Accurate ammonia spraying and sampling system for SCR (selective catalytic reduction) denitration process

Abstract

The invention discloses a precise ammonia spraying and sampling system for an SCR (Selective Catalytic Reduction) denitration process, and relates to the technical field of industrial flue gas treatment. Comprising a flow field optimization module, a distributed sampling module, an intelligent analysis control module, a round-robin measurement module, an execution module and a remote monitoring module, and all the modules form a simulation-evaluation-sampling-control-feedback closed loop. Through dynamic self-adaptive multi-dimensional simulation of a flow field optimization module, and in combination with a multi-scale grid pool and an ammonia injection valve mechanical delay compensation mechanism, the velocity distribution uniformity of the flue gas section is greater than or equal to 95%; the intelligent analysis control module can flexibly switch an adjustment mode according to NOx concentration deviation by means of a PID neural network algorithm and triple control logic, stabilize the deviation to be less than or equal to 10 mg / m, and can start a compensation mode in combination with a catalyst activity state, so that the denitration efficiency is remarkably improved, resource waste caused by excessive ammonia injection in a traditional process is avoided, and the system is suitable for large-scale industrial production. And the denitration does not reach the standard due to insufficient ammonia amount.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of industrial flue gas treatment, in particular to a SCR denitration process precise ammonia injection and sampling system. BACKGROUND

[0002] In the field of industrial flue gas treatment, the selective catalytic reduction (SCR) denitration process is the core technology for controlling nitrogen oxide (NOx) emissions and is widely used in the power, chemical, metallurgical and other industries. However, with the increasingly stringent environmental standards and the increasing complexity of operating conditions, the traditional SCR denitration process has some problems. Firstly, due to the influence of load fluctuations and flue gas duct structures, the flow field in the traditional reactor is unevenly distributed, and vortex regions are formed in some areas. Secondly, the traditional sampling system uses a uniform distribution of sampling points without considering the concentration gradient differences. This results in insufficient sampling in high gradient areas and redundancy in low gradient areas. Thirdly, the traditional sampling system does not eliminate the ±30% flue gas turbulence pulsation disturbance, and the sampling hardware lacks heat tracing, anti-deposition and automatic purging designs. This leads to a deviation of more than 10% between the data and the true concentration, making it impossible to provide reliable basis for ammonia injection adjustment, and forming a vicious cycle of "blind ammonia injection - inaccurate data - ineffective regulation and control". In addition, the traditional system relies on fixed PID algorithms and does not consider AMI and catalyst activity parameters. When the catalyst activity is less than 80%, it cannot be compensated in time. The fixed grid used in flow field simulation has low accuracy when the operating conditions fluctuate greatly, and there is no self-learning mechanism. The long-term operation adjustment accuracy decays, and the remote control lacks permission grading and double confirmation, which can easily lead to safety accidents caused by misoperation. SUMMARY

[0003] The present application provides a SCR denitration process precise ammonia injection and sampling system, which solves the technical problems raised in the background art.

[0004] To achieve the above-mentioned purpose, the present application provides the following technical solution: a SCR denitration process precise ammonia injection and sampling system, which comprises a flow field optimization module, a distributed sampling module, an intelligent analysis and control module, a round-robin measurement module, an execution module and a remote monitoring module. Each module forms an analog-evaluation-sampling-control-feedback closed loop, and the feedback priority is to first correct the ammonia injection instruction and then adjust the flow field simulation parameters.

[0005] The flow field optimization module optimizes the flue gas flow field, ammonia gas flow field and gas-gas mixed flow field in the reactor through dynamic self-adaptive multi-dimensional simulation. The simulation includes operating condition sensing, multi-scale grid dynamic switching and ammonia injection valve mechanical delay compensation mechanisms.

[0006] The distributed sampling module is arranged at the inlet section, mixing section and outlet section of the reactor. The sampling points are arranged according to the principle of "high gradient area encryption - low gradient area sparseness". The high gradient area is a region with a concentration gradient , and the low gradient area is a region with a concentration gradient . The "multi-point synchronization + polling" sampling mode is used, and the turbulence pulsation disturbance is eliminated through wavelet transform.

[0007] The intelligent analysis control module receives sampling data, generates an ammonia injection adjustment instruction based on an ammonia and flue gas mixing efficiency quantitative index, a PID neural network algorithm model, and a catalyst activity state;

[0008] The round-robin measurement module is configured at the catalyst outlet and monitors the gas parameters according to a concentration gradient priority;

[0009] The execution module executes ammonia injection amount adjustment by regulating the valve;

[0010] The remote monitoring module displays system parameters in real time, distributed sampling data is transmitted to the intelligent analysis control module in real time, round-robin measurement data is used for verification, and the intelligent analysis control module sends instructions according to the verification data.

[0011] Preferably, the flow field optimization module comprises:

[0012] The flue gas guide sub-module adopts a dynamic self-adaptive computational fluid dynamics technology, defines a "working condition fluctuation intensity factor" to quantify real-time fluctuations of flue gas flow rate, temperature, NOx and NH3 concentration, and the working condition fluctuation intensity factor calculation formula is wherein is a flow rate fluctuation value, is a reference flow rate, is a temperature fluctuation value, is a reference temperature, is a NOx concentration fluctuation value, is a reference concentration, a "multi-scale grid pool" containing coarse grids, medium grids and fine grids is established, the coarse grids correspond to working conditions of F≤0.1, the medium grids correspond to working conditions of 0.1

[0013] The ammonia gas injection optimization sub-module plans the injection hole diameter gradient distribution and spatial density according to the ammonia gas flow field simulation results and the AMI index;

[0014] The mixing enhancement sub-module is configured with a detachable static mixer, the blade angle is adjusted in real time according to the mixing section sampling data and the AMI index feedback, and the AMI is maintained at ≥0.85.

