney outlet. It also constructs a feedforward unit, a predictive feedback unit, and a weight adjustment unit. The weight adjustment unit reasonably configures the weights of the feedforward unit and the predictive feedback unit, which can accurately calculate the total ammonia injection, improve the reliability of flue gas emissions, and adapt to different operating conditions.

[0013] In one example, the data processing module further includes a soft sensor model, which can select air volume series, total coal volume, air-to-coal ratio, unit load, and coal mill start-up / shutdown status that are highly correlated with the NOx concentration at the denitrification inlet as starting features. Considering the influence of coupling between features, the model is subjected to dimensionality reduction by principal component analysis (KPCA) and then trained using an LSTM or LSSVM network to calculate the NOx concentration at the denitrification inlet. The correction unit is used to correct the NOx concentration at the denitrification inlet output by the soft sensor model.

[0014] In one example, the weighting adjustment unit includes an emission control quality assessment function. When the difference between the current NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet is greater than the standard deviation, the emission control quality assessment function evaluates the impact of the first total ammonia injection and the second total ammonia injection on the flue gas emissions, and adjusts the weight values ​​of the first total ammonia injection and the second total ammonia injection to suppress the increase in the difference between the NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet.

[0015] In one example, the flue gas denitrification control system further includes a zoned ammonia injection collaborative optimization module, which can generate the initial zoned ammonia injection weights when the system starts up through an expert knowledge base, and dynamically adjust the zoned ammonia injection weights in real time after the system starts up, so that the NOx concentration CV value and ammonia slip rate of the zone under the current operating conditions are both within a preset range, wherein the zoned NOx concentration CV value is calculated based on the current state zone outlet NOx concentration and the denitrification outlet NOx concentration control value.

[0016] In one example, the zoned ammonia injection collaborative optimization module includes an intelligent optimization unit. The intelligent optimization unit uses the XGBoost algorithm to construct a prediction model of the zoned ammonia injection amount and the zoned outlet NOx concentration under different operating conditions. It can also construct an evaluation function with the zoned NOx concentration CV value and the ammonia slip rate as parameters. The calculated value of the prediction model and the zoned ammonia slip rate are substituted into the evaluation function and the optimal zone weight that meets the flue gas emission standards is selected through multiple iterations.

[0017] Alternatively / and, the data processing module can also use the K-Means clustering algorithm to stratify various operating parameters such as unit load and coal nitrogen content, and construct a dynamic operating condition feature library; the zoned ammonia injection collaborative optimization module includes a cold start unit, which is used to combine the power supply operating condition feature library and the expert knowledge base to generate the initial zoned ammonia injection weights when the system starts.

[0018] In one example, the partitioned ammonia injection co-optimization module further includes a partitioned ammonia slip rate optimization module, which is used to reduce the amount of ammonia injected into the partition when the removal efficiency of the current working condition partition is lower than a predetermined value, provided that the CV value of the NOx concentration in the partition meets the conditions, so as to keep the removal efficiency within a preset range.

[0019] In one example, a total amount-zone coordination module is also included, which is used to reduce the total amount of ammonia injection to meet emission standards when the NOx concentration at the chimney outlet is lower than the target chimney outlet NOx concentration, and the average value of the NOx concentration at all zone outlets is lower than the NOx concentration control value at the denitrification outlet for a predetermined period of time.

[0020] In one example, a catalyst degradation analysis module is also included. Based on the probabilistic statistical analysis of the frequency of historical operating conditions, a weight allocation matrix is ​​constructed. Combining the weighted accumulation algorithm and dimensional normalization processing technology, spatiotemporal dimensional fusion calculations are performed on multi-operating condition monitoring data. Finally, a catalyst activity decay characteristic analysis curve based on weight allocation is established to achieve scientific characterization and quantitative evaluation of catalyst degradation laws.

[0021] In one example, when the catalyst activity decreases by more than a predetermined value, the coefficient of the total ammonia injection control module is adjusted to increase the total ammonia injection amount, thereby compensating for the impact of activity decay on denitrification efficiency.

[0022] In one example, a three-dimensional digitization module is also included, which uses digital twin technology to construct a 3D model to display the temperature field diagram, flow field diagram, ammonia injection grid flow balance diagram, or system parameters of the SCR system on a human-machine interface, which supports interactive operation.

