A denitration control method of in-situ NOx actual measurement feedforward and collaborative prediction feedback

Abstract

The application provides a denitration control method of in-situ NOx actual measurement feedforward and collaborative prediction feedback. The application takes the in-situ NOx concentration at the furnace outlet as a feedforward signal, takes the in-situ NOx concentration at the SCR outlet as a feedback signal, calculates a basic compensation value of ammonia injection based on the in-situ NOx concentration at the furnace outlet, establishes a nonlinear dynamic prediction model and rolls optimization to output a prediction correction amount, takes the set value of the outlet NOx as a target to perform closed-loop fine tuning and output a feedback correction amount, and superimposes the three to generate an ammonia injection adjustment instruction; the method realizes disturbance advance compensation, dynamic optimization and accurate correction, improves the response speed and control precision of the denitration system, reduces ammonia consumption and ammonia escape, adapts to large delay and nonlinear time-varying characteristics, and guarantees that the outlet NOx meets the standard stably.

Description

Technical Field

[0001] This invention relates to the field of denitrification control technology, and in particular to a denitrification control method based on in-situ NOx measurement feedforward and synergistic predictive feedback. Background Technology

[0002] Selective catalytic reduction (SCR) denitrification technology plays a crucial role in nitrogen oxide emission control, serving as a core technology for achieving ultra-low emissions in coal-fired power plants. With increasingly stringent national environmental policies, ensuring stable NOx concentration at the denitrification system outlet and optimizing ammonia consumption have become critical issues for the safe and economical operation of coal-fired power plants. Among related technologies, a cascade control system based on PID control has been constructed through the coordinated operation of operating parameters such as unit load and total coal consumption with feedback adjustment of outlet NOx concentration. Specifically, this control system covers the entire process from flue gas parameter acquisition and ammonia injection calculation to actuator action, including key aspects such as feedforward signal acquisition, feedback deviation calculation, and control command output.

[0003] However, existing denitrification control methods directly use indirect parameters such as load and mill combination as feedforwards, without directly measuring the original amount of NOx generated at the furnace outlet. This may lead to a serious lag in the feedforward response, making it impossible to capture major disturbances such as sudden changes in coal quality in a timely manner, or causing over- or under-adjustment due to the complexity of the object characteristics. This affects the control accuracy and stability of the outlet NOx concentration, while also increasing the risk of ammonia escape and operating costs. Summary of the Invention

[0004] The present invention aims to at least partially solve one of the technical problems in the related art.

[0005] Therefore, the first objective of this invention is to propose a denitrification control method based on in-situ measured NOx feedforward and synergistic predictive feedback.

[0006] Another objective of this invention is to propose a denitrification control device based on in-situ NOx measured feedforward and synergistic predictive feedback.

[0007] The third objective of this invention is to provide a computer device.

[0008] A fourth objective of this invention is to provide a non-transitory computer-readable storage medium.

[0009] To achieve the above objectives, a first aspect of the present invention proposes a denitrification control method based on in-situ measured NOx feedforward and synergistic predictive feedback, comprising:

[0010] S1, collect the in-situ NOx concentration at the furnace outlet as a feedforward measurement signal, and collect the in-situ NOx concentration at the SCR outlet as a feedback measurement signal. S2, Based on the feedforward measurement signal, calculate the basic compensation value of the ammonia injection adjustment amount according to the preset dynamic feedforward function; S3. Based on historical and real-time data, a nonlinear dynamic prediction model is established. The ammonia injection rate at future moments is optimized online using the basic compensation value as a benchmark, and the prediction correction amount is output. S4. Using the SCR outlet NOx concentration setpoint as the target and the measured SCR outlet NOx concentration as the feedback quantity, perform closed-loop fine-tuning, output feedback correction quantity, and superimpose the basic compensation value, predicted correction quantity, and feedback correction quantity to generate the final ammonia injection adjustment command.