[0015] Preferably, the distributed sampling module comprises:

[0016] The inlet section sampling unit is arranged at 3-5D before the reactor inlet, collects the original flue gas NOx concentration and oxygen content, and the sampling point avoids the vortex area with a flow rate of <0.5 m / s;

[0017] The mixing section sampling unit is arranged 1-2 m downstream of the ammonia injection grid, collects the NOx concentration and the ammonia slip concentration, and is provided with one sampling point every 0.3 m in the high-gradient zone and one sampling point every 1.5 m in the low-gradient zone;

[0018] The outlet section sampling unit is located 0.5-1 m downstream of the catalyst layer, collects the flue gas parameters after the reaction, adopts "multi-point synchronous sampling + polling sampling", collects at least 100 groups of data, and then uses wavelet transform to eliminate the ±30% turbulent fluctuation interference, the wavelet transform adopts db4 wavelet basis and 3-layer decomposition, and the threshold is 1.5 times the standard deviation to obtain the time-averaged concentration;

[0019] Each sampling unit is provided with a heat tracing sampling pipe, an automatic purging device with purging pressure of 0.6-0.8 MPa and a tapered anti-ash accumulation probe, the probe is made of Inconel625 with temperature resistance of ≤600°C, is provided with a built-in water cooling jacket, the cooling medium is deionized water with a flow rate of 0.5-1 L / min, and the deviation of the sampling data from the true ammonia slip is less than 3%.

[0020] Preferably, the sampling points of the sampling unit are arranged to meet the following conditions:

[0021] According to the double standards of the reactor cross-sectional area and the concentration gradient, the high-gradient zone is a zone with a concentration gradient , and the low-gradient zone is a zone with a concentration gradient , one sampling point is arranged every 0.3 m in the high-gradient zone, and one sampling point is arranged every 1.5 m in the low-gradient zone, and the vortex zone with a flow rate of less than 0.5 m / s is avoided;

[0022] The sampling pipe is inserted to a depth of 1 / 3-2 / 3 of the pipe diameter, the sampling direction is arranged at an angle of 45° opposite to the flue gas flow direction, and the same section is arranged in a matrix type and uniformly distributed;

[0023] The sampling data are processed by wavelet transform to eliminate the turbulent fluctuation interference, the wavelet transform adopts db4 wavelet basis and 3-layer decomposition, and the threshold is 1.5 times the standard deviation to ensure the data authenticity.

[0024] Preferably, the intelligent analysis and control module comprises:

[0025] The data preprocessing unit performs temperature and humidity compensation, 3σ criterion abnormal value elimination and wavelet transform to eliminate ±30% turbulent fluctuation interference on the sampling data, the wavelet transform adopts db4 wavelet basis and 3-layer decomposition, and the threshold is 1.5 times the standard deviation;

[0026] The concentration field reconstruction unit adopts the Kriging interpolation algorithm to construct the NOx concentration distribution cloud picture of the reactor cross section in combination with the dynamic self-adaptive CFD simulation results, and the spatial resolution is ≤0.5 m×0.5 m;

[0027] The adjustment decision unit adopts a PID neural network algorithm to output an instruction according to a NOx concentration deviation, a deviation change rate, in combination with an AMI index and a catalyst activity state, and to start a compensation mode when a catalyst activity retention rate is less than 80%, and to synchronously introduce a 0.5-2s mechanical delay compensation of an ammonia injection valve;

[0028] The self-learning unit trains a model based on historical data of NOx concentration, ammonia slip, AMI index and catalyst activity in the last 30 days, updates parameters every 24 hours, optimizes adjustment accuracy and catalyst-ammonia injection strategy matching degree, and synchronously feeds back the updated parameters to grid switching logic of the flow field optimization module.

[0029] Preferably, the adjustment decision unit is provided with triple control logic.

[0030] When the NOx concentration deviation is greater than 50mg / m³, a fast adjustment mode is started, and the valve adjustment amplitude is 5%-10% / s until the deviation is less than or equal to 30mg / m³.

[0031] When the deviation is less than or equal to 50mg / m³, a fine adjustment mode is enabled, and the adjustment amplitude is 0.5%-2% / s to stabilize the deviation at less than or equal to 10mg / m³.

[0032] When the AMI is less than 0.85, a mixing-ammonia injection collaborative adjustment is triggered, the static mixer blade angle is first adjusted, the adjustment amplitude is 5°-10° / time, and when the AMI does not rise to greater than or equal to 0.85 within 30s, the ammonia injection amount is corrected according to the fine adjustment mode.

[0033] Preferably, the round-robin measurement module comprises:

[0034] A movable sampling probe switches points along a preset track, the switching response time is less than or equal to 3s, the track covers high and low gradient areas of the outlet section, and the probe structure, material and distributed sampling unit are consistent;

[0035] A measurement sequence generation unit sets the sampling frequency of the high gradient area to be 2-3 times that of the low gradient area according to the NOx concentration gradient of each point in the last 1 hour, and completes a full-section polling every 1 minute;

[0036] A backflushing and cleaning unit automatically performs a 3s high-pressure purging after each point switching, the purging pressure is 0.6-0.8MPa, and the probe particulate matter is removed by additional purging every 5 minutes;

[0037] A data storage unit stores time-averaged concentration data processed by wavelet transform in real time, which is used for AMI calculation and trend analysis.

[0038] Preferably, the round-robin measurement module and the distributed sampling module form a triple data verification mechanism.

[0039] When the NOx concentration deviation in the same region is greater than 10%, the ammonia escape concentration deviation is greater than 8%, or the AMI index deviation is greater than 5%, the self-checking program is automatically started;

[0040] The self-checking successively checks the sampling pipeline patency, the analyzer calibration state, the probe position accuracy, and the dynamic self-adaptive CFD grid switching logic. The sampling pipeline patency is judged by the decay rate of the purge pressure. The decay is less than or equal to 10% per minute.

[0041] Preferably, the execution module comprises:

[0042] The control valve is electrically and hydraulically servo-driven, has a working pressure of 10-15 MPa, a built-in high-frequency response valve core, a full-stroke regulation time of less than or equal to 0.5 s, and a flow characteristic matching the ammonia injection grid aperture gradient.

[0043] The intelligent positioner adopts a digital signal processing technology, has a positioning accuracy of less than or equal to 0.1% FS, supports a high-speed PWM control signal with a frequency of 10-20 kHz, and receives the AMI index feedback in real time.

[0044] The double-channel feedback unit feeds back through the absolute value encoder and the pressure sensor, accesses the catalyst outlet NOx degradation rate, and forms the “opening-flow-reaction effect” verification.

