Denitration optimization control method and device based on dynamic ammonia injection

Abstract

The application discloses a denitration optimization control method and device based on dynamic ammonia injection, relates to the technical field of ammonia injection optimization control inspection, and comprises ammonia injection coupling analysis, feedforward prediction correction, escape feedback correction and ammonia injection execution monitoring. According to the application, the characteristic vector of the multiple-node parameters at the inlet of a denitration reactor is constructed and similarity matching is performed, and the basic theoretical ammonia injection amount is calculated in combination with the multiple-parameter coupling analysis; the time-space double trend parameters are extracted based on historical data, the parameter change in the next control cycle is predicted through a trend analysis model, and feedforward correction is performed; the principal component analysis method is used to comprehensively evaluate the multiple-dimensional ammonia escape parameters, and feedback correction is realized through similarity matching of the characteristic vectors; finally, the valve opening is secondarily corrected in combination with the change rates of the ammonia injection amounts in all stages, and the system operation state is monitored in real time. The application can significantly improve the denitration efficiency and reduce the operation cost.

Description

Technical Field

[0001] This invention relates to the field of optimized control and inspection technology for ammonia injection, specifically to a denitrification optimized control method and device based on dynamic ammonia injection. Background Technology

[0002] With increasingly stringent environmental protection requirements, flue gas denitrification technology has become an indispensable environmental protection measure in industries such as thermal power generation, steel, cement, and glass. Selective catalytic reduction denitrification technology has become the most widely used flue gas denitrification technology due to its advantages such as high denitrification efficiency, mature technology, and stable operation. Ammonia injection control is the core link of the denitrification system, which directly affects denitrification efficiency, ammonia slip rate, and system operating costs. Therefore, denitrification optimization control methods and devices based on dynamic ammonia injection are needed.

[0003] In existing technologies, ammonia injection quantity control mostly adopts the traditional PID control method, which only performs single feedback adjustment based on the nitrogen oxide concentration at the reactor outlet.

[0004] The existing technology has the following drawbacks: 1. Traditional PID control relies solely on the nitrogen oxide concentration at the reactor outlet for single feedback adjustment. Due to the significant volumetric inertia and reaction time delay of the denitrification reactor, coupled with the inherent lag in flue gas transmission and sensor response, the control signal often lags behind changes in operating conditions by tens of seconds. When the boiler load fluctuates rapidly, this lag can lead to instantaneous exceedances of nitrogen oxide emissions or a sharp increase in ammonia slip rate, failing to meet increasingly stringent environmental protection requirements and causing ammonia waste and equipment corrosion.

[0005] 2. Existing technologies generally employ a single-parameter step-by-step correction method, independently correcting for temperature, oxygen content, and pressure, completely ignoring the coupling effect between parameters. In reality, temperature changes simultaneously affect the denitrification reaction rate and flue gas oxygen content distribution, while pressure changes alter flue gas volume and reactant concentration. Single corrections cannot accurately reflect this complex interplay, leading to significant deviations in ammonia injection calculations.

[0006] 3. The existing solution only monitors the overall ammonia slip rate, which cannot comprehensively reflect the operating status of the ammonia injection system. The ammonia slip rate only reflects the average ammonia concentration at the reactor outlet and cannot reflect the uniformity of ammonia injection through the injection grid or local ammonia slip issues. When the concentration gradient at the edge is large, the local ammonia slip rate may far exceed the average value, leading to local blockage and corrosion of the air preheater. Monitoring a single parameter cannot detect this potential problem in a timely manner.

[0007] 4. Existing technologies lack a comprehensive operational status monitoring and evaluation mechanism, focusing only on the final control result without real-time monitoring of the ammonia regulating valve's operational status. Long-term valve operation can lead to problems such as characteristic drift, jamming, and wear, resulting in significant deviations between the actual ammonia injection rate and the theoretical value. Traditional methods cannot detect and compensate for these anomalies in a timely manner, severely impacting the long-term stable operation of the system. Summary of the Invention

[0008] To address the aforementioned technical shortcomings, the present invention aims to provide a denitrification optimization control method and apparatus based on dynamic ammonia injection.

[0009] To solve the above technical problems, the present invention adopts the following technical solution: The present invention provides a denitrification optimization control method based on dynamic ammonia injection, including the following steps: Step 1, ammonia injection coupling analysis: collect the current inlet parameters of the denitrification reactor, analyze the current inlet parameters of the denitrification reactor, and obtain the basic theoretical ammonia injection amount.

[0010] Step 2, Feedforward Prediction Correction: Based on the historically collected inlet parameters of the denitrification reactor, obtain the trend data of the inlet parameters of the denitrification reactor, analyze the trend data of the inlet parameters of the denitrification reactor, predict the parameter changes in the next control cycle, and perform feedforward correction on the basic theoretical ammonia injection rate according to the prediction results to obtain the feedforward corrected ammonia injection rate.

