A predictive adaptive ammonia injection control method, system, equipment, and medium for multi-field coupling in a coal-fired power unit SCR system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUIZHOU CHUANGXING ELECTRIC POWER RES INST CO LTD
- Filing Date
- 2026-01-06
- Publication Date
- 2026-05-26
Smart Images
Figure CN122085658A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of denitrification control technology in thermal power generation, specifically to a predictive adaptive ammonia injection control method, system, equipment, and medium for multi-field coupling in the SCR system of a coal-fired unit. Background Technology
[0002] Under multiple constraints on coal-fired power, the NO of existing coal-fired units x Emissions ≤50mg / Nm³, and selective catalytic reduction (SCR) is the core technology of denitrification system in coal-fired units. Its ammonia injection control precision directly determines denitrification efficiency, ammonia slip level and operating economy.
[0003] With the increasing penetration rate of new energy sources, coal-fired power units face higher requirements for supply security and frequency regulation / peak shaving. The operating conditions of coal-fired power units have also undergone profound changes, leading to complex challenges in the ammonia injection control of the SCR system in coal-fired units: wide load fluctuation range (20%–100% rated load), rapid load change rate, frequent changes in the quality of coal fed into the furnace, and strong coupling between combustion conditions and flue gas flow characteristics within the boiler furnace, resulting in increased NOx emissions. x The generation and transport processes exhibit significant nonlinearity and large time lag. Meanwhile, traditional ammonia injection control technology suffers from numerous bottlenecks: relying solely on SCR inlet and outlet NO... x Concentration feedback is not correlated with combustion system and operating parameters; transmission lag is treated as a fixed value, failing to dynamically adapt to changes in flue gas velocity and combustion state; the predictive model lacks multi-field feature fusion capabilities and has poor generalization; and the constraints are rigid, unable to adapt to the optimization requirements of all operating conditions. These problems directly lead to traditional control methods exhibiting issues such as "excessive ammonia injection → ammonia slip exceeding the standard (>5ppm), catalyst poisoning" or "insufficient ammonia injection → NO..." x The vicious cycle of "excessive emissions" will lead to waste of urea / ammonia water, blockage of the cold end of the air preheater due to ammonium bisulfate crystallization, and increased operating costs.
[0004] Therefore, there is an urgent need to develop a method that integrates multi-field coupling characteristics, dynamic quantization lag, and accurate NO prediction. x The ammonia injection control technology with concentration and adaptive adjustment constraints overcomes the pain points of traditional control, such as "large lag, low precision, and poor adaptability," and achieves synergistic optimization of denitrification efficiency, operational economy, and environmental compliance. Summary of the Invention
[0005] In view of the above-mentioned problems, the present invention is proposed.
[0006] Therefore, this invention aims to solve the problems of NO caused by the strong lag, low prediction accuracy, and poor adaptability of traditional ammonia injection control. x The problem is that emissions exceed standards or ammonia escape is too high.
[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired power unit SCR system, comprising, Multi-field coupling data is collected and corresponding features are extracted to construct a preprocessed feature set. Combustion-flue gas transport hysteresis is dynamically quantified based on the feature set to obtain hysteresis parameters. A custom prediction model is used to fuse the feature set and hysteresis parameters for multi-field coupling. Accurate prediction; calculation of ammonia injection rate, combined with PID feedback correction for feedforward-feedback fusion ammonia injection control to obtain the total ammonia injection rate; dynamic adjustment of PID parameters based on load change rate, based on outlet... The ammonia injection rate is adjusted based on the ammonia escape concentration under adaptive operating conditions; the ammonia injection rate is converted into a current signal to control the ammonia injection valve, and the valve feedback deviation is checked in real time. When the deviation exceeds the limit, the system switches to manual mode.
[0008] As a preferred embodiment of the predictive adaptive ammonia injection control method for multi-field coupling of SCR system of coal-fired unit described in this invention, the construction of the preprocessed feature set includes collecting combustion field, flue gas field, operating condition and coal quality parameter signals. The DCS is used to perform preprocessing in the order of elimination, discretization, and filtering, and to perform signal redundancy switching. When the main signal triggers the bad point mark, it switches to the backup measurement point or the historical average value. After preprocessing, the features are constructed.
[0009] As a preferred embodiment of the predictive adaptive ammonia injection control method for multi-field coupling of a coal-fired unit SCR system described in this invention, the method for obtaining lag parameters includes calculating combustion lag based on correction coefficients and air-coal ratio correction coefficients. The flue gas flow hysteresis is calculated based on the flue volume and flue gas velocity, and then configured and implemented.
[0010] As a preferred embodiment of the predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired power unit SCR system described in this invention, wherein: the multi-field coupling Accurate prediction includes establishing relationships between various signals and... The correlation model of the generated quantity is used to perform multi-field coupling feature fusion, and the feature set and hysteresis parameters are fused to obtain the multi-field coupling feature fusion value; The prediction model employs a mechanistic model and data-driven correction to output the SCR input. Predicted value.