[0045] The pre-action compensation unit predicts the regulation trend and the 0.5-2 s mechanical delay, pre-regulates the valve by 0.1-0.2 s in advance, and offsets the mechanical inertia.

[0046] When the actual opening deviation is greater than 1%, the deviation compensation is completed within 50 ms. When the AMI does not meet the standard after the valve is regulated, the intelligent analysis control module is automatically fed back to adjust the grid simulation parameters.

[0047] Preferably, the remote monitoring module comprises:

[0048] The data display unit displays the dynamic self-adaptive CFD flow field simulation, the sampling point concentration, the AMI index, the valve state, and the catalyst activity in real time.

[0049] The data query unit supports the query and trend analysis of nearly one year of historical data, and the dimensions include the NOx concentration, the ammonia escape, the AMI index, the catalyst activity, and the valve regulation record.

[0050] The warning unit pushes the warning to the specified terminal within 10 s when the NOx outlet concentration is greater than 10% of the set value, the ammonia escape is greater than 8 ppm, the AMI is less than 0.8, or the valve opening deviation is greater than 2%, the probe purge pressure is abnormal, the analyzer calibration fails, or the catalyst NOx degradation rate decreases by more than 5% per 24 hours.

[0051] The remote control unit supports the authorized user to issue a grid template adjustment and a purging parameter modification instruction, and the authority is set in stages, that is, an administrator can modify all parameters, and an operator can only query data and modify purging parameters, the instruction needs to be executed after being confirmed by two persons, and the real-time feedback result is obtained after the instruction is executed.

[0052] Compared with the related art, the SCR denitration process precise ammonia injection and sampling system has the following beneficial effects:

[0053] 1. The SCR denitration process precise ammonia injection and sampling system, through dynamic self-adaptive multi-dimensional simulation of the flow field optimization module, combined with multi-scale grid pool and ammonia injection valve mechanical delay compensation mechanism, makes the uniformity of flue gas cross-section velocity distribution ≥95%; at the same time, the distributed sampling module accurately distributes points according to the principle of high gradient area encryption-low gradient area sparseness, cooperates with wavelet transform to eliminate turbulent pulsation interference, ensures that the deviation of sampling data and real ammonia escape is less than 3%, and on this basis, the intelligent analysis control module relies on the PID neural network algorithm and triple control logic, can flexibly switch the adjustment mode according to the NOx concentration deviation, and stabilizes the deviation at ≤10mg / m³, and can start the compensation mode combined with the catalyst activity state, not only significantly improves the denitration efficiency, but also avoids the problems of resource waste caused by excessive ammonia injection and substandard denitration caused by insufficient ammonia in the traditional process, realizes the precise and efficient use of ammonia resources.

[0054] 2. The SCR denitration process precise ammonia injection and sampling system, through the flow field optimization module to define "working condition fluctuation intensity factor F", can quantize the fluctuation of flue gas flow rate, temperature, NOx and NH3 concentration in real time, and trigger the automatic switching of coarse, medium and fine grids through online sensors, adapt to different fluctuation intensity working conditions; the detachable static mixer of the hybrid enhancement submodule can adjust the blade angle in real time according to the sampling data and AMI index of the mixing section to maintain AMI ≥0.85. In addition, the round-robin measurement module and the distributed sampling module form a triple data verification mechanism, which automatically starts self-checking when the data deviation is out of limit, checks the pipeline smoothness, analyzer calibration state and other key links, and guarantees the data reliability. At the same time, the self-learning unit of the intelligent analysis control module updates the model parameters based on the historical data of the last 30 days every day, continuously optimizes the adjustment precision and the matching degree of catalyst-ammonia injection strategy, so that the system can not only cope with short-term working condition fluctuations, but also adapt to catalyst aging and other changes, and the long-term operation stability is greatly enhanced.

[0055] 3, The present application provides SCR denitration process precise ammonia injection and sampling system, through remote monitoring module realizes real-time visual display of system core parameters, supports multi-dimensional historical data query and trend analysis for more than 1 year, facilitates operation and maintenance personnel to quickly master the operation condition and trace the problem source; the early warning unit can push the early warning to the specified terminal within 10s for the risk scenarios such as NOx outlet concentration exceeding standard, ammonia escape anomaly, valve opening deviation, etc., reduce the lag time of abnormal disposal, and the remote control unit adopts permission grading design, the administrator and the operator perform their respective functions, the instruction needs to be confirmed by two people before execution, avoid the risk of misoperation, and the control valve of the execution module is adjusted within 0.5s, the pre-action compensation unit can offset the mechanical inertia in advance, complete the opening deviation compensation within 50ms, reduce the frequency of manual intervention; each sampling unit is equipped with automatic purging device and anti-accumulation probe, reduce the probability of probe blockage and equipment damage, reduce the on-site maintenance workload and equipment maintenance cost of operation and maintenance personnel, improve the overall operation and maintenance efficiency and the management is more safe and standardized. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 The flowchart of the present application;

[0057] Figure 2 The extension flowchart of the flow field optimization module of the present application;

[0058] Figure 3 The extension flowchart of the distributed sampling module of the present application;

[0059] Figure 4 The extension flowchart of the intelligent analysis and control module of the present application;

[0060] Figure 5 The extension flowchart of the round-robin measurement module of the present application;

[0061] Figure 6 The extension flowchart of the execution module of the present application;

[0062] Figure 7 The extension flowchart of the remote monitoring module of the present application. DETAILED DESCRIPTION

[0063] The technical solutions in the embodiments of the present application will be described in detail below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all embodiments; based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0064] Embodiment one:

[0065] Please refer to Figures 1-7The application provides a technical scheme: an SCR denitration process precise ammonia injection and sampling system, comprising a flow field optimization module, a distributed sampling module, an intelligent analysis control module, a round-robin measurement module, an execution module and a remote monitoring module, each module forms an analog-evaluation-sampling-control-feedback closed loop, and the feedback priority is to first correct the ammonia injection instruction and then adjust the flow field simulation parameters.

