Emission control method and device, electronic equipment and storage medium

Abstract

The invention discloses an emission control method and device, electronic equipment and a storage medium, and relates to the technical field of data processing, because an internal model control compensation loop is constructed based on an emission average value and mathematical transfer model parameters, the large inertia and long delay characteristics of a selective catalytic reduction system can be dynamically compensated, and the reliability of the system is improved. A concentration prediction model and a compensation loop are integrated into a denitration control loop, rapid adjustment of the ammonia spraying amount and stable control of emission are achieved, the problems of system measurement delay, concentration fluctuation, sampling distortion and the like are fully considered, a traditional control strategy which does not fully consider the problems is not adopted, and therefore the control method has the advantages of being high in practicability. The technical problems that in the prior art, due to the fact that a traditional control strategy is adopted, control lags, ammonia spraying is excessive, emission exceeds the standard, an air pre-heater is blocked, potential safety hazards are caused by the too high ammonia concentration, and then the environmental protection compliance, the economic operation efficiency and the system safety and stability of a unit are affected can be solved. The technical effects of reducing control lag, accurately adjusting the ammonia spraying amount, stabilizing NOx emission, avoiding blockage and potential safety hazards of the air pre-heater, guaranteeing environmental protection and compliance of the unit, improving the economic operation efficiency and enhancing the safety and stability of the system are achieved.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, and in particular to an emission control method and apparatus, electronic equipment and storage medium. Background Technology

[0002] Selective catalytic reduction (SCR) technology is a core means for thermal power units to achieve ultra-low emissions of nitrogen oxides (NOx) and is widely used in the flue gas denitrification system of coal-fired power plants.

[0003] With the increasing stringent environmental protection requirements for the thermal power industry, NOx emission limits have been strictly controlled within the emission standards for gas turbine units. In related technologies, SCR systems typically regulate the amount of ammonia injected, thereby controlling the outlet NOx concentration, through the coordinated operation of a Continuous Emission Monitoring System (CEMS) and a Distributed Control System (DCS) control loop.

[0004] Existing SCR control methods directly adopt traditional feedforward, cascade, or single-loop control strategies without fully considering issues such as system measurement delay, concentration fluctuation, and sampling distortion. This may lead to control lag, excessive ammonia injection, excessive emissions, or air preheater blockage and excessively high ammonia concentration, causing safety hazards and affecting the unit's environmental compliance, economic operating efficiency, and system safety and stability. Summary of the Invention

[0005] This disclosure provides an emission control method, apparatus, electronic device, and storage medium. Its main purpose is to address issues affecting the environmental compliance, economic operating efficiency, and system safety and stability of power units.

[0006] According to a first aspect of this disclosure, an emission control method is provided, comprising:

[0007] Obtain the average emission value output by the concentration prediction model;

[0008] Based on the average emission values ​​and mathematical transfer model parameters, an internal model control compensation loop is constructed to achieve dynamic compensation for the large inertia and long delay characteristics of the selective catalytic reduction system.

[0009] The concentration prediction model and the internal model control compensation loop are integrated into the denitrification control loop to achieve rapid adjustment of ammonia injection and stable control of emissions.

[0010] Optionally, before obtaining the average emissions output from the concentration prediction model, the method further includes:

[0011] A neural network algorithm is used to train historical data on furnace operating parameters and selective catalytic reduction inlet emission concentrations to establish the concentration prediction model; wherein, the furnace operating parameters include at least one of the following: unit load, total air volume, total coal volume, secondary air damper opening, burnout damper opening, and oxygen content.

[0012] Optionally, before obtaining the average emissions output from the concentration prediction model, the method further includes:

[0013] Based on step test data, a transfer model of the selective catalytic reduction system under different load ranges was constructed to obtain its inertia time and delay time parameters; wherein, the step test data are step test data of valve, ammonia injection rate and ammonia injection outlet emission concentration;

[0014] Wavelet transform was used to denoise the step test data while preserving key dynamic response features;

[0015] Based on the ammonia injection response characteristics at different load levels, the least squares method is used to fit the transfer function, and transfer models are established for low load, medium load and high load respectively.

[0016] Optionally, integrating the concentration prediction model and the internal model control compensation loop into the denitrification control loop to achieve rapid adjustment of ammonia injection and stable control of emissions includes:

[0017] The weighting coefficients of the prediction model and the internal model control loop are dynamically adjusted based on real-time feedback of the selective catalytic reduction outlet emission concentration.

[0018] Optionally, the method further includes:

[0019] The backflush cycle is identified by analyzing the sampled signal, and signal interpolation is performed using historical data and the output of the prediction model during the backflush period.

