An intelligent control algorithm for SCR partition ammonia injection based on ammonia escape dynamic compensation

Abstract

The application discloses an intelligent control algorithm for SCR partition ammonia injection based on ammonia escape dynamic compensation, which comprises the following steps: collecting the NOx concentration and flue gas flow rate of each partition of the inlet flue, and the NOx and ammonia concentration of each partition of the outlet flue; taking the NOx concentration, total amount of NOx and its change rate of the inlet flue as feedforward signals, and obtaining the feedforward ammonia injection amount; taking the NOx concentration, ammonia escape concentration and ammonia escape field uniformity index of the outlet flue as feedback signals, and dynamically compensating the feedforward deviation according to the feedback signals; dividing the ammonia injection grid into multiple independent control zones, and dynamically distributing the ammonia amount according to the feedforward ammonia injection amount, feedforward signals and feedback signals of each partition; adopting model predictive control or machine learning method to adaptively adjust the nonlinear factors; and judging whether to trigger the material interruption protection according to the catalyst temperature. The application realizes the rapid and accurate sensing of inlet and outlet gases, and generates the optimal partition ammonia injection instruction by comprehensively considering the feedforward and feedback information.

Description

[0001] This invention relates to the field of flue gas denitrification technology for coal-fired power plants, industrial boilers and kilns, and particularly to an intelligent control algorithm for SCR zone ammonia injection based on dynamic compensation for ammonia escape. Background Technology

[0002] SCR (Selective Catalytic Reduction) denitrification technology is currently the mainstream technology for controlling nitrogen oxide (NOx) emissions. Traditional control methods typically rely on conventional emission monitoring systems (CEMS) to obtain inlet and outlet NOx concentrations, and use a combination of various prediction methods, state compensators, and PID controllers to adjust the ammonia injection rate. For example, Chinese patent application CN121944776A (application date 2026.01.07, publication date 2026.05.01) discloses a denitrification NOx control method, device, and electronic equipment based on an intelligent control strategy. This method obtains the corresponding process reaction parameters of the reactor based on the status signals of the reactor's monitoring instruments; predicts the first NOx concentration of the reactor based on the process reaction parameters using different concentration prediction methods; determines the second NOx concentration of the reactor based on the first NOx concentration and process reaction parameters through collaborative calculations using a concentration prediction model and a state compensator; and uses a first PID controller to perform closed-loop regulation of the ammonia flow rate based on the second NOx concentration and ammonia flow rate feedforward control commands to generate the ammonia flow rate control command for the reactor. This method utilizes a first PID controller combined with feedforward commands for NOx concentration and ammonia flow rate to perform closed-loop regulation, thereby offsetting the large hysteresis characteristics of the denitrification system and improving the accuracy of ammonia injection flow rate control.

[0003] However, such methods have the following inherent defects: (1) Inaccurate feedforward: The feedforward signal is based only on the predicted value of NOx concentration at the inlet, without considering the real-time changes in flue gas velocity, and cannot accurately calculate the total amount of NOx entering the reactor, resulting in low accuracy of feedforward open-loop control; (2) Ammonia escape cannot be locally identified: It only relies on the ammonia escape value at the outlet or does not use ammonia escape as feedback at all, and does not realize the spatial distribution measurement and zoned closed-loop control of ammonia escape, which leads to the inability to identify local over-spray, causing ammonium bisulfate blockage in the air preheater; (3) Weak adaptive capability of control algorithm: Although the existing technology uses neural network prediction, it still relies on the PID architecture as a whole, and its adaptability to nonlinear and large lag processes such as coal type change and catalyst decay is insufficient.

[0004] It is evident that traditional SCR denitrification systems exhibit prominent problems in actual operation, such as measurement lag, high ammonia slip, and uneven ammonia injection, making it difficult to meet the dual goals of "ultra-low emissions" and "energy saving and consumption reduction." Therefore, it is imperative to upgrade the technology through advanced sensing and intelligent control methods. Summary of the Invention

[0005] This invention provides an intelligent control algorithm for SCR zone ammonia injection based on dynamic compensation for ammonia slip, aiming to solve the following technical problems: eliminating CEMS measurement lag and realizing real-time feedforward of total NOx at the inlet; incorporating the spatial distribution of ammonia slip into closed-loop control to achieve field matching between zone ammonia injection and zone ammonia slip; and improving the adaptive capability of the control system to sudden load changes and coal type changes.