[0066] The flow field optimization module optimizes the flue gas flow field, ammonia gas flow field and gas-gas mixed flow field in the reactor through dynamic self-adaptive multi-dimensional simulation, simulates the working condition perception, multi-scale grid dynamic switching and ammonia injection valve mechanical delay compensation mechanism;

[0067] The flow field optimization module comprises:

[0068] The flue gas flow guide sub-module adopts a dynamic self-adaptive computational fluid dynamics technology, defines a "working condition fluctuation intensity factor" to quantify the real-time fluctuations of the flue gas flow rate, temperature, NOx and NH3 concentration, and the working condition fluctuation intensity factor calculation formula is , wherein is the flow rate fluctuation value, is the reference flow rate, is the temperature fluctuation value, is the reference temperature, is the NOx concentration fluctuation value, is the reference concentration, a "multi-scale grid pool" comprising a coarse grid, a medium grid and a fine grid is established, the coarse grid corresponds to a working condition of F≤0.1, the medium grid corresponds to a working condition of 0.1<F≤0.3, and the fine grid corresponds to a working condition of F>0.3, the grid automatic switching is triggered through a Pitot tube speed measurement, an infrared temperature measurement and other online sensors, a 0.5~2s ammonia injection valve mechanical delay compensation is introduced, the three-dimensional layout of the flow guide plate and the 30°-60° inclination angle are designed based on the simulation results, and the uniformity of the flue gas cross-sectional velocity distribution is ≥95%;

[0069] The ammonia gas injection optimization sub-module plans the injection hole aperture gradient distribution and spatial density according to the ammonia gas flow field simulation results and AMI index;

[0070] The mixed enhancement sub-module is configured with a detachable static mixer, the blade angle is adjusted in real time according to the sampling data of the mixing section and the AMI index feedback, and the AMI is maintained to be ≥0.85;

[0071] In this embodiment, through the synergistic effect of the flue gas flow guide module, the ammonia gas injection optimization module and the mixing enhancement module, the technical pain points of uneven flue gas flow field distribution, low ammonia injection and flue gas mixing efficiency and poor working condition fluctuation adaptability in traditional SCR denitration process are solved. The "working condition fluctuation intensity factor F" defined by the flue gas flow guide module can quantify the real-time fluctuations of flue gas flow rate, temperature, NOx and NH3 concentration. Combined with the "multi-scale grid pool", the grid is automatically switched. When the working condition is stable, coarse grid is used to improve the simulation efficiency. When the working condition fluctuates greatly, fine grid is switched to ensure the simulation accuracy. Cooperate with pitot tube speed measurement, infrared temperature measurement and other online sensors to ensure that the flow field simulation is highly matched with the actual working condition. At the same time, introduce 0.5-2s ammonia injection valve mechanical delay compensation to avoid ammonia excess or deficiency caused by valve action lag. Finally, through the three-dimensional layout of the 30°-60° inclined guide plate, the uniformity of flue gas cross-section velocity distribution is ≥95%, which lays the foundation for subsequent ammonia and flue gas mixing. The ammonia gas injection optimization module plans the injection hole diameter gradient distribution and spatial density based on the ammonia gas flow field simulation results and ammonia and flue gas mixing efficiency indicators to avoid local ammonia concentration being too high or too low. The detachable static mixer of the mixing enhancement module can adjust the blade angle in real time according to the sampling data and AMI indicators of the mixing section to forcibly maintain AMI≥0.85, effectively solving the problem of fixed blade angle of traditional mixer and the problem of decreasing mixing efficiency with working condition change.

[0072] The distributed sampling module is arranged at the inlet section, mixing section and outlet section of the reactor. The sampling points are arranged according to the principle of "high gradient area encryption-low gradient area sparseness". The high gradient area is a region with concentration gradient , and the low gradient area is a region with concentration gradient . A "multi-point synchronization + polling" sampling mode is adopted, and wavelet transform is used to eliminate turbulent pulsation interference.

[0073] The sampling unit sampling point setting satisfies:

[0074] According to the double standards of reactor cross-sectional area and concentration gradient, the high gradient area is a region with concentration gradient , and the low gradient area is a region with concentration gradient . One sampling point is arranged every 0.3m in the high gradient area, and one sampling point is arranged every 1.5m in the low gradient area, avoiding the vortex area with flow rate <0.5m / s.

[0075] The insertion depth of the sampling tube is 1 / 3-2 / 3 of the pipe diameter, and the sampling direction is arranged at 45° against the flue gas flow direction. The same section is arranged in a matrix.

[0076] The sampling data are processed by wavelet transform to eliminate turbulent pulsation interference. The wavelet transform uses db4 wavelet basis and 3-layer decomposition, and the threshold value is 1.5 times the standard deviation to ensure data authenticity.

[0077] The sampling unit sampling point setting satisfies:

[0078] According to the reactor cross-sectional area and the concentration gradient, the high gradient area is a region with a concentration gradient , and the low gradient area is a region with a concentration gradient ; one sampling point is arranged every 0.3 m in the high gradient area, and one sampling point is arranged every 1.5 m in the low gradient area, and the vortex area with a flow rate of less than 0.5 m / s is avoided;

[0079] The insertion depth of the sampling pipe is 1 / 3-2 / 3 of the pipe diameter, the sampling direction is arranged at an angle of 45° opposite to the flue gas flow direction, and the same section is uniformly distributed in a matrix form;

[0080] The sampling data is processed by wavelet transform to eliminate turbulent fluctuation interference, the wavelet transform adopts db4 wavelet basis and 3-layer decomposition, and the threshold value is 1.5 times the standard deviation to ensure data authenticity;

[0081] In the embodiment, through the design of “segmented sampling + gradient point distribution + data accurate processing + hardware anti-interference”, the problems of unreasonable sampling point distribution, large data interference from turbulent fluctuation and ash deposition, and poor representativeness of the traditional sampling system are solved; the sampling unit of the inlet section is arranged at 3-5D before the reactor inlet to avoid the vortex area with a flow rate of less than 0.5 m / s, so as to ensure that the collected original flue gas NOx concentration and oxygen content data truly reflect the inlet working condition; the sampling unit of the mixing section is arranged at a key position 1-2 m downstream of the ammonia injection grid, and is arranged according to the encryption-sparse principle of “one point every 0.3 m in the high gradient area and one point every 1.5 m in the low gradient area” to accurately capture the concentration gradient change after ammonia and flue gas mixing; the sampling unit of the outlet section collects at least 100 groups of data through the “multi-point synchronization + polling” mode, and then processes the data through wavelet transform with db4 wavelet basis, 3-layer decomposition and a threshold value of 1.5 times the standard deviation to eliminate ±30% turbulent fluctuation interference and obtain reliable time-averaged concentration; in terms of hardware, the heating type sampling pipe arranged in each sampling unit can prevent flue gas condensation, the automatic purging device with a pressure of 0.6-0.8 MPa once every 5 minutes, and the tapered anti-ash deposition probe can effectively prevent probe ash deposition and high-temperature damage, so as to finally ensure that the deviation of the sampling data from the real ammonia escape is less than 3%, and provide high-quality data input for subsequent intelligent analysis.

