A method and system for controlling SNCR denitrification in a municipal solid waste incinerator
By collecting and preprocessing data from municipal solid waste incinerators, dividing injection sub-regions and performing collaborative optimization, the problem of injection volume deviation in existing technologies has been solved, achieving more efficient NOx control and equipment safety.
Patent Information
- Application Number
- CN202611194342.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-07
- Publication Date
- 2026-11-17
AI Technical Summary
The existing SNCR denitrification control system for municipal solid waste incinerators is unable to accurately reflect the true reaction status of each injection zone under complex operating conditions, resulting in deviations in injection volume and affecting control performance and equipment safety.
By uniformly collecting and preprocessing data on furnace temperature, flue gas emissions, and boiler operating conditions, injection sub-regions are divided and mapping relationships are established to form sub-region state vectors. This allows for zoned collaborative optimization of injection quantity control. Combined with safety constraints and branch fluid coupling compensation, the final injection setpoint is generated.
This enables precise control of each injection sub-region, improves NOx emission stability, reduces ammonia escape risk, and enhances the safety and reliability of system operation.
Smart Images

Figure CN123102134A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of waste treatment, and specifically to a method and system for controlling SNCR denitrification in a municipal solid waste incinerator. Background Technology
[0002] With the increasing volume of municipal solid waste and limited landfill space, waste-to-energy incineration has become an important method for the final treatment of municipal solid waste. The incineration process produces flue gas containing nitrogen oxides, acidic gases, particulate matter, and dioxins. Therefore, incineration facilities typically require in-furnace combustion control and external flue gas purification systems, and continuous flue gas monitoring devices are used to monitor the emissions of major pollutants.
[0003] In terms of nitrogen oxide control, municipal solid waste incinerators often employ selective non-catalytic reduction (SNCR) denitrification technology. This process typically uses urea solution or ammonia water as a reducing agent, injected into a suitable temperature zone within the furnace. The ammonia-containing components released from the decomposition of the reducing agent react with NOx in the flue gas, thereby reducing the concentration of nitrogen oxides in the flue gas. SNCR does not require a catalyst, the system is relatively simple, and investment and operating costs are low, making it widely used in waste incineration flue gas denitrification.
[0004] Existing engineering studies have introduced acoustic temperature measurement systems into SNCR systems of waste incinerators to acquire two-dimensional temperature fields in the furnace, enabling zoned injection. Simultaneously, support vector machines are used to establish NOx emission prediction models, and PLC control logic is used to adjust the amount of reducing agent injected. Compared to control methods that solely rely on outlet NOx feedback, this approach improves the timeliness of temperature window identification and injection control.
[0005] CN114811589A discloses a dynamic control method and system for denitrification reaction (SNCR) in waste incineration based on big data technology. This method collects incinerator operating data, identifies optimal operating parameters under different conditions, and calls upon corresponding parameters based on real-time furnace temperature and outlet ammonia escape. This allows for the adjustment of nozzle operating parameters, reducing agent type, and reducing agent concentration, thereby achieving dynamic control of outlet NOx and NH3 concentrations. This approach illustrates a technical direction for improving SNCR control using historical operating data and real-time operating parameters.
[0006] CN120361695B discloses a dynamic collaborative optimization control method for denitrification in waste incinerators. This method collects and preprocesses operating parameters of the waste incinerator, establishes a dynamic model of the denitrification process based on historical operating data, identifies and classifies operating conditions, and then uses multivariate model predictive control to formulate a collaborative control strategy. The control strategy is then output to the actuators of the denitrification and incineration systems. This approach further demonstrates the application trend of modeling, operating condition identification, and collaborative control in waste incineration denitrification.
[0007] Publication No. CN104056541A discloses a combined SNCR and cryogenic SCR denitrification system for municipal solid waste incinerators. By combining the SNCR system with the cryogenic SCR system, it aims to meet NOx emission requirements while balancing denitrification efficiency and reducing agent consumption. This type of solution mainly improves the denitrification process route and system combination, with less focus on the fine-grained control between the injection branches within the SNCR subsystem.
[0008] While the above technologies can improve the efficiency of SNCR denitrification and enhance the dynamic control of NOx emission concentration to some extent, in actual engineering operation, the high temperature, high dust, high corrosion, and ash adhesion environment inside the furnace can affect the stability of temperature measurement and monitoring data.
[0009] Thermocouple temperature measurement is easily affected by installation location and probe condition. While acoustic temperature measurement can provide temperature field information, it remains difficult to obtain a consistently high-quality temperature field under complex flue gas flow, dust deposition, and acoustic path interference conditions. Therefore, even when existing solutions incorporate acoustic temperature measurement, thermocouple temperature measurement, or historical operating condition data, it is usually difficult to accurately reflect the true reaction state of each injection area using only a single measurement method.
[0010] Existing control schemes primarily focus on predicting outlet NOx concentrations, adjusting overall ammonia injection rates, or controlling the activation and deactivation of several injection layers. Even those schemes that introduce temperature zoning or operating condition identification typically emphasize selecting suitable injection zones or adjusting overall injection parameters. They lack a unified and calculable state representation for the differences between injection sub-zones in terms of temperature window matching, unit ammonia injection denitrification response, ammonia escape risk, and actual branch execution deviations. Consequently, the control system struggles to determine which injection sub-zone should receive increased ammonia injection under current operating conditions, which should have its injection restricted, and whether the injection demand of a particular sub-zone should be transferred to an adjacent area.
[0011] SNCR injection systems typically include a reducing agent main pipe, atomizing air main pipe, and multiple injection gun circuits. When multiple injection gun circuits share a main pipe, any change in the valve opening or target flow rate in any branch can cause changes in the main pipe pressure and the actual flow rates in other branches. Existing model-based or big data-driven control methods often focus on generating upper-level control variables, while insufficiently considering the fluid coupling between injection branches, the deviation between the set flow rate and the actual flow rate, and the impact of this deviation on subsequent zone injection allocation. Therefore, even if the upper-level algorithm provides the zone ammonia injection quantity, the actual execution end may experience injection quantity deviations due to branch coupling, leading to a decrease in control effectiveness.
[0012] Existing DCS / PLC control systems in waste incineration plants typically handle multiple control tasks, including combustion, waste heat boilers, flue gas purification, and safety interlocks. If the intelligent optimization module is directly used as the main control output, it may issue inappropriate ammonia injection commands when there are communication anomalies, distorted monitoring data, model mismatch, or injection branch failures, affecting emission stability and equipment safety.
[0013] Therefore, the SNCR denitrification control of municipal solid waste incinerators still requires an intelligent control scheme that can adapt to engineering conditions such as limited furnace measurement information, large fluctuations in operating conditions, and obvious coupling of injection branches. Summary of the Invention
[0014] Therefore, the purpose of this invention is to provide a method and system for controlling SNCR denitrification in a municipal solid waste incinerator. This method maps furnace temperature-related data, flue gas emission monitoring data, boiler operation data, and reductant injection execution data to each injection sub-zone, forming a sub-zone state characterization that includes temperature window matching, denitrification response, ammonia escape risk, and injection execution deviation. Based on this, the reductant injection quantity of each injection sub-zone is collaboratively optimized, and the feasibility and operational safety of the zoned injection commands at the physical execution end are improved through safety constraints, DCS / PLC interlock verification, and branch fluid coupling compensation.
[0015] To achieve the above objectives, the present invention provides the following technical solution:
[0016] A method for controlling SNCR denitrification in a municipal solid waste incinerator includes the following steps:
[0017] S1. Obtain multi-source operational data of the SNCR denitrification process in the municipal solid waste incinerator. The multi-source operational data includes furnace temperature-related data, flue gas emission monitoring data, boiler operation process data, and reducing agent injection execution data.
[0018] S2. Based on the arrangement of the SNCR spray gun in the furnace, the SNCR spray area is divided into multiple spray sub-areas, and a mapping relationship between multi-source operating data and each spray sub-area is established.
[0019] S3. Based on the mapping relationship, the multi-source operation data is fused to obtain the sub-region state vector of each injection sub-region. The sub-region state vector includes at least the temperature window matching parameter, denitrification response parameter, ammonia escape risk parameter and injection execution deviation parameter.
[0020] S4. Based on the state vector of each injection sub-region, determine the injection priority of each injection sub-region, and construct a zoned collaborative optimization process with the objectives of NOx emission control, ammonia slip risk control and reductant consumption control to obtain the recommended reductant injection amount for each injection sub-region.
[0021] S5. Apply safety constraints to the recommended reducing agent injection amount for each injection sub-zone to generate the final reducing agent injection setting value that meets the execution requirements of the DCS / PLC control system.
[0022] S6. Output the final reducing agent injection setpoint to the DCS / PLC control system, which then controls the corresponding spray gun paths of each injection sub-zone to perform reducing agent injection.
[0023] The present invention is further configured such that: in step S2, the multiple injection sub-zones are divided according to the elevation of the SNCR spray gun, the arrangement position of the spray gun on the cross-section of the furnace, and the main direction of the flue gas flow;
[0024] Each injection sub-region corresponds to at least one injection gun path, and the injection gun paths within the same injection sub-region have the same or adjacent furnace height position, the same or adjacent flue gas flow influence area, or similar historical injection response characteristics.