[0006] To achieve the above objectives, the specific solution of the present invention is as follows: A smart control algorithm for SCR zone ammonia injection based on dynamic compensation for ammonia slip includes the following steps: S1: Multiple sampling points are set up in the inlet and outlet flues of the SCR reactor to collect NOx concentration and flue gas velocity in each section of the inlet flue and NOx and ammonia concentration in each section of the outlet flue in real time. S2: Calculate the total amount of NOx and the rate of change of the total amount of NOx based on the NOx concentration and flue gas velocity in the inlet flue. Use the NOx concentration, the total amount of NOx, and the rate of change of the total amount of NOx as feedforward signals. Calculate the theoretically required total amount of ammonia injection based on the total amount of NOx in the inlet flue and the target denitrification efficiency. Add the rate of change compensation to the theoretical total amount of ammonia injection to obtain the feedforward ammonia injection amount. S3: Calculate the ammonia slip field uniformity index based on the ammonia slip concentration in the outlet flue, use the NOx concentration, ammonia slip concentration and ammonia slip field uniformity index as feedback signals, and dynamically compensate for the feedforward bias based on the feedback signals. S4: Divide the ammonia injection grid into multiple independent control zones and dynamically allocate the ammonia quantity based on the feedforward ammonia injection quantity, feedforward signal and feedback signal of each zone; S5: Adaptive adjustment of nonlinear factors using model predictive control or machine learning methods; S6: Determine whether to trigger the feed cut-off protection based on the catalyst temperature.

[0007] Preferably, in step S1, the method for setting sampling points is to divide the cross-section of the inlet flue and the cross-section of the outlet flue into multiple partitions, with each partition serving as a sampling point.

[0008] Preferably, in step S1, sampling is performed at each sampling point in the inlet flue and the outlet flue using a TDLAS analyzer.

[0009] Preferably, in step S2, the rate of change of the total NOx in the inlet flue includes a first-order rate of change and a second-order rate of change.

[0010] Preferably, in step S3, the ammonia escape field uniformity index is the ratio of the standard deviation to the average value of the ammonia escape concentration in each zone of the outlet flue; when the ammonia escape field uniformity index of a certain zone exceeds a set threshold, the ammonia injection weight of that zone is reduced, and the ammonia injection weight of adjacent zones is increased, so that the ammonia escape concentration in each zone tends to be uniform.

[0011] Preferably, in steps S1 to S4, the partitions of the inlet flue, the partitions of the outlet flue, and the partitions of the ammonia injection grid correspond in number and relationship.

[0012] Preferably, in step S4, the method for dynamically allocating ammonia is to take the current ammonia injection amount, inlet flue gas temperature, oxygen content, feedforward signal and feedback signal as inputs, and take the NOx concentration, total ammonia escape, ammonia escape field uniformity index and total ammonia injection amount of each zone of the outlet flue gas as targets to be close to preset values, and to solve the ammonia injection amount of each zone based on model predictive control or machine learning methods for rolling optimization.

[0013] Preferably, in step S5, the nonlinear factors include changes in coal type and / or catalyst degradation.

[0014] Preferably, the specific method of step S6 is to monitor the flue gas temperature at the inlet of the SCR reactor in real time, and immediately cut off all ammonia injection in all zones when the temperature is too high or too low.

[0015] Preferably, the hardware system required by the algorithm includes: SCR reactor, used for flue gas denitrification; The TDLAS analyzer is installed in each section of the inlet flue and the outlet flue to collect NOx concentration and flue gas velocity in each section of the inlet flue in real time, as well as NOx and ammonia concentration in each section of the outlet flue. The feedforward calculation module, connected to the MPC controller and the TDLAS analyzer in the inlet flue, is used to calculate the total amount of NOx and the rate of change of the total amount of NOx based on the NOx concentration and flue gas velocity in the inlet flue, and transmits the NOx concentration, the total amount of NOx, and the rate of change of the total amount of NOx as feedforward signals to the MPC controller. The feedback calculation module, connected to the MPC controller and the TDLAS analyzer in the outlet flue, is used to calculate the ammonia slip field uniformity index based on the ammonia slip concentration in the outlet flue, and transmits the NOx concentration, ammonia slip concentration and ammonia slip field uniformity index as feedback signals to the MPC controller. The MPC controller is connected to the ammonia injection actuator and is used to dynamically allocate the ammonia quantity and transmit the ammonia injection command to the ammonia injection actuator. The ammonia injection actuator, including a high-precision regulating valve, an ammonia injection assembly, and a mass flow meter, is used to inject ammonia gas into the SCR reactor. The safety interlock module connects the TDLAS analyzer and MPC controller in the inlet flue, and is used to cut off the ammonia injection command and send alarm information in an emergency. The display platform is used to display system data.