[0082] The intelligent analysis control module receives the sampling data, generates an ammonia injection adjustment instruction based on the ammonia and flue gas mixing efficiency quantitative index, a PID neural network algorithm model and a catalyst activity state;

[0083] The intelligent analysis control module comprises:

[0084] The data preprocessing unit performs temperature and humidity compensation, 3σ criterion outlier rejection and wavelet transform to eliminate ±30% turbulent fluctuation interference on the sampling data, the wavelet transform adopts db4 wavelet basis, 3-layer decomposition and a threshold value of 1.5 times the standard deviation;

[0085] The concentration field reconstruction unit, combined with the dynamic adaptive CFD simulation results, adopts the Kriging interpolation algorithm to construct the NOx concentration distribution cloud picture of the reactor cross section, with a spatial resolution of ≤0.5 m x 0.5 m;

[0086] The adjustment decision unit adopts the PID neural network algorithm to output instructions according to the NOx concentration deviation, the deviation change rate, combined with the AMI index and the catalyst activity state, and starts the compensation mode when the catalyst activity retention rate is <80%, and synchronously introduces the 0.5~2s ammonia injection valve mechanical delay compensation;

[0087] The self-learning unit trains the model based on the historical data of NOx concentration, ammonia slip, AMI index and catalyst activity in the past 30 days, updates the parameters every 24 hours, optimizes the adjustment accuracy and the matching degree of catalyst-ammonia injection strategy, and synchronously feeds back the updated parameters to the grid switching logic of the flow field optimization module;

[0088] The adjustment decision unit has triple control logic:

[0089] When the NOx concentration deviation is >50 mg / m³, the fast adjustment mode is started, the valve adjustment amplitude is 5%-10% / s, and the deviation is ≤30 mg / m³ until the deviation is ≤30 mg / m³;

[0090] When the deviation is ≤50 mg / m³, the fine adjustment mode is enabled, the adjustment amplitude is 0.5%-2% / s, and the deviation is stabilized at ≤10 mg / m³;

[0091] When AMI is <0.85, the mixing-ammonia injection collaborative adjustment is triggered, the static mixer blade angle is adjusted first, the adjustment amplitude is 5°-10° / time, and if AMI does not rise to ≥0.85 within 30s, the ammonia injection amount is corrected according to the fine adjustment mode;

[0092] This implementation scheme addresses the problems of traditional SCR control systems, such as reliance on human experience, low adjustment accuracy, slow response, and lack of adaptive operation, through a complete process design of "data preprocessing - concentration field reconstruction - dynamic decision-making - self-learning optimization". The data preprocessing unit uses temperature and humidity compensation to eliminate environmental influences and the 3σ criterion to remove outliers. Combined with wavelet transform technology consistent with the distributed sampling module, the reliability of the sampled data is further guaranteed. The concentration field reconstruction unit combines the results of dynamic adaptive CFD simulation and uses the Kriging interpolation algorithm to construct a NOx concentration distribution cloud map with a spatial resolution of ≤0.5m×0.5m. This visually presents the uneven concentration areas within the reactor, providing a visual basis for precise adjustment and decision-making. The unit is based on a PID neural network algorithm. According to the NOx concentration deviation, the rate of change of deviation, combined with the AMI index and the catalyst activity status, the output command is: when the catalyst activity retention rate is <80%, the compensation mode is activated to avoid the decrease in denitrification efficiency caused by catalyst aging; a mechanical delay compensation of 0.5~2s for the ammonia injection valve is introduced simultaneously to offset the effect of valve action lag; the triple control logic can be flexibly switched according to the actual working conditions to ensure that the NOx concentration deviation is stable at ≤10mg / m³. The self-learning unit updates the model parameters daily based on the historical data of the past 30 days, optimizes the adjustment accuracy and the matching degree of the catalyst-ammonia injection strategy, and feeds the parameters back to the flow field optimization module to adjust the grid switching logic, so as to achieve continuous optimization of the overall system performance.

[0093] The circulating measurement module is located at the catalyst outlet and monitors gas parameters according to the concentration gradient priority.

[0094] The cyclic measurement module includes:

[0095] The movable sampling probe can switch points along a preset track with a switching response time of ≤3s. The track covers the high and low gradient areas of the exit section. The probe structure and materials are consistent with the distributed sampling unit.

[0096] The measurement sequence generation unit, based on the NOx concentration gradient of each point in the past hour, sets the sampling frequency of the high gradient area to be 2-3 times that of the low gradient area, and completes a full cross-section polling once every 1 minute;

[0097] The backflush cleaning unit automatically performs a 3-second high-pressure purge after each point switch, with a purge pressure of 0.6-0.8 MPa. It also performs an additional purge every 5 minutes to remove particles from the probe.

[0098] The data storage unit stores time-averaged concentration data processed by wavelet transform in real time for AMI calculation and trend analysis.

[0099] The polling measurement module and the distributed sampling module form a triple data verification mechanism:

[0100] When the NOx concentration deviation in the same area is greater than 10%, the ammonia escape concentration deviation is greater than 8%, or the AMI index deviation is greater than 5%, the self-checking program is automatically started;

[0101] The self-checking successively checks the sampling pipeline patency, the analyzer calibration state, the probe position accuracy, and the dynamic self-adaptive CFD grid switching logic. The sampling pipeline patency is judged by the decay rate of the purging pressure. The decay is less than or equal to 10% per minute.