[0025] When the spray coverage of a spray gun path spans more than two spray sub-regions, the spray gun path is assigned a branch affiliation weight to the corresponding spray sub-region based on the spray coverage ratio of the spray gun path to each spray sub-region or its historical spray response contribution.
[0026] The present invention is further configured such that: multi-source operational data is preprocessed before being used for fusion processing;
[0027] Preprocessing includes:
[0028] The furnace temperature-related data, flue gas emission monitoring data, boiler operation process data, and reducing agent injection execution data are time-aligned according to a unified control cycle.
[0029] Based on the transmission time of flue gas from the injection area to the flue gas emission monitoring point, lag compensation is applied to NOx concentration data and ammonia slip monitoring data.
[0030] Bad values such as abrupt changes, distorted values, and long-term unchanged values are removed from furnace temperature-related data.
[0031] The dimensions of boiler load, flue gas oxygen content, furnace negative pressure, reducing agent main pipe pressure, nozzle valve opening and actual nozzle flow rate are standardized or normalized.
[0032] The present invention is further configured such that: the mapping relationship includes temperature mapping weight, emission response mapping weight, and execution mapping weight;
[0033] Temperature mapping weights are used to characterize the contribution of furnace temperature-related data to the temperature estimation of each injection sub-region;
[0034] The emission response mapping weight is used to characterize the hysteretic response contribution of changes in the reducing agent injection rate in each injection sub-region to changes in flue gas emission monitoring data;
[0035] The execution mapping weight is used to characterize the contribution of the set injection quantity and actual injection quantity of the spray gun path to the injection state of the corresponding injection sub-region;
[0036] The temperature mapping weight, emission response mapping weight, and execution mapping weight are determined based on at least two of the following: the physical location of the spray gun, the direction of flue gas flow in the furnace, the location of the temperature measuring point, the location of the flue gas emission monitoring point, and historical injection response data.
[0037] The present invention is further configured such that: the temperature window matching parameter is determined based on the relationship between the estimated temperature of the injection sub-region and the preset effective reaction temperature window of SNCR;
[0038] When the estimated temperature of the sub-region is within the preset effective reaction temperature window of SNCR, the temperature window matching parameter increases as the estimated temperature of the sub-region approaches the preset optimal reaction temperature.
[0039] When the estimated temperature of the sub-region is lower than the lower limit of the preset effective SNCR reaction temperature window, or higher than the upper limit of the preset effective SNCR reaction temperature window, the temperature window matching parameter is reduced.
[0040] The denitrification response parameters are determined based on the change in the amount of reducing agent injected in the corresponding injection sub-zone during the historical control period, as well as the change in NOx concentration after hysteresis compensation, and are corrected according to the boiler load, flue gas oxygen content, and sub-zone temperature change trends.
[0041] The present invention is further configured such that: the ammonia slip risk parameter is determined based on the outlet ammonia slip monitoring value, the duration of the temperature of the injection sub-region being lower than the preset SNCR effective reaction temperature window, the cumulative amount of reducing agent injected into the corresponding injection sub-region, the cumulative amount of reducing agent injected into the upstream adjacent injection sub-region, and the NOx concentration change trend;
[0042] When the temperature window matching parameter of the injection sub-region is lower than the preset matching threshold, or the ammonia escape risk parameter is higher than the preset risk threshold, the injection sub-region is marked as a restricted injection sub-region.
[0043] For restricted spray sub-regions, reduce their spray allocation weight, limit the increase in their spray volume, or stop increasing the amount of reducing agent sprayed into them during the partitioned collaborative optimization process.
[0044] The present invention is further configured such that: the injection execution deviation parameter is determined based on the deviation between the target flow rate, actual flow rate, valve opening degree, and reducing agent main pipe pressure in the spray gun path;
[0045] When the deviation between the target flow rate and the actual flow rate of the spray gun path exceeds the preset deviation threshold, the spray allocation weight of the corresponding spray sub-region of the spray gun path in the next control cycle is reduced.
[0046] When the valve opening of the spray gun circuit reaches the upper limit and the actual flow rate is lower than the target flow rate, or when the valve opening of the spray gun circuit reaches the lower limit and the actual flow rate is higher than the target flow rate, the spray sub-region corresponding to the spray gun circuit is marked as an execution restricted sub-region.
[0047] For restricted sub-regions, during the partitioned collaborative optimization process, the increase of the recommended reducing agent injection amount is limited, or the corresponding injection demand is transferred to an injection sub-region adjacent to the area affected by its flue gas flow.
[0048] The present invention is further configured such that the partitioned collaborative optimization process includes:
[0049] The injection priority of each injection sub-zone is determined based on the temperature window matching parameters, denitrification response parameters, ammonia slip risk parameters, and injection execution deviation parameters of each injection sub-zone.
[0050] Prioritize meeting NOx emission target constraints, and select the injection quantity combination with the lowest overall risk of ammonia slip and reducing agent consumption from the feasible injection quantity set that meets NOx emission target constraints;
[0051] Among them, the injection allocation weight is increased for injection sub-regions with higher temperature window matching parameters and higher denitrification response parameters;
[0052] Reduce the injection allocation weight for injection sub-regions with high ammonia slip risk parameters, large injection execution deviation parameters, or those marked as restricted injection sub-regions or execution-limited sub-regions;
[0053] The feasible injection quantity set satisfies the upper and lower limits of the injection quantity of a single injection sub-zone, the upper limit of the total injection quantity of reducing agent, the upper limit of the injection quantity change rate between adjacent control cycles, and the DCS / PLC interlock permission status.
[0054] The present invention is further configured such that, before generating the final reducing agent injection setpoint, a branch fluid coupling compensation step is included, the branch fluid coupling compensation step comprising:
[0055] During the commissioning phase, the valve opening or target flow rate of a single spray gun circuit is changed sequentially while keeping the set values of other spray gun circuits unchanged. The changes in the actual flow rate of other spray gun circuits and the pressure of the reducing agent main pipe are recorded to establish a branch coupling characteristic matrix.
[0056] During operation, when the recommended reducing agent injection quantity of any spray gun path changes, the disturbance of this change to the actual flow of other spray gun paths is predicted based on the branch coupling characteristic matrix.
[0057] Reverse compensation is performed on the recommended reducing agent injection amount or the final reducing agent injection set value for other spray gun routes based on the disturbance amount, so as to reduce the impact of reducing agent main pipe pressure fluctuation on the actual injection amount of each spray sub-zone;
[0058] The injection setpoint after reverse compensation still needs to be limited by single branch amplitude, total amplitude and adjacent control cycle change rate before it is output to the DCS / PLC control system as the final reducing agent injection setpoint.
[0059] The present invention is further configured to include:
[0060] The data acquisition interface module is used to acquire multi-source operational data of the SNCR denitrification process of the municipal solid waste incinerator. The multi-source operational data includes furnace temperature-related data, flue gas emission monitoring data, boiler operation process data, and reducing agent injection execution data.
[0061] The injection sub-region mapping module is used to divide the SNCR injection area into multiple injection sub-regions according to the arrangement of the SNCR nozzles in the furnace, and to establish a mapping relationship between multi-source operating data and each injection sub-region.
[0062] The multi-source fusion state characterization module is used to fuse multi-source operating data based on mapping relationships to obtain the sub-region state vector of each injection sub-region. The sub-region state vector includes at least temperature window matching parameters, denitrification response parameters, ammonia escape risk parameters, and injection execution deviation parameters.
[0063] The partitioned collaborative optimization control module is used to determine the injection priority of each injection sub-region based on the sub-region state vector of each injection sub-region, and to construct a partitioned collaborative optimization process with the objectives of NOx emission control, ammonia slip risk control and reductant consumption control, so as to obtain the recommended reductant injection amount for each injection sub-region.
[0064] The safety hybrid integration module is used to perform safety constraint processing on the recommended reducing agent injection amount of each injection sub-zone, generate the final reducing agent injection setting value that meets the execution requirements of the DCS / PLC control system, and output the final reducing agent injection setting value to the DCS / PLC control system.
[0065] The branch fluid coupling compensation module is used to perform feedforward compensation on the suggested reducing agent injection amount or the final reducing agent injection set value of each injection sub-region based on the pre-established branch coupling characteristic matrix.
[0066] The DCS / PLC control system is used to control the corresponding nozzle path of each injection sub-zone to perform reducing agent injection according to the final reducing agent injection set value, and to switch to the standby injection control curve based on boiler load and flue gas oxygen content when abnormal retreat conditions are triggered.
[0067] Compared with the shortcomings of the prior art, the beneficial effects of the present invention are as follows:
[0068] This invention unifies the collection, preprocessing, and sub-region mapping of furnace temperature, flue gas emissions, boiler operating conditions, and injection execution data to form initial feature sets and sub-region state vectors for each injection sub-region. This allows SNCR control to no longer rely solely on outlet NOx concentration or total ammonia injection, but to reflect the actual reaction state of different injection sub-regions. It can distinguish the temperature suitability, denitrification contribution, ammonia slip risk, and execution reliability of each injection sub-region, providing a calculable basis for zoned, refined injection control. While meeting NOx emission targets, it also considers ammonia slip risk and reducing agent consumption, avoiding blindly increasing the reducing agent injection amount in injection sub-regions with low temperatures, high ammonia slip risk, or abnormal execution.