[0016] The technical solution of this invention has the following beneficial effects: This invention achieves rapid and accurate perception of flue gas composition through TDLAS, providing high-quality input signals for the algorithm; the algorithm integrates feedforward and feedback information to generate optimal ammonia injection commands for different zones, driving the actuators to move precisely; the clean flue gas output by the system is monitored again, forming a closed-loop feedback to ensure control accuracy; the material shortage protection mechanism ensures reliable operation of the system under abnormal conditions. Attached Figure Description

[0017] Figure 1 This is a flowchart of the algorithm of the present invention; Figure 2 This is a functional block diagram of the system of the present invention; Figure 3 This is a schematic diagram illustrating the working principle of the present invention. Detailed Implementation

[0018] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0019] Reference Figures 1 to 3 This invention provides an intelligent control algorithm for SCR zone ammonia injection based on dynamic compensation for ammonia slip, comprising the following steps: S1: Multiple sampling points are set up in the inlet and outlet flues of the SCR reactor to collect NOx concentration and flue gas velocity in each section of the inlet flue and NOx and ammonia concentration in each section of the outlet flue in real time. S2: Calculate the total amount of NOx and the rate of change of the total amount of NOx based on the NOx concentration and flue gas velocity in the inlet flue. Use the NOx concentration, the total amount of NOx, and the rate of change of the total amount of NOx as feedforward signals. Calculate the theoretically required total amount of ammonia injection based on the total amount of NOx in the inlet flue and the target denitrification efficiency. Add the rate of change compensation to the theoretical total amount of ammonia injection to obtain the feedforward ammonia injection amount. S3: Calculate the ammonia slip field uniformity index based on the ammonia slip concentration in the outlet flue, use the NOx concentration, ammonia slip concentration and ammonia slip field uniformity index as feedback signals, and dynamically compensate for the feedforward bias based on the feedback signals. S4: Divide the ammonia injection grid into multiple independent control zones and dynamically allocate the ammonia quantity based on the feedforward ammonia injection quantity, feedforward signal and feedback signal of each zone; S5: Adaptive adjustment of nonlinear factors using model predictive control or machine learning methods; S6: Determine whether to trigger the feed cut-off protection based on the catalyst temperature.

[0020] in: In step S1, the method for setting sampling points is to divide the cross-section of the inlet flue and the cross-section of the outlet flue into multiple zones, with each zone serving as a sampling point. Each sampling point is measured and sampled in real time and continuously using a TDLAS analyzer, which greatly shortens the response time, eliminates the lag problem of traditional sampling systems, and achieves accurate perception of the concentration distribution across the entire cross-section. In step S2, the rate of change of the total NOx in the inlet flue includes the first-order rate of change and the second-order rate of change. The larger the first-order rate of change, the higher the total NOx is, and the more ammonia injection is needed. The larger the second-order rate of change, the faster the total NOx is increasing, and the more ammonia injection is needed. Using the total NOx in the inlet flue and its rate of change as feedforward input, the required ammonia injection can be predicted one transmission delay time in advance, achieving "zero overshoot" control when the load changes abruptly. In step S3, the uniformity index of ammonia slip field is the ratio of the standard deviation to the average value of ammonia slip concentration in each zone of the outlet flue. When the uniformity index of ammonia slip field in a certain zone exceeds the set threshold, the ammonia injection weight of that zone is reduced and the ammonia injection weight of the adjacent zones is increased so that the ammonia slip concentration in each zone tends to be uniform. In step S4, the method for dynamically allocating ammonia is to use the feedforward ammonia injection rate, inlet flue gas temperature, oxygen content, feedforward signal, and feedback signal as inputs, and output them in advance before the flue gas actually reaches the catalyst to overcome transmission delay. The goal is to make the NOx concentration, total ammonia slip, ammonia slip field uniformity index, and total ammonia injection rate of each zone of the outlet flue close to preset values. The ammonia injection rate of each zone is solved by rolling optimization based on model predictive control or machine learning methods. By using the NOx total change rate to compensate in advance, the outlet NOx can be reduced to no overshoot or minimal overshoot during sudden load changes. In steps S1 to S4, the partitions of the inlet flue, the partitions of the outlet flue, and the partitions of the ammonia injection grid correspond in number and relationship. By measuring the partitions and implementing the closed loop, the uniformity of ammonia injection is improved, and the reducing agent is effectively saved. In step S5, nonlinear factors include changes in coal type and catalyst decay. Adaptive adjustment of nonlinear factors can help improve ammonia utilization and system stability. The specific method of step S6 is to monitor the flue gas temperature at the inlet of the SCR reactor in real time. When the temperature is too high or too low (such as below 300℃ or above 420℃), the ammonia injection is automatically cut off outside the catalyst activity window, and an audible and visual alarm is issued to prevent catalyst poisoning and ammonia waste, and to ensure safety.