[0023] In addition, this application also provides a flue gas denitrification control method, including:

[0024] Based on the correlation analysis algorithm of dynamic time warping, the dynamic time warping distance of the time series between coal consumption and NOx concentration at the denitrification inlet, and between NOx concentration at the denitrification inlet and NOx concentration at the outlet is calculated, thereby correcting the time lag problem of NOx concentration measurement at the denitrification inlet.

[0025] A nonlinear mapping model between the NOx concentration at the denitrification outlet and the NOx concentration at the chimney outlet is established using historical data. The control value for the NOx concentration at the denitrification outlet is calculated based on the target NOx concentration at the chimney outlet and the nonlinear mapping model. The first total ammonia injection is calculated based at least on the corrected NOx concentration at the denitrification inlet, unit load, and the start / stop status of the coal mill. The predicted NOx concentration at the denitrification outlet for a future time period is extrapolated based on historical data, and the second total ammonia injection is then calculated. The historical data includes unit load, NOx concentration at the denitrification inlet, and NOx concentration at the denitrification outlet.

[0026] The weights of the first and second total ammonia injections are adjusted to calculate the total ammonia injection amount, so that the flue gas meets the emission standards.

[0027] In one example, the weighting of the first total ammonia injection and the second total ammonia injection is achieved in the following way:

[0028] An emission control quality assessment function is constructed. When the difference between the current NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet is greater than the standard deviation, the weight values ​​of the first total ammonia injection and the second total ammonia injection are adjusted according to the emission control quality assessment function to suppress the increase in the difference between the NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet.

[0029] In one example, the control method can also generate an initial ammonia injection weight for each zone when the system starts up using an expert knowledge base, and dynamically adjust the ammonia injection weight of each zone in real time after the system starts up, so that the NOx concentration CV value and the ammonia escape rate of each zone under the current operating conditions are both within a preset range, wherein the NOx concentration CV value of each zone is calculated based on the NOx concentration at the outlet of the zone under the current state and the NOx concentration control value at the denitrification outlet.

[0030] The ammonia injection weight is adjusted in the following way: Using the XGBoost algorithm, a prediction model of the ammonia injection amount and NOx concentration at the zone outlet under different operating conditions is constructed, eliminating ammonia injection amounts from zones that deviate from the emission standard range; an evaluation function is constructed using the zone NOx concentration CV value and the ammonia slip rate as parameters; the calculated value of the zone outlet NOx concentration prediction model and the zone ammonia slip rate are substituted into the evaluation function, and the optimal zone weight that meets the flue gas emission standard is selected through multiple iterations; and, provided that the zone NOx concentration CV value meets the condition, when the removal efficiency of the zone under the current operating condition is lower than a predetermined value, the ammonia injection amount of that zone is reduced to keep the removal efficiency within the preset range.

[0031] In one example, when the NOx concentration at the chimney outlet is lower than the target chimney outlet NOx concentration, and the average NOx concentration at all the zone outlets is lower than the NOx concentration control value at the denitrification outlet for a predetermined period of time, the total ammonia injection is reduced to meet the emission standards.

[0032] The flue gas denitrification control method provided by the present invention is based on the above-mentioned flue gas denitrification control system, and therefore also has the above-mentioned technical effects of the flue gas denitrification control system. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the structure of the flue gas denitrification control system provided in one embodiment of this application;

[0034] Figure 2 This is a scatter plot of the NOx mass at the denitrification inlet versus the actual ammonia nitrogen ratio in the embodiments of this application.

[0035] Figure 3 This is a thermogram of inlet NOx amount - ammonia slip rate in an embodiment of this application;

[0036] Figure 4 This is a diagram showing the unit load and inlet SO2 distribution in an embodiment of this application;

[0037] Figure 5 This is a schematic diagram of the predicted ammonia injection rate, actual ammonia injection rate, NOx concentration at the denitrification outlet, and ammonia injection rate curve when the denitrification outlet NOx is set to 35 mg / Nm3 in the embodiments of this application.