[0011] In one embodiment of the present invention, the step of collecting the in-situ NOx concentration at the furnace outlet as a feedforward measurement signal and collecting the in-situ NOx concentration at the SCR outlet as a feedback measurement signal includes: The original NOx concentration at the furnace outlet is collected in real time by an in-situ NOx online analyzer installed at the furnace outlet or SCR inlet flue. The NOx concentration at the SCR outlet is collected in real time using an in-situ online NOx analyzer installed at the SCR outlet. Synchronously collect sensor data on process parameters such as unit load, flue gas flow rate, and flue gas temperature.

[0012] In one embodiment of the present invention, the step of calculating the basic compensation value of the ammonia injection adjustment amount based on the feedforward measurement signal and according to a preset dynamic feedforward function includes: When a change in NOx generation is detected due to sudden changes in coal quality or load fluctuations, the basic compensation value for the ammonia injection adjustment is calculated based on the preset dynamic feedforward function or intelligent feedforward model. The calculation weight of the basic compensation value is dynamically adjusted based on the rate of change of NOx concentration at the furnace outlet to proactively overcome inlet disturbances.

[0013] In one embodiment of the present invention, the dynamic feedforward function takes the following form:

[0014] in, This is the base compensation value at time k. The deviation of the NOx concentration at the furnace outlet from the baseline value. The rate of change in NOx concentration. and These are dynamic feedforward coefficients.

[0015] In one embodiment of the present invention, the step of establishing a nonlinear dynamic prediction model based on historical data and real-time data, using the basic compensation value as a benchmark to continuously optimize the ammonia injection rate at future times, and outputting a prediction correction amount includes: A nonlinear dynamic prediction model is established by comprehensively considering multiple factors such as unit load, flue gas flow rate, and flue gas temperature. Based on the aforementioned basic compensation value, the ammonia injection rate at future times is continuously optimized within the prediction time domain. It provides auxiliary compensation for model mismatch or unmodeled dynamic disturbances and outputs the prediction correction.

[0016] In one embodiment of the present invention, the rolling optimization strategy for the ammonia injection rate at future times in the prediction time domain is adopted, and the ammonia injection rate setting value at several future times is updated based on the latest prediction model output and measured data in each control cycle.

[0017] In one embodiment of the present invention, the nonlinear dynamic prediction model adopts the following state-space form:

[0018]

[0019] in, For state variables, To control the input, For measurable disturbances, For system output, and It is a nonlinear function.

[0020] In one embodiment of the present invention, the step of using the SCR outlet NOx concentration setpoint as the target and the measured SCR outlet NOx concentration as the feedback quantity for closed-loop fine-tuning, and outputting a feedback correction quantity, includes: When the NOx concentration at the SCR outlet deviates from the set value in a steady state, an adaptive PID controller or other zero steady-state error regulator is used for fine closed-loop adjustment. The output feedback correction value is used to eliminate steady-state deviation and ensure that the NOx concentration at the SCR outlet is accurately and stably maintained near the set value.

[0021] In one embodiment of the present invention, the step of superimposing the basic compensation value, the predicted correction amount, and the feedback correction amount to generate the final ammonia injection adjustment command includes: The basic compensation value output by the feedforward controller, the prediction correction value output by the prediction controller, and the feedback correction value output by the feedback controller are weighted and summed. The control variables are dynamically adjusted according to their weight coefficients to achieve a coordinated control effect where feedforward is responsible for speed, prediction is responsible for quality, and feedback is responsible for accuracy.

[0022] To achieve the above objectives, a second aspect of the present invention provides a denitrification control device based on in-situ NOx measured feedforward and synergistic predictive feedback, comprising: The data acquisition module is used to acquire the in-situ NOx concentration at the furnace outlet as a feedforward measurement signal and to acquire the in-situ NOx concentration at the SCR outlet as a feedback measurement signal. The feedforward calculation module is used to calculate the basic compensation value of the ammonia injection adjustment amount based on the feedforward measurement signal and according to the preset dynamic feedforward function. The prediction and optimization module is used to establish a nonlinear dynamic prediction model based on historical and real-time data, and to continuously optimize the ammonia injection amount at future moments online based on the basic compensation value, and output the prediction correction amount. The closed-loop control module is used to perform closed-loop fine-tuning with the SCR outlet NOx concentration setpoint as the target and the measured SCR outlet NOx concentration as the feedback quantity. It outputs a feedback correction quantity and superimposes the basic compensation value, the predicted correction quantity, and the feedback correction quantity to generate the final ammonia injection control command.