[0021] The hardware system required by the algorithm includes: SCR reactor 1 is used for flue gas denitrification; The TDLAS analyzer 2 in the inlet flue and the TDLAS analyzer 3 in the outlet flue are respectively set in each section of the inlet flue and the outlet flue to collect the NOx concentration and flue gas velocity of each section of the inlet flue and the NOx and ammonia concentration of each section of the outlet flue in real time. The feedforward calculation module 4 is connected to the MPC controller 6 and the TDLAS analyzer 2 in the inlet flue. It is used to calculate the total amount of NOx and the rate of change of the total amount of NOx based on the NOx concentration and flue gas velocity in the inlet flue, and transmit the NOx concentration, the total amount of NOx, and the rate of change of the total amount of NOx as feedforward signals to the MPC controller. Feedback calculation module 5, connected to MPC controller 6 and TDLAS analyzer 3 in outlet flue, is used to calculate the uniformity index of ammonia escape field based on the ammonia escape concentration in outlet flue, and transmits the NOx concentration, ammonia escape concentration and ammonia escape field uniformity index as feedback signals to MPC controller. MPC controller 6 is connected to ammonia injection actuator 7. It is used to calculate the feedforward ammonia injection amount and dynamically allocate ammonia amount, and transmit the ammonia injection command to the ammonia injection actuator. The ammonia injection actuator 7, including a high-precision regulating valve, an ammonia injection assembly, and a mass flow meter, is used to inject ammonia into the SCR reactor, achieving millisecond-level response and accurate metering. It optimizes the nozzle arrangement density to 10–12 nozzles / m², ensuring uniform spatial distribution of ammonia, and works with a static mixer to enhance the mixing effect of ammonia and flue gas. Safety interlock module 8 connects the TDLAS analyzer 2 and MPC controller 6 in the inlet flue, and is used to cut off the ammonia injection command in an emergency and send alarm information; Display platform 9 is used to display system data, including key parameters such as NOx concentration cloud map, ammonia escape trend, branch opening and flow rate.

[0022] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. All equivalent structural transformations made under the inventive concept of the present invention using the contents of the present invention specification and drawings, or direct / indirect applications in other related technical fields, are included within the protection scope of the present invention.

Claims

1. A smart control algorithm for SCR zone ammonia injection based on dynamic compensation for ammonia slip, characterized in that, Includes the following steps: S1: Multiple sampling points are set up in the inlet and outlet flues of the SCR reactor to collect NOx concentration and flue gas velocity in each section of the inlet flue and NOx and ammonia concentration in each section of the outlet flue in real time. S2: Calculate the total amount of NOx and the rate of change of the total amount of NOx based on the NOx concentration and flue gas velocity in the inlet flue. Use the NOx concentration, the total amount of NOx, and the rate of change of the total amount of NOx as feedforward signals. Calculate the theoretically required total amount of ammonia injection based on the total amount of NOx in the inlet flue and the target denitrification efficiency. Add the rate of change compensation to the theoretical total amount of ammonia injection to obtain the feedforward ammonia injection amount. S3: Calculate the ammonia slip field uniformity index based on the ammonia slip concentration in the outlet flue, use the NOx concentration, ammonia slip concentration and ammonia slip field uniformity index as feedback signals, and dynamically compensate for the feedforward bias based on the feedback signals. S4: Divide the ammonia injection grid into multiple independent control zones and dynamically allocate the ammonia quantity based on the feedforward ammonia injection quantity, feedforward signal and feedback signal of each zone; S5: Adaptive adjustment of nonlinear factors using model predictive control or machine learning methods; S6: Determine whether to trigger the feed cut-off protection based on the catalyst temperature.