[0020] According to a second aspect of this disclosure, an emission control device is provided, comprising:

[0021] The first acquisition unit is used to acquire the average emission value output by the concentration prediction model;

[0022] The construction unit is used to construct an internal model control compensation loop based on the average emission value and mathematical transfer model parameters, so as to realize dynamic compensation for the large inertia and long delay characteristics of the selective catalytic reduction system.

[0023] An integrated unit is used to integrate the concentration prediction model and the internal model control compensation loop into the denitrification control loop, so as to realize rapid adjustment of ammonia injection and stable control of emissions.

[0024] Optionally, the device further includes:

[0025] The second acquisition unit is used to train the historical data of furnace operating parameters and selective catalytic reduction inlet emission concentrations using a neural network algorithm before the first acquisition unit acquires the average emission output of the concentration prediction model, thereby establishing the concentration prediction model; wherein, the furnace operating parameters include at least one of the following: unit load, total air volume, total coal volume, secondary air damper opening, burnout air damper opening, and oxygen content.

[0026] Optionally, the device further includes:

[0027] The third acquisition unit is used to construct a transfer model of the selective catalytic reduction system under different load ranges based on step test data before the first acquisition unit acquires the average emission output of the concentration prediction model, and to acquire its inertia time and delay time parameters; wherein, the step test data are step test data of valve, ammonia injection rate and ammonia injection rate outlet emission concentration;

[0028] The processing unit is used to perform noise reduction processing on the step test data using wavelet transform, while retaining key dynamic response features;

[0029] A unit is established to fit the transfer function using the least squares method based on the ammonia injection response characteristics of different load segments, and to establish the transfer models under low load, medium load and high load respectively.

[0030] Optionally, the integration unit is further configured to:

[0031] The weighting coefficients of the prediction model and the internal model control loop are dynamically adjusted based on real-time feedback of the selective catalytic reduction outlet emission concentration.

[0032] Optionally, the device further includes:

[0033] The identification unit is used to identify the backflush cycle of the analyzed sampled signal and to interpolate the signal using historical data and the output of the prediction model during the backflush period.

[0034] According to a third aspect of this disclosure, an electronic device is provided, comprising:

[0035] At least one processor; and

[0036] A memory communicatively connected to the at least one processor; wherein,

[0037] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect above.

[0038] According to a fourth aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are configured to cause the computer to perform the method described in the first aspect above.

[0039] According to a fifth aspect of this disclosure, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the method described in the first aspect above.

[0040] The emission control method, apparatus, electronic equipment, and storage medium disclosed herein, through the construction of an internal model control compensation loop based on the emission average value and mathematical transfer model parameters, can dynamically compensate for the large inertia and long delay characteristics of the selective catalytic reduction system. Furthermore, by integrating the concentration prediction model with this compensation loop into the denitrification control loop, it achieves rapid adjustment of ammonia injection and stable emission control. It fully considers issues such as system measurement delay, concentration fluctuation, and sampling distortion, rather than employing traditional control strategies that do not adequately address these problems. Therefore, it can solve the technical problems in the prior art caused by traditional control strategies, such as control lag, excessive ammonia injection, excessive emissions, air preheater blockage, and safety hazards caused by excessively high ammonia concentrations, which in turn affect the unit's environmental compliance, economic operating efficiency, and system safety and stability. It achieves the technical effects of reducing control lag, accurately adjusting ammonia injection, stabilizing NOx emissions, avoiding air preheater blockage and safety hazards, ensuring unit environmental compliance, improving economic operating efficiency, and enhancing system safety and stability.

[0041] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0042] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0043] Figure 1 This is a schematic flowchart of an emission control method provided in an embodiment of the present disclosure;

[0044] Figure 2 This is a schematic diagram of the structure of an emission control device provided in an embodiment of the present disclosure;

[0045] Figure 3 This is a schematic diagram of the structure of an emission control device provided in an embodiment of the present disclosure;

[0046] Figure 4 A schematic block diagram of an example electronic device provided for embodiments of this disclosure. Detailed Implementation

[0047] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0048] The emission control method, apparatus, electronic device, and storage medium of this disclosure are described below with reference to the accompanying drawings.

[0049] Figure 1 This is a schematic flowchart of an emission control method provided in an embodiment of the present disclosure.