[0023] This invention discloses a denitrification control method and apparatus based on in-situ measured NOx feedforward and collaborative predictive feedback. By introducing the measured in-situ NOx value at the furnace outlet as the core feedforward and combining it with a three-level collaborative architecture of predictive control and feedback control, it achieves advanced pre-adjustment and precise control of ammonia injection quantity, effectively solving the problems of lag response and insufficient control precision in traditional denitrification systems.

[0024] To achieve the above objectives, a third aspect of this application provides a computer device, including a processor and a memory; wherein the processor runs a program corresponding to the executable program code by reading executable program code stored in the memory, for implementing the method described in the first aspect embodiment.

[0025] To achieve the above objectives, a fourth aspect of this application provides a non-transitory computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method described in the first aspect.

[0026] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0027] Figure 1 This is a flowchart of a denitrification control method based on in-situ NOx measured feedforward and synergistic predictive feedback according to an embodiment of the present invention; Figure 2 This is an architecture diagram of an in-situ NOx measured feedforward and collaborative predictive feedback denitrification control system according to an embodiment of the present invention; Figure 3 This is a structural diagram of a denitrification control device based on in-situ NOx measurement feedforward and collaborative predictive feedback according to an embodiment of the present invention. Detailed Implementation

[0028] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0029] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0030] The following description, with reference to the accompanying drawings, describes an in-situ NOx measured feedforward and synergistic predictive feedback denitrification control method and apparatus according to an embodiment of the present invention.

[0031] Example 1 Figure 1 This is a flowchart of an in-situ NOx measured feedforward and synergistic predictive feedback denitrification control method according to an embodiment of the present invention, as shown below. Figure 1 and Figure 2 As shown, it includes: S1, collect the in-situ NOx concentration at the furnace outlet as a feedforward measurement signal, and collect the in-situ NOx concentration at the SCR outlet as a feedback measurement signal. Specifically, in intelligent denitrification control systems, data acquisition is a fundamental step towards achieving precise control. The system uses an in-situ NOx online analyzer located at the furnace outlet or SCR inlet flue to collect the NOx concentration at the furnace outlet in real time, serving as the core feedforward measurement signal to directly sense the initial NOx generation during combustion within the furnace. Simultaneously, an in-situ NOx online analyzer located at the SCR outlet collects the NOx concentration at the SCR reactor outlet in real time, serving as a feedback measurement signal to monitor the actual emission concentration after denitrification treatment. These feedforward and feedback measurement signals together constitute the dual-channel measurement architecture of the control system, providing real-time and accurate input data for subsequent feedforward compensation, predictive optimization, and feedback correction. As a specific implementation, the in-situ NOx online analyzer at the furnace outlet can employ ultraviolet differential absorption spectroscopy or laser scattering technology to achieve continuous online monitoring of NOx concentration in the flue gas. Similarly, the in-situ NOx online analyzer at the SCR outlet can employ corresponding high-precision measurement technologies to ensure the reliability and real-time nature of the measurement data.

[0032] By using an in-situ online NOx analyzer to directly measure the NOx concentration at the furnace outlet and SCR outlet, the system can obtain real-time data on changes in combustion conditions and denitrification effects in the first instance. This avoids the lag and inaccuracy problems caused by traditional indirect parameter measurements, laying a data foundation for achieving rapid response and precise control.