2. The intelligent control algorithm for SCR zone ammonia injection based on dynamic compensation for ammonia slip according to claim 1, characterized in that, In step S1, the method for setting sampling points is to divide the cross-section of the inlet flue and the cross-section of the outlet flue into multiple partitions, with each partition serving as a sampling point.

3. The SCR zone ammonia injection intelligent control algorithm based on ammonia slip dynamic compensation according to claim 1, characterized in that, In step S1, samples are taken at each sampling point in the inlet flue and the outlet flue using a TDLAS analyzer.

4. The SCR zone ammonia injection intelligent control algorithm based on ammonia slip dynamic compensation according to claim 1, characterized in that, In step S2, the rate of change of the total amount of NOx in the inlet flue includes the first-order rate of change and the second-order rate of change.

5. The SCR zone ammonia injection intelligent control algorithm based on ammonia slip dynamic compensation according to claim 1, characterized in that, In step S3, the uniformity index of ammonia escape field is the ratio of the standard deviation to the average value of ammonia escape concentration in each zone of the outlet flue. When the uniformity index of ammonia escape field in a certain zone exceeds a set threshold, the ammonia injection weight of that zone is reduced and the ammonia injection weight of adjacent zones is increased so that the ammonia escape concentration in each zone tends to be uniform.

6. The intelligent control algorithm for SCR zone ammonia injection based on dynamic compensation for ammonia slip according to claim 1, characterized in that, In steps S1 to S4, the partitions of the inlet flue, the partitions of the outlet flue, and the partitions of the ammonia injection grid correspond in number and relationship.

7. The SCR zone ammonia injection intelligent control algorithm based on ammonia slip dynamic compensation according to claim 1, characterized in that, In step S4, the method for dynamically allocating ammonia is to take the current ammonia injection amount, inlet flue gas temperature, oxygen content, feedforward signal and feedback signal as inputs, and take the NOx concentration, total ammonia escape, ammonia escape field uniformity index and total ammonia injection amount of each zone of the outlet flue as targets to be close to preset values. The ammonia injection amount of each zone is solved by rolling optimization based on model predictive control or machine learning methods.

8. The intelligent control algorithm for SCR zone ammonia injection based on dynamic compensation for ammonia slip according to claim 1, characterized in that, In step S5, nonlinear factors include changes in coal type and / or catalyst degradation.

9. The SCR zone ammonia injection intelligent control algorithm based on ammonia slip dynamic compensation according to claim 1, characterized in that, The specific method of step S6 is to monitor the flue gas temperature at the inlet of the SCR reactor in real time, and immediately cut off the ammonia injection in all zones when the temperature is too high or too low.

10. The SCR zone ammonia injection intelligent control algorithm based on ammonia slip dynamic compensation according to claim 1, characterized in that, The hardware system required by the algorithm includes: SCR reactor, used for flue gas denitrification; The TDLAS analyzer is installed in each section of the inlet flue and the outlet flue to collect NOx concentration and flue gas velocity in each section of the inlet flue in real time, as well as NOx and ammonia concentration in each section of the outlet flue. The feedforward calculation module, connected to the MPC controller and the TDLAS analyzer in the inlet flue, is used to calculate the total amount of NOx and the rate of change of the total amount of NOx based on the NOx concentration and flue gas velocity in the inlet flue, and transmits the NOx concentration, the total amount of NOx, and the rate of change of the total amount of NOx as feedforward signals to the MPC controller. The feedback calculation module, connected to the MPC controller and the TDLAS analyzer in the outlet flue, is used to calculate the ammonia slip field uniformity index based on the ammonia slip concentration in the outlet flue, and transmits the NOx concentration, ammonia slip concentration and ammonia slip field uniformity index as feedback signals to the MPC controller. The MPC controller is connected to the ammonia injection actuator and is used to dynamically allocate the ammonia quantity and transmit the ammonia injection command to the ammonia injection actuator. The ammonia injection actuator, including a high-precision regulating valve, an ammonia injection assembly, and a mass flow meter, is used to inject ammonia gas into the SCR reactor. The safety interlock module connects the TDLAS analyzer and MPC controller in the inlet flue, and is used to cut off the ammonia injection command and send alarm information in an emergency. The display platform is used to display system data.

Images