[0038] Figure 6 This is a flowchart illustrating the calculation of NOx concentration at the denitrification inlet using a soft-sensing model in this application embodiment;

[0039] Figure 7 The figure shows the catalyst activity decay characteristic analysis curve in the embodiments of this application, where the vertical axis represents the catalyst concentration and the horizontal axis represents time, in seconds. Detailed Implementation

[0040] To enable those skilled in the art to better understand the technical solutions of the embodiments of this application, the embodiments of this application will be further described in detail below with reference to the accompanying drawings and specific examples.

[0041] Please refer to Figures 1 to 6 , Figure 1 This is a schematic diagram of the structure of the flue gas denitrification control system provided in one embodiment of this application; Figure 2 This is a scatter plot of the NOx mass at the denitrification inlet versus the actual ammonia nitrogen ratio in the embodiments of this application. Figure 3 This is a thermogram of inlet NOx amount - ammonia slip rate in an embodiment of this application; Figure 4 This is a diagram showing the unit load and inlet SO2 distribution in an embodiment of this application; Figure 5 This is a schematic diagram of the predicted ammonia injection rate, actual ammonia injection rate, NOx concentration at the denitrification outlet, and ammonia injection rate curve when the denitrification outlet NOx is set to 35 mg / Nm3 in the embodiments of this application. Figure 6 This is a flowchart illustrating the calculation of NOx concentration at the denitrification inlet using a soft measurement model in this embodiment of the application.

[0042] This application provides a flue gas denitrification control system, including a data processing module, an ammonia injection total control module, a zoned ammonia injection collaborative optimization module, a total injection-zone collaborative module, a catalyst degradation analysis module, and a three-dimensional digitization module.

[0043] The data processing module includes a correction unit that, based on a dynamic time warping correlation analysis algorithm, calculates the dynamic time warping distance (Euclidean distance) between the time series of coal consumption and NOx concentration at the denitrification inlet, and between the NOx concentration at the denitrification inlet and the NOx concentration at the outlet. It then finds the optimal nonlinear alignment between the two sets of data, minimizing the cumulative distance between them, thereby correcting the time mismatch, i.e., correcting the time lag problem in the NOx concentration measurement at the denitrification inlet. This correlation analysis corrects the measurement lag problem of NOx concentration at the denitrification inlet, improving data accuracy and providing a reliable data foundation for intelligent control.

[0044] The data processing module also includes a soft sensing model, capable of selecting air volume series (including total air volume, primary air, and secondary air), total coal volume, air-to-coal ratio, unit load, and pulverizer start-up / shutdown status—all highly correlated with the NOx concentration at the denitrification inlet—as starting features. Considering the impact of feature coupling, dimensionality reduction is performed using Principal Component Analysis (KPCA), followed by training with an LSTM or LSSVM network to calculate the NOx concentration at the denitrification inlet. Please refer to [link / reference needed]. Figure 6 understand, Figure 6 The operational logic is shown.

[0045] Data can also be displayed on a screen using a visualization module, which can improve the efficiency of anomaly handling.

[0046] For example, a scatter plot of NOx removal amount - actual ammonia nitrogen ratio ( Figure 2 (As shown), inlet NOx amount - ammonia slip rate thermogram ( Figure 3 All data (as shown) can be displayed on the screen through the visualization module. Through precise machine recognition and final manual confirmation, abnormal operating conditions such as ammonia nitrogen ratio greater than 1.5 and ammonia escape greater than 3 ppm can be efficiently handled, improving the accuracy of data anomaly correction.

[0047] The data processing module can also use the K-Means clustering algorithm to stratify more than 20 operating condition parameters, such as unit load (high load / medium load / low load) and coal nitrogen content (>1.5% / 1.0%-1.5% / <1.0%), to build a dynamic operating condition feature library. This supports rapid matching of operating conditions and control strategies, providing a basis for the initial system cold start. Please refer to [link / reference]. Figure 4 , Figure 4 The horizontal axis represents the unit load, and the vertical axis represents the inlet SO2. Different bubble colors in the figure represent the numerical distribution of slurry flow rate. Figure 4 In the middle, X represents points that do not conform to the distribution or clustering and need to be removed.