[0050] like Figure 1 As shown, the method includes the following steps:

[0051] Step 101: Obtain the average emission value output by the concentration prediction model;

[0052] In the process of denitrification control of thermal power units, SCR systems often face problems such as delayed NOx concentration measurement, frequent concentration fluctuations and sampling distortion. Conventional control methods are difficult to achieve stable control of NOx emissions. Therefore, a prediction model of NOx concentration at the SCR inlet is first constructed based on the main operating parameters in the furnace.

[0053] The SCR inlet NOx concentration prediction model is constructed based on the actual operating conditions of the unit. Its input parameters cover the main in-furnace operating parameters that have a direct impact on the generation and change of NOx concentration, including unit load, total air volume, total coal volume, secondary air damper opening, burnout damper opening, and unit oxygen content. These parameters can comprehensively reflect the combustion state and flue gas characteristics in the furnace. By dynamically analyzing and calculating these real-time collected operating parameters, the model can accurately capture the changing pattern of SCR inlet NOx concentration.

[0054] During unit operation, the predictive model continuously outputs real-time data related to the NOx concentration at the SCR inlet. The average emission value is not a single instantaneous concentration data, but a stable value obtained by processing the continuous concentration data over a period of time output by the model. This processing can effectively filter out high-frequency oscillations in NOx concentration caused by factors such as coal quality fluctuations and air volume changes, avoid interference from instantaneous data fluctuations to subsequent control, and ensure that the output average emission value has good stability and reliability.

[0055] Obtaining this average emission value essentially provides an advanced and accurate concentration reference for ammonia injection control in the denitrification system. Compared to the lagging data that relies on CEMS sampling analysis, this average value can predict the changing trend of NOx concentration at the SCR inlet in advance, creating conditions for the early action of the ammonia injection valve. This helps overcome the large delay characteristics of the SCR system and lays a key data foundation for achieving stable control of NOx emissions under wide loads, reducing urea over-injection, and ensuring the safe and efficient operation of the unit.

[0056] Step 102: Based on the average emission value and the mathematical transfer model parameters, construct an internal model control compensation loop to achieve dynamic compensation for the large inertia and long delay characteristics of the selective catalytic reduction system;

[0057] The core lies in forming a precise compensation logic based on average emission values ​​and mathematical transfer model parameters. These mathematical transfer model parameters are obtained through field experiments and data processing: First, field step tests are conducted on valve-ammonia injection rate and ammonia injection rate-outlet NOx. This involves gradually adjusting the ammonia injection valve opening to change the ammonia injection rate, while simultaneously monitoring the impact of ammonia injection rate changes on the outlet NOx concentration in real time, recording the complete dynamic process of the system from input adjustment to output response. Then, the raw data collected during the experiments is filtered and noise-reduced to remove invalid data caused by equipment vibration, signal interference, and other factors. Finally, a complete mathematical transfer model of the SCR system covering all operating conditions is constructed for different load segments of the unit. From this model, key parameters characterizing the system's dynamic characteristics, such as inertia time and lag time, are extracted. These parameters accurately reflect the SCR system's response speed and lag time to ammonia injection adjustments under different loads.

[0058] The average emission value is a stable data output from the SCR inlet NOx concentration prediction model. It has filtered out high-frequency oscillation interference and accurately reflects the actual trend of NOx concentration changes at the SCR inlet, providing a reliable concentration benchmark for the compensation loop. When constructing the internal model control compensation loop, the average emission value is used as the core input signal. Combined with the mathematical transfer model parameters of the corresponding load segment, a dynamic correlation between concentration changes and control adjustments is established through the internal model control algorithm. When the average emission value indicates that the SCR inlet NOx concentration will change, the loop can calculate the magnitude and timing of the ammonia injection control adjustment in advance based on the inertia time and delay time parameters, avoiding control lag due to system inertia and delay. This internal model control compensation loop, constructed based on key data and model parameters, can specifically offset the large inertia and long delay characteristics of the SCR system, making the denitrification control more closely match the actual system response law. This lays an important foundation for achieving stable NOx emission control under wide loads, reducing urea waste, and avoiding safety hazards such as air preheater blockage.

[0059] Step 103: Integrate the concentration prediction model and the internal model control compensation loop into the denitrification control loop to achieve rapid adjustment of ammonia injection and stable control of emissions.