[0011] As a preferred embodiment of the predictive adaptive ammonia injection control method for multi-field coupling of SCR system of coal-fired unit described in this invention, the feedforward-feedback fusion ammonia injection control includes: establishing a fusion architecture of feedforward prediction and feedback correction; calculating the ammonia injection quantity through DCS calculation block and PID adjustment block, including feedforward ammonia injection quantity and reference ammonia injection quantity; and obtaining the total ammonia injection quantity by combining PID feedback correction quantity. The feedforward ammonia injection amount ,in, It is a stoichiometric ratio. This represents the fusion value of multi-field coupling features; Reference ammonia injection volume ,in, The catalyst activity coefficient is... For SCR entry Predicted value; Feedback correction of ammonia injection quantity For process values, Output correction value as setpoint initial parameters =1.0, Ti=70 seconds, Td=18 seconds, where, Actual test of SCR entry point ; Total ammonia injection ; Set constraints and limitations: Exporting SCR Limit values should be considered, with redundancy taken into account. .
[0012] The beneficial effects of the preferred technical solution in the embodiments of the present invention are as follows: by integrating feedforward and feedback control, the accurate calculation and dynamic compensation of ammonia injection amount are realized, thereby improving the overall accuracy and response speed of denitrification control.
[0013] As a preferred embodiment of the predictive adaptive ammonia injection control method for multi-field coupling of a coal-fired unit SCR system described in this invention, the method of adjusting the ammonia injection quantity according to the operating condition adaptive constraint includes dynamically adjusting the control parameters and constraint thresholds, and achieving full-condition optimization through DCS selection block and limiting block. Adaptive switching of operating conditions is implemented, including steady-state, dynamic, and low-load operating conditions. The steady-state operating condition is primarily driven by feedforward and prediction; the dynamic operating condition feedback accounts for 40% and is amplified. Shorten Ti and increase load change rate compensation; the low load condition ,reduce ; Adaptive operating condition constraints include PID parameter adaptation, Constraint adaptive, ammonia escape constraint adaptive, valve position constraint adaptive; The PID parameters are adaptively used to calculate the load change rate. ,when When the speed is greater than 5 MW / min, switch the PID parameters. When the flow rate is ≤5MW / min, restore the PID parameters; The Constraint Adaptation for Monitoring SCR Outlet Concentration, when Ammonia levels increase over time. < Reduce ammonia levels; The ammonia escape constraint is adaptively configured to calculate the ammonia reduction amount. ×0.02×( ) / 0.5, when When the concentration is >3ppm, the ammonia reduction amount is added to the SEL block; The valve position constraint adaptively limits the ammonia injection valve position to 5% to 90%, and adjusts the lower limit of the valve position through the SEL block when under low load.
[0014] The beneficial effects of the preferred technical solution in the embodiments of the present invention are as follows: the control logic and constraint thresholds are adaptively adjusted based on operating parameters such as load, which ensures the stability of the system and compliance with environmental protection standards under different operating conditions.
[0015] As a preferred embodiment of the predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired power unit SCR system described in this invention, wherein: the real-time verification of valve feedback deviation includes valve feedback deviation... The calculation is as follows: in, This is the moving average value of the valve output command. This is the moving average value of the valve feedback signal. To obtain the absolute value, the deviation must always be positive; An alarm is triggered and manual control is switched when the deviation exceeds 3%.
[0016] The beneficial effects of the preferred technical solution in the embodiments of the present invention are: real-time verification of the deviation between valve commands and feedback and switching to manual mode when the limit is exceeded, which enhances the reliability of the execution link and avoids control failure caused by valve failure.
[0017] Another objective of this invention is to provide a predictive adaptive ammonia injection control system with multi-field coupling for a coal-fired power unit's SCR system.
[0018] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a predictive adaptive ammonia injection control system for multi-field coupling of a coal-fired unit SCR system, comprising: a signal acquisition unit including an AI module to acquire temperature, concentration, and flow rate; a DI module to acquire load plans; and a MAN manual setting block to set coal nitrogen content and catalyst activity coefficient; wherein, the AI module adopts a 3-point average + 3σ bad point elimination preprocessing, and in the event of a fault, switches to a backup measuring point or historical average value within 2 seconds through the FAIL_SEL block; The dynamic hysteresis quantization unit includes a CALC operation block, a MUL_DIV multiplication and division block, and an ADD_SUB addition and subtraction block, which are used to calculate the hysteresis parameters. The prediction unit includes a UDFB custom block for deploying the LSTM residual model and a MEM cache block for time synchronization; the model file of the UDFB custom block is NOX_PRED_LSTM.mdl, the trigger period is 10s, and the output residual compensation value ε(t) ranges from -3 to 3mg / m³. The fusion control unit includes a PID_REG adjustment block, a MUL_DIV feedforward calculation block, and an ADD_SUB total ammonia injection quantity summation block; The adaptive constraint unit includes a RATE_CALC rate of change calculation block, a SEL selection block, and a LIMIT limit block; The execution and feedback unit includes an AO output module, a valve feedback AI acquisition block, an ALM alarm block, and a MAN manual control block; among them, the AO output module and the MAN manual control block are interlocked, and the AO output tracks the manual value when in manual mode; All units are deployed on the DCS platform of the coal-fired unit, supporting redundant configuration and fault switching.