[0069] It can participate in SNCR injection control without disrupting the existing DCS / PLC safety interlock logic, thereby improving the safety, compatibility, and engineering maintainability of the system. Attached Figure Description
[0070] Figure 1 This is a flowchart illustrating the overall process of the method of the present invention.
[0071] Figure 2 This is a schematic diagram illustrating the system architecture and control relationships of the present invention;
[0072] Figure 3 This is a schematic diagram of the injection sub-region division and data mapping of the present invention;
[0073] Figure 4 This is a schematic diagram of the data flow and control closed loop of the present invention. Detailed Implementation
[0074] Reference Figures 1 to 4 The embodiments of the present invention will be further described below.
[0075] The SNCR denitrification control for municipal solid waste incinerators is applied to the SNCR denitrification system. The system includes a reducing agent storage and supply unit, a metering and conveying unit, a reducing agent main pipe, atomizing air piping, multiple spray nozzle paths, a DCS / PLC control system, and a control unit. The control unit communicates with the DCS / PLC to receive multi-source operating data and generate injection setting suggestions; the final injection setting value is executed by the DCS / PLC after safety constraints.
[0076] The DCS / PLC control system is used to control the metering pump, spray gun path regulating valve, atomizing air valve, or spray gun path deployment / retraction mechanism to perform spraying actions based on the final reducing agent spraying setpoint after safety constraint processing.
[0077] The control unit includes a data acquisition interface, a data preprocessing module, a jet sub-region mapping module, a multi-source fusion state characterization module, a zoned collaborative optimization control module, a safe hybrid integration module, and a branch fluid coupling compensation module.
[0078] The data acquisition interface is used to obtain operational data from furnace temperature detection equipment, continuous flue gas monitoring system, boiler DCS system, and reducing agent injection actuator. Furnace temperature detection equipment includes acoustic temperature measuring devices, thermocouple temperature measuring devices, and infrared temperature measuring devices; the continuous flue gas monitoring system is used to acquire at least some of the data from NOx concentration, flue gas oxygen content, and ammonia slip monitoring values; the boiler DCS system is used to provide boiler operation data such as boiler load, furnace negative pressure, primary air volume, secondary air volume, flue gas temperature, grate speed, or steam volume; the reducing agent injection actuator is used to provide injection execution data such as reducing agent main pipe pressure, atomizing air pressure, injection gun line valve opening, injection gun line target flow rate, and injection gun line actual flow rate.
[0079] The spray gun path is divided into multiple spray sub-zones according to the installation elevation, cross-sectional position, and main flue gas direction of the spray gun in the furnace. Each spray sub-zone corresponds to one spray gun path, or more than two spray gun paths. When the spray coverage of a spray gun path spans more than two spray sub-zones, the spray gun path is assigned a branch weight to different spray sub-zones based on the spray coverage ratio, spray direction, or historical spray response contribution of the spray gun path.
[0080] For municipal solid waste incinerators with a two-layer SNCR spray nozzle arrangement, the initial division is based on the upper and lower spray nozzle areas. Within each spray nozzle area, it is further divided into multiple injection sub-zones according to the distribution of the furnace front wall, rear wall, left wall, right wall, or cross-section. For incinerators where the mainstream flue gas direction in the furnace is significantly biased to one side or where the furnace cross-section temperature distribution has been uneven for a long time, the boundaries of the injection sub-zones are also adjusted based on historical temperature distribution and injection response results.
[0081] Step S1: Multi-source running data acquisition and preprocessing. Data from different devices, different sampling periods, and different physical response times are uniformly converted into a preprocessed running dataset D(t) that can be used for jet sub-region mapping and state characterization. The preprocessed running dataset D(t) serves as the input to the subsequent initial feature set Fp(t) of the jet sub-region.
[0082] During the t-th control cycle, raw multi-source operational data of the SNCR denitrification process of the municipal solid waste incinerator are acquired through the data acquisition interface. This multi-source operational data includes furnace temperature-related data, flue gas emission monitoring data, boiler operation process data, and reductant injection execution data. The furnace temperature-related data originates from one or more of acoustic temperature measurement, thermocouple, and infrared temperature measurement methods; the flue gas emission monitoring data includes at least a portion of NOx concentration, flue gas oxygen content, and ammonia slip monitoring values; the boiler operation process data includes at least a portion of boiler load, steam output, furnace negative pressure, air volume, flue gas temperature, and grate speed; and the reductant injection execution data includes at least a portion of main pipe pressure, atomized air pressure, branch valve opening, target flow rate, actual flow rate, and activation / deactivation status.
[0083] After obtaining the raw multi-source operating data, the control unit first performs time alignment on various types of data according to a unified control cycle. The control unit's optimization calculation cycle is used as the unified control cycle. Within each control cycle, data with sampling frequencies higher than the unified control cycle are processed using the average, final, or median values; data with sampling frequencies lower than the unified control cycle are supplemented using the previous effective value, linear interpolation, or the latest effective value. For NOx and ammonia slip data, lag compensation is performed based on the transmission time of flue gas from the injection zone to the CEMS sampling point. Data exceeding the range, exhibiting abrupt changes, remaining unchanged for a long period, or showing abnormal deviations from adjacent measuring points are marked and replaced with the previous effective value, the weighted value of adjacent measuring points, or historical effective values. Temperature, concentration, pressure, flow rate, opening degree, and load are dimensionlessly standardized to form a preprocessed operating dataset D(t). This preprocessed operating dataset D(t) is then used in the subsequent injection sub-region division and mapping relationship establishment steps to generate the initial feature set Fp(t) for each injection sub-region.
[0084] Step S2: Distribute the preprocessed running dataset D(t) to different injection sub-regions according to the physical arrangement of the SNCR nozzles, the direction of flue gas flow in the furnace, the location of temperature measuring points, and the historical injection response relationship, forming the initial feature set Fp(t) for each injection sub-region. The initial feature set Fp(t) serves as the input to the subsequent sub-region state vector Xp(t).
[0085] The input for this step is the preprocessed operation dataset D(t) for the t-th control cycle. The preprocessed operation dataset D(t) includes at least furnace temperature-related data, flue gas emission monitoring data, boiler operation process data, and reducing agent injection execution data.
[0086] The output of this step is the initial feature set Fp(t) of each jet sub-region, where p represents the p-th jet sub-region.
[0087] Based on the installation elevation of the SNCR spray nozzles in the furnace, the area within the furnace where reducing agent can be injected is divided into one or more injection layers. For municipal solid waste incinerators with two layers of spray nozzles, they are divided into upper and lower injection zones; for incinerators with three or more layers of spray nozzles, they are divided sequentially according to the elevation of the spray nozzles into high-level, mid-level, and low-level injection zones. The division of injection layers reflects the differences in SNCR reaction temperature windows and flue gas residence time at different furnace heights.
[0088] After the spray layer division is completed, each spray layer is further divided laterally or directionally according to the arrangement of the spray guns on the cross-section of the furnace. For incinerators with spray guns installed on the front and rear walls, the division is based on the front wall spray zone and the rear wall spray zone; for incinerators with spray guns installed on the left and right side walls, the division is based on the left side spray zone and the right side spray zone; for incinerators where the temperature distribution of the furnace cross-section has long been inconsistent, the division is further divided into a left sub-zone, a middle sub-zone, and a right sub-zone.
[0089] When dividing the injection sub-zones, adjustments are made based on the mainstream flue gas direction in the furnace. The flue gas flow in a municipal solid waste incinerator is affected by grate operation, primary and secondary air distribution, furnace negative pressure, and the state of waste combustion. Based on the furnace structure, mainstream flue gas direction, injection nozzle direction, and historical operating data, injection nozzle paths within the same flue gas flow influence range are grouped into the same injection sub-zone, or a single injection nozzle path is associated with two or more injection sub-zones with different weightings.
[0090] Each injection sub-region corresponds to at least one injection gun path. Injection gun paths within the same injection sub-region share the same or adjacent furnace height positions, the same or adjacent flue gas flow influence areas, and similar historical injection response characteristics. Historical injection response characteristics refer to the response characteristics of furnace temperature, NOx concentration, or ammonia slip monitoring values after a change in the injection volume of a particular injection gun path during past operation or commissioning.
[0091] Let there be P spray sub-regions, and let the p-th spray sub-region be denoted as P. The number of furnace temperature measuring points or temperature grids is M, and the temperature value of the i-th temperature measuring point or temperature grid in the t-th control cycle is denoted as... The number of spray gun paths is R, and the target flow and actual flow of the r-th spray gun path are denoted as , ... and .
[0092] The control unit establishes temperature mapping weights, emission response mapping weights, and execution mapping weights respectively. These three types of mapping weights are used to solve the following problems: how to assign temperature data to the injection sub-region; how to establish a hysteresis response relationship between outlet NOx and ammonia slip data and the injection sub-region; and how to convert the target flow and actual flow of the spray gun path into the injection state of the injection sub-region.