[0048] In this embodiment, the ammonia injection total control module is mainly used to calculate the total ammonia injection volume of the system. The ammonia injection total control module includes a nonlinear mapping unit, a feedforward unit, a predictive feedback unit, and a weight adjustment unit. The nonlinear mapping unit is used to establish a nonlinear mapping model between the NOx concentration at the denitrification outlet and the NOx concentration at the chimney outlet using historical data. Based on the target NOx concentration at the chimney outlet and the nonlinear mapping model, it calculates the control value for the NOx concentration at the denitrification outlet. In other words, the control system of this application establishes a model that correlates the NOx concentration at the denitrification outlet with the NOx concentration at the chimney outlet, thereby improving the control accuracy. Operators can set the target NOx concentration at the chimney outlet (hourly average target NOx concentration at the chimney outlet, for example, 45 mg / Nm³) through the human-machine interface. The control system analyzes the data characteristics within a predetermined time window, establishes a nonlinear mapping model of the deviation between the NOx concentration at the denitrification outlet and the NOx concentration at the chimney outlet, and intelligently generates the control value for the NOx concentration at the denitrification outlet, achieving a low-latency conversion between the "chimney target" and the "denitrification control."

[0049] In this embodiment, the feedforward unit can calculate the first total ammonia injection based at least on the corrected denitrification inlet NOx concentration (the value calculated by the corrected soft measurement model), unit load, and pulverizer start-up and shutdown status. The parameters on which the feedforward model calculates the first total ammonia injection are not limited to the above parameters, but may also include other parameters such as the distance of the ammonia injection grid. The first total ammonia injection is output through the value and change slope of each parameter in a minute-level time window.

[0050] The prediction feedback unit can extrapolate the predicted NOx concentration at the denitrification outlet within a future time period based on historical data, and then calculate the total amount of the second ammonia injection. The historical data includes unit load, NOx concentration at the denitrification inlet, and NOx at the denitrification outlet. The historical time can be a minute-level time window, predicting the NOx concentration at the denitrification outlet 3-5 minutes later.

[0051] In this embodiment, the weighting adjustment unit adjusts the weights of the first and second total ammonia injections to calculate the total ammonia injection, ensuring the flue gas meets emission standards. The emission standards include the target chimney outlet NOx concentration. When the difference between the current chimney outlet NOx concentration and the target chimney outlet NOx concentration is within a predetermined range, the current flue gas is considered to meet the emission standards. In one example, the weighting adjustment unit includes an emission control quality assessment function. When the difference between the current chimney outlet NOx concentration and the target chimney outlet NOx concentration is greater than the standard deviation, the impact of the first and second total ammonia injections on the emitted flue gas is assessed, and the weight values ​​of the first and second total ammonia injections are adjusted to suppress the increase in the difference between the chimney outlet NOx concentration and the target chimney outlet NOx concentration. For example, when the standard deviation increases, the feedforward weight (γ1=0.4-0.7) and the feedback weight (γ2=0.3-0.6) are dynamically adjusted in reverse to suppress emission oscillations caused by sudden load changes and improve system flexibility. The feedforward weight is the weight of the feedforward unit, and the feedback weight is the weight of the predictive feedback unit.

[0052] In this embodiment, the data processing module can correct the lag problem in the measurement of NOx concentration at the denitrification inlet through correlation analysis, thereby improving data accuracy and providing a reliable data foundation for intelligent control. The ammonia injection total control module achieves low-delay closed-loop control by coordinating the NOx concentration at the denitrification outlet and the NOx concentration at the chimney outlet. It also constructs a feedforward unit, a predictive feedback unit, and a weight adjustment unit. The weight adjustment unit reasonably configures the weights of the feedforward unit and the predictive feedback unit, which can accurately calculate the total ammonia injection, improve the reliability of flue gas emissions, and adapt to different operating conditions.

[0053] In this embodiment, the zoned ammonia injection collaborative optimization module can generate the initial zoned ammonia injection weights at system startup through an expert knowledge base, and dynamically adjust the zoned ammonia injection weights in real time after system startup, so that the NOx concentration CV value and ammonia slip rate of the current operating zone are both within a preset range. The zoned NOx concentration CV value is calculated based on the NOx concentration at the current zone outlet and the NOx concentration control value at the denitrification outlet. CV stands for Coefficient of Variation, which is a statistical indicator that measures the dispersion (i.e., uniformity) of NOx concentration data at multiple zone outlets.