[0033] S2, Based on the feedforward measurement signal, calculate the basic compensation value of the ammonia injection adjustment amount according to the preset dynamic feedforward function; Specifically, the method for calculating the basic compensation value of ammonia injection adjustment based on the in-situ NOx concentration at the furnace outlet relies on using the real-time NOx concentration signal collected by an in-situ NOx online analyzer located at the furnace outlet or SCR inlet flue as the core input for feedforward control. The feedforward controller establishes a mapping relationship from the furnace outlet NOx concentration to the ammonia injection adjustment amount through a preset dynamic feedforward function. Based on the real-time detected changes in NOx concentration, it calculates the basic compensation value of the ammonia injection amount in advance to overcome the large delay and inertia characteristics of the denitrification system. This dynamic feedforward function can be an empirical model built based on the process mechanism or an intelligent feedforward model trained using historical operating data. Its design needs to comprehensively consider the influence of multiple factors such as unit load, coal quality changes, and flue gas flow on the NOx generation law, thereby achieving rapid prediction and advance compensation for inlet disturbances. As a specific implementation method, when the feedforward controller detects a significant change in the furnace outlet NOx concentration due to sudden changes in coal quality or load, it can immediately output the corresponding basic compensation value of the ammonia injection adjustment amount according to the preset dynamic feedforward function, thus overcoming inlet disturbances in advance.

[0034] Furthermore, when a change in NOx generation is detected due to abrupt changes in coal quality or load variations, the basic compensation value for ammonia injection adjustment is calculated based on a preset dynamic feedforward function or intelligent feedforward model. Specifically, the in-situ NOx online analyzer at the furnace outlet in the data acquisition layer outputs the current NOx concentration measurement value in real time. The system compares this measurement value with a preset benchmark NOx concentration value to obtain the concentration deviation. Simultaneously, the system calculates the derivative of this concentration deviation with respect to time to obtain the NOx concentration change rate. In one possible implementation, the dynamic feedforward function takes the following form: ,in This is the base compensation value at time k. and This is the dynamic feedforward coefficient. This basic compensation value is used as the output of the feedforward controller and transmitted to the signal superposition and processing unit.

[0035] Furthermore, the calculation weight of the basic compensation value is dynamically adjusted based on the rate of change of NOx concentration at the furnace outlet, thereby proactively overcoming inlet disturbances. Specifically, when the NOx concentration change rate... A larger value indicates that the ingress disturbance is changing rapidly, at which point the system automatically increases the dynamic feedforward coefficient. The weighting of the values ​​increases the predictive component in the base compensation value, allowing for earlier adjustments to ammonia injection; when the NOx concentration change rate plateaus, the system appropriately increases the weighting. The weighting of the parameters makes the basic compensation value focus more on the static compensation of the current deviation. This dynamic weighting adjustment mechanism based on the rate of change enables the feedforward controller to proactively adjust the compensation strategy according to the development trend of the disturbance, truly achieving proactive overcoming of the ingress disturbance.

[0036] This invention directly uses the in-situ measured NOx value at the furnace outlet as the feedforward signal, abandoning the traditional feedforward method that relies on indirect parameters such as load and coal quantity. This enables the system to sense the real changes in inlet disturbances in real time and act in advance, significantly improving the response speed of the denitrification system to inlet disturbances and effectively solving the technical problems of lagging and inaccurate feedforward response in traditional control.

[0037] S3. Based on historical and real-time data, a nonlinear dynamic prediction model is established. The ammonia injection rate at future moments is optimized online using the basic compensation value as a benchmark, and the prediction correction amount is output. Furthermore, a nonlinear dynamic prediction model is established based on historical and real-time data. Using a baseline compensation value as a benchmark, the ammonia injection rate for future moments is optimized online in a rolling manner, outputting a prediction correction. This step corresponds to the second level of the intelligent control layer, namely the auxiliary prediction optimization stage. The predictive controller is connected to both the data acquisition layer and the signal overlay unit. Its core function is to establish a prediction model that reflects the dynamic characteristics of the SCR denitrification process. This model comprehensively considers multiple factors such as unit load, flue gas flow, and temperature, and optimizes the ammonia injection rate for future moments online in a rolling manner based on the feedforward compensation calculated at the first level. When there is a mismatch between the prediction model and the actual situation, or when there are unmodeled random disturbances, the predictive controller outputs a prediction correction to compensate for the aforementioned deviations, thereby continuously correcting the prediction trajectory and improving control accuracy during the rolling optimization process. As a specific implementation method, the predictive controller can adopt a prediction model based on a nonlinear model predictive control algorithm. The parameter structure of the model is trained using historical operating data, and the model is updated in conjunction with real-time acquired process data. Within each control cycle, the optimal ammonia injection rate for multiple future moments is solved in a rolling manner.