[0093] Temperature mapping weights are used to characterize the contribution of furnace temperature-related data to the temperature estimation of each injection sub-region. Based on the spatial distance between the temperature measuring point or temperature grid and the injection sub-region, the proximity of the furnace height, and the correlation of the flue gas mainstream direction, the temperature contribution coefficient of the i-th temperature measuring point or temperature grid to the p-th injection sub-region is determined. .
[0094] Temperature contribution coefficient Determine as follows:
[0095]
[0096] in, This represents the spatial distance between the i-th temperature measuring point or temperature grid and the center of the p-th injection sub-region; This indicates the relative positional relationship between the i-th temperature measuring point or temperature grid and the p-th injection sub-region in the furnace height direction; This indicates the degree of correlation between the i-th temperature measuring point or temperature grid and the p-th injection sub-region in the direction of the mainstream flue gas flow. , and These are the spatial distance evaluation function, the elevation proximity evaluation function, and the flue gas flow direction correlation evaluation function, respectively. , and These are the corresponding weighting coefficients.
[0097] Set as a function that decreases as spatial distance increases; Set up a piecewise function that assigns higher values to measuring points on the same floor, medium values to measuring points on adjacent floors, and lower values to measuring points on distant floors. The values are assigned based on whether the temperature measuring point or temperature grid is located upstream, downstream, or within the same flue gas flow influence range of the p-th injection sub-region. The above functions and weighting coefficients are determined based on the spray gun layout diagram, measuring point layout diagram, furnace structure, and on-site commissioning data.
[0098] Contribution coefficient to temperature After normalization, the temperature mapping weights of the i-th temperature measurement point or temperature grid to the p-th injection sub-region are obtained:
[0099] Therefore, the baseline value of the estimated sub-region temperature of the p-th injection sub-region in the t-th control cycle. Calculate according to the following formula:
[0100] When the furnace temperature data comes from a two-dimensional temperature grid formed by an acoustic temperature measurement device... This represents the temperature value at the i-th temperature grid point;
[0101] When the furnace temperature data comes from a thermocouple This represents the temperature value at the i-th thermocouple measuring point; when the furnace temperature data comes from an infrared temperature measuring device, This indicates the temperature value of the corresponding temperature measurement area within the infrared field of view.
[0102] When multiple temperature data exist simultaneously, the validity of different temperature sources is first assessed, and a reliability coefficient for each temperature source is set. Temperature sources that are valid, have a clear spatial relationship with the injection sub-region, and exhibit good historical stability have a higher reliability coefficient; temperature sources that are in a state of dust accumulation, drift, malfunction, or long-term abnormality have a lower reliability coefficient or are temporarily excluded from the calculation.
[0103] Since NOx concentration and ammonia slip monitoring values are usually collected uniformly by the downstream continuous flue gas monitoring system, they cannot be directly associated with a specific injection sub-region. Therefore, emission response mapping weights are established through hysteresis compensation and historical injection responses.
[0104] During the commissioning or historical data analysis phase, within safe limits, a small variation is applied to the reducing agent injection rate in the p-th injection sub-region. The changes in NOx concentration and ammonia slip monitoring value after hysteresis compensation are recorded. The NOx response contribution coefficient b_p of the p-th injection sub-region is determined according to the following formula:
[0105]
[0106] in, This represents the change in the amount of reducing agent injected into the p-th injection sub-region; This represents the NOx reduction corresponding to the injection change after hysteresis compensation; ε is a very small positive number to prevent the denominator from being zero.
[0107] Contribution coefficient of NOx response in each injection sub-region After normalization, the NOx emission response mapping weights for the p-th injection sub-region are obtained:
[0108] in, The larger the value, the greater the contribution of the injection rate change in the p-th injection sub-region to the historical response of NOx reduction at the outlet. In subsequent calculations of denitrification response parameters, This serves as the base weight for the NOx response capability of the injection sub-region.
[0109] For the ammonia slip response, the contribution coefficient of the ammonia slip response is:
[0110] in, This represents the increase in the ammonia slip monitoring value corresponding to the change in the injection volume of the p-th injection sub-region after hysteresis compensation. The larger the value, the higher the risk of ammonia escape when the amount of reducing agent injected into that injection sub-region continues to increase.
[0111] The emission response mapping weights are not recalibrated in each control cycle, but are updated within a sliding history window based on valid injection disturbance samples. Valid injection disturbance samples refer to samples where the CEMS is in an effective state during that time period, boiler load changes do not exceed a preset range, injection volume changes in other injection sub-zones are small, or the main response source can be distinguished through data processing. The emission response mapping weights are not updated when the CEMS is in calibration, backflushing, maintenance, malfunction, or data invalidation state.
[0112] To avoid sudden changes in weights due to accidental disturbances, a smooth update method is used to update the NOx response contribution coefficient and the ammonia escape response contribution coefficient.
[0113] The execution mapping weight is used to characterize the affiliation between the spray gun path and the spray sub-region. Let the execution mapping weight of the r-th spray gun path to the p-th spray sub-region be... This weight is determined based on the spray coverage ratio, spray direction, design flow rate, actual flow rate contribution, and historical response contribution of the spray gun path.
[0114] For a spray gun path that covers only one spray sub-region, the execution mapping weight for that spray sub-region is set to 1, and the execution mapping weight for other spray sub-regions is set to 0. For a spray gun path that covers more than two spray sub-regions, the allocation is normalized according to the spray coverage ratio or historical response contribution, and it satisfies the following conditions:
[0115] Based on this, the target jet volume of the p-th jet sub-region The result is obtained by weighting the target flow rate of each spray gun path:
[0116] The actual jet volume of the p-th jet sub-region The result is obtained by weighting the actual flow rates of each spray gun path:
[0117] When a spray gun path is under maintenance, blocked, experiencing flow meter malfunction, valve feedback anomaly, or is disabled by the DCS / PLC, the execution mapping weight of that spray gun path to the corresponding spray sub-region is reduced, or the corresponding spray sub-region is marked as an execution-restricted sub-region. This setting prevents subsequent partitioned collaborative optimization processes from continuing to assign the main spraying tasks to spray gun paths that cannot execute stably.
[0118] By using temperature mapping weights, emission response mapping weights, and execution mapping weights, the control unit converts the preprocessed running dataset D(t) into the initial feature set Fp(t) for each injection sub-region.
[0119] For the p-th injection sub-region, the initial feature set Fp(t) is expressed as:
[0120]
[0121] in, The baseline value for the sub-region temperature estimation of the p-th injection sub-region; The target injection quantity for the p-th injection sub-region; This represents the actual injection volume of the p-th injection sub-region; Map the NOx emission response weights for the p-th injection sub-region; The contribution coefficient for the ammonia slip response of the p-th injection sub-region; The oxygen content of the flue gas is L(t); the boiler load is L(t). The pressure in the reducing agent main pipe; For atomizing air pressure; This refers to the engagement / disengagement status, fault status, or execution permission status of the spray gun path corresponding to the p-th spray sub-region.
[0122] The initial feature set Fp(t) is not directly used as the injection control quantity, but rather as the input to the subsequent sub-region state vector Xp(t). Based on Fp(t), the control unit further calculates the temperature window matching parameters, denitrification response parameters, ammonia slip risk parameters, and injection execution deviation parameters, thereby obtaining the sub-region state vector Xp(t) of the p-th injection sub-region.
[0123] The division of injection sub-regions, branch assignment weights, temperature mapping weights, emission response mapping weights, and execution mapping weights are pre-established during the system commissioning phase and periodically corrected based on historical data during subsequent operation. For incinerators with CFD analysis capabilities, the initial mapping relationship is determined by combining the furnace CFD flow field simulation results; for incinerators without CFD analysis capabilities, the mapping relationship is also established based on the spray gun layout diagram, measuring point layout diagram, flue gas mainstream direction judgment, on-site commissioning response data, and historical operating data.
[0124] Step S3: The input to this step is the initial feature set Fp(t) of the p-th injection sub-region in the t-th control cycle. The output of this step is the sub-region state vector Xp(t) of the p-th injection sub-region, which includes the temperature window matching parameters. Denitrification response parameters Ammonia escape risk parameters and injection execution deviation parameters The sub-region state vector is represented as:
[0125] Among them, the state marking of the injection sub-region This includes states such as normal sub-region, restricted injection sub-region, execution restricted sub-region, or prohibition of adding injection sub-regions.
[0126] The SNCR reaction is significantly dependent on furnace temperature. At excessively low temperatures, the decomposition and reduction reaction of the reducing agent are incomplete, easily leading to ammonia escape; at excessively high temperatures, the reducing agent may undergo oxidation side reactions, reducing denitrification efficiency. The baseline value is estimated based on the sub-zone temperature of the p-th injection sub-zone. Calculate temperature window matching parameters The preset effective reaction temperature window for SNCR is... The preset optimal reaction temperature is .in, This indicates the lower limit of the effective reaction temperature window. Indicates the upper limit of the effective reaction temperature window lie in and between.
[0127] Temperature window matching parameters Calculation using a segmented approach:
[0128] when Below If the temperature in the injection sub-region is too low, the temperature window matching parameter should be set to a lower value.
[0129] when Higher than If the temperature in the injection sub-region is too high, the temperature window matching parameter should also be set to a lower value.
[0130] when lie in and Between these times, the temperature window matching parameters vary. near And increase.