[0019] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the aforementioned predictive adaptive ammonia injection control method for multi-field coupling of a coal-fired unit SCR system.
[0020] The present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the aforementioned predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired unit SCR system.
[0021] The beneficial effects of this invention are as follows: This invention achieves high prediction accuracy through multi-field coupled modeling, integrates multi-dimensional features of combustion, flue gas, operating conditions, and coal quality, and combines dynamic hysteresis quantization and residual compensation to reduce NO at the SCR inlet. x The prediction deviation is ≤5mg / m³ in steady state and ≤10mg / m³ in dynamic state, which is significantly better than the traditional model (deviation >15mg / m³). This invention features dynamic hysteresis adaptation, a fast response speed, and real-time calculation of both combustion and flue gas flow hysteresis, enabling precise matching of the prediction time window with actual operating conditions. This reduces NO at the SCR outlet after a load step. x Concentration recovery time ≤60s, with no overshoot; This invention, through an adaptive constraint mechanism, exhibits strong adaptability to operating conditions and dynamically adjusts PID parameters and NO. x Threshold and ammonia slip limit, adaptable to 20% to 100% rated load fluctuations and coal quality changes, low load denitrification efficiency ≥90%, high load ammonia slip ≤2.5ppm; This invention constructs a fusion control architecture with good robustness. Feedforward prediction reduces the impact of lag, feedback correction compensates for model errors, and with the help of fault switching and bad spot elimination mechanisms, the system's fault-free operation time is >8000h. This invention has strong engineering applicability, is developed entirely based on the DCS system of coal-fired power units, and its functional block naming, interface definition, and parameter configuration are closely aligned with actual field operations. It can be directly configured in the DCS system without the need for a large amount of new hardware, resulting in low modification costs. Attached Figure Description
[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0023] Figure 1 The above is a flowchart of a predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired power unit SCR system, provided as an embodiment of the present invention. Detailed Implementation
[0024] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.
[0025] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides a predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired unit SCR system, comprising: S100: Collect multi-field coupled data and extract corresponding features to construct a preprocessed feature set; S200, Based on feature set dynamic quantization of combustion-flue gas delivery hysteresis, obtain hysteresis parameters; S300: Multi-field coupling is achieved by fusing feature sets and lag parameters through a custom prediction model. Accurate prediction; S400: Calculate the ammonia injection quantity, and combine it with the PID feedback correction quantity to perform feedforward-feedback fusion ammonia injection control to obtain the total ammonia injection quantity; S500 dynamically adjusts PID parameters based on load change rate, based on outlet... The ammonia injection rate is adjusted based on the ammonia slip concentration under adaptive operating conditions. S600 converts the ammonia injection quantity into a current signal to control the ammonia injection valve, verifies the valve feedback deviation in real time, and switches to manual mode when the limit is exceeded. It should be noted that existing SCR ammonia injection control methods typically rely on a single feedback signal or simple feedforward, failing to effectively integrate coupled data from multiple fields such as the combustion field and flue gas field, and neglecting the hysteresis effect of combustion and flue gas transmission, leading to... Insufficient prediction accuracy, delayed ammonia injection control response, poor adaptability to dynamic operating conditions such as load changes, and prone to problems. Excessive emissions or high ammonia escape, coupled with a lack of real-time deviation verification in valve control, result in low reliability.
[0026] Therefore, to address the aforementioned problems, steps S100-S600 were used to achieve a solution based on multi-field coupling characteristics and dynamic hysteresis compensation. Accurate prediction, combined with feedforward-feedback fusion control and adaptive constraints, significantly improves the accuracy, response speed, and overall stability of ammonia injection control. Simultaneously, real-time valve deviation verification ensures the reliability of the execution process, thereby effectively reducing... Emissions are controlled and ammonia escape is prevented, thus meeting environmental protection requirements and improving system operational safety.
[0027] Example 2, refer to Figure 1 This is one embodiment of the present invention, which provides a predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired unit SCR system, comprising: In this embodiment of the invention, S100 involves collecting multi-field coupled data and extracting corresponding features to construct a preprocessed feature set, including the following steps S101-S103: S101, Collect combustion field, flue gas field, operating conditions, and coal quality parameter signals; Collect the temperature of the main combustion zone and the excess air coefficient of the combustion field; Collect flue gas velocity and ammonia slip from the flue gas field; Collect data on load and air volume under operating conditions; Collect coal quality parameter signals, such as coal nitrogen content and volatile matter.
[0028] In addition, analog signals, including but not limited to temperature, pressure, concentration, flow rate, and wind speed, are collected through the AI_IN block, digital load planning signals are collected through the DI_IN block, and coal quality parameters and catalyst activity coefficients are set through the MAN block.