[0038] Furthermore, the predictive controller comprehensively considers multiple factors such as unit load, flue gas flow rate, and flue gas temperature to establish a nonlinear dynamic prediction model. Specifically, the predictive controller acquires real-time unit load signals, flue gas flow rate sensor measurements, and flue gas temperature sensor measurements from the data acquisition layer, while also retrieving historical operating data as modeling data support. Based on these multidimensional input parameters, the predictive controller constructs a nonlinear dynamic prediction model in a state-space form, where the state variables... This includes key states reflecting the dynamic characteristics of the system, such as catalyst activity state and ammonia storage state; control inputs. This represents the current ammonia injection rate; measurable disturbances. Including unit load, flue gas flow rate, flue gas temperature, etc.; system output The predicted NOx concentration at the SCR outlet. The model's state transition function. and output function It is implemented by using neural networks or nonlinear autoregressive models trained on historical data to describe the nonlinear dynamic characteristics of the SCR denitrification process.

[0039] Furthermore, the predictive controller uses the aforementioned basic compensation value as a benchmark to continuously optimize the ammonia injection rate for future times within the prediction time domain. Specifically, the predictive controller uses the basic compensation value as a feedforward benchmark input to the prediction model and continuously calculates the ammonia injection rate setpoint for each future time within the prediction time domain N. Within each control cycle, the predictive controller updates the ammonia injection rate setpoint for several future times using a continuous optimization strategy based on the latest prediction model output and measured data. The continuous optimization performance index comprehensively considers the deviation between the predicted output and the target value, the amplitude of ammonia injection rate changes, and control smoothness. It solves for the optimal ammonia injection rate sequence for the next N times using an optimization method within a finite time domain, and outputs the first value of this sequence as a prediction correction to the signal superposition and processing unit.

[0040] The predictive controller provides auxiliary compensation for model mismatch or unmodeled dynamic disturbances, outputting a prediction correction. When there is a deviation between the model's predicted output and the actual SCR outlet NOx concentration, the predictive controller calculates the prediction error and introduces an error correction term. It then adjusts the model parameters online or through adaptive gain to compensate for unmodeled dynamic disturbances such as sudden changes in coal quality and catalyst activity. Finally, the prediction correction is superimposed on the basic compensation value and sent to the signal superposition and processing unit to achieve coordinated optimization control of the ammonia injection rate.

[0041] This technical step, by introducing a rolling optimization mechanism of predictive control, can effectively compensate for model mismatch and unmodeled dynamic disturbances, improve the system's adaptability to complex operating conditions, and work in synergy with feedforward control to further optimize the dynamic adjustment performance of ammonia injection.

[0042] S4. Using the SCR outlet NOx concentration setpoint as the target and the measured SCR outlet NOx concentration as the feedback quantity, perform closed-loop fine-tuning, output feedback correction quantity, and superimpose the basic compensation value, predicted correction quantity, and feedback correction quantity to generate the final ammonia injection adjustment command.