[0011] Step 3, Escape Feedback Correction: Obtain multidimensional ammonia escape parameters, analyze the multidimensional ammonia escape parameters, and perform feedback correction on the feedforward correction ammonia injection amount based on the analysis results to obtain the final corrected ammonia injection amount.

[0012] Step 4: Ammonia Injection Execution Monitoring: The final corrected ammonia injection quantity is converted into a control signal to drive the ammonia water regulating valve to adjust its opening. At the same time, the operating status data of the ammonia water regulating valve is monitored in real time, and the operating status data is analyzed to generate a real-time ammonia water control scheme.

[0013] Preferably, the analysis of the current inlet parameters of the denitrification reactor is carried out as follows: The current inlet parameters of the denitrification reactor include the flue gas temperature, oxygen content and flue gas pressure of each node. The temperature feature vector, oxygen content feature vector and pressure feature vector of the current denitrification reactor inlet are obtained by using the preset node order as the feature vector order.

[0014] The similarity between the temperature feature vector at the current inlet of the denitrification reactor and the temperature feature vectors corresponding to each temperature correction coefficient is calculated to obtain the similarity of each temperature correction coefficient at the current inlet of the denitrification reactor. The temperature correction coefficient corresponding to the maximum similarity is selected as the temperature correction coefficient at the current inlet of the denitrification reactor.

[0015] Based on the process of obtaining the temperature correction coefficient, the oxygen content correction coefficient and pressure correction coefficient at the current inlet of the denitrification reactor are obtained.

[0016] Based on the flue gas temperature and oxygen content of each node, mapping table matching and interpolation calculations are performed to obtain the first coupling coefficient of each node. The average value of the first coupling coefficients of all nodes is calculated to obtain the current first coupling correction coefficient. Based on the flue gas temperature and flue gas pressure of each node, coupling analysis is performed to obtain the current second coupling correction coefficient.

[0017] Multiply the current first coupling correction coefficient by the current second coupling correction coefficient to obtain the current comprehensive coupling correction coefficient. Multiply the initial basic ammonia injection quantity by the temperature correction coefficient, oxygen content correction coefficient, pressure correction coefficient and comprehensive coupling correction coefficient in sequence to obtain the basic theoretical ammonia injection quantity.

[0018] Preferably, the feedback correction of the feedforward correction ammonia injection amount is performed as follows: the similarity between the ammonia escape feature vector and the escape feature vector corresponding to each escape correction coefficient is calculated to obtain the similarity of each escape correction coefficient, and the escape correction coefficient corresponding to the maximum similarity is set as the ammonia escape correction coefficient.

[0019] Multiply the feedforward corrected ammonia injection rate by the escape correction factor to obtain the final corrected ammonia injection rate.

[0020] On the other hand, the present invention provides a denitrification optimization control device based on dynamic ammonia injection, including the following modules: an ammonia injection coupling analysis module, used to collect the current inlet parameters of the denitrification reactor, analyze the current inlet parameters of the denitrification reactor, and obtain the basic theoretical ammonia injection rate.

[0021] The feedforward prediction and correction module is used to obtain the trend data of the inlet parameters of the denitrification reactor based on the historically collected inlet parameters, analyze the trend data of the inlet parameters of the denitrification reactor, predict the parameter changes in the next control cycle, and perform feedforward correction on the basic theoretical ammonia injection rate based on the prediction results to obtain the feedforward corrected ammonia injection rate.

[0022] The escape feedback correction module is used to acquire multidimensional ammonia escape parameters, analyze these parameters, and correct the feedforward ammonia injection amount based on the analysis results to obtain the final corrected ammonia injection amount.

[0023] The ammonia injection execution monitoring module is used to convert the final corrected ammonia injection quantity into a control signal to drive the ammonia water regulating valve to adjust its opening. At the same time, it monitors the operating status data of the ammonia water regulating valve in real time, analyzes the operating status data, and generates a real-time ammonia water control scheme.

[0024] The beneficial effects of this invention are as follows: 1. This invention uses a multi-parameter coupling analysis method to calculate the basic theoretical ammonia injection rate, comprehensively considering parameters such as temperature, oxygen content, and pressure, as well as their coupling relationships. By constructing multi-node feature vectors and performing similarity matching to select correction coefficients, the characteristics of the denitrification reaction under different operating conditions can be accurately reflected, thus improving the accuracy of ammonia injection rate calculation.

[0025] 2. This invention employs a feedforward-feedback composite control architecture. By extracting time and spatial trend parameters from historical data, a joint vector input trend analysis model is constructed, enabling the prediction of parameter changes in the next control cycle. This advance feedforward correction alleviates the lag problem of traditional feedback control, shortens system response time, and better adapts to rapidly changing boiler load conditions.

[0026] 3. This invention employs a multidimensional ammonia slip parameter analysis method, comprehensively considering three parameters: ammonia slip rate, ammonia edge concentration gradient, and ammonia jet velocity. Principal component analysis is used to extract a comprehensive evaluation value, constructing an ammonia slip feature vector and selecting feedback correction coefficients through similarity matching. This approach ensures denitrification efficiency while keeping the ammonia slip rate at a low level, thus reducing ammonia water consumption.