[0029] In the embodiment of the present invention, S102, the preprocessing is performed by using the built-in AI and DI function blocks of DCS, employing a 3-point average + 3σ bad point elimination, 10-minute discretization of the DI_IN block, and first-order filtering with a filtering time of 5~10s, to ensure signal reliability. In an optional implementation, the filtering in S102 can be a moving average filter, but this implementation has a lag in response to signal abrupt changes, which weakens the tracking capability under rapidly changing operating conditions.
[0030] In another alternative implementation, the filtering in S102 can also be median filtering, but this implementation is less effective at suppressing low-frequency noise in stable signals and retains some small fluctuations.
[0031] In the embodiment of the present invention, S103, signal redundancy switching is realized through the built-in FAIL_SEL function block of DCS. When the main signal triggers the bad point mark, it switches to the backup measurement point or historical average value within 2 seconds to ensure the continuity of the feature set. In an optional implementation, the redundancy switching in S103 can be based on alarm delay and manual confirmation. When the main signal is marked as abnormal by the bad point elimination algorithm, the system will not switch immediately. The DCS operation interface triggers an audible and visual alarm and keeps the signal value at the valid value at the last moment before the fault, or keeps its output changing linearly. The alarm information prompts the operator to confirm. The operator needs to check the relevant process parameters on the monitoring screen to determine whether it is a real fault or a temporary disturbance. If the operator confirms the fault within the preset delay window, he can manually click the switching command to switch the control loop to the pre-configured backup measurement point. If no confirmation is made within the time limit, in order to prevent the long-term use of unreliable data, the system automatically switches to a conservative and safe fixed value instead of the optimal backup value. However, this implementation has a serious response lag and relies on manual efficiency. When the operating conditions change rapidly, it cannot guarantee the timeliness and continuity of control and is not suitable for this invention.
[0032] In this embodiment of the invention, step S200, which involves dynamically quantifying the combustion-flue gas transport hysteresis based on a feature set to obtain hysteresis parameters, includes the following steps S201-S202: Based on the principles of fluid mechanics and combustion dynamics, two types of lag times are calculated in real time through the DCS computing block, and the prediction time window is dynamically matched.
[0033] S201. Calculate combustion lag based on correction coefficients and air-coal ratio correction coefficients; Specifically, combustion lag is quantified by calculating the residence time of pulverized coal. : Based on the designed residence time in the furnace As the baseline lag value, a primary wind correction factor is considered. Correction coefficient for wind coal ratio The formula is: Where V is the furnace volume, To design the excess air coefficient, To design the flue gas density, To design fuel quantity, The lower calorific value of coal, The specific heat capacity of flue gas at constant pressure. Design temperature for the main combustion zone; in, For primary wind speed, =25m / s; This is the design baseline value for air volume. =1.5×10 5 m³ / h; For primary wind speed, This is the actual air volume value. The actual excess air coefficient is calculated from the total amount of primary and secondary air plus the amount of coal. The value is 1.2.
[0034] S202. Calculate flue gas flow hysteresis based on flue volume and flue gas velocity; Specifically, the hysteresis of flue gas flow is quantified by calculating the hysteresis of SCR inlet detection. : Based on the flue volume and real-time flue gas velocity, the formula is: in, The total volume of the flue from the furnace outlet to the SCR inlet is determined by the boiler structure and is taken as the design value of 800 m³. The lower calorific value of coal, The value here is 23 MJ / kg, and B is the real-time coal feed rate. , These are the actual values of primary and secondary air volumes; In embodiments of the present invention, multiplication and division operations are performed by the MUL_DIV block, addition and subtraction operations are performed by the ADD_SUB block, and the actual excess air coefficient α and flue gas velocity are calculated by the CALC block. The calculation results are cached in 10-second cycles in the MEM block.
[0035] In an optional implementation, the configuration in S202 can be based on a simplified calculation configuration of the flue gas flow meter signal. The AI_IN function block is used to collect the flue gas flow meter signal installed at the economizer outlet or SCR inlet flue. The signal is then subjected to first-order filtering and bad pixel removal. The DIV function block is used to perform a division operation with the pre-set fixed parameter "total volume of flue gas after furnace V flue" in the MEM function block as the divisor and the processed real-time flue gas velocity signal as the divisor. The result of the division operation is the real-time estimated flue gas flow lag time. The MEM function block is used to buffer the result with a period of 10 seconds and output it to the downstream control module. However, this implementation is highly dependent on the accuracy and reliability of the flue gas flow meter. When the flow meter fails or the measurement is distorted, it will directly lead to errors in the lag time calculation. In addition, the installation and maintenance of a dedicated flow meter increases the cost.