[0131] Determined according to the following rules:
[0132]
[0133]
[0134] in, The value of is limited to between 0 and 1. The closer it is to 1, the closer the ejector sub-region is to the optimal SNCR reaction temperature; The closer it is to 0, the less suitable it is for the injection sub-region to continue increasing the amount of reducing agent injected.
[0135] The control unit selects valid injection response samples within a sliding historical time window. Valid injection response samples refer to samples where CEMS data is valid within that time period, boiler load changes are within acceptable limits, injection volume changes in other injection sub-zones are small, or the main response source can be distinguished by mapping weights.
[0136] For the p-th spray sub-region, within a valid sample, let the change in the amount of reducing agent sprayed be . The NOx reduction after lag compensation is The unit jet denitrification response value for this sample is expressed as:
[0137] Here, ε is a very small positive number used to prevent the denominator from being zero. When the injection rate increases but NOx does not decrease, the effective denitrification response for this sample is 0 or a lower value.
[0138] To ensure that the denitrification response parameters reflect the current operating conditions, adjustments are made based on the trends in boiler load, flue gas oxygen content, and sub-zone temperature. Corrections are made to the denitrification response parameters. Represented as:
[0139] in, Map the NOx emission response weights for the p-th injection sub-region; The average or weighted average of the unit jet denitrification response of valid samples within the sliding historical time window; This is the boiler load correction factor; This is the correction factor for the oxygen content of flue gas; This is a correction factor for the temperature change trend in the sub-region.
[0140] The correction factor is determined using table lookup, piecewise functions, or empirical rules. When the boiler load variation is small and the oxygen content in the flue gas is within the normal range... and Take a value close to 1; when the boiler load changes rapidly or the oxygen content in the flue gas deviates significantly from the normal range, reduce the reliability of the denitrification response parameters to avoid over-reliance on historical response results when the operating conditions fluctuate drastically.
[0141] Further normalize to between 0 and 1. The higher the value, the more significant the NOx reduction effect of the unit reduction agent injection change in the injection sub-region under the current operating conditions; The lower the value, the lower the marginal denitrification benefit of continuing to increase the amount of reducing agent injected in that injection sub-region.
[0142] Ammonia escape risk parameters This parameter is used to characterize the likelihood of ammonia slip when the reducing agent injection rate continues to increase in the p-th injection sub-region. It is not solely determined by the outlet ammonia slip monitoring value, but rather by a combination of the injection sub-region temperature, cumulative injection rate, injection status of adjacent upstream sub-regions, and NOx decreasing trend.
[0143] Ammonia escape risk parameters It consists of the following factors:
[0144] Exported ammonia slip monitoring value or estimated ammonia slip value;
[0145] The duration for which the temperature of the p-th injector sub-region is below the lower limit of the effective reaction temperature window for SNCR;
[0146] The cumulative amount of reducing agent injected into the p-th injection sub-region within the sliding time window;
[0147] The cumulative amount of reducing agent injected into the upstream adjacent injection sub-region of the p-th injection sub-region;
[0148] Does the decreasing trend of NOx concentration with increasing injection volume weaken?
[0149] Construct it in the following form:
[0150]
[0151] in, This is a risk item related to the escape of exported ammonia. This is a risk item related to persistent low temperatures; Accumulated spray risk items for this sub-region; The cumulative spray risk item for the upstream adjacent sub-region; The risk term is the weakening of the NOx decline trend; α1, α2, α3, α4 and α5 are the weighting coefficients of each risk term.
[0152] Exported ammonia escape risk item The ammonia slip threshold is determined based on the relationship between online ammonia slip monitoring values and preset ammonia slip control thresholds. When an online ammonia slip meter is not configured, an ammonia slip estimate is also constructed based on historical ammonia slip monitoring data, the trend of ammonia injection rate changes, the NOx decrease trend, and the degree of temperature window deviation.
[0153] Low temperature continuous risk item According to the temperature of the p-th jet sub-region being lower than The duration is determined. When Below And if the duration exceeds the preset time, Increase; when When returning to the effective reaction temperature window, Gradually decrease.
[0154] This sub-area has a cumulative spray risk item. The amount of reducing agent injected into the p-th injection sub-region within the sliding time window is determined. When the cumulative injection amount in this sub-region remains high but the NOx decrease is not significant, it indicates that the marginal benefit of continuing to increase ammonia injection is decreasing, and the risk of ammonia escape is increasing.
[0155] Upstream adjacent sub-region cumulative spray risk item This parameter reflects the potential impact of excessive upstream injection on unreacted ammonia migration in downstream areas. For adjacent injection sub-regions located upstream of the p-th injection sub-region along the main flue gas flow direction, the ammonia escape risk parameter of the p-th injection sub-region increases accordingly if its cumulative injection volume is high.
[0156] Risk of weakening NOx decline trend This is used to determine whether increasing ammonia injection still leads to an effective reduction in NOx. When the ammonia injection rate in the p-th injection sub-region or its adjacent sub-region increases, if the NOx reduction after hysteresis compensation gradually decreases, or if NOx increases instead of decreasing, then... Increase.
[0157] Normalize to between 0 and 1. The higher the value, the greater the risk of ammonia escape from continuing to increase the amount of reducing agent injected into the p-th injection sub-region. The lower the value, the lower the risk of ammonia escape in that sub-region under the current operating conditions.
[0158] when The temperature exceeds the preset risk threshold, or the temperature window matching parameter. When the value is below a preset matching threshold, the control unit marks the p-th injection sub-region as a restricted injection sub-region. For restricted injection sub-regions, their injection allocation weight is reduced during subsequent partition co-optimization processes, limiting the increase in their injection volume, or the increase in reducing agent injection volume to them is stopped.
[0159] Injection execution deviation parameters This parameter is used to characterize the deviation between the target injection quantity and the actual injection quantity of the corresponding spray gun path in the p-th injection sub-region, and whether this deviation is related to the main pipe pressure, atomizing air pressure, valve opening, or branch fault. This parameter is used to address the problem that the results of the upper-level optimization calculation cannot be accurately reproduced at the physical execution end.
[0160] The flow deviation term of the p-th injection sub-region is represented as:
[0161] Let p be the target injection quantity for the p-th injection sub-region. The actual injection volume of the p-th injection sub-region ε represents the maximum allowable injection volume for the p-th injection sub-region, where ε is a very small positive number.
[0162] A comprehensive injection execution deviation parameter is formed by combining the valve opening state and pressure state. :
[0163]
[0164] in, This is an abnormal valve opening condition. This is an abnormal pressure item in the reducing agent main pipe. β1 represents the atomized air pressure anomaly term; β2, β3, and β4 are the corresponding weighting coefficients.
[0165] Determined based on whether the opening of the spray gun circuit valve is close to the upper or lower limit.
[0166] The determination is based on whether the pressure in the reducing agent main pipe is lower or higher than the design operating range. The determination is based on whether the atomizing air pressure meets the atomization requirements of the spray gun.
[0167] Normalize to between 0 and 1. The higher the value, the worse the execution reliability of the corresponding spray gun path in that spray sub-region; The lower the value, the more consistent the target injection volume is with the actual injection volume, indicating a better execution status.
[0168] when When the preset execution deviation threshold is exceeded, the control unit reduces the injection allocation weight of the p-th injection sub-region in the next control cycle. If the corresponding spray gun path of the injection sub-region simultaneously experiences valve opening reaching its limit, actual flow rate significantly deviating from the target flow rate, abnormal main pipe pressure, or abnormal atomizing air pressure, then the p-th injection sub-region is marked as an execution-restricted sub-region.
[0169] For restricted sub-regions, the increase of their recommended reducing agent injection amount is limited during subsequent partitioned collaborative optimization, or the corresponding injection demand is transferred to an injection sub-region adjacent to its flue gas flow influence area, with a higher temperature window matching parameter and normal execution status.
[0170] In obtaining , , and Then, the control unit determines the status flag of the p-th injection sub-region according to the preset judgment rules. .
[0171] when When the temperature is below the preset matching threshold, it indicates that the temperature conditions of the p-th injection sub-region are not suitable for further increasing the amount of reducing agent injected, and it is marked as a restricted injection sub-region; when When the risk level exceeds a preset risk threshold, it indicates a high risk of ammonia escape in that sub-region, and it is also marked as a restricted injection sub-region; when When the deviation exceeds the preset execution threshold, it is marked as an execution-restricted sub-region; when the DCS / PLC reports that the corresponding spray gun circuit of the sub-region is in a state of prohibition from operation, maintenance or failure, it is marked as a prohibited spray sub-region.
[0172] For injection sub-regions that do not trigger the above-mentioned constraints, they are marked as normal sub-regions. Normal sub-regions do not necessarily increase the injection rate. Instead, the decision to increase, maintain, or decrease the injection rate is made during subsequent zonal co-optimization based on NOx emission targets, denitrification response parameters, and reducing agent consumption constraints.
[0173] The sub-region state vector Xp(t) of the p-th injection sub-region is fully determined. This sub-region state vector includes temperature window matching parameters reflecting reaction conditions, denitrification response parameters reflecting control benefits, ammonia slip risk parameters reflecting side effect risks, and injection execution deviation parameters reflecting execution reliability. Therefore, subsequent zone-based collaborative optimization processes can be based on the actual executable state of each injection sub-region, rather than solely on outlet NOx concentration deviation or total ammonia injection volume as the control basis.