[0060] The concentration prediction model is built based on key furnace operating parameters such as unit load, total air volume, total coal volume, secondary air damper opening, burnout damper opening, and unit oxygen content. Its core advantage lies in its ability to detect the changing trend of NOx concentration at the SCR inlet in advance. During steady-state operation of the unit, the concentration data output by the model is consistent with the actual SCR inlet concentration. When the unit's operating conditions change, the model can output concentration change signals in advance, providing an advanced concentration reference for the denitrification control loop. The internal model control compensation loop is built based on the mathematical transfer model of the SCR system at different load segments. This mathematical transfer model is obtained through on-site step tests of valve-ammonia injection rate and ammonia injection rate-outlet NOx, as well as filtering and noise reduction processing. It can accurately extract the system's inertia time and delay time parameters, thereby specifically compensating for the large inertia and long delay characteristics of the SCR system.

[0061] During the integration process, the output signal of the concentration prediction model serves as a key pre-input for the denitrification control loop, enabling the control loop to acquire trend information before actual NOx concentration changes, thus reserving response time for ammonia injection adjustment. The internal model control compensation loop is deeply integrated with the ammonia injection adjustment unit in the denitrification control loop. When the control loop issues an ammonia injection adjustment command based on the concentration prediction model signal, the internal model control compensation loop can dynamically correct the amplitude and timing of ammonia injection adjustment based on system inertia and delay parameters, avoiding excessive or insufficient ammonia injection due to system lag. After integration, the denitrification control loop can achieve "advanced perception" through concentration prediction and solve the "response lag" problem through internal model compensation, significantly improving the timeliness and accuracy of ammonia injection adjustment. When unit load fluctuations, coal quality changes, or air volume adjustments cause NOx concentration to change, the control loop can quickly adjust the ammonia injection, avoiding excessive NOx emissions and preventing waste caused by excessive urea injection, as well as safety hazards such as air preheater blockage and ammonia explosion. Ultimately, it achieves stable control of NOx emissions and safe and efficient operation of the unit under wide load conditions.

[0062] In some embodiments, before obtaining the average emissions output from the concentration prediction model, the method further includes:

[0063] A neural network algorithm is used to train historical data on furnace operating parameters and selective catalytic reduction inlet emission concentrations to establish the concentration prediction model; wherein, the furnace operating parameters include at least one of the following: unit load, total air volume, total coal volume, secondary air damper opening, burnout damper opening, and oxygen content.

[0064] Before obtaining the average emission value output by the concentration prediction model, a model that can reliably predict the NOx concentration at the SCR inlet is constructed. The core reason for choosing a neural network algorithm to complete this construction is that the algorithm is good at mining and capturing the complex nonlinear relationship between furnace operating parameters and SCR inlet emission concentration. This relationship is not a simple linear correspondence, but a dynamic relationship affected by the interaction of multiple parameters. Traditional linear models are difficult to accurately characterize, while neural network algorithms can effectively learn this complex mapping law through a multi-layer data processing structure.

[0065] When constructing this concentration prediction model, it is necessary to first collect historical data accumulated during the actual operation of thermal power units. This data must be temporally relevant and comprehensive in terms of operating conditions, specifically including two core components: one is the furnace operating parameters, which must cover at least one of the following: unit load, total air volume, total coal volume, secondary air damper opening, burnout damper opening, and oxygen content. These parameters directly determine the intensity of combustion in the furnace, the oxygen supply, and the fuel combustion efficiency, thereby fundamentally affecting the amount of NOx generated and the emission concentration. For example, when the unit load increases, if the total coal volume and air volume are not adjusted in tandem, the NOx concentration will often fluctuate accordingly. The other component is the actual monitoring data of the SCR inlet emission concentration corresponding to the above operating parameters within the same time period, ensuring that each set of operating parameters can match the actual concentration results. Furthermore, the historical data must cover different operating conditions such as low load, medium load, and high load of the unit, as well as scenarios under different coal quality combustion and different air volume configurations, to avoid the model being only applicable to specific operating conditions and unable to cope with the demand for peak load regulation due to limited data.

[0066] The collected historical data is preprocessed to remove abnormal data caused by equipment failure or signal interference, and then divided into training and validation sets according to a reasonable ratio. The training set is used to train the neural network algorithm: the neural network analyzes the operating parameters and concentration data in the training set layer by layer, gradually learning the changing patterns of NOx concentration under different parameter combinations. For example, when the total coal quantity increases by a certain amount, combined with changes in oxygen content, the NOx concentration at the SCR inlet typically shows a certain trend; after adjusting the secondary damper opening, the lag time and magnitude of concentration response are also observed. During training, the algorithm continuously compares the predicted output with the actual concentration value, constantly optimizing the internal weights and threshold parameters to gradually reduce the prediction error. After training, the model performance is tested using the validation set to ensure that the model can stably output predicted SCR inlet emission concentration values ​​with minimal deviation from the actual situation when receiving real-time input furnace operating parameters. This concentration prediction model, built using a neural network algorithm, not only provides an accurate data source for obtaining average emission values, but also enables early detection of NOx concentration trends in denitrification control. This lays the foundation for overcoming the large delay characteristics of the SCR system and reduces problems such as excessive NOx emissions and over-spraying of urea in conventional control from the source.