[0043] Specifically, in the third-level collaborative control, a feedback control strategy is employed to finely correct the ammonia injection rate after feedforward compensation and prediction optimization, ensuring that the outlet NOx concentration can be accurately and stably maintained near the setpoint. The feedback controller uses the SCR outlet NOx concentration setpoint as the control target and the measured concentration collected by the in-situ NOx online analyzer installed at the SCR outlet flue as the feedback signal. By comparing the deviation between the setpoint and the measured value, it performs closed-loop fine-tuning calculations according to a preset control law and outputs the feedback correction. This feedback control loop primarily eliminates steady-state deviations. When the system experiences long-term deviations between the outlet NOx concentration and the setpoint due to factors such as model mismatch, unmodeled disturbances, or changes in catalyst activity, the feedback controller can continuously perform fine-tuning corrections to ensure control accuracy. The feedback controller can employ an adaptive PID control algorithm or other control algorithms with zero steady-state error to adapt to the nonlinear and time-varying characteristics of the SCR denitrification system. The signal superposition and processing unit receives the basic compensation value from the feedforward controller, the predictive correction value from the predictive controller, and the feedback correction value from the feedback controller. It then integrates these values ​​according to a preset collaborative superposition rule to generate the final ammonia injection regulation command, which is then sent to the actuator layer. Through this collaborative superposition of feedforward, predictive, and feedback control quantities, the system can achieve a combination of rapid response, rolling optimization, and precise correction when facing complex disturbances.

[0044] This technical step effectively eliminates the steady-state deviation of the system through fine closed-loop adjustment of the feedback control loop, ensuring that the outlet NOx concentration can remain stable within the set value range for a long time, and improving the steady-state accuracy and robustness of the control system.

[0045] This invention proposes an iSCR intelligent denitrification control system based on in-situ NOx measured feedforward and collaborative predictive feedback. Figure 2 As shown, it includes a data acquisition layer, an intelligent control layer, and an actuator layer. The data acquisition layer includes an in-situ NOx online analyzer (core feedforward measurement) installed at the furnace outlet or SCR inlet flue, an in-situ NOx online analyzer at the SCR outlet (feedback measurement), and process parameter sensors. The intelligent control layer includes a feedforward controller, a predictive controller, a feedback controller, and a signal superposition and processing unit. The intelligent control layer adopts a three-level collaborative control architecture. The first level (core feedforward compensation): The feedforward controller is directly connected to the in-situ NOx online analyzer at the furnace outlet to sense changes in the original NOx generation at the furnace outlet in real time. When a change in NOx generation caused by sudden changes in coal quality or load fluctuations is detected, the feedforward controller immediately calculates the basic compensation value for the ammonia injection adjustment based on a preset dynamic feedforward function or intelligent feedforward model, thus overcoming inlet disturbances and the large delay characteristics of the denitrification system in advance.

[0046] The second level (assisted prediction optimization): The prediction controller is connected to the data acquisition layer and the signal overlay unit respectively, and is used to establish a nonlinear dynamic prediction model based on historical data and real-time data. This model comprehensively considers multi-dimensional factors such as unit load, flue gas flow, and temperature. Based on the feedforward compensation calculated in the first level, it continuously optimizes the ammonia injection amount at future moments online, provides auxiliary compensation for dynamic disturbances that are mismatched or not modeled, and outputs the prediction correction amount.

[0047] The third level (fine feedback correction): The feedback controller is connected to the SCR outlet in-situ NOx online analyzer, using the setpoint of the outlet NOx concentration as the target and the measured outlet NOx concentration as the feedback quantity. When a steady-state deviation occurs between the outlet NOx concentration and the setpoint, the feedback controller (such as an adaptive PID or other zero steady-state error regulator) performs fine closed-loop adjustments and outputs a feedback correction quantity, ultimately ensuring that the outlet NOx concentration is accurately and stably maintained near the setpoint.

[0048] The signal superposition and processing unit organically superimposes the basic compensation value output by the feedforward controller, the prediction correction amount output by the prediction controller, and the feedback correction amount output by the feedback controller to generate the final ammonia injection adjustment command and send it to the actuator layer.