[0027] 4. This invention adds comprehensive operation status monitoring and evaluation functions, collects valve hardware status and ammonia injection control process status data in real time, constructs operation status feature vectors and performs similarity matching to evaluate system operation status, and adopts corresponding control strategies according to different operation status levels. It can promptly detect system anomalies and take measures, thereby improving the reliability and safety of system operation.

[0028] 5. This invention adds a secondary correction step for valve opening based on the rate of change of ammonia injection volume at each stage, building upon the traditional valve position-flow velocity mapping. By constructing an opening feature vector and selecting the opening correction coefficient through similarity matching, it can effectively compensate for valve characteristic drift and nonlinear errors, thereby improving valve control accuracy. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a schematic diagram of the implementation steps of the method of the present invention.

[0031] Figure 2 This is a schematic diagram of the device structure of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0033] according to Figure 1 As shown, the present invention provides a denitrification optimization control method based on dynamic ammonia injection, including the following steps: Step 1, ammonia injection coupling analysis: collect the current inlet parameters of the denitrification reactor, analyze the current inlet parameters of the denitrification reactor, and obtain the basic theoretical ammonia injection rate.

[0034] In one specific embodiment, the acquisition of the current inlet parameters of the denitrification reactor is carried out as follows: The current inlet parameters of the denitrification reactor include the flue gas temperature, oxygen content and flue gas pressure of each node. On the cross-section of the inlet flue of the denitrification reactor, multiple measuring points are arranged according to the equal area grid method. Temperature sensor, oxygen content sensor and pressure sensor are installed at each measuring point to collect the flue gas temperature, oxygen content and flue gas pressure parameters of that point respectively.

[0035] In one specific embodiment, the analysis of the current denitrification reactor inlet parameters is carried out as follows: using a preset node order as the order of feature vectors, the temperature feature vector, oxygen content feature vector, and pressure feature vector of the current denitrification reactor inlet are obtained.

[0036] The similarity between the temperature feature vector at the current inlet of the denitrification reactor and the temperature feature vectors corresponding to each temperature correction coefficient is calculated to obtain the similarity of each temperature correction coefficient at the current inlet of the denitrification reactor. The temperature correction coefficient corresponding to the maximum similarity is selected as the temperature correction coefficient at the current inlet of the denitrification reactor.

[0037] Based on the process of obtaining the temperature correction coefficient, the oxygen content correction coefficient and pressure correction coefficient at the current inlet of the denitrification reactor are obtained.

[0038] It should be noted that the larger the magnitude of the temperature feature vector, the smaller the corresponding temperature correction coefficient; the larger the magnitude of the oxygen content feature vector, the smaller the corresponding oxygen content correction coefficient; and the larger the magnitude of the pressure feature vector, the smaller the corresponding pressure correction coefficient.

[0039] Based on the flue gas temperature and oxygen content of each node, mapping table matching and interpolation calculations are performed to obtain the first coupling coefficient of each node. The average value of the first coupling coefficients of all nodes is calculated to obtain the current first coupling correction coefficient. Based on the flue gas temperature and flue gas pressure of each node, coupling analysis is performed to obtain the current second coupling correction coefficient.

[0040] It should be noted that the mapping table matching and interpolation calculation process is as follows: obtain a preset two-dimensional mapping table, which contains two independent variable dimensions and one dependent variable dimension; determine the values ​​of the two independent variables; find four reference data points adjacent to the current independent variable value in the mapping table; perform interpolation calculation based on the positional and numerical relationships of the four reference data points; and obtain the dependent variable value corresponding to the current independent variable value.

[0041] The coupling analysis based on the flue gas temperature and pressure of each node is performed as follows: Multiple sets of preset temperature-pressure data are acquired, each set containing a temperature value, a pressure value, and a corresponding baseline coupling coefficient; a temperature-pressure coupling relationship mapping table is established based on these multiple sets of data; the flue gas temperature and pressure of each node are matched with the coupling relationship mapping table; the second coupling coefficient of each node is calculated by interpolation based on the matching results; and the average of the second coupling coefficients of all nodes is calculated to obtain the current second coupling correction coefficient.

[0042] Multiply the current first coupling correction coefficient by the current second coupling correction coefficient to obtain the current comprehensive coupling correction coefficient. Multiply the initial basic ammonia injection quantity by the temperature correction coefficient, oxygen content correction coefficient, pressure correction coefficient and comprehensive coupling correction coefficient in sequence to obtain the basic theoretical ammonia injection quantity.