[0102] In the embodiment, the problems of limited coverage, poor data timeliness, and no self-checking mechanism of traditional fixed-point measurement are solved by the design of "mobile monitoring + gradient frequency conversion + active cleaning + triple verification". The movable sampling probe switches points along the preset track with a switching response time less than or equal to 3 seconds. The track covers the high and low gradient areas of the catalyst outlet section. The probe structure and material are consistent with the distributed sampling unit, ensuring the uniformity of the measurement standard. The measurement sequence generation unit sets the sampling frequency of the high gradient area to 2-3 times that of the low gradient area according to the NOx concentration gradient of each point in nearly 1 hour. The full-section polling is completed every 1 minute, which ensures the monitoring density of the high gradient key area and avoids redundant measurement in the low gradient area, improving the monitoring efficiency. The backflushing cleaning unit automatically performs high-pressure purging of 3 seconds and 0.6-0.8 MPa after each point switching, and additionally purges every 5 minutes, effectively removing particles on the probe surface and preventing blockage from affecting measurement accuracy. The data storage unit stores the time-averaged concentration data processed by wavelet transform in real time, providing continuous data support for AMI calculation and denitration trend analysis. In addition, the triple data verification mechanism formed by the round-robin measurement module and the distributed sampling module automatically starts the self-checking when the NOx concentration deviation in the same area is greater than 10%, the ammonia escape concentration deviation is greater than 8%, or the AMI index deviation is greater than 5%. The sampling pipeline patency, the analyzer calibration state, the probe position accuracy, and the CFD grid switching logic are successively checked to ensure the reliability of the measurement data and provide a reliable verification basis for the intelligent analysis and control module.

[0103] The execution module adjusts the ammonia injection amount by regulating the valve;

[0104] The execution module includes:

[0105] The valve is electrically and hydraulically servo-driven, with a working pressure of 10-15 MPa, a built-in high-frequency response valve core, and a full-stroke regulation time less than or equal to 0.5 seconds. The flow characteristics match the ammonia injection grid aperture gradient.

[0106] The intelligent positioner adopts digital signal processing technology, with a positioning accuracy less than or equal to 0.1% FS and supports high-speed PWM control signals with a frequency of 10-20 kHz. The intelligent positioner receives AMI index feedback in real time.

[0107] Double-channel feedback unit, through absolute value encoder and pressure sensor double-parameter feedback, access to catalyst outlet NOx degradation rate, form "opening-flow-reaction effect" verification;

[0108] Pre-action compensation unit, according to the adjustment trend prediction and 0.5~2s mechanical delay, 0.1-0.2s pre-adjustment valve, offset mechanical inertia;

[0109] When the actual opening and command deviation is >1%, the deviation compensation is completed within 50ms; When AMI does not meet the standard after valve adjustment, it is automatically fed back to the intelligent analysis control module to adjust the grid simulation parameters;

[0110] In this embodiment, through the design of "high-precision control + real-time feedback + pre-compensation", the problems of slow response, low positioning accuracy, large adjustment deviation and no effect verification of traditional ammonia injection actuator are solved. The control valve adopts electro-hydraulic servo drive, the working pressure reaches 10-15MPa, the built-in high-frequency response valve core, the full stroke adjustment time is ≤0.5s, and the flow characteristic is matched with the ammonia injection grid aperture gradient, which ensures that the ammonia injection amount can quickly and accurately follow the change of the adjustment command. The intelligent positioner adopts digital signal processing technology, the positioning accuracy is ≤0.1%FS, supports 10-20kHz high-speed PWM control signal, and receives AMI index feedback in real time, which can dynamically correct the valve positioning according to the mixing efficiency, avoid invalid adjustment, and the double-channel feedback unit through absolute value encoder and pressure sensor double-parameter feedback, access to catalyst outlet NOx degradation rate, form "opening-flow-reaction effect" closed loop verification, ensure that the adjustment action can effectively improve the denitration effect; The pre-action compensation unit according to the adjustment trend prediction and 0.5~2s valve mechanical delay, 0.1-0.2s pre-adjustment valve, offset mechanical inertia influence; When the actual opening and command deviation is >1%, the deviation compensation is completed within 50ms, which further improves the adjustment accuracy. If AMI still does not meet the standard after valve adjustment, the execution module will automatically feed back to the intelligent analysis control module, trigger the flow field simulation parameter adjustment, form "execution-checking-optimization" secondary closed loop.

[0111] The remote monitoring module displays system parameters in real time, and the distributed sampling data is transmitted to the intelligent analysis control module in real time. The round-robin measurement data is used for verification, and the intelligent analysis control module sends instructions according to the verification data;

[0112] The remote monitoring module comprises:

[0113] The data display unit displays dynamic adaptive CFD flow field simulation, sampling point concentration, AMI index, valve state and catalyst activity in real time;

[0114] The data query unit supports query and trend analysis of nearly 1 year of historical data, including NOx concentration, ammonia slip, AMI index, catalyst activity and valve adjustment record;

[0115] Early warning unit, when the NOx outlet concentration exceeds the set value of 10%, ammonia escape > 8 ppm, AMI < 0.8, or valve opening deviation > 2%, probe purge pressure is abnormal, analyzer calibration failure, or catalyst NOx degradation rate decreases > 5% / 24h, push the early warning to the designated terminal within 10s;

[0116] Remote control unit, support authorized users to issue grid template adjustment, purge parameter modification instructions, permission level setting is that administrators can modify all parameters, operation and maintenance personnel can only query data and modify purge parameters, instructions need to be executed after two people confirm, and the results are fed back in real time after the execution of the instructions;

[0117] In the embodiment, through the design of "real-time visualization + data traceability + active early warning + hierarchical control", the problems of traditional monitoring system information lag, data query difficulty, slow abnormal response and operation permission confusion are solved. The data display unit displays the dynamic adaptive CFD flow field simulation results, the concentration data of each sampling point, the AMI index, the valve working state and the catalyst activity in real time, presents the overall operation condition of the system with a visual interface, and facilitates the operation and maintenance personnel to master the core parameters in real time; the data query unit supports near 1 year historical data query and trend analysis, covering multi-dimensional data such as NOx concentration, ammonia escape, AMI index, catalyst activity and valve adjustment record, which can assist operation and maintenance personnel to trace the root cause of the problem, optimize operation strategy, and the early warning unit pushes the early warning information to the designated terminal within 10s for multiple key abnormal scenes, ensures that the abnormal situation is found and disposed in time, and the remote control unit adopts the permission level design: administrators can modify all parameters, operation and maintenance personnel can only query data and modify purge parameters, and all instructions need to be executed after two people confirm to avoid misoperation; the results are fed back in real time after the execution of the instructions, ensuring the safety and effectiveness of remote operation, realizing the remote and efficient operation and maintenance of the SCR denitration system.