[0049] The embodiments of this invention also have the following technical effects: Truly achieving proactive control: By introducing the in-situ measured NOx value at the furnace outlet as the core feedforward, the traditional feedforward method relying on indirect parameters such as load and coal quantity is abandoned. Regardless of sudden changes in coal quality, the system can "see" the real value of the inlet disturbance in real time, achieving proactive and precise pre-adjustment of the ammonia injection quantity, fundamentally solving the response lag problem of traditional control. Multi-scale collaboration, adaptability to all operating conditions: A three-level collaborative architecture of "feedforward + prediction + feedback" is adopted. Feedforward is responsible for "speed" and solving major disturbances; prediction is responsible for "optimization" and solving dynamic accuracy; feedback is responsible for "accuracy" and eliminating steady-state deviations. The three work together to achieve full-chain control of inlet disturbances from proactive response and rolling optimization to fine correction, significantly improving the system's adaptability and robustness under all operating conditions. Improved economic and environmental performance: By precisely controlling the ammonia injection rate, while ensuring stable NOx compliance at the outlet (improved hourly average compliance rate), ammonia consumption is effectively reduced, ammonia escape is minimized, and the risk of air preheater blockage is avoided, achieving a win-win situation for both environmental protection and economic benefits. High model maintainability: The predictive model can be updated online or offline using actual operating data, and can adapt to changes in characteristics caused by long-term unit operation (such as decreased catalyst activity), demonstrating strong engineering applicability.

[0050] To achieve the above embodiments, such as Figure 3 As shown, this embodiment also provides an in-situ NOx measurement feedforward and collaborative predictive feedback denitrification control device 10, including: The data acquisition module 100 is used to acquire the in-situ NOx concentration at the furnace outlet as a feedforward measurement signal and to acquire the in-situ NOx concentration at the SCR outlet as a feedback measurement signal. The feedforward calculation module 200 is used to calculate the basic compensation value of the ammonia injection adjustment amount based on the feedforward measurement signal and according to the preset dynamic feedforward function. The prediction and optimization module 300 is used to establish a nonlinear dynamic prediction model based on historical data and real-time data, and to optimize the ammonia injection amount at future moments online based on the basic compensation value, and output the prediction correction amount. The closed-loop control module 400 is used to perform closed-loop fine-tuning with the SCR outlet NOx concentration setpoint as the target and the measured SCR outlet NOx concentration as the feedback quantity, output the feedback correction quantity, and superimpose the basic compensation value, the predicted correction quantity, and the feedback correction quantity to generate the final ammonia injection control command.

[0051] This invention discloses an in-situ NOx measured feedforward and collaborative predictive feedback denitrification control device. By introducing the in-situ measured NOx value at the furnace outlet as the core feedforward and combining it with a three-level collaborative architecture of predictive control and feedback control, it achieves advanced pre-adjustment and precise control of ammonia injection quantity, effectively solving the problems of response lag and insufficient control accuracy in traditional denitrification systems.

[0052] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0053] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A denitrification control method based on in-situ measured NOx feedforward and synergistic predictive feedback, characterized in that, include: S1, collect the in-situ NOx concentration at the furnace outlet as a feedforward measurement signal, and collect the in-situ NOx concentration at the SCR outlet as a feedback measurement signal. S2, Based on the feedforward measurement signal, calculate the basic compensation value of the ammonia injection adjustment amount according to the preset dynamic feedforward function; S3. Based on historical and real-time data, a nonlinear dynamic prediction model is established. The ammonia injection rate at future moments is optimized online using the basic compensation value as a benchmark, and the prediction correction amount is output. S4. Using the SCR outlet NOx concentration setpoint as the target and the measured SCR outlet NOx concentration as the feedback quantity, perform closed-loop fine-tuning, output feedback correction quantity, and superimpose the basic compensation value, predicted correction quantity, and feedback correction quantity to generate the final ammonia injection adjustment command.

2. The method as described in claim 1, characterized in that, The method of collecting the in-situ NOx concentration at the furnace outlet as a feedforward measurement signal and collecting the in-situ NOx concentration at the SCR outlet as a feedback measurement signal includes: The original NOx concentration at the furnace outlet is collected in real time by an in-situ NOx online analyzer installed at the furnace outlet or SCR inlet flue. The NOx concentration at the SCR outlet is collected in real time using an in-situ online NOx analyzer installed at the SCR outlet. Synchronously collect sensor data on process parameters such as unit load, flue gas flow rate, and flue gas temperature.