[0043] It should be noted that the initial basic ammonia injection rate is obtained by multiplying the nitrogen oxide mass concentration at the inlet of the denitrification reactor by the nitrogen oxide mass concentration, dry flue gas volume flow rate, ammonia-nitrogen molar ratio, and ammonia molar mass in sequence, and then dividing by the average molar mass of nitrogen oxides. The nitrogen oxide mass concentration and dry flue gas volume flow rate at the inlet of the denitrification reactor are collected by concentration sensors and flow sensors, respectively.

[0044] Step 2, Feedforward Prediction Correction: Based on the historically collected inlet parameters of the denitrification reactor, obtain the trend data of the inlet parameters of the denitrification reactor, analyze the trend data of the inlet parameters of the denitrification reactor, predict the parameter changes in the next control cycle, and perform feedforward correction on the basic theoretical ammonia injection rate according to the prediction results to obtain the feedforward corrected ammonia injection rate.

[0045] In one specific embodiment, the process of acquiring the trend data of the inlet parameters of the denitrification reactor is as follows: the inlet parameters of the denitrification reactor for the most recent preset number of control cycles are collected. The inlet parameters of the denitrification reactor for each control cycle include the flue gas temperature, oxygen content and flue gas pressure at each node.

[0046] Using the control cycle corresponding to the farthest timestamp as the starting point of the horizontal axis, the horizontal axis represents the time from the starting point of the horizontal axis, and the vertical axis represents the average flue gas temperature of the corresponding control cycle, a time-temperature variation graph is obtained. Based on the time-temperature variation graph, the temperature-time slope of each node in each control cycle is calculated. According to the method of obtaining the temperature-time slope of each node in each control cycle, the oxygen content time slope and pressure time slope are then obtained and recorded as the time trend parameters of the denitrification reactor inlet parameter variation trend data.

[0047] Preset initial node coordinates, determine the distance between each node and the initial coordinate point, use the distance as the abscissa, and use the temperature, oxygen content and flue gas pressure of the current denitrification reactor inlet parameters as the ordinates to obtain the spatial slope of temperature, spatial slope of oxygen content and spatial slope of pressure of each node, which are recorded as the spatial trend parameters of the denitrification reactor inlet parameter change trend data.

[0048] In one specific embodiment, the analysis of the trend data of the inlet parameters of the denitrification reactor is carried out as follows: the historical operation data of the denitrification reactor is used as the training set, wherein each sample contains the measured values ​​of flue gas temperature, oxygen content and flue gas pressure of each node within a control cycle.

[0049] For each sample, calculate its corresponding temporal slope and spatial slope to form temporal trend parameters and spatial trend parameters. Concatenate the temporal trend parameters and spatial trend parameters into a joint vector, which is used as the model input. For the next cycle of the same control cycle, obtain the actual flue gas temperature, oxygen content and flue gas pressure of each node, which are used as the expected output of the model.

[0050] The training set is divided into a training subset, a validation subset, and a test subset. The training subset is used to supervise the training of the multilayer perceptron network. The training objective is to minimize the mean square error between the predicted output and the expected output. Training is terminated when the error of the validation subset fails to decrease for a preset number of consecutive times, thus obtaining the trend analysis model.

[0051] During model operation, the input joint vector is linearly weighted in the input layer and then enters the hidden layer. The hidden layer performs linear weighting and activation function operations, and the output layer outputs the prediction parameters.

[0052] Input the trend data of the inlet parameters of the denitrification reactor to obtain the predicted inlet parameters of the denitrification reactor. The predicted inlet parameters of the denitrification reactor include the predicted flue gas temperature, oxygen content and flue gas pressure of each node.

[0053] In one specific embodiment, the feedforward correction of the basic theoretical ammonia injection rate is performed as follows: based on the methods for obtaining the temperature correction coefficient, oxygen content correction coefficient, pressure correction coefficient, and comprehensive coupling correction coefficient, the predicted parameters of the denitrification reactor inlet are analyzed to obtain the predicted temperature correction coefficient, predicted oxygen content correction coefficient, predicted pressure correction coefficient, and predicted comprehensive coupling correction coefficient.

[0054] The feedforward corrected ammonia injection rate is obtained by multiplying the basic theoretical ammonia injection rate by the predicted temperature correction factor, the predicted oxygen content correction factor, the predicted pressure correction factor, and the predicted comprehensive coupling correction factor in sequence.

[0055] Step 3, Escape Feedback Correction: Obtain multidimensional ammonia escape parameters, analyze the multidimensional ammonia escape parameters, and perform feedback correction on the feedforward correction ammonia injection amount based on the analysis results to obtain the final corrected ammonia injection amount.

[0056] In one specific embodiment, the acquisition process of the multidimensional ammonia escape parameters is as follows: the multidimensional ammonia escape parameters include the ammonia escape rate, the ammonia edge concentration gradient, and the ammonia jet velocity.

[0057] The ammonia escape rate at the outlet of the denitrification reactor is obtained in real time using a laser ammonia analyzer.