[0174] Step S4: Using the sub-region state vector Xp(t) of each injection sub-region as input, and under the conditions of satisfying NOx emission control requirements, ammonia slip risk control requirements, reductant consumption control requirements, and DCS / PLC execution constraints, calculate the recommended reductant injection quantity Qp*(t) for each injection sub-region.
[0175] During the t-th control cycle, the control unit first determines the current control requirements based on the outlet NOx concentration, the NOx emission target value, and the NOx change trend. The NOx emission target value is the plant control target that is lower than the statutory emission limit, and it is also the preset operation control target in the DCS / PLC system.
[0176] Let the current NOx concentration after effectiveness assessment be N(t), and the NOx target value be... Then NOx control deviation Represented as:
[0177] when When the value exceeds the preset positive deviation threshold, it indicates that the NOx concentration is higher than the target value or has a tendency to approach the upper limit of the target, and the zonal collaborative optimization process enters the injection optimization state.
[0178] when When the deviation is within the allowable range, the partitioned collaborative optimization process enters the optimization maintenance state;
[0179] when When the deviation is less than the preset negative deviation threshold and the risk of ammonia escape or the consumption of reducing agent is high, the zone collaborative optimization process enters the spray reduction optimization state.
[0180] To avoid frequent changes in injection volume due to short-term fluctuations in NOx levels at the export site, Perform moving averages or trend analysis.
[0181] After determining the NOx control requirements, the control unit calculates the injection priority based on the sub-region state vector Xp(t) of each injection sub-region. The injection priority is used to determine which injection sub-regions should receive the injection increment first when the reducing agent injection quantity needs to be increased, and which injection sub-regions should reduce the injection quantity first when the reducing agent injection quantity needs to be reduced.
[0182] The jet priority score of the p-th jet sub-region Matching parameters based on temperature window Denitrification response parameters Ammonia escape risk parameters and injection execution deviation parameters Sure:
[0183]
[0184] in, , , , and These are the corresponding weighting coefficients; This is a penalty item for the state of the spray sub-region.
[0185] When the p-th injection sub-region is a normal sub-region Take the lower value or 0; when the p-th injection sub-region is marked as a restricted injection sub-region, an execution restricted sub-region, or an application to add injection sub-regions is prohibited. A higher value is chosen to reduce its injection priority.
[0186] In the above scoring, the temperature window matching parameter The higher the value, the closer the injection sub-region is to the effective reaction temperature window of SNCR, making it suitable for participating in the denitrification reaction; denitrification response parameters The higher the value, the more significant the NOx reduction contribution per unit ammonia injection in that injection sub-region; ammonia slip risk parameter The higher the value, the more likely it is that further increasing ammonia injection may lead to ammonia escape; injection execution deviation parameter The higher the value, the worse the reliability of the corresponding branch in the injection sub-region.
[0187] In booster spray optimization mode, the control unit prioritizes selecting... Higher injection sub-regions allocate injection increments; under reduced injection optimization, priority is given to... Reduce the injection volume in injection sub-regions with lower ammonia escape risk, higher ammonia escape risk, or larger execution deviations.
[0188] After calculating the injection priority, the control unit constructs a set of feasible injection quantities for each injection sub-region. The set of feasible injection quantities is used to limit the recommended reducing agent injection quantity Qp*(t) to be within the allowable range of field equipment, safety interlocks, and process controls.
[0189] For the p-th injection sub-region, the recommended reducing agent injection amount Qp*(t) should satisfy the following constraint:
[0190] in, Let p be the minimum allowable injection volume for the p-th injection sub-region. This represents the maximum allowable spray volume for the p-th spray sub-zone. The upper and lower limits are determined based on the spray gun circuit design flow rate, metering pump capacity, spray gun atomization requirements, and DCS / PLC setting range.
[0191] The total recommended reducing agent injection amount for each injection sub-region should meet the total quantity constraint:
[0192] in, This represents the upper limit of the total amount of reducing agent that the SNCR system can inject under current operating conditions.
[0193] To avoid frequent and significant changes in the injection volume, it is recommended that the reducing agent injection volume also meet the constraint of the rate of change between adjacent control cycles:
[0194] This represents the actual or final setpoint value of the p-th injection sub-region in the previous control cycle. This represents the maximum allowable variation of the injection sub-region within a single control cycle.
[0195] For jet sub-regions marked as restricted jet sub-regions, set:
[0196] Alternatively, the increase can be limited to a smaller threshold. For injection sub-regions marked as prohibited from increasing injection volume, the control unit will not allocate injection increments to them. For injection sub-regions marked as restricted in execution, the recommended increase in reducing agent injection volume will be limited, or their injection demand will be transferred to an adjacent injection sub-region that is in normal condition and is adjacent to their flue gas flow influence area.
[0197] When the DCS / PLC reports that a certain spraying sub-area's corresponding branch is in a prohibited operation, maintenance, fault, interlock shutdown, or manual mode, that spraying sub-area will not be included in the feasible set for increasing the spraying volume.
[0198] After constructing a feasible set of injection quantities, the control unit allocates the recommended reducing agent injection quantity based on NOx control requirements and the injection priority of each injection sub-zone.
[0199] When the system is in boosted injection optimization mode, first calculate the total injection increment required for the current control cycle. The total injection increment is based on the NOx control deviation. The NOx upward trend, historical total ammonia injection response, and upper limit of total reducing agent were determined. Then, based on the injection priority score of each injection sub-region, [the following was determined]. Assign to the injection sub-region that allows for increased injection volume.
[0200] First, for the injection sub-regions that allow increased injection volume... After nonnegation and normalization, the allocation coefficients are obtained. :
[0201]
[0202] Ω represents the set of sub-regions where the injection volume is currently allowed to increase.
[0203] The suggested injection increment for the p-th injection sub-region is expressed as:
[0204] The corresponding recommended reducing agent injection amount is:
[0205] During the calculation process, if a certain injection sub-region reaches the upper limit of single-region injection volume, the upper limit of change rate, or is restricted by safety rules, the injection increment that it could not be allocated will be reallocated to other injection sub-regions that are allowed to increase injection volume and have higher injection priority.
[0206] When the system is in optimized mode, the control unit maintains the current injection rate and slightly reduces the injection rate in injection sub-regions with high ammonia slip risk, low denitrification response, or large execution deviation. Simultaneously, the corresponding injection rate is compensated to injection sub-regions with better temperature window matching, higher denitrification response, and normal execution. This method can reduce ammonia slip risk and reducing agent consumption while achieving NOx compliance.
[0207] When the system is in reduced injection optimization mode, the control unit prioritizes reducing the ammonia slip risk parameter. High denitrification response parameters Lower or injection execution deviation parameters The injection volume in a larger injection sub-region is also limited by the lower limit of the injection volume and the rate of change of the injection volume in a single region, so as to avoid a sudden increase in NOx concentration due to rapid injection reduction.
[0208] When the p-th injection sub-region is marked as a restricted injection sub-region or an execution restricted sub-region, the control unit searches for injection sub-regions that are adjacent to or downstream of the p-th injection sub-region in the flue gas flow influence area. If the temperature window matching parameters of the adjacent injection sub-regions are... The denitrification response parameters are higher than the preset matching threshold. Ammonia escape risk parameter exceeding the preset response threshold Below the preset risk threshold, and the injection execution deviation parameter If the deviation is below the preset execution deviation threshold, the portion of the injection increment that should have been allocated to the p-th injection sub-region will be transferred to the adjacent injection sub-region.
[0209] The transfer of injection demand is not unconditional. The transferred injection quantity must still meet the single-zone injection quantity limit, change rate limit, and DCS / PLC interlock permission status of the transferred sub-zone. If there is no adjacent injection sub-zone that meets the conditions, the control unit does not forcibly transfer the injection demand, but limits the total injection increment for this control cycle, and the DCS / PLC system handles it according to the backup control strategy or alarm logic. After injection priority calculation, feasible injection quantity set construction, optimized allocation, and injection demand transfer processing, the control unit outputs the suggested reducing agent injection quantity Qp*(t) for each injection sub-zone in the t-th control cycle.
[0210] The recommended reducing agent injection rate Qp*(t) includes the recommended target flow rate for each injection sub-zone, the recommended direction of increase or decrease, and the corresponding status basis for this recommended value. The status basis includes the temperature window matching parameters, denitrification response parameters, ammonia slip risk parameters, injection execution deviation parameters, and injection priority score for that injection sub-zone. The above status basis is used by operators to view the reasons for control and also for tracing anomalies in subsequent control processes.
[0211] Step S5, Branch fluid coupling compensation:
[0212] After the partitioned collaborative optimization process outputs the suggested reducing agent injection amount Qp*(t) for each injection sub-region, and before generating the final reducing agent injection setpoint, a branch fluid coupling compensation step is further performed. This step is used to correct for main pipe pressure fluctuations and other branch actual flow deviations caused by changes in flow rate in a single branch when multiple spray gun branches share a reducing agent main pipe or atomizing air main pipe.