[0118] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (or computer- readable storage media) having computer-usable program code embodied in the medium. The medium can be any available medium or combination thereof that is accessible by a general purpose or special purpose computer. By way of example, such computer-usable storage media can include a volatile memory, a non-volatile memory, a hard disk drive, a solid state drive, a magnetic diskette, an optical disk (e.g., a compact disk or a DVD), a magnetic tape, a flash memory device, or in general any storage medium that can be used to store and / or transfer data or computer Figure 1 one or more functions specified in the flow or flows and / or blocks Figure 1 one or more functions specified in the flow or flows and / or blocks

[0119] It should be noted that the above-mentioned embodiments are only used to illustrate but not to limit the technical solutions of the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or changed without departing from the spirit and scope of the technical solutions of the present application, and all modifications and changes should be covered in the scope of the claims of the present application.

Claims

1. A precision ammonia injection and sampling system for SCR denitrification process, characterized in that, It includes a flow field optimization module, a distributed sampling module, an intelligent analysis and control module, a cyclic measurement module, an execution module, and a remote monitoring module. Each module forms a simulation-evaluation-sampling-control-feedback closed loop. The feedback priority is to first correct the ammonia injection command, and then adjust the flow field simulation parameters. The flow field optimization module optimizes the flue gas flow field, ammonia flow field and gas-gas mixture flow field in the reactor through dynamic adaptive multi-dimensional simulation, including operating condition perception, dynamic switching of multi-scale grids and mechanical delay compensation mechanism of ammonia injection valve. Distributed sampling modules are deployed at the reactor inlet, mixing, and outlet sections, with sampling points arranged according to the principle of "intensified sampling in high gradient regions and sparse sampling in low gradient regions," where the high gradient regions are the concentration gradients. In the region where the gradient is low, the region is a concentration gradient. In the region, a "multi-point synchronization + polling" sampling mode is adopted and wavelet transform is used to eliminate turbulent pulsation interference; The intelligent analysis and control module receives sampled data and generates ammonia injection adjustment commands based on the quantitative index of ammonia-fume mixing efficiency, the PID neural network algorithm model, and the catalyst activity state. The circulating measurement module is located at the catalyst outlet and monitors gas parameters according to the concentration gradient priority. The execution module regulates the ammonia injection rate by controlling the valves; The remote monitoring module displays system parameters in real time, and the distributed sampling data is transmitted to the intelligent analysis and control module in real time. The cyclic measurement data is used for verification, and the intelligent analysis and control module sends instructions based on the verification data.

2. The precision ammonia injection and sampling system for SCR denitrification process according to claim 1, characterized in that: The flow field optimization module includes: The flue gas flow guiding submodule employs dynamic adaptive computational fluid dynamics technology, defining an "operating condition fluctuation intensity factor" to quantify real-time fluctuations in flue gas velocity, temperature, NOx, and NH3 concentrations. The calculation formula for the operating condition fluctuation intensity factor is as follows: ,in For flow velocity fluctuation value, As the reference flow rate, For temperature fluctuation values, As the reference temperature, NOx concentration fluctuation value, As a baseline concentration, a "multi-scale grid pool" containing coarse, medium, and fine grids was established. The coarse grid corresponds to the operating condition of F≤0.1, the medium grid corresponds to the operating condition of 0.1<F≤0.3, and the fine grid corresponds to the operating condition of F>0.

3. The grid switching is automatically triggered by online sensors such as Pitot tube velocity measurement and infrared temperature measurement. A mechanical delay compensation of 0.5~2s for the ammonia injection valve is introduced. Based on the simulation results, the three-dimensional layout of the guide vane and the tilt angle of 30°-60° are designed to make the uniformity of flue gas cross-section velocity distribution ≥95%. The ammonia injection optimization submodule plans the orifice diameter gradient distribution and spatial density based on the ammonia flow field simulation results and AMI index. The hybrid enhancement submodule is equipped with a detachable static mixer. The blade angle is adjusted in real time based on the sampling data of the mixing section and the feedback of the AMI index to maintain AMI≥0.

85.

3. The precision ammonia injection and sampling system for SCR denitrification process according to claim 1, characterized in that: The distributed sampling module includes: The inlet sampling unit is located 3-5D before the reactor inlet to collect the original flue gas NOx concentration and oxygen content. The sampling point avoids the eddy region with a flow velocity of <0.5m / s. The mixed section sampling unit is located 1-2m downstream of the ammonia injection grid to collect NOx concentration and ammonia slip concentration. One sampling point is set every 0.3m in the high gradient zone and one sampling point is set every 1.5m in the low gradient zone. The sampling unit at the outlet section is located 0.5-1m downstream of the catalyst layer. It collects flue gas parameters after the reaction by using "multi-point synchronous sampling + polling sampling". After collecting at least 100 sets of data, wavelet transform is used to eliminate ±30% turbulent fluctuation interference. The wavelet transform uses db4 wavelet basis and 3-level decomposition. The threshold is 1.5 times the standard deviation to obtain the time-averaged concentration. Each sampling unit is equipped with a heated sampling tube, an automatic purging device that purges every 5 minutes at a pressure of 0.6-0.8 MPa, and a tapered anti-dust probe made of Inconel 625 with a temperature resistance of ≤600℃. The probe has a built-in water-cooled jacket, and the cooling medium is deionized water with a flow rate of 0.5-1 L / min, ensuring that the sampling data deviates from the actual ammonia escape by less than 3%.