[0036] In an embodiment of the present invention, in S300, a custom prediction model is used to fuse feature sets and hysteresis parameters to perform multi-field coupling. Accurate prediction includes the following steps S301-S302: S301, Establish the relationship between each signal and... through a custom prediction model The correlation model of the generated quantity is used to perform multi-field coupling feature fusion, and the feature set and hysteresis parameters are fused to obtain the multi-field coupling feature fusion value; Construct a multi-field coupled prediction model and deploy it through a DCS custom function block (UDFB) to realize the SCR entry point. Accurate concentration prediction: Multi-field coupling feature fusion modeling: Multi-field coupling feature fusion value (Weighted summation of load, coal quality, and temperature characteristics) to select the daily load plan for dispatching. ), coal quality (Nar, Var, Aar), coal quantity (B), main combustion zone temperature ( ), furnace outlet temperature ( As a feedforward signal, establish the relationship between each signal and... The correlation model of the generated quantity: Load planning association: Where a and b are the fitting coefficients for historical data; Coal quality and quantity correlation: Where c, d, e, and f are the correlation coefficients for fitting historical data of coal types; Temperature correlation: Where g, h, i, and j are the fitting coefficients for historical temperature data; Multi-field coupling feature fusion value: in, , , The different weight coefficients assigned are summed to 1 and calibrated by multiple linear regression.
[0037] S302. Employing a mechanistic model and a data-driven corrected prediction model, the SCR input is output. Predicted value; SCR Entry NO x The prediction model employs a "mechanism model + data-driven correction" approach to predict... The entrance after The concentration prediction model calculates it using the following formula: Using the DCS built-in UDFB function block, compile the UDFB block into a DCS-compatible file (NOX_PRED.mdl) and output the residual compensation values. , within the range of ±3mg / m³.
[0038] The calculation formula of the prediction model is executed through the CALC block within the DCS, where, For combustion correction factor, For flue gas delay correction factor: in, To provide real-time furnace negative pressure, Designed for negative pressure; in, For real-time flue gas flow rate, To design flue gas flow rate; For residual compensation in LSTM networks, input for nearly 10 minutes , , Data, output model bias.
[0039] Time synchronization: via MEM block by pressing " "Time offset output prediction value to ensure that the prediction time window matches the actual time when flue gas arrives at the SCR inlet. When the prediction deviation is ≤5mg / m³, it is steady state, i.e., load fluctuation ±3%; when the prediction deviation is ≤10mg / m³, it is dynamic, i.e., load step ±10% or coal type switching."
[0040] In an embodiment of the present invention, the ammonia injection quantity is calculated in step S400, and feedforward-feedback fusion ammonia injection control is performed in combination with the PID feedback correction quantity to obtain the total ammonia injection quantity, including the following steps S401-S402: S401. Establish a fusion architecture of feedforward prediction and feedback correction. The ammonia injection quantity is calculated through DCS calculation block and PID control block, including feedforward ammonia injection quantity and reference ammonia injection quantity. Combined with PID feedback correction quantity, the total ammonia injection quantity is obtained.
[0041] S402, the feedforward ammonia injection quantity ,in, It is a stoichiometric ratio. This is the fusion value of multi-field coupling characteristics; calculated through the MUL_DIV block in DCS.
[0042] Reference ammonia injection volume ,in, The catalyst activity coefficient is... For SCR entry Predicted value; calculated via the MUL_DIV block.
[0043] Feedback correction of ammonia injection quantity Process value (PV) Output correction value (SV) initial parameters =1.0, Ti=70 seconds, Td=18 seconds, where, Actual test of SCR entry point ; Total ammonia injection The summation is performed through the ADD_SUB block, and the MEM block is filtered and smoothed with a 5-second period.
[0044] Set constraints and limitations: Exporting SCR The limit is 50, based on ultra-low emission requirements. Therefore, it is necessary to consider reserving redundancy. .
[0045] In an embodiment of the present invention, the PID parameters are dynamically adjusted in S500 according to the load change rate, based on the outlet... Adjusting the ammonia injection rate under operating condition adaptive constraints based on ammonia slip concentration includes the following steps S501-S502: S501: Dynamically adjust control parameters and constraint thresholds, and achieve full-condition optimization through DCS selection block and limiting block; Adaptive switching of operating conditions is performed, including steady-state operating conditions, dynamic operating conditions, and low-load operating conditions; The steady-state operating condition refers to a load fluctuation within ±3%, dominated by feedforward and prediction, with feedback accounting for 20%. The dynamic operating conditions include load step change ±10% and coal type switching; feedback accounts for 40%, amplification. Shorten Ti and increase load change rate compensation; The low-load condition is defined as P < 30% of the rated load. ,reduce .
[0046] S502, adaptive operating condition constraints include PID parameter adaptation, Constraint adaptive, ammonia escape constraint adaptive, valve position constraint adaptive; The PID parameters are adaptively used to calculate the load change rate using the RATE_CALC block. (Load fluctuation, the change in unit load per unit time), when When the speed is greater than 5 MW / min, switch the PID parameters as follows: =1.2, Ti=56 seconds, improving dynamic response; when When the flow rate is ≤5MW / min, restore the PID parameters; The Constraint Adaptation for Monitoring SCR Outlet Concentration, when Ammonia increase of 10%–15% / min, < Reduce ammonia concentration by 5% / min. The ammonia escape constraint is adaptively configured to calculate the ammonia reduction amount. ×0.02×( ) / 0.5, when When the ammonia slip is greater than 3ppm (according to the standard and operating procedures, the ammonia slip during the operation of thermal power units must be less than 3ppm), the SEL block will be used to reduce the ammonia amount. The valve position constraint adaptively limits the ammonia injection valve position to 5% to 90%, and adjusts the lower limit of the valve position to 8% through the SEL block when under low load (<30% of rated load).