[0213] For SNCR systems where each spray gun path has an independent flow closed-loop control and the main pipe pressure is stable, this step is not enabled; for SNCR systems where the reducing agent main pipe has significant flow diversion and changes in the valve opening of one branch will affect the actual flow of other branches, this step is preferred.
[0214] The output of this step is the compensated recommended reducing agent injection quantity Qp,c(t), or the compensated target flow rate Qr,c(t) of the spray gun path, output via the DCS / PLC interface. This compensation result is still a recommended control quantity and needs to be processed with safety constraints before it can be output as the final reducing agent injection setpoint.
[0215] During the system commissioning phase, select a period when the boiler load, reducing agent main pipe pressure, atomizing air pressure, and spray gun line operation status are relatively stable. Apply small target flow disturbances to a single spray gun line in turn, while keeping the target flow of other spray gun lines unchanged, and record the changes in the actual flow of other spray gun lines.
[0216] Suppose the SNCR system includes R spray gun paths, and the target flow rate change in the r-th spray gun path is... The actual flow rate change of the j-th spray gun path is Then, the coupling coefficient between the change in the r-th spray gun path and the actual flow rate of the j-th spray gun path is... Represented as:
[0217]
[0218] ε is a very small positive number used to prevent the denominator from being zero. Cj,r is used to characterize the degree of influence of the target flow rate change of the r-th spray gun path on the actual flow rate of the j-th spray gun path. When j=r, This indicates the response relationship between the target flow change and the actual flow change of the branch; when j≠r, This indicates the flow perturbation relationship between different spray gun paths.
[0219] After performing disturbance tests on all spray gun paths sequentially, the branch coupling characteristic matrix C is formed:
[0220]
[0221] Establish corresponding branch coupling characteristic matrices according to different boiler load ranges, different reducing agent main pipe pressure ranges, or different nozzle circuit operation combinations. During operation, the control unit calls the corresponding matrix based on the current boiler load, main pipe pressure, and branch operation status; if the current operating condition differs significantly from the operating condition corresponding to the calibrated matrix, branch fluid coupling compensation is not activated, and the zonal collaborative optimization results are directly input into the safety constraint processing.
[0222] Since the output of the partitioned collaborative optimization process is the suggested reducing agent injection quantity Qp*(t) at the injection sub-region level, while the actual execution object of the DCS / PLC is usually the spray gun path, metering pump, or branch regulating valve, before performing coupling compensation, it is necessary to convert Qp*(t) into the suggested target flow rate at the spray gun path level according to the execution mapping weight between the injection sub-region and the spray gun path.
[0223] Let the execution mapping weight of the r-th spray gun path to the p-th spray sub-region be... The suggested target flow rate for the r-th spray gun path is... Represented as:
[0224]
[0225] P represents the number of spray sub-regions. When a spray gun path corresponds to only one spray sub-region, the spray gun path directly receives the suggested spray volume of the corresponding spray sub-region; when a spray gun path is associated with multiple spray sub-regions, a weighted transformation is performed according to the execution mapping weight.
[0226] During the t-th control cycle, the control unit suggests the target flow rate based on the current branch. Compared with the final target flow of the branch in the previous control cycle Calculate the recommended changes for each spray gun path:
[0227] The proposed changes for each branch are combined into a vector:
[0228]
[0229] Based on the branch coupling characteristic matrix C, predict the actual flow disturbance of each spray gun path caused by changes in other branches:
[0230] For the j-th spray gun path, the predicted perturbation is:
[0231]
[0232] When the predicted disturbance of a certain spray gun path exceeds the preset allowable disturbance threshold, it indicates that the change in the current recommended spray volume may cause the actual flow of the branch to deviate from the target value, and reverse compensation is required; when the predicted disturbance does not exceed the allowable disturbance threshold, no compensation is made for the branch.
[0233] For the j-th spray gun path that requires compensation, the control unit determines the target flow rate of the branch after reverse compensation based on the predicted disturbance. :
[0234]
[0235] Wherein, γj is the compensation coefficient of the j-th spray gun path, and its value is determined based on the flow control response speed, flow meter accuracy, valve adjustment characteristics, and on-site commissioning results of that branch. γj takes a value between 0 and 1; when the branch coupling effect is significant and the flow measurement is stable, γj takes a higher value; when the flow measurement noise is large or the branch control response is slow, γj takes a lower value.
[0236] If the target flow of the compensated branch If the flow rate exceeds the upper limit or falls below the lower limit of the allowable flow rate for that branch, it will be limited to the allowable range of that branch. No additional directional compensation will be applied to spray gun routes that are under maintenance, experiencing malfunctions, blockages, flow meter abnormalities, valve feedback abnormalities, or are prohibited from operation by the DCS / PLC.
[0237] After completing the reverse compensation, the compensation result is further corrected by combining the reducing agent main pipe pressure Pm(t) and the atomizing air pressure Pa(t).
[0238] When the pressure in the main reducing agent pipe is lower than the preset lower pressure limit, it indicates that the current liquid supply capacity or pressure reserve is insufficient. The control unit restricts multiple spray gun paths from increasing the spray volume at the same time, or reduces the compensated target flow rate of the corresponding branch of the low-priority spray sub-area. When the pressure in the main reducing agent pipe is higher than the preset upper pressure limit, the target flow rate of the branch is stopped from being increased, and the current state is sent to the subsequent safety constraint processing.
[0239] When the atomizing air pressure is lower than the atomization requirement of the spray gun, even if the reducing agent flow rate reaches the target value, it may lead to poor atomization and insufficient mixing. Therefore, the control unit limits the increase of the spray volume of the corresponding spray gun path, or marks the corresponding spray sub-area of the spray gun path as an execution restricted sub-area.
[0240] After branch disturbance prediction, reverse compensation, and correction for main pipe pressure and atomization conditions, the compensated target flow rate of the spray gun path is obtained. If the DCS / PLC interface uses the target flow rate of the spray gun path as input, then As input for subsequent safety constraint processing.
[0241] If the DCS / PLC interface uses the target value of the spray sub-region or the target value of the spray layer as input, then the compensated branch target flow rate will be recalculated back to the compensated recommended reducing agent spraying amount of the corresponding spray sub-region based on the execution mapping weight. :
[0242]
[0243] As input for subsequent safety constraint processing.
[0244] or This is still not the final execution value. It still needs to go through single branch limit, total limit, adjacent control cycle change rate limit, DCS / PLC interlock status verification and abnormal backoff judgment before the final reducing agent injection setting value can be formed.
[0245] The branch coupling characteristic matrix does not need to be re-established in every control cycle. An initial matrix is established during the system commissioning phase and periodically corrected based on effective disturbance samples during subsequent operation and maintenance. Effective disturbance samples refer to samples where the boiler load change is within the allowable range, the reducing agent main pipe pressure is controllable, the feedback from relevant flow meters and valves is effective, and the DCS / PLC is not in a manual intervention or interlocking restriction state during the disturbance period.
[0246] When the branch flow meter fails, the reducing agent main pipe pressure sensor fails, the branch valve feedback is abnormal, the spray gun branch activation / deactivation status changes frequently, or the current operating condition differs significantly from the operating condition corresponding to the calibrated branch coupling characteristic matrix, the control unit stops using the branch coupling compensation result and directly inputs the suggested reducing agent injection quantity Qp*(t) output by the partitioned collaborative optimization process into the subsequent safety constraint processing, and adopts a more conservative limit and rate of change limit in the safety constraint processing.
[0247] Through the above-mentioned branch fluid coupling compensation process, the injection quantity recommendation output by the partitioned collaborative optimization process can be further corrected by combining the main pipe diversion, the actual flow response of the branch, and the pressure conditions at the execution end. This reduces the impact of the actual flow deviation of other branches caused by the adjustment of a single branch, and improves the consistency between the final reducing agent injection setpoint and the actual injection quantity.
[0248] Step S6, Safety Hybrid Integration with DCS / PLC Output:
[0249] The output of this step is the final reducing agent injection setpoint for each injection sub-region. Alternatively, the final target flow rate of each spray gun path can be obtained by converting the DCS / PLC interface. .
[0250] The injection volume that enters the safety restraint process is uniformly recorded as... When branch fluid coupling compensation is enabled, Recommended reducing agent injection volume after compensation When branch fluid coupling compensation is not enabled, Qp,in(t) is the suggested reducing agent injection amount Qp*(t) output by the partitioned collaborative optimization process, i.e.: or:
[0251] It is still an intermediate output value of the control unit, and it needs to be processed by safety constraints to form the final reducing agent injection setting value.
[0252] The safety hybrid integration module first performs single-sub-region limiting processing on the input injection quantity Qp,in(t) of each injection sub-region to ensure that it does not exceed the equipment's allowable range for the corresponding injection sub-region.
[0253] Let p be the minimum allowable injection volume for the p-th injection sub-region. The maximum allowable injection volume for the p-th injection sub-region is... This is the injection volume after limiting the amplitude of a single injection sub-region.
[0254] Single-injection sub-region limiting is performed as follows:
[0255]
[0256] and The parameters are determined based on the design flow rate of the spray gun circuit, the adjustment range of the metering pump, the atomization requirements of the spray gun, the setting range of the DCS / PLC, and the permitted range of on-site operation.