[0067] In some embodiments, before obtaining the average emissions output from the concentration prediction model, the method further includes:

[0068] Based on step test data, a transfer model of the selective catalytic reduction system under different load ranges was constructed to obtain its inertia time and delay time parameters; wherein, the step test data are step test data of valve, ammonia injection rate and ammonia injection outlet emission concentration;

[0069] Wavelet transform was used to denoise the step test data while preserving key dynamic response features;

[0070] Based on the ammonia injection response characteristics at different load levels, the least squares method is used to fit the transfer function, and transfer models are established for low load, medium load and high load respectively.

[0071] In the denitrification control of thermal power units, the large inertia and long delay characteristics of the selective catalytic reduction system (SCR system) will show significant differences with the changes in unit load. Under low load, medium load and high load, the different flue gas volume, catalyst activity and reaction conditions will lead to significant differences in the system's response speed and lag time to ammonia injection adjustment. If a uniform control model is used, it is difficult to adapt to the dynamic characteristics under full load conditions. Therefore, before obtaining the average emission value output by the concentration prediction model, it is necessary to first construct a load-segmented SCR system transfer model.

[0072] The construction process begins with acquiring step test data, which is achieved through on-site step tests: During the test, the opening of the ammonia injection valve is gradually adjusted to change the ammonia injection rate, while real-time data on the changes in the ammonia injection rate and the dynamic response data of the NOx emission concentration at the SCR outlet after the ammonia injection rate adjustment are collected. This forms complete step test data covering valve action, changes in ammonia injection rate, and the response of the outlet concentration. These data directly reflect the input-output dynamic relationship of the SCR system under specific loads and are the core foundation for building the transfer model.

[0073] Due to factors such as equipment vibration and signal interference in the field test environment, the original step test data will contain noise. If used directly for modeling, it will lead to a decrease in model accuracy. Therefore, wavelet transform is required for noise reduction. Wavelet transform can decompose the original test signal into components of different frequencies. By identifying and removing high-frequency components that represent noise, it can fully retain the signal components that reflect the key dynamic response characteristics of the SCR system, such as the inflection point time of the outlet NOx concentration after the ammonia injection rate adjustment, the rate of concentration change, and the value after stabilization. This ensures that the processed data can truly reflect the dynamic characteristics of the system.

[0074] After obtaining the noise-reduced step test data, the least squares method was used to fit the transfer function based on the differences in ammonia injection response characteristics of the SCR system under different load conditions. For the characteristics of small flue gas volume and relatively concentrated reaction space in the low-load section, large flue gas volume and long reaction path in the high-load section, and the characteristics of the medium-load section which fall between the two, appropriate transfer function forms were selected for each. The least squares method was used to calculate and minimize the sum of squared errors between the experimental data and the theoretical output value of the transfer function, thereby determining the specific parameters of the transfer function for each load section. Finally, independent SCR system transfer models were established for low, medium, and high loads, and the inertia time and delay time parameters corresponding to each load section were extracted from the models. These load-segment transfer models and parameters can accurately characterize the dynamic laws of the SCR system under different operating conditions, providing a reliable basis for the subsequent construction of internal model control compensation loops. This ensures that compensation measures can adapt to different load scenarios, laying a crucial foundation for achieving precise adjustment of ammonia injection volume and stable NOx emission control across the entire load range.

[0075] In some embodiments, integrating the concentration prediction model and the internal model control compensation loop into the denitrification control loop to achieve rapid adjustment of ammonia injection and stable control of emissions includes:

[0076] The weighting coefficients of the prediction model and the internal model control loop are dynamically adjusted based on real-time feedback of the selective catalytic reduction outlet emission concentration.

[0077] In the denitrification control of thermal power units, although the concentration prediction model and the internal model control compensation loop have been integrated into the denitrification control loop, the accuracy of the leading prediction of the concentration prediction model and the lag compensation effect of the internal model control compensation loop may dynamically change with the operating conditions due to the influence of actual operating conditions such as coal quality fluctuations, air volume changes, and combustion state adjustments during unit operation. If the weight coefficients of the two are kept fixed, it is easy to cause a decrease in the adjustment accuracy of the ammonia injection quantity in the denitrification control loop, and even the NOx emission concentration at the SCR outlet may deviate from the ultra-low emission target (50 mg / Nm³). 3 Therefore, it is necessary to rely on the real-time feedback of NOx emission concentration at the SCR outlet to dynamically adjust the weight coefficients of the prediction model and the internal model control loop, so as to ensure that the two can work together efficiently under different operating conditions, and to ensure the accuracy of ammonia injection regulation and the stability of emission control.