[0047] In an embodiment of the present invention, in S600, the ammonia injection quantity is converted into a current signal to control the ammonia injection valve, the valve feedback deviation is checked in real time, and the switch is made to manual when the limit is exceeded. In an embodiment of the present invention, valve feedback deviation The calculation is as follows: in, This is the moving average value of the valve output command. This is the moving average value of the valve feedback signal. To obtain the absolute value, the deviation must always be positive; An alarm is triggered and manual control is switched when the deviation exceeds 3%.
[0048] In an optional implementation, the valve feedback deviation in S600 can be directly read from the current instantaneous value of the ammonia injection valve output command and the current instantaneous value of the valve feedback signal. Using the DCS's CALC calculation function block, the absolute difference between the two instantaneous values is calculated. The absolute difference is divided by the instantaneous value of the valve output command and then multiplied by 100% to obtain the instantaneous relative deviation percentage. The calculated instantaneous relative deviation percentage is compared with a preset threshold. When the deviation continues to exceed the threshold for a preset time, the DCS triggers an alarm and switches the ammonia injection control loop to manual mode. This scheme is simple and direct to calculate, but it is easily affected by instantaneous signal fluctuations or noise interference, resulting in poor stability and unnecessary false alarms and frequent switching to manual mode.
[0049] In another optional implementation, the valve feedback deviation in S600 can also be verified by combining the incremental change rate of the deviation dead zone. Using the DCS's ADD_SUB function block, the real-time difference between the ammonia injection valve's output command value and the valve feedback value is calculated. A small dead zone is set for this difference. When the difference is within the dead zone, the valve is considered to be operating normally, and no further verification is performed. When the difference exceeds the dead zone, the DCS's RATE_CALC or similar function block is used to calculate the rate of change of this difference, i.e., the rate at which the deviation increases per unit time. The calculated rate of change is compared with a preset safe rate of change threshold. If the rate of change exceeds the threshold, an alarm is immediately triggered, and manual control is switched. This scheme can more sensitively capture abnormal valve development trends, but the dead zone setting and rate of change threshold need fine-tuning; otherwise, small, continuous deviations may be missed, or normal rapid adjustment processes may be falsely reported.
[0050] Example 3 is an embodiment of the present invention, which provides a predictive adaptive ammonia injection control method for multi-field coupling of SCR system of coal-fired unit. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.
[0051] 1. Configure the hardware as shown in Table 1: Table 1 Hardware Configuration Table
[0052] 2. Software module configuration implementation based on DCS: (1) Signal acquisition unit (AI_IN / DI_IN / MAN / FAIL_SEL block); (2) Dynamic hysteresis quantization unit (CALC / MUL_DIV / ADD_SUB / MEM block); (3) Prediction module (CALC / UDFB / MEM block); (4) Fusion control unit (MUL_DIV / PID_REG / ADD_SUB / MEM block); (5) Adaptive constraint unit (RATE_CALC / SEL / LIMIT / CALC block); (6) Execution and feedback unit (AO_OUT / CALC / ALM block); 3. Debugging and Verification Steps (1) Static debugging, no-load operation stage: Signal verification: Input a standard signal to the AI_IN block (e.g., 25m / s corresponds to 12mA for AI_IN_01_V1), and verify that the deviation of the output variable (AI_V1_MPS=25) from the theoretical value is ≤0.5%; Lag calculation verification: fixed primary air parameters ( =25m / s, =1.5×10 5 (m³ / h), confirm CAL_K1=1, CAL_TAU1_S=2.5s, deviation ≤0.1s; Static validation of the prediction model: fixed input =300mg / m³, TZR=1400K, Q smoke=1.2×10 6 m³ / h, verifying CAL_NOX_IN_PRED_MGM3=300±2mg / m³; Ammonia injection rate calculation verification: setting =300mg / m³ =0.95, confirming CAL_AM_BASE_KGH=300×0.567×0.95≈160.5kg / h, with a deviation ≤1%.
[0053] (2) Dynamic commissioning (low load 50% of rated operating condition) Load step simulation: Input load from 240MW to 300MW (step rate 5MW / min) through DI_IN_01_P_PLAN and record: lag time τ1 from 2.5s to 2.8s, τ2 from 0.67s to 0.58s, dynamic response ≤2s; The concentration of the drug increased from 300 mg / m³ to 360 mg / m³, with a deviation from the measured value of ≤8 mg / m³.
[0054] Adaptive PID parameter verification: When the load change rate dP / dt = 5MW / min, confirm that the PID parameters are switched to Kp = 1.2 and T. i =56s, no overshoot.
[0055] Constraint logic verification: Force AI_NOX_OUT_MGM3=50mg / m³, confirm that SEL_02_NOX_LIMIT outputs ammonia increase of 15%, and the response time is ≤1s.