[0054] Specifically, the zoned ammonia injection collaborative optimization module can include a cold start unit and an intelligent optimization unit. The cold start unit, based on the dynamic operating condition feature library of the components in the aforementioned data processing module and combined with an expert knowledge base of effective control rules, uses a fuzzy logic reasoning mechanism to automatically generate an initial zoned ammonia injection weight combination (e.g., adding 15% to the weight of side A in high-load condition) when a new operating condition is introduced. It then uses a reinforcement learning algorithm to correct the rules online (the update frequency can be reasonably set, e.g., once per minute), reducing manual trial-and-error time and forming a rich dataset.

[0055] The zoned ammonia injection collaborative optimization module includes an intelligent optimization unit that uses the XGBoost algorithm to construct predictive models of zoned ammonia injection rates and zoned outlet NOx concentrations under different operating conditions. This allows for advance prediction of potential emission fluctuations from multiple recommended strategies, eliminating strategies that deviate from the target setpoint range to ensure the safety of the control system. The intelligent optimization unit can also construct an evaluation function using the zoned NOx concentration CV value and ammonia slip rate as parameters. By substituting the calculated values ​​from the predictive model and the zoned ammonia slip rate into the evaluation function, the optimal zone weight that meets the flue gas emission standards is selected through multiple iterations. Depending on actual needs, penalty factors can be added for significant adjustments to the ammonia injection rate or for CV values ​​or ammonia slip rates exceeding preset targets, further improving the reliability of the recommended strategies.

[0056] In this embodiment, the zoned ammonia injection synergistic optimization module further includes a zoned ammonia slip rate optimization module. This module, under the premise that the zoned NOx concentration CV value meets the requirements, reduces the ammonia injection amount in the zone when the removal efficiency of the current operating zone is lower than a predetermined value, so that the removal efficiency remains within a preset range. Specifically, while ensuring that the zoned NOx concentration CV value meets the standards, the module monitors the actual ammonia-nitrogen ratio and removal efficiency of each zone online, dynamically plans the ammonia injection and emission control intensity for each zone, and reduces ammonia usage by 5%-10% in zones with relatively insufficient removal efficiency, thereby reducing excessive local ammonia injection and further lowering the ammonia slip rate.

[0057] In this embodiment, the total ammonia injection control module aims to ensure that the NOx concentration at the denitrification outlet and the NOx concentration at the chimney outlet meet the standards and stabilizes, and recommends the benchmark value of total ammonia injection in real time; the zoned ammonia injection collaborative optimization module dynamically adjusts the weight of each zone as needed based on real-time changes in operating conditions (such as coal quality and load), so as to promote the NOx concentration CV value of each zone to meet the standards, reduce ammonia escape in each zone, and improve the overall removal efficiency.

[0058] In this embodiment of the application, the flue gas denitrification control system further includes a total amount-zone coordination module, which is used to reduce the total amount of ammonia injection to meet emission standards when the NOx concentration at the chimney outlet is lower than the target chimney outlet NOx concentration and the average value of the NOx concentration at all zone outlets is lower than the denitrification outlet NOx concentration control value for a predetermined period of time.

[0059] For example, when it is determined that the NOx concentration at the denitrification and chimney outlet decreases by more than 5% for 5 consecutive minutes, the total amount model is automatically triggered to reduce the total ammonia injection by 2%-5%, thereby reducing the overall ammonia consumption and forming a positive synergistic mechanism between total amount and zone.

[0060] This allows for the establishment of a closed-loop mechanism of "continuous control of total ammonia injection - zoned adjustment as needed - positive collaborative optimization".

[0061] In this embodiment of the application, the flue gas denitrification control system also includes a catalyst degradation analysis module. Based on the probability statistical analysis of the frequency of historical operating conditions, a weight allocation matrix is ​​constructed. Combining the weighted accumulation algorithm and dimensional normalization processing technology, spatiotemporal dimensional fusion calculations are performed on the monitoring data of multiple operating conditions. Finally, a catalyst activity decay characteristic analysis curve based on weight allocation is established to achieve scientific characterization and quantitative evaluation of the catalyst degradation law.

[0062] In one example, when the catalyst activity decreases by more than a predetermined value, the correction coefficient of the total ammonia injection control module is increased to compensate for the impact of activity decay on denitrification efficiency and ensure the long-term compliance rate of the system.