[0078] Real-time feedback of SCR outlet NOx emission concentration is obtained through CEMS sampling and analysis. This data directly reflects the actual effect of denitrification control and is the core basis for judging the synergistic effect of the current prediction model and internal model control loop. When the real-time feedback outlet concentration is stable within the target range, it indicates that the weight coefficients of the two are suitable for the current operating conditions. If the outlet concentration fluctuates or exceeds the target value, the problem needs to be located through this real-time feedback data to determine whether the concentration prediction model has not accurately captured the trend of NOx concentration change at the SCR inlet or whether the internal model control compensation loop has not fully offset the large inertia and long delay characteristics of the system.

[0079] The dynamic adjustment of the weighting coefficients needs to be combined with the real-time feedback of the outlet concentration deviation: if the outlet concentration exceeds the target value due to a sudden change in the NOx concentration at the SCR inlet, and the concentration prediction model does not output a trend signal in advance, it indicates that the predictive model's adaptability to the current operating conditions has decreased. In this case, the weighting coefficient of the internal model control loop needs to be increased to enhance its ability to compensate for system lag and quickly adjust the ammonia injection rate to suppress the concentration rise. If the outlet concentration fluctuates due to a lag in ammonia injection adjustment caused by system inertia, it indicates that the internal model control compensation is insufficient. The weighting coefficient of the prediction model can be appropriately increased to rely on its predictive ability to adjust the ammonia injection rate in advance and avoid adjustment lag. Through this dynamic weighting adjustment based on real-time feedback, the proactive adjustment of the concentration prediction model and the lag compensation of the internal model control compensation loop can form a dynamic balance, enabling the denitrification control loop to always optimize the adjustment logic according to the actual operating conditions. This achieves rapid response of the ammonia injection rate and stable control of NOx emissions at the SCR outlet, while avoiding waste caused by excessive urea injection and safety hazards such as air preheater blockage and ammonia explosion.

[0080] In some embodiments, the method further includes:

[0081] The backflush cycle is identified by analyzing the sampled signal, and signal interpolation is performed using historical data and the output of the prediction model during the backflush period.

[0082] In the SCR denitrification control process of thermal power units, NOx concentration monitoring relies on the continuous flue gas monitoring system (CEMS) to sample and analyze the inlet and outlet flue gas. To prevent the CEMS sampling tubes from being clogged by dust and other impurities in the flue gas, the sampling tubes are periodically backflushed according to a fixed schedule. During the backflushing, the CEMS sampling process is temporarily interrupted, and the sampling signal remains at the value before the backflushing began and does not update. This signal stagnation will cause the denitrification control loop to be unable to obtain accurate NOx concentration change information, resulting in control process distortion. This may lead to delayed or excessive ammonia injection, increasing the risk of NOx emissions exceeding standards. Therefore, this problem needs to be solved by identifying the backflushing cycle of the analyzed sampling signal and interpolating the signal during the backflushing.

[0083] First, backflush cycle identification is performed by real-time monitoring of the operating status parameters of the CEMS sampling system. For example, the opening and closing signals of the backflush control valve are tracked. When the valve opens, backflush is determined to begin, and when the valve closes, backflush is determined to end. Alternatively, the operating parameters of the sampling pump, such as current and pressure, are monitored. These parameters will show characteristic changes when backflush starts, and the specific time node of backflush is determined accordingly. Then, the duration and interval of the backflush cycle are analyzed to ensure that the time range of each backflush can be accurately captured.

[0084] After confirming the start of the backflushing phase, signal interpolation is performed using historical data and the output of the concentration prediction model. Historical sampling data from non-backflushing periods that closely match the current unit operating conditions (such as unit load, total air volume, total coal volume, and secondary damper opening) are selected. Because of the similar operating conditions, the NOx concentration trend of this data is strongly correlated with the actual concentration change during the current backflushing period, serving as the basis for interpolation. Simultaneously, relying on the previously constructed concentration prediction model, which continuously outputs predicted values ​​of SCR inlet NOx concentration based on real-time collected furnace operating parameters, reflecting the dynamic trend of concentration change. Historical data from similar operating conditions are fused with the model prediction values ​​to generate a continuous and realistic NOx concentration replacement signal during backflushing. This signal replaces the stagnant original sampling signal input to the denitrification control loop, ensuring that the control loop can still obtain a reliable concentration reference during backflushing. This avoids malfunctions in ammonia injection control due to signal distortion, ensuring the continuity and accuracy of ammonia injection regulation, thereby maintaining stable NOx emissions and meeting the requirements for ultra-low emissions and safe operation of thermal power units.