[0058] On the flue cross-section at a predetermined distance downstream of the ammonia injection grid at the inlet of the denitrification reactor, multiple ammonia concentration sensors are arranged using the equal-area grid method. The ammonia concentration values ​​at each measuring point are acquired synchronously by the multiple ammonia concentration sensors. The average concentration difference between the edge region and the center region of the flue cross-section is calculated based on the ammonia concentration values ​​at each measuring point. The average concentration difference is then normalized to obtain the ammonia edge concentration gradient.

[0059] A jet velocity sensor is installed on the outlet pipe of each ammonia water regulating valve; the actual ammonia jet velocity of each ammonia water regulating valve is obtained in real time through the jet velocity sensor; the average value of the actual jet velocity of all ammonia water regulating valves is calculated to obtain the ammonia jet velocity of the system.

[0060] In one specific embodiment, the analysis of the multidimensional ammonia escape parameters is carried out as follows: principal component analysis is performed on the ammonia escape rate, ammonia edge concentration gradient, and ammonia jet velocity; the first principal component score is extracted; and the first principal component score is used as the comprehensive evaluation value of the multidimensional ammonia escape parameters.

[0061] The ammonia escape rate, ammonia edge concentration gradient, ammonia jet velocity, and comprehensive evaluation value are normalized and arranged in a preset order to form an ammonia escape feature vector.

[0062] In one specific embodiment, the feedback correction of the feedforward correction ammonia injection amount is performed as follows: the similarity between the ammonia escape feature vector and the escape feature vector corresponding to each escape correction coefficient is calculated to obtain the similarity of each escape correction coefficient, and the escape correction coefficient corresponding to the maximum similarity is set as the ammonia escape correction coefficient.

[0063] Multiply the feedforward corrected ammonia injection rate by the escape correction factor to obtain the final corrected ammonia injection rate.

[0064] Step 4: Ammonia Injection Execution Monitoring: The final corrected ammonia injection quantity is converted into a control signal to drive the ammonia water regulating valve to adjust its opening. At the same time, the operating status data of the ammonia water regulating valve is monitored in real time, and the operating status data is analyzed to generate a real-time ammonia water control scheme.

[0065] In one specific embodiment, the ammonia water regulating valve is adjusted in the following process: the final corrected ammonia injection quantity is converted into a target injection flow rate command, and the target injection flow rate command is converted into a valve opening control signal according to the preset valve position-flow rate mapping relationship to obtain the basic valve opening.

[0066] The difference between the initial basic ammonia injection rate, the basic theoretical ammonia injection rate, the feedforward correction ammonia injection rate, and the final correction ammonia injection rate is calculated to obtain the rate of change for each stage. The rate of change for each stage is used as the opening feature vector for the current cycle. The similarity between the opening feature vector for the current cycle and the opening feature vectors corresponding to each opening correction coefficient is calculated to obtain the similarity of each opening correction coefficient for the current cycle. The opening correction coefficient corresponding to the maximum similarity is set as the opening correction coefficient for the current cycle.

[0067] Multiply the valve's base opening by the opening correction factor to obtain the corrected valve opening, and then drive the ammonia water regulating valve to adjust to the corresponding opening.

[0068] In one specific embodiment, the analysis of the operating status data is carried out as follows: the valve position feedback signal, actual injection flow rate signal, and valve drive current signal of the ammonia water regulating valve are acquired in real time, as well as the initial basic ammonia injection quantity, basic theoretical ammonia injection quantity, feedforward correction ammonia injection quantity, final correction ammonia injection quantity, and valve correction opening degree data of the current cycle, which are used as operating status data.

[0069] The operational status data is normalized, and the normalized operational status data is used to form a current operational status feature vector in a preset order. The current operational status feature vector is then compared with the standard status feature vectors corresponding to multiple preset operational status levels to obtain multiple similarity values. The operational status level corresponding to the maximum similarity value is selected as the current system operational status evaluation level.

[0070] When the operational status assessment level is greater than or equal to the safety assessment level, the current valve position-flow velocity mapping relationship, the feature vector library of each correction coefficient, and the coupling relationship mapping table remain unchanged.

[0071] When the operational status assessment level is lower than the safety assessment level but greater than or equal to the hazard assessment level, the valve position-flow velocity mapping relationship is gradually adjusted, and the feature vector library of each correction coefficient is updated.

[0072] It should be noted that the gradual adjustment process is as follows: when the adjustment triggering condition is met, multiple sets of data on valve opening and actual jet flow rate within a preset time period are collected.

[0073] Using the difference between valve opening and actual jet velocity in each set of corresponding data as elements, an adjustment feature vector is constructed. The similarity between the adjustment feature vector and the standard adjustment feature vector corresponding to multiple preset adjustment coefficients is calculated to obtain multiple similarity values. The adjustment coefficient corresponding to the maximum similarity value is selected as the current adjustment coefficient.

[0074] Make a small adjustment to the valve position-flow velocity mapping relationship based on the current adjustment coefficient; collect data on the valve opening and actual jet velocity after adjustment to verify the adjustment effect; if the adjustment effect does not meet the preset requirements, repeat the above process for the next adjustment until the adjustment effect meets the preset requirements or the preset maximum number of adjustments is reached.