[0063] In this embodiment of the application, the flue gas denitrification control system also includes a three-dimensional digitization module, which uses digital twin technology to construct a 3D model to display the temperature field diagram, flow field diagram, ammonia injection grid flow balance diagram or system parameters of the SCR system on the human-machine interface. The human-machine interface supports interactive operation.

[0064] System parameters may include one or more of the following: real-time display of CEMS status, zone NOx concentration, ammonia slip rate, ammonia injection rate, etc., providing historical curve display, daily / monthly ammonia saving statistics, abnormal alarms (such as audible and visual alarms when ammonia slip is >3ppm), and supporting customization of page elements to improve the monitoring efficiency of operators.

[0065] Based on the above-mentioned flue gas denitrification control system, the present invention also provides a flue gas denitrification control method, comprising:

[0066] Based on the correlation analysis algorithm of dynamic time warping, the dynamic time warping distance of the time series between coal consumption and NOx concentration at the denitrification inlet, and between NOx concentration at the denitrification inlet and NOx concentration at the outlet is calculated, thereby correcting the time lag problem of NOx concentration measurement at the denitrification inlet.

[0067] A nonlinear mapping model between NOx concentration at the denitrification outlet and NOx concentration at the chimney outlet is established using historical data. The control value of NOx concentration at the denitrification outlet is calculated based on the target NOx concentration at the chimney outlet and the nonlinear mapping model. The total amount of ammonia injected for the first time is calculated based on at least the corrected NOx concentration at the denitrification inlet, unit load, and coal mill start-up and shutdown status. The predicted NOx concentration at the denitrification outlet in the future time period is estimated based on historical data, and then the total amount of ammonia injected for the second time is calculated. The historical data includes unit load, NOx concentration at the denitrification inlet, and NOx concentration at the denitrification outlet.

[0068] The weights of the first and second total ammonia injections are adjusted to calculate the total ammonia injection amount, so that the flue gas meets the emission standards.

[0069] The weights of the first and second total ammonia injections are adjusted in the following way:

[0070] An emission control quality assessment function is constructed. When the difference between the current NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet is greater than the standard deviation, the weight values ​​of the first and second total ammonia injections are adjusted according to the emission control quality assessment function to suppress the increase in the difference between the NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet.

[0071] The above control method can also generate the initial ammonia injection weight of the zone when the system starts up through the expert knowledge base, and dynamically adjust the ammonia injection weight of the zone in real time after the system starts up, so that the NOx concentration CV value and the ammonia slip rate of the zone under the current operating conditions are both within the preset range. The NOx concentration CV value of the zone is calculated based on the NOx concentration at the zone outlet and the NOx concentration control value at the denitrification outlet under the current state.

[0072] Specifically, the ammonia injection weight is adjusted in the following ways: using the XGBoost algorithm, a prediction model of the ammonia injection amount and NOx concentration at the zone outlet under different operating conditions is constructed, and strategies that deviate from the emission standard range are eliminated; an evaluation function is constructed with the zone NOx concentration CV value and ammonia slip rate as parameters, and the calculated value of the prediction model of the zone outlet NOx concentration and the zone ammonia slip rate are substituted into the evaluation function to select the optimal zone weight that meets the flue gas emission standard through multiple iterations. Furthermore, under the premise that the zone NOx concentration CV value meets the condition, when the removal efficiency of the zone under the current operating condition is lower than the predetermined value, the ammonia injection amount of that zone is reduced so that the removal efficiency is within the preset range.

[0073] In the flue gas denitrification control method of this application, when the NOx concentration at the chimney outlet is lower than the target chimney outlet NOx concentration, and the average NOx concentration at all zone outlets is lower than the target denitrification outlet NOx concentration for a predetermined period of time, the total amount of ammonia injected is reduced to meet the emission standards.

[0074] The above-mentioned flue gas denitrification control method is implemented based on the above-mentioned flue gas denitrification control system, and therefore also has the above-mentioned technical effects of the flue gas denitrification control system.