[0085] Corresponding to the emission control method described above, this invention also proposes an emission control device. Since the device embodiments of this invention correspond to the method embodiments described above, details not disclosed in the device embodiments can be referred to in the method embodiments described above, and will not be repeated here.

[0086] Figure 2 This is a schematic diagram of the structure of an emission control device provided in an embodiment of this disclosure, as shown below. Figure 2 As shown, it includes:

[0087] The first acquisition unit 21 is used to acquire the average emission value output by the concentration prediction model;

[0088] Construction unit 22 is used to construct an internal model control compensation loop based on the average emission value and mathematical transfer model parameters to achieve dynamic compensation for the large inertia and long delay characteristics of the selective catalytic reduction system.

[0089] The integration unit 23 is used to integrate the concentration prediction model and the internal model control compensation loop into the denitrification control loop to achieve rapid adjustment of ammonia injection and stable control of emissions.

[0090] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the device further includes:

[0091] The second acquisition unit 24 is used to train the historical data of furnace operating parameters and selective catalytic reduction inlet emission concentration using a neural network algorithm before the first acquisition unit 21 acquires the average emission output of the concentration prediction model, and to establish the concentration prediction model; wherein, the furnace operating parameters include at least one of unit load, total air volume, total coal volume, secondary air damper opening, burnout air damper opening and oxygen content.

[0092] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the device further includes:

[0093] The third acquisition unit 25 is used to construct a transfer model of the selective catalytic reduction system under different load segments based on step test data before the first acquisition unit 21 acquires the average emission output of the concentration prediction model, and to acquire its inertia time and delay time parameters; wherein, the step test data are step test data of valve, ammonia injection quantity and ammonia injection quantity outlet emission concentration.

[0094] Processing unit 26 is used to perform noise reduction processing on the step test data using wavelet transform, while retaining key dynamic response features;

[0095] Unit 27 is established to fit the transfer function using the least squares method based on the ammonia injection response characteristics of different load segments, and to establish the transfer models under low load, medium load and high load respectively.

[0096] Furthermore, in one possible implementation of this disclosure, the integration unit 23 is further configured to:

[0097] The weighting coefficients of the prediction model and the internal model control loop are dynamically adjusted based on real-time feedback of the selective catalytic reduction outlet emission concentration.

[0098] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the device further includes:

[0099] The identification unit 28 is used to identify the backflush cycle of the analyzed sampling signal and to interpolate the signal using historical data and the output of the prediction model during the backflush period.

[0100] It should be noted that the foregoing explanation of the method embodiments also applies to the apparatus of the embodiments of this disclosure, and the principle is the same. Therefore, the embodiments of this disclosure are not limited thereto.

[0101] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0102] Figure 4 A schematic block diagram of an example electronic device 300 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0103] like Figure 4 As shown, device 300 includes a computing unit 301, which can perform various appropriate actions and processes based on a computer program stored in ROM (Read-Only Memory) 302 or a computer program loaded from storage unit 308 into RAM (Random Access Memory) 303. RAM 303 can also store various programs and data required for the operation of device 300. The computing unit 301, ROM 302, and RAM 303 are interconnected via bus 304. I / O (Input / Output) interface 305 is also connected to bus 304.

[0104] Multiple components in device 300 are connected to I / O interface 305, including: input unit 306, such as keyboard, mouse, etc.; output unit 307, such as various types of monitors, speakers, etc.; storage unit 308, such as disk, optical disk, etc.; and communication unit 309, such as network card, modem, wireless transceiver, etc. Communication unit 309 allows device 300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0105] The computing unit 301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, CPUs (Central Processing Units), GPUs (Graphics Processing Units), various special-purpose AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, DSPs (Digital Signal Processors), and any suitable processor, controller, microcontroller, etc. The computing unit 301 performs the various methods and processes described above, such as emission control methods. For example, in some embodiments, the emission control method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 308. In some embodiments, part or all of the computer program may be loaded and / or installed on device 300 via ROM 302 and / or communication unit 309. When the computer program is loaded into RAM 303 and executed by the computing unit 301, one or more steps of the methods described above may be performed. Alternatively, in other embodiments, the computing unit 301 may be configured to perform the aforementioned emission control method by any other suitable means (e.g., by means of firmware).