[0075] An early warning will be issued when the operational status assessment level is lower than the hazard assessment level.

[0076] according to Figure 2 As shown, the present invention provides a denitrification optimization control device based on dynamic ammonia injection, comprising the following modules: an ammonia injection coupling analysis module, a feedforward prediction correction module, an escape feedback correction module, and an ammonia injection execution monitoring module.

[0077] The feedforward prediction correction module is connected to the ammonia injection coupling analysis module and the escape feedback correction module, respectively, and the ammonia injection execution monitoring module is connected to the escape feedback correction module.

[0078] The ammonia injection coupling analysis module is used to collect the current inlet parameters of the denitrification reactor, analyze the current inlet parameters of the denitrification reactor, and obtain the basic theoretical ammonia injection rate.

[0079] The feedforward prediction and correction module is used to obtain the trend data of the inlet parameters of the denitrification reactor based on the historically collected inlet parameters, analyze the trend data of the inlet parameters of the denitrification reactor, predict the parameter changes in the next control cycle, and perform feedforward correction on the basic theoretical ammonia injection rate based on the prediction results to obtain the feedforward corrected ammonia injection rate.

[0080] The escape feedback correction module is used to acquire multidimensional ammonia escape parameters, analyze these parameters, and correct the feedforward ammonia injection amount based on the analysis results to obtain the final corrected ammonia injection amount.

[0081] The ammonia injection execution monitoring module is used to convert the final corrected ammonia injection quantity into a control signal to drive the ammonia water regulating valve to adjust its opening. At the same time, it monitors the operating status data of the ammonia water regulating valve in real time, analyzes the operating status data, and generates a real-time ammonia water control scheme.

[0082] The temperature sensor, oxygen content sensor, pressure sensor, nitrogen oxide analyzer, flow transmitter, laser ammonia analyzer, ammonia concentration sensor, jet flow rate sensor, ammonia water regulating valve, and multilayer sensing network described in this invention are all existing technologies. Those skilled in the art can select appropriate models and parameters according to actual working conditions, so they will not be described in detail here.

[0083] The following preset values ​​mentioned in this embodiment can be set according to actual business needs: preset number of nodes, for example, can be set to 9 or 16, arranged in a 3×3 or 4×4 equal area grid; preset control cycle, for example, can be set to 1 to 10 seconds; preset number of historical data collection cycles, for example, can be set to 10 to 100 control cycles; preset ammonia-nitrogen molar ratio, for example, can be set to 0.8 to 1.2; preset training set division ratio, for example, can be divided into training subset, validation subset and test subset according to a ratio of 7:2:1; preset number of consecutive times the verification error does not decrease, for example, can be set to 10 to 50 times; preset distance of downstream measuring point of ammonia injection grid, for example, can be set to 0.5 meters to 2 meters; preset safety assessment level threshold, for example, can be set to 0.7; preset hazard assessment level threshold, for example, can be set to 0.3; preset adjustment effect verification cycle, for example, can be set to 5 to 20 control cycles; preset maximum number of adjustments, for example, can be set to 5 to 20 times.

[0084] The examples described in this invention are not limited to the specific embodiments listed above. The examples are merely illustrative to facilitate understanding of the invention and do not constitute a limitation on the scope of protection of this invention. Any modifications, equivalent substitutions, etc., made within the spirit and principles of this invention should be included within the scope of protection.

[0085] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the protection scope of the present invention.

Claims

1. A denitrification optimization control method based on dynamic ammonia injection, characterized in that, Includes the following steps: Step 1: Ammonia injection coupling analysis: Collect the current inlet parameters of the denitrification reactor, analyze the current inlet parameters of the denitrification reactor, and obtain the basic theoretical ammonia injection rate; Step 2, Feedforward Prediction Correction: Based on the historically collected inlet parameters of the denitrification reactor, obtain the trend data of the inlet parameters of the denitrification reactor, analyze the trend data of the inlet parameters of the denitrification reactor, predict the parameter changes in the next control cycle, and make feedforward corrections to the basic theoretical ammonia injection rate based on the prediction results to obtain the feedforward corrected ammonia injection rate. Step 3, Escape Feedback Correction: Obtain multidimensional ammonia escape parameters, analyze the multidimensional ammonia escape parameters, and perform feedback correction on the feedforward correction ammonia injection amount based on the analysis results to obtain the final corrected ammonia injection amount; Step 4: Ammonia Injection Execution Monitoring: The final corrected ammonia injection quantity is converted into a control signal to drive the ammonia water regulating valve to adjust its opening. At the same time, the operating status data of the ammonia water regulating valve is monitored in real time, and the operating status data is analyzed to generate a real-time ammonia water control scheme.