[0075] In the description of embodiments of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0076] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A flue gas denitrification control system, characterized in that, include: The data processing module, including the correction unit, is capable of calculating the dynamic time warping distance between the time series of coal consumption and the NOx concentration at the denitrification inlet, and between the NOx concentration at the denitrification inlet and the NOx concentration at the outlet, based on a dynamic time warping correlation analysis algorithm, thereby correcting the time lag problem in the measurement of the NOx concentration at the denitrification inlet. The ammonia injection total control module includes a nonlinear mapping unit, a feedforward unit, a predictive feedback unit, and a weight adjustment unit. The nonlinear mapping unit is used to establish a nonlinear mapping model between the NOx concentration at the denitrification outlet and the NOx concentration at the chimney outlet using historical data, so as to generate a control value for the NOx concentration at the denitrification outlet. The feedforward unit can calculate the first total ammonia injection based at least on the corrected NOx concentration at the denitrification inlet, the unit load, and the start-up and shutdown status of the coal mill. The prediction feedback unit can estimate the predicted NOx concentration at the denitrification outlet in the future time period based on historical data, and then calculate the total amount of the second ammonia injection. The historical data includes unit load, NOx concentration at the denitrification inlet, and NOx concentration at the denitrification outlet. The weighting adjustment unit adjusts the weights of the first total ammonia injection and the second total ammonia injection, and calculates the total ammonia injection to ensure that the flue gas meets the emission standards. The flue gas denitrification control system also includes a zoned ammonia injection collaborative optimization module, which can generate the initial zoned ammonia injection weights when the system starts up through an expert knowledge base, and dynamically adjust the zoned ammonia injection weights in real time after the system starts up, so that the NOx concentration CV value and ammonia slip rate of the zone under the current operating conditions are both within the preset range. The zoned NOx concentration CV value is calculated based on the current zone outlet NOx concentration and the denitrification outlet NOx concentration control value. The zoned ammonia injection collaborative optimization module includes an intelligent optimization unit. The intelligent optimization unit uses the XGBoost algorithm to construct a prediction model of the zoned ammonia injection amount and the zoned outlet NOx concentration under different operating conditions. It can also construct an evaluation function with the zoned NOx concentration CV value and the ammonia slip rate as parameters. The calculated value of the prediction model and the zoned ammonia slip rate are substituted into the evaluation function and the optimal zone weight that meets the flue gas emission standards is selected through multiple iterations.

2. The flue gas denitrification control system according to claim 1, characterized in that, The data processing module also includes a soft measurement model, which can select air volume series, total coal volume, air-coal ratio, unit load, and coal mill start-up and shutdown status that are highly correlated with the NOx concentration at the denitrification inlet as starting features. Considering the influence of coupling between features, the model is subjected to dimensionality reduction by principal component analysis (KPCA) and then trained using an LSTM or LSSVM network to calculate the NOx concentration at the denitrification inlet. The correction unit is used to correct the NOx concentration at the denitrification inlet output by the soft measurement model.

3. The flue gas denitrification control system according to claim 2, characterized in that, The weighting adjustment unit includes an emission control quality assessment function. When the difference between the current NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet is greater than the standard deviation, the emission control quality assessment function evaluates the impact of the first total ammonia injection and the second total ammonia injection on the flue gas emissions, and adjusts the weight values ​​of the first total ammonia injection and the second total ammonia injection to suppress the increase in the difference between the NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet.

4. The flue gas denitrification control system according to claim 1, characterized in that, The data processing module can also use the K-Means clustering algorithm to stratify various operating parameters such as unit load and coal nitrogen content, and construct a dynamic operating condition feature library; the zoned ammonia injection collaborative optimization module includes a cold start unit, which is used to generate the initial zoned ammonia injection weights when the system starts by combining the dynamic operating condition feature library and the expert knowledge base.

5. The flue gas denitrification control system according to claim 4, characterized in that, The partitioned ammonia injection co-optimization module also includes a partitioned ammonia slip rate optimization module, which is used to reduce the amount of ammonia injected into the partition when the removal efficiency of the current working condition partition is lower than a predetermined value, provided that the CV value of the NOx concentration in the partition meets the conditions, so as to keep the removal efficiency within a preset range.

6. The flue gas denitrification control system according to claim 1, characterized in that, It also includes a total amount-zone coordination module, which is used to reduce the total amount of ammonia injected to meet emission standards when the NOx concentration at the chimney outlet is lower than the target chimney outlet NOx concentration, and the average value of the NOx concentration at all zone outlets is lower than the denitrification outlet NOx concentration control value for a predetermined period of time.