[0106] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application-Specific Standard Products), SOCs (System-on-Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0107] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0108] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, RAM, ROM, EPROM (Electrically Programmable Read-Only Memory) or flash memory, optical fiber, CD-ROM (Compact Disc Read-Only Memory), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0109] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0110] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include LANs (Local Area Networks), WANs (Wide Area Networks), the Internet, and blockchain networks.

[0111] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. A server can be a cloud server, also known as a cloud computing server or cloud host, a hosting product within the cloud computing service ecosystem, addressing the shortcomings of traditional physical hosts and VPS (Virtual Private Server, or simply "VPS") services, such as high management difficulty and weak business scalability. Servers can also be servers for distributed systems or servers incorporating blockchain technology.

[0112] It's important to note that artificial intelligence (AI) is the study of enabling computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, and planning). It encompasses both hardware and software technologies. AI hardware technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, and big data processing. AI software technologies primarily include computer vision, speech recognition, natural language processing, machine learning / deep learning, big data processing, and knowledge graph technologies.

[0113] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0114] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. An emission control method characterized by, The method comprises the following steps: obtaining an average value of emissions output by a concentration prediction model; constructing an internal model control compensation loop based on the average value of emissions and parameters of a mathematical transfer model, to realize dynamic compensation for large inertia and long delay characteristics of a selective catalytic reduction system; integrating the concentration prediction model and the internal model control compensation loop into a denitration control loop, to realize rapid adjustment of ammonia injection amount and stable control of emissions.

2. The method of claim 1, wherein, Before obtaining the average value of emissions output by the concentration prediction model, the method further comprises the following steps: training historical data of in-furnace operating parameters and selective catalytic reduction inlet emission concentration by using a neural network algorithm, to establish the concentration prediction model; wherein the in-furnace operating parameters include at least one of unit load, total air volume, total coal volume, secondary air damper opening, overfire damper opening and oxygen content.

3. The method of claim 1, wherein, Before obtaining the average value of emissions output by the concentration prediction model, the method further comprises the following steps: constructing a transfer model of the selective catalytic reduction system under different load sections based on step test data, to obtain inertia time and delay time parameters; wherein the step test data are step test data of a valve, ammonia injection amount and ammonia injection amount outlet emission concentration; performing noise reduction processing on the step test data by using wavelet transform, to retain key dynamic response characteristics; according to ammonia injection response characteristics under different load sections, fitting a transfer function by using a least square method, to respectively establish the transfer model under low load, medium load and high load.

4. The method of claim 1, wherein, The step of integrating the concentration prediction model and the internal model control compensation loop into the denitration control loop, to realize rapid adjustment of ammonia injection amount and stable control of emissions, comprises the following step: dynamically adjusting weight coefficients of the prediction model and the internal model control loop according to real-time feedback of selective catalytic reduction outlet emission concentration.

5. The method of claim 1, wherein, The method further comprises the following step: identifying a blowback period for an analysis sampling signal, and performing signal interpolation by using historical data and prediction model output during the blowback period.

6. An emissions control device characterized by, The method comprises the following steps: a first obtaining unit, configured to obtain an average value of emissions output by a concentration prediction model; a constructing unit, configured to construct an internal model control compensation loop based on the average value of emissions and parameters of a mathematical transfer model, to realize dynamic compensation for large inertia and long delay characteristics of a selective catalytic reduction system; an integrating unit, configured to integrate the concentration prediction model and the internal model control compensation loop into a denitration control loop, to realize rapid adjustment of ammonia injection amount and stable control of emissions.

7. The apparatus of claim 6, wherein, The device further comprises: a second obtaining unit, configured to, before the first obtaining unit obtains the average value of emissions output by the concentration prediction model, train historical data of in-furnace operating parameters and selective catalytic reduction inlet emission concentration by using a neural network algorithm, to establish the concentration prediction model; wherein the in-furnace operating parameters include at least one of unit load, total air volume, total coal volume, secondary air damper opening, overfire damper opening and oxygen content.

8. An electronic device, comprising: The device comprises at least one processor, and a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method in any one of claims 1-5. ​ ​ ​ 9. A non-transitory computer-readable storage medium having stored thereon computer instructions, wherein, The computer instructions are for causing the computer to perform the method according to any one of claims 1-5.

10. A computer program product, characterised in that, A computer program comprising instructions which, when executed by a processor, implement the method according to any one of claims 1-5.

Images