4. The precision ammonia injection and sampling system for SCR denitrification process according to claim 3, characterized in that: The sampling points of the sampling unit are set to satisfy the following: Based on both reactor cross-sectional area and concentration gradient, the high gradient region represents the concentration gradient. In the region where the gradient is low, the region is a concentration gradient. In the high gradient region, one sampling point is set every 0.3m, and in the low gradient region, one sampling point is set every 1.5m, avoiding the eddy current region with a flow velocity <0.5m / s; The sampling tube is inserted to a depth of 1 / 3 to 2 / 3 of the pipe diameter, and the sampling direction is arranged at a 45° angle to the flue gas flow direction, with the same cross section being uniformly distributed in a matrix pattern. The sampled data were processed by wavelet transform to eliminate turbulence and fluctuation interference. The wavelet transform used a db4 wavelet basis and a 3-level decomposition, with a threshold of 1.5 times the standard deviation to ensure data authenticity.

5. The precision ammonia injection and sampling system for SCR denitrification process according to claim 1, characterized in that: The intelligent analysis and control module includes: The data preprocessing unit performs temperature and humidity compensation, 3σ criterion outlier removal, and wavelet transform to eliminate ±30% turbulent fluctuation interference on the sampled data. The wavelet transform uses the db4 wavelet basis and 3-level decomposition, with a threshold of 1.5 times the standard deviation. The concentration field reconstruction unit, combined with the dynamic adaptive CFD simulation results, uses the Kriging interpolation algorithm to construct a NOx concentration distribution cloud map of the reactor cross section with a spatial resolution of ≤0.5m×0.5m; The adjustment decision unit adopts a PID neural network algorithm. Based on the NOx concentration deviation, deviation change rate, AMI index and catalyst activity status, it outputs instructions. When the catalyst activity retention rate is <80%, the compensation mode is activated, and a mechanical delay compensation of 0.5~2s for the ammonia injection valve is introduced simultaneously. The self-learning unit trains the model based on historical data of NOx concentration, ammonia slip, AMI index, and catalyst activity over the past 30 days. The parameters are updated every 24 hours to optimize the adjustment accuracy and the matching degree of catalyst-ammonia injection strategy. The updated parameters are then synchronously fed back to the grid switching logic of the flow field optimization module.

6. The precision ammonia injection and sampling system for SCR denitrification process according to claim 5, characterized in that: The adjustment decision unit is equipped with triple control logic: When the NOx concentration deviation is >50mg / m³, the rapid adjustment mode is activated, and the valve is adjusted by 5%-10% / s until the deviation is ≤30mg / m³. When the deviation is ≤50mg / m³, the fine adjustment mode is activated, with an adjustment range of 0.5%-2% / s, to stabilize the deviation at ≤10mg / m³. When AMI < 0.85, the mixing-ammonia injection coordinated adjustment is triggered. First, the angle of the static mixer blades is adjusted, with an adjustment range of 5°-10° / time. If AMI does not rise to ≥ 0.85 within 30 seconds, the ammonia injection amount is corrected according to the fine adjustment mode.

7. The precision ammonia injection and sampling system for SCR denitrification process according to claim 1, characterized in that: The polling measurement module includes: The movable sampling probe can switch points along a preset track with a switching response time of ≤3s. The track covers the high and low gradient areas of the exit section. The probe structure and materials are consistent with the distributed sampling unit. The measurement sequence generation unit, based on the NOx concentration gradient of each point in the past hour, sets the sampling frequency of the high gradient area to be 2-3 times that of the low gradient area, and completes a full cross-section polling once every 1 minute; The backflush cleaning unit automatically performs a 3-second high-pressure purge after each point switch, with a purge pressure of 0.6-0.8 MPa. It also performs an additional purge every 5 minutes to remove particles from the probe. The data storage unit stores time-averaged concentration data processed by wavelet transform in real time for AMI calculation and trend analysis.

8. The SCR denitrification process precision ammonia injection and sampling system according to claim 7, characterized in that: The polling measurement module and the distributed sampling module form a triple data verification mechanism: When the NOx concentration deviation is greater than 10%, the ammonia slip concentration deviation is greater than 8%, or the AMI index deviation is greater than 5% in the same area, the self-test program will be automatically started. The self-test sequentially checks the patency of the sampling pipeline, the calibration status of the analyzer, the accuracy of the probe position, and the dynamic adaptive CFD grid switching logic. The patency of the sampling pipeline is judged by the purge pressure decay rate; decay ≤10% / min indicates patency.

9. The precision ammonia injection and sampling system for SCR denitrification process according to claim 1, characterized in that: The execution module includes: The control valve is electro-hydraulic servo driven, with a working pressure of 10-15MPa. It has a built-in high-frequency response valve core, and the full stroke adjustment time is ≤0.5s. The flow characteristics are matched with the orifice gradient of the ammonia injection grid. The intelligent locator uses digital signal processing technology, with a positioning accuracy of ≤0.1%FS. It supports high-speed PWM control signals with a frequency of 10-20kHz and receives AMI index feedback in real time. The dual-channel feedback unit uses a dual-parameter feedback mechanism of an absolute encoder and a pressure sensor to input the NOx degradation rate at the catalyst outlet, forming a "opening degree-flow rate-reaction effect" verification. The pre-action compensation unit, based on the predicted adjustment trend and a mechanical delay of 0.5~2s, pre-adjusts the valve 0.1-0.2s in advance to counteract mechanical inertia; When the actual opening deviation from the command is greater than 1%, the deviation compensation is completed within 50ms; when the AMI does not meet the standard after valve adjustment, it is automatically fed back to the intelligent analysis and control module to adjust the grid simulation parameters.

10. The SCR denitrification process precision ammonia injection and sampling system according to claim 1, characterized in that: The remote monitoring module includes: The data display unit shows in real time dynamic adaptive CFD flow field simulation, sampling point concentration, AMI index, valve status, and catalyst activity; The data query unit supports querying and trend analysis of historical data for the past year, with dimensions including NOx concentration, ammonia slip, AMI index, catalyst activity, and valve adjustment records. The early warning unit will push an early warning to the designated terminal within 10 seconds when the NOx outlet concentration exceeds the set value by 10%, ammonia slip is >8ppm, AMI is <0.8, or valve opening deviation is >2%, probe purging pressure is abnormal, analyzer calibration fails, or catalyst NOx degradation rate decreases by >5% / 24h. The remote control unit supports authorized users to issue commands for adjusting grid templates and modifying purging parameters. The permission hierarchy is set so that administrators can modify all parameters, while maintenance personnel can only query data and modify purging parameters. Commands must be confirmed by two people before execution, and the results are fed back in real time after the command is executed.

Images