3. The method as described in claim 2, characterized in that, The calculation of the basic compensation value for the ammonia injection adjustment based on the feedforward measurement signal and according to a preset dynamic feedforward function includes: When a change in NOx generation is detected due to sudden changes in coal quality or load fluctuations, the basic compensation value for the ammonia injection adjustment is calculated based on the preset dynamic feedforward function or intelligent feedforward model. The calculation weight of the basic compensation value is dynamically adjusted based on the rate of change of NOx concentration at the furnace outlet to proactively overcome inlet disturbances.

4. The method as described in claim 3, characterized in that, The dynamic feedforward function takes the following form: in, This is the base compensation value at time k. The deviation of the NOx concentration at the furnace outlet from the baseline value. The rate of change in NOx concentration. and These are dynamic feedforward coefficients.

5. The method as described in claim 1, characterized in that, The nonlinear dynamic prediction model established based on historical and real-time data uses the basic compensation value as a benchmark to continuously optimize the ammonia injection rate at future times, and outputs the prediction correction amount, including: A nonlinear dynamic prediction model is established by comprehensively considering multiple factors such as unit load, flue gas flow rate, and flue gas temperature. Based on the aforementioned basic compensation value, the ammonia injection rate at future times is continuously optimized within the prediction time domain. It provides auxiliary compensation for model mismatch or unmodeled dynamic disturbances and outputs the prediction correction.

6. The method as described in claim 5, characterized in that, The rolling optimization strategy for the ammonia injection rate in the future time domain is adopted, which updates the ammonia injection rate setpoint for several future time periods based on the latest prediction model output and measured data in each control cycle.

7. The method as described in claim 5, characterized in that, The nonlinear dynamic prediction model adopts the following state-space form: in, For state variables, To control the input, For measurable disturbances, For system output, and It is a nonlinear function.

8. The method as described in claim 1, characterized in that, The process involves using the SCR outlet NOx concentration setpoint as the target and the measured SCR outlet NOx concentration as the feedback quantity for closed-loop fine-tuning, outputting a feedback correction quantity, including: When the NOx concentration at the SCR outlet deviates from the set value in a steady state, an adaptive PID controller or other zero steady-state error regulator is used for fine closed-loop adjustment. The output feedback correction value is used to eliminate steady-state deviation and ensure that the NOx concentration at the SCR outlet is accurately and stably maintained near the set value.

9. The method as described in claim 1, characterized in that, The step of superimposing the basic compensation value, the predicted correction amount, and the feedback correction amount to generate the final ammonia injection adjustment command includes: The basic compensation value output by the feedforward controller, the prediction correction value output by the prediction controller, and the feedback correction value output by the feedback controller are weighted and summed. The control variables are dynamically adjusted according to their weight coefficients to achieve a coordinated control effect where feedforward is responsible for speed, prediction is responsible for quality, and feedback is responsible for accuracy.

10. A denitrification control device based on in-situ NOx measurement feedforward and collaborative predictive feedback, characterized in that, include: The data acquisition module is used to acquire the in-situ NOx concentration at the furnace outlet as a feedforward measurement signal and to acquire the in-situ NOx concentration at the SCR outlet as a feedback measurement signal. The feedforward calculation module is used to calculate the basic compensation value of the ammonia injection adjustment amount based on the feedforward measurement signal and according to the preset dynamic feedforward function. The prediction and optimization module is used to establish a nonlinear dynamic prediction model based on historical and real-time data, and to continuously optimize the ammonia injection amount at future moments online based on the basic compensation value, and output the prediction correction amount. The closed-loop control module is used to perform closed-loop fine-tuning with the SCR outlet NOx concentration setpoint as the target and the measured SCR outlet NOx concentration as the feedback quantity. It outputs a feedback correction quantity and superimposes the basic compensation value, the predicted correction quantity, and the feedback correction quantity to generate the final ammonia injection control command.

Images