[0056] (3) Joint commissioning verification (100% rated load) Full-condition control performance: Continuous operation for 72 hours, key indicators recorded: SCR Export NO x Concentration: 40~45 mg / m³, compliance rate 100%; Ammonia slip concentration: 1.5~2.5 ppm, not exceeding the standard; Ammonia injection valve position: 20%~70%, no jamming; Fault redundancy verification: Disconnect the AI_IN_07_NOX_IN main signal and confirm that FAIL_SEL_01_AI switches to the backup measurement point within 2 seconds without control disturbance; Economic verification: Compared with traditional control, urea / ammonia consumption is reduced by 8% to 12%, and catalyst life is expected to be extended by 1 to 2 years.
[0057] Example 4 is an embodiment of the present invention. The above is a schematic scheme of a predictive adaptive ammonia injection control method with multi-field coupling in a coal-fired power unit SCR system. It should be noted that the technical solution of a predictive adaptive ammonia injection control system with multi-field coupling in a coal-fired power unit SCR system belongs to the same concept as the above-described predictive adaptive ammonia injection control method with multi-field coupling in a coal-fired power unit SCR system. Details not described in detail in this embodiment of the predictive adaptive ammonia injection control system with multi-field coupling in a coal-fired power unit SCR system can be found in the description of the above-described predictive adaptive ammonia injection control method with multi-field coupling in a coal-fired power unit SCR system.
[0058] This embodiment provides a predictive adaptive ammonia injection control system for a coal-fired unit SCR system with multi-field coupling, including: a signal acquisition unit comprising an AI module to acquire temperature, concentration, and flow rate; a DI module to acquire load plans; and a MAN manual setting block to set coal nitrogen content and catalyst activity coefficient; wherein, the AI module adopts a 3-point average + 3σ bad point elimination preprocessing, and in the event of a fault, switches to a backup measuring point or historical average value within 2 seconds through the FAIL_SEL block; The dynamic hysteresis quantization unit includes a CALC operation block, a MUL_DIV multiplication and division block, and an ADD_SUB addition and subtraction block, which are used to calculate the hysteresis parameters. The prediction unit includes a UDFB custom block for deploying the LSTM residual model and a MEM cache block for time synchronization; the model file of the UDFB custom block is NOX_PRED_LSTM.mdl, the trigger period is 10s, and the output residual compensation value ε(t) ranges from -3 to 3mg / m³. The fusion control unit includes a PID_REG adjustment block, a MUL_DIV feedforward calculation block, and an ADD_SUB total ammonia injection quantity summation block; The adaptive constraint unit includes a RATE_CALC rate of change calculation block, a SEL selection block, and a LIMIT limit block; The execution and feedback unit includes an AO output module, a valve feedback AI acquisition block, an ALM alarm block, and a MAN manual control block; among them, the AO output module and the MAN manual control block are interlocked, and the AO output tracks the manual value when in manual mode; All units are deployed on the DCS platform of the coal-fired unit, supporting redundant configuration and fault switching.
[0059] This embodiment also provides an electronic device applicable to a predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired power unit SCR system, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired power unit SCR system as proposed in the above embodiment.
[0060] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired unit SCR system as proposed in the above embodiment.
[0061] The storage medium proposed in this embodiment belongs to the same inventive concept as the predictive adaptive ammonia injection control method for multi-field coupling of a coal-fired unit SCR system proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.
[0062] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.
[0063] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. A method for predictive ammonia injection control of a coal-fired unit SCR system with multi-field coupling, characterized in that: include, Collect multi-field coupled data and extract corresponding features to construct a preprocessed feature set; Dynamic quantization of combustion-flue gas transport hysteresis based on feature set to obtain hysteresis parameters; Fusing feature sets and lag parameters through a custom predictive model for multi-field coupling Accurate prediction; Calculate the ammonia injection rate, and combine it with the PID feedback correction to perform feedforward-feedback fusion ammonia injection control to obtain the total ammonia injection rate; Adjusting PID parameters dynamically according to load change rate, adjusting ammonia injection amount based on outlet Conducting self-adaptive constraint adjustment of ammonia injection amount according to ammonia escape concentration The ammonia injection volume is converted into a current signal to control the ammonia injection valve, and the valve feedback deviation is checked in real time. If the deviation exceeds the limit, the system switches to manual mode.
2. The predictive ammonia injection control method for a coal-fired unit SCR system with multi-field coupling according to claim 1, characterized in that: The construction of the preprocessed feature set includes collecting combustion field, flue gas field, operating conditions, and coal quality parameter signals; The DCS is used to perform preprocessing in the order of elimination, discretization, and filtering, and to perform signal redundancy switching. When the main signal triggers the bad point mark, it switches to the backup measurement point or the historical average value. After preprocessing, the features are constructed.
3. The predictive ammonia injection control method for a coal-fired unit SCR system with multi-field coupling according to claim 2, characterized in that: The acquisition of lag parameters includes calculating combustion lag parameters based on correction coefficients and air-coal ratio correction coefficients. The flue gas flow hysteresis is calculated based on the flue volume and flue gas velocity, and then configured and implemented.