2. The denitrification optimization control method based on dynamic ammonia injection according to claim 1, characterized in that, The analysis of the current inlet parameters of the denitrification reactor is carried out in the following specific process: The current inlet parameters of the denitrification reactor include the flue gas temperature, oxygen content, and flue gas pressure at each node. Using the preset node order as the order of feature vectors, the temperature feature vector, oxygen content feature vector, and pressure feature vector at the current inlet of the denitrification reactor are obtained. The similarity between the temperature feature vector at the current inlet of the denitrification reactor and the temperature feature vectors corresponding to each temperature correction coefficient is calculated to obtain the similarity between each temperature correction coefficient at the current inlet of the denitrification reactor. The temperature correction coefficient corresponding to the maximum similarity is selected as the temperature correction coefficient at the current inlet of the denitrification reactor. Based on the process of obtaining the temperature correction coefficient, the current oxygen content correction coefficient and pressure correction coefficient at the inlet of the denitrification reactor are obtained. Based on the flue gas temperature and oxygen content of each node, mapping table matching and interpolation calculation are performed to obtain the first coupling coefficient of each node. The average value of the first coupling coefficients of all nodes is calculated to obtain the current first coupling correction coefficient. Based on the flue gas temperature and flue gas pressure of each node, coupling analysis is performed to obtain the current second coupling correction coefficient. Multiply the current first coupling correction coefficient by the current second coupling correction coefficient to obtain the current comprehensive coupling correction coefficient. Multiply the initial basic ammonia injection quantity by the temperature correction coefficient, oxygen content correction coefficient, pressure correction coefficient and comprehensive coupling correction coefficient in sequence to obtain the basic theoretical ammonia injection quantity.

3. The denitrification optimization control method based on dynamic ammonia injection according to claim 2, characterized in that, The specific process for obtaining the trend data of inlet parameters of the denitrification reactor is as follows: Collect the inlet parameters of the denitrification reactor for the most recent preset number of control cycles. The inlet parameters of the denitrification reactor for each control cycle include the flue gas temperature, oxygen content and flue gas pressure at each node. Using the control cycle corresponding to the farthest timestamp as the starting point of the horizontal axis, the horizontal axis represents the time from the starting point of the horizontal axis, and the vertical axis represents the average flue gas temperature of the corresponding control cycle, a time-temperature variation graph is obtained. The temperature-time slope of each node in each control cycle is calculated based on the time-temperature variation graph. Based on the method of obtaining the temperature-time slope of each node in each control cycle, the oxygen content time slope and pressure time slope are obtained and recorded as the time trend parameters of the denitrification reactor inlet parameter variation trend data. Preset initial node coordinates, determine the distance between each node and the initial coordinate point, use the distance as the abscissa, and use the temperature, oxygen content and flue gas pressure of the current denitrification reactor inlet parameters as the ordinates to obtain the spatial slope of temperature, spatial slope of oxygen content and spatial slope of pressure of each node, which are recorded as the spatial trend parameters of the denitrification reactor inlet parameter change trend data.

4. The denitrification optimization control method based on dynamic ammonia injection according to claim 3, characterized in that, The analysis of the trend data of inlet parameter changes in the denitrification reactor is as follows: The training set is based on the historical operating data of the denitrification reactor, where each sample contains the measured values ​​of flue gas temperature, oxygen content and flue gas pressure at each node within a control cycle. For each sample, calculate its corresponding temporal slope and spatial slope to form temporal trend parameters and spatial trend parameters. Concatenate the temporal trend parameters and spatial trend parameters into a joint vector as the model input. For the next cycle of the same control cycle, obtain the actual flue gas temperature, oxygen content and flue gas pressure of each node as the expected output of the model. The training set is divided into a training subset, a validation subset, and a test subset. The training subset is used to supervise the training of the multilayer perceptron network. The training objective is to minimize the mean square error between the predicted output and the expected output. Training is terminated when the error of the validation subset does not decrease for a preset number of consecutive times, thus obtaining the trend analysis model. During model operation, the input joint vector is linearly weighted in the input layer and then enters the hidden layer. The hidden layer performs linear weighting and activation function operations, and the output layer outputs the predicted parameters. Input the trend data of the inlet parameters of the denitrification reactor to obtain the predicted inlet parameters of the denitrification reactor. The predicted inlet parameters of the denitrification reactor include the predicted flue gas temperature, oxygen content and flue gas pressure of each node.

5. The denitrification optimization control method based on dynamic ammonia injection according to claim 2, characterized in that, The feedforward correction of the basic theoretical ammonia injection rate is performed as follows: Based on the methods for obtaining the temperature correction coefficient, oxygen content correction coefficient, pressure correction coefficient, and comprehensive coupling correction coefficient, the predicted parameters of the denitrification reactor inlet are analyzed to obtain the predicted temperature correction coefficient, predicted oxygen content correction coefficient, predicted pressure correction coefficient, and predicted comprehensive coupling correction coefficient. The feedforward corrected ammonia injection rate is obtained by multiplying the basic theoretical ammonia injection rate by the predicted temperature correction factor, the predicted oxygen content correction factor, the predicted pressure correction factor, and the predicted comprehensive coupling correction factor in sequence.