7. The flue gas denitrification control system according to any one of claims 1 to 6, characterized in that, It also includes a catalyst degradation analysis module, which constructs a weight allocation matrix based on the probability statistical analysis of the frequency of historical operating conditions. Combining a weighted accumulation algorithm and dimensional normalization processing technology, it performs spatiotemporal dimensional fusion calculations on multi-operating condition monitoring data, and finally establishes a catalyst activity decay characteristic analysis curve based on weight allocation, so as to realize the scientific characterization and quantitative evaluation of catalyst degradation law.

8. The flue gas denitrification control system according to claim 7, characterized in that, When the catalyst activity decreases by more than a predetermined value, the coefficient of the total ammonia injection control module is adjusted to increase the total ammonia injection amount and thus compensate for the impact of activity decay on denitrification efficiency.

9. The flue gas denitrification control system according to any one of claims 1 to 6, characterized in that, It also includes a 3D digitization module, which uses digital twin technology to build a 3D model to display the temperature field diagram, flow field diagram, ammonia injection grid flow balance diagram or system parameters of the SCR system on the human-machine interface. The human-machine interface supports interactive operation.

10. A method for controlling flue gas denitrification, characterized in that, include: Based on the correlation analysis algorithm of dynamic time warping, the dynamic time warping distance of the time series between coal consumption and NOx concentration at the denitrification inlet, and between NOx concentration at the denitrification inlet and NOx concentration at the outlet is calculated, thereby correcting the time lag problem of NOx concentration measurement at the denitrification inlet. A nonlinear mapping model between the NOx concentration at the denitrification outlet and the NOx concentration at the chimney outlet is established using historical data. The control value for the NOx concentration at the denitrification outlet is calculated based on the target NOx concentration at the chimney outlet and the nonlinear mapping model. The first total ammonia injection is calculated based at least on the corrected NOx concentration at the denitrification inlet, unit load, and the start / stop status of the coal mill. The predicted NOx concentration at the denitrification outlet for a future time period is extrapolated based on historical data, and the second total ammonia injection is then calculated. The historical data includes unit load, NOx concentration at the denitrification inlet, and NOx concentration at the denitrification outlet. Adjust the weights of the first and second total ammonia injection amounts to calculate the total ammonia injection amount so that the flue gas meets emission standards; The control method can also generate the initial ammonia injection weight of the partition when the system starts by using an expert knowledge base, and dynamically adjust the ammonia injection weight of the partition in real time after the system starts, so that the NOx concentration CV value and the ammonia escape rate of the partition under the current operating condition are both within the preset range, wherein the partition NOx concentration CV value is calculated based on the NOx concentration at the current state of the partition outlet and the NOx concentration control value at the denitrification outlet. The ammonia injection weight is adjusted in the following way: Using the XGBoost algorithm, a prediction model of the ammonia injection amount and NOx concentration at the zone outlet under different operating conditions is constructed, eliminating ammonia injection amounts from zones that deviate from the emission standard range; an evaluation function is constructed using the zone NOx concentration CV value and the ammonia slip rate as parameters; the calculated value of the zone outlet NOx concentration prediction model and the zone ammonia slip rate are substituted into the evaluation function, and the optimal zone weight that meets the flue gas emission standard is selected through multiple iterations; and, provided that the zone NOx concentration CV value meets the condition, when the removal efficiency of the zone under the current operating condition is lower than a predetermined value, the ammonia injection amount of that zone is reduced to ensure that the removal efficiency is within the preset range.

11. The flue gas denitrification control method according to claim 10, characterized in that, The weights of the first total ammonia injection and the second total ammonia injection are determined in the following way: An emission control quality assessment function is constructed. When the difference between the current NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet is greater than the standard deviation, the weight values ​​of the first total ammonia injection and the second total ammonia injection are adjusted according to the emission control quality assessment function to suppress the increase in the difference between the NOx concentration at the chimney outlet and the target NOx concentration at the chimney outlet.

12. The flue gas denitrification control method according to claim 10 or 11, characterized in that, When the NOx concentration at the chimney outlet is lower than the target chimney outlet NOx concentration, and the average NOx concentration at all zone outlets is lower than the NOx concentration control value at the denitrification outlet for a predetermined period of time, the total ammonia injection will be reduced to meet the emission standards.

Images