4. The predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired unit SCR system as described in claim 3, characterized in that: The multi-field coupling The precise prediction comprises establishing the correlation between each signal and the generation amount by a self-defined prediction model A correlation model of the generation amount is generated, multi-field coupling feature fusion is performed, and a multi-field coupling feature fusion value is obtained by fusing the feature set and the lag parameter. The mechanism model and the data-driven correction prediction model are used to output the SCR inlet predicted value.
5. The predictive ammonia injection control method for a coal-fired unit SCR system with multi-field coupling of claim 4, wherein: The aforementioned feedforward-feedback fusion ammonia injection control includes establishing a fusion architecture of feedforward prediction and feedback correction, and calculating the ammonia injection amount through DCS calculation block and PID adjustment block, including feedforward ammonia injection amount and reference ammonia injection amount, and obtaining the total ammonia injection amount by combining PID feedback correction amount. The feedforward ammonia injection amount wherein, is the stoichiometric ratio, is a multi-field coupling characteristic fusion value; Reference ammonia injection wherein, is the catalyst activity coefficient, SCR inlet predicted value; Feedback correction of the amount of ammonia injected to the process value, the set value, the output correction with initial parameters = 1.0, Ti = 70 seconds, Td = 18 seconds, wherein is the measured SCR inlet value. Total amount of ammonia injected ; Set constraint limits: For SCR export Limit value, considering reserved redundancy taken .
6. The predictive self-adaptive ammonia injection control method of a coal-fired unit SCR system multi-field coupling according to claim 5, characterized in that: The adaptive constraint adjustment of ammonia injection quantity under working conditions includes dynamically adjusting control parameters and constraint thresholds, and achieving full-condition optimization through DCS selection blocks and limiting blocks. Adaptive switching of working conditions is carried out, and the working conditions include a steady state working condition, a dynamic working condition and a low load working condition, wherein the steady state working condition is dominated by feedforward + prediction; the dynamic working condition has a feedback proportion of 40%, and is amplified and Ti is shortened, and the load change rate compensation is increased; the low load working condition is reduced . The working condition self-adapting constraint includes PID parameter self-adapting, constraint self-adapting, ammonia escape constraint self-adapting, valve position constraint self-adapting; PID parameter self-adaptation is a calculation load change rate When > 5 MW / min, switch PID parameters, when ≤ 5 MW / min, restore PID parameters; The Constraint Adaptation for Monitoring SCR Outlet Concentration, when Ammonia levels increase over time. < Reduce ammonia levels; The ammonia escape constraint is adaptively configured to calculate the ammonia reduction amount. ×0.02×( ) / 0.5, when When the concentration is >3ppm, the ammonia reduction amount is added to the SEL block; The valve position constraint adaptively limits the ammonia injection valve position to 5% to 90%, and adjusts the lower limit of the valve position through the SEL block when under low load.
7. The predictive adaptive ammonia injection control method for multi-field coupling in a coal-fired unit SCR system as described in claim 6, characterized in that: The real-time verification valve feedback deviation includes valve feedback deviation. The calculation is as follows: in, This is the moving average value of the valve output command. This is the moving average value of the valve feedback signal. To obtain the absolute value, the deviation must always be positive; An alarm is triggered and manual control is switched when the deviation exceeds 3%.
8. A predictive adaptive ammonia injection control system for a multi-field coupled SCR system of a coal-fired power unit, employing the predictive adaptive ammonia injection control method for a multi-field coupled SCR system of a coal-fired power unit as described in any one of claims 1 to 7, characterized in that, include: The signal acquisition unit includes an AI module to acquire temperature, concentration, and flow rate; a DI module to acquire load plans; and a MAN manual setting block to set coal nitrogen content and catalyst activity coefficient. The AI module uses a 3-point average + 3σ bad point elimination preprocessing. In case of failure, it switches to a backup measurement point or historical average value within 2 seconds through the FAIL_SEL block. The dynamic hysteresis quantization unit includes a CALC operation block, a MUL_DIV multiplication and division block, and an ADD_SUB addition and subtraction block, which are used to calculate the hysteresis parameters. The prediction unit includes a UDFB custom block for deploying the LSTM residual model and a MEM cache block for time synchronization; the model file of the UDFB custom block is NOX_PRED_LSTM.mdl, the trigger period is 10s, and the output residual compensation value ε(t) ranges from -3 to 3mg / m³. The fusion control unit includes a PID_REG adjustment block, a MUL_DIV feedforward calculation block, and an ADD_SUB total ammonia injection quantity summation block; The adaptive constraint unit includes a RATE_CALC rate of change calculation block, a SEL selection block, and a LIMIT limit block; The execution and feedback unit includes an AO output module, a valve feedback AI acquisition block, an ALM alarm block, and a MAN manual control block; among them, the AO output module and the MAN manual control block are interlocked, and the AO output tracks the manual value when in manual mode; All units are deployed on the DCS platform of the coal-fired unit, supporting redundant configuration and fault switching.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the predictive adaptive ammonia injection control method for multi-field coupling of a coal-fired unit SCR system according to any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the predictive adaptive ammonia injection control method for multi-field coupling of a coal-fired unit SCR system according to any one of claims 1 to 7.