6. The denitrification optimization control method based on dynamic ammonia injection according to claim 1, characterized in that, The analysis of the multidimensional ammonia slip parameters is as follows: Multidimensional ammonia escape parameters include ammonia escape rate, ammonia edge concentration gradient, and ammonia jet velocity; Principal component analysis was performed on ammonia escape rate, ammonia edge concentration gradient and ammonia jet velocity, and the first principal component score was extracted. The first principal component score was used as the comprehensive evaluation value of the multidimensional ammonia escape parameters. The ammonia escape rate, ammonia edge concentration gradient, ammonia jet velocity, and comprehensive evaluation value are normalized and arranged in a preset order to form an ammonia escape feature vector.

7. The denitrification optimization control method based on dynamic ammonia injection according to claim 6, characterized in that, The specific correction process for the feedforward correction ammonia injection amount is as follows: The similarity between the ammonia escape feature vector and the escape feature vector corresponding to each escape correction coefficient is calculated to obtain the similarity of each escape correction coefficient. The escape correction coefficient corresponding to the maximum similarity is set as the ammonia escape correction coefficient. Multiply the feedforward corrected ammonia injection rate by the escape correction factor to obtain the final corrected ammonia injection rate.

8. The denitrification optimization control method based on dynamic ammonia injection according to claim 1, characterized in that, The ammonia water regulating valve is adjusted to change its opening degree. The specific adjustment process is as follows: The final corrected ammonia injection quantity is converted into a target injection velocity command. Based on the preset valve position-velocity mapping relationship, the target injection velocity command is converted into a valve opening control signal to obtain the basic valve opening. The difference between the initial basic ammonia injection rate, the basic theoretical ammonia injection rate, the feedforward correction ammonia injection rate, and the final correction ammonia injection rate is calculated to obtain the rate of change for each stage. The rate of change for each stage is used as the opening feature vector for the current cycle. The similarity between the opening feature vector for the current cycle and the opening feature vectors corresponding to each opening correction coefficient is calculated to obtain the similarity of each opening correction coefficient for the current cycle. The opening correction coefficient corresponding to the maximum similarity is set as the opening correction coefficient for the current cycle. Multiply the valve's base opening by the opening correction factor to obtain the corrected valve opening, and then drive the ammonia water regulating valve to adjust to the corresponding opening.

9. The denitrification optimization control method based on dynamic ammonia injection according to claim 1, characterized in that, The analysis of the operational status data is performed as follows: The valve position feedback signal, actual injection flow rate signal, and valve drive current signal of the ammonia water regulating valve are acquired in real time, as well as the initial basic ammonia injection quantity, basic theoretical ammonia injection quantity, feedforward correction ammonia injection quantity, final correction ammonia injection quantity, and valve correction opening data of the current cycle, as operating status data. The operational status data is normalized, and the normalized operational status data is used to form a current operational status feature vector in a preset order. The current operational status feature vector is then compared with the standard status feature vectors corresponding to multiple preset operational status levels to obtain multiple similarity values. The operational status level corresponding to the maximum similarity value is selected as the current system operational status evaluation level. When the operational status assessment level is greater than or equal to the safety assessment level, the current valve position-flow velocity mapping relationship, the feature vector library of each correction coefficient, and the coupling relationship mapping table remain unchanged. When the operational status assessment level is lower than the safety assessment level but greater than or equal to the hazard assessment level, the valve position-flow velocity mapping relationship is gradually adjusted, and the feature vector library of each correction coefficient is updated. An early warning will be issued when the operational status assessment level is lower than the hazard assessment level.

10. A control device utilizing the denitrification optimization control method based on dynamic ammonia injection as described in any one of claims 1-9, characterized in that, Includes the following modules: The ammonia injection coupling analysis module is used to collect the current inlet parameters of the denitrification reactor, analyze the current inlet parameters of the denitrification reactor, and obtain the basic theoretical ammonia injection rate. The feedforward prediction and correction module is used to obtain the trend data of the inlet parameters of the denitrification reactor based on the historically collected inlet parameters, analyze the trend data of the inlet parameters of the denitrification reactor, predict the parameter changes in the next control cycle, and make feedforward corrections to the basic theoretical ammonia injection rate based on the prediction results to obtain the feedforward corrected ammonia injection rate. The escape feedback correction module is used to acquire multidimensional ammonia escape parameters, analyze the multidimensional ammonia escape parameters, and correct the feedforward correction ammonia injection amount based on the analysis results to obtain the final corrected ammonia injection amount. The ammonia injection execution monitoring module is used to convert the final corrected ammonia injection quantity into a control signal to drive the ammonia water regulating valve to adjust its opening. At the same time, it monitors the operating status data of the ammonia water regulating valve in real time, analyzes the operating status data, and generates a real-time ammonia water control scheme.

Images