[0257] When the corresponding spray gun circuit of a certain spray sub-zone is under maintenance, malfunction, prohibited from operation, or DCS / PLC interlock closed, the spray sub-zone The value is temporarily set to the current safe value or zero to prevent the control unit from continuing to allocate injection increments to this injection sub-region.
[0258] After limiting the spray volume of a single spray sub-region, the safety mixing integration module further determines whether the total spray volume of all spray sub-regions exceeds the upper limit of the total allowable reducing agent spray volume under the current operating conditions. The total constraint is expressed as:
[0259]
[0260] When the total jet volume after limiting the amplitude of each jet sub-region does not exceed When the total injection volume exceeds a certain threshold, the rate of change limit step is entered; when the total injection volume exceeds a certain threshold... At the same time, the safety hybrid integration module reduces the spray volume according to the spray priority of each spray sub-zone.
[0261] The injection volume of injection sub-regions with high temperature window matching parameters, high denitrification response parameters, low ammonia slip risk parameters, and low injection execution deviation parameters is retained; the injection volume of sub-regions with high ammonia slip risk parameters, large injection execution deviation parameters, and marked as restricted or execution-limited sub-regions is preferentially reduced. After total volume limiting processing, the total volume-constrained injection volume is obtained. .
[0262] To prevent drastic changes in injection volume between adjacent control cycles, which could cause fluctuations in furnace reaction, main pipe pressure, or frequent actuator actions, the safety hybrid integrated module... Apply a rate of change limit.
[0263] For the p-th injection sub-region, its rate of change constraint is expressed as:
[0264]
[0265] This is the final reducing agent injection setpoint for the p-th injection sub-region in the previous control cycle. This represents the maximum allowable variation in injection quantity for this injection sub-region within a single control cycle. This is the injection volume after the rate of change is limited.
[0266] when Greater than + At that time, Restricted to + ;
[0267] when Less than - At that time, Restricted to - ;
[0268] when When within the allowable range of variation, make = .
[0269] The pressure was determined based on the response speed of the spray gun circuit regulating valve, the regulating capacity of the metering pump, the stability of the reducing agent main pipe pressure, the DCS / PLC control cycle, and field operating experience. For operating conditions with a high risk of ammonia escape or significant fluctuations in the main pipe pressure, the pressure was temporarily reduced. This makes the injection volume adjustment more conservative.
[0270] After the amplitude and rate of change limits are completed, the safety hybrid integration module reads the interlock status and equipment status of the DCS / PLC control system. The interlock status and equipment status include at least some of the following: spray gun circuit activation / deactivation status, metering pump operation status, spray gun circuit regulating valve status, DCS / PLC manual / automatic status, reducing agent main pipe pressure alarm status, atomizing air pressure alarm status, boiler main protection status, and flue gas purification system interlock status.
[0271] When the DCS / PLC reports that a certain spraying sub-zone's corresponding spray gun circuit is in a prohibited operation, maintenance, fault, manual lockout, valve abnormality, or interlock closed state, the safety hybrid integrated module does not allow the final reducing agent injection setting value of that spraying sub-zone to be increased. For interlock states that require immediate cessation of spraying, the final reducing agent injection setting value of that spraying sub-zone is set to zero or the safety value specified by the DCS / PLC.
[0272] When the DCS / PLC is in a state where external optimization settings are not allowed to be written, the control unit stops outputting new optimization settings, and the DCS / PLC control system continues to use its original injection control logic or backup injection control curve.
[0273] The safety hybrid integrated module also verifies the reducing agent main pipe pressure Pm(t), atomizing air pressure Pa(t), CEMS data validity status, and control unit communication status.
[0274] When the pressure in the reducing agent main pipe is lower than the preset lower pressure limit, it indicates that the current liquid supply capacity or pressure reserve is insufficient. The safety mixing integrated module restricts multiple injection sub-zones from increasing the injection volume at the same time, or reduces the injection setting value of low priority injection sub-zones. When the pressure in the reducing agent main pipe is higher than the preset upper pressure limit, it prohibits further increasing the injection setting value and maintains the safety setting value of the previous control cycle or switches to standby control mode.
[0275] When the atomizing air pressure is lower than the atomization requirement of the spray gun, the safety mixing integrated module prohibits increasing the reducing agent injection setting value of the corresponding spray sub-zone, or marks the corresponding spray sub-zone as an execution restricted sub-zone.
[0276] When the CEMS is in calibration, backflushing, maintenance, fault, invalid data, or communication interruption state, the control unit will no longer generate new injection boosting settings based on the current NOx concentration or ammonia slip monitoring value. It will maintain the safety settings from the previous control cycle or switch to the backup injection control curve built into the DCS / PLC.
[0277] When the control unit communicates abnormally with the DCS / PLC, key input data is missing, the model calculation results exceed the safety boundary, or the recommended injection quantity is significantly inconsistent with the actual operating response in multiple consecutive control cycles, the safety hybrid integrated module stops using the recommended injection quantity output by the control unit and triggers abnormal rollback.
[0278] After single-injection sub-region limiting, total volume limiting, rate of change limiting, DCS / PLC interlock status verification, pressure and atomization condition verification, and data validity verification, if no abnormal rollback condition is triggered, the safety mixing integrated module generates the final reducing agent injection setpoint for the t-th control cycle:
[0279] If a certain injection sub-zone triggers the limit increase condition, the final reducing agent injection setting value of that injection sub-zone will not be higher than the final setting value of the previous control cycle, or will be limited to the safe range allowed by the DCS / PLC.
[0280] If a serious abnormal condition is triggered, such as DCS / PLC interlocking prohibition, spray gun circuit failure, abnormal reducing agent main pipe pressure, severely insufficient atomizing air pressure, invalid CEMS data and inability to make safety judgments, abnormal control unit communication, or model calculation results exceeding the safety boundary, the safety hybrid integrated module will stop outputting new optimized settings and switch the DCS / PLC control system to the backup injection control curve.
[0281] The backup injection control curve is an empirical control curve based on boiler load, flue gas oxygen content, outlet NOx concentration, or manually set ammonia injection rate in the existing DCS / PLC system of the incineration plant. The backup control mode does not rely on the zonal optimization results of the control unit and is used to ensure that the SNCR system can still operate according to the existing control logic when the control unit malfunctions or data conditions are insufficient.
[0282] The DCS / PLC control system directly receives the final reducing agent injection setpoint for each injection sub-zone. .
[0283] The DCS / PLC control system executes actions based on the target flow rate of the spray gun path. At this point, the execution mapping weights between the spray sub-region and the spray gun path are used to... Converted to the final target flow rate of each spray gun path:
[0284]
[0285] in, The mapping weight for the r-th spray gun path to the p-th spray sub-region.
[0286] After receiving the final setpoint, the DCS / PLC control system controls the metering pump, spray gun path regulating valve, atomizing air valve, or spray gun path deployment / retraction mechanism to perform spraying actions according to its existing control logic. During execution, the DCS / PLC control system retains its original manual control, automatic control, equipment protection, and safety interlock logic.
[0287] After the injection action is executed, the actual flow rate of the injection gun path, valve opening, reducing agent main pipe pressure, atomizing air pressure, injection gun path activation / deactivation status, equipment alarm status, NOx concentration, flue gas oxygen content, ammonia slip monitoring value, and furnace temperature change are used as feedback data to enter the next control cycle. Based on the new feedback data, the control unit re-performs data preprocessing, injection sub-region mapping, sub-region state vector construction, zone collaborative optimization, branch fluid coupling compensation, and safety constraint processing, thereby forming a closed-loop control.
[0288] Through safe hybrid integration and DCS / PLC output process, the optimization results of the control unit are converted into the final reducing agent injection setpoint that meets the field execution conditions and safety interlock requirements. This avoids the optimization results from directly affecting the actuator and causing safety risks, and also ensures that this method is compatible with the existing DCS / PLC control system of the municipal solid waste incinerator.
[0289] Example 1: A municipal solid waste incinerator is equipped with two layers of SNCR spray guns. Each layer is divided into three regions—left, middle, and right—according to the cross-sectional position of the furnace, forming a total of six spray sub-regions, denoted as Z1, Z2, Z3, Z4, Z5, and Z6. Z1, Z2, and Z3 are the lower spray sub-regions, and Z4, Z5, and Z6 are the upper spray sub-regions. Each spray sub-region corresponds to at least one spray gun path.
[0290] The control unit's optimization calculation cycle is 30 seconds. The preset effective reaction temperature window for SNCR is 850℃ to 1050℃, the preset optimal reaction temperature is 950℃, and the NOx control target value is 120 mg / Nm³. During the t-th control cycle, the control unit acquires data such as furnace temperature, outlet NOx concentration, flue gas oxygen content, boiler load, reducing agent main pipe pressure, atomizing air pressure, target flow rate, actual flow rate, and valve opening for each nozzle path. After time alignment, lag compensation, out-of-range value processing, and dimension unification, a preprocessed operating dataset D(t) is formed.
[0291] After mapping the injection sub-regions, the estimated base values of the sub-region temperature for each injection sub-region are shown in the table below:
[0292]
[0293] The control unit calculates the temperature window matching parameter M_p(t) based on the relationship between the estimated base value of the sub-zone temperature and the preset effective reaction temperature window of SNCR. Simultaneously, it calculates the denitrification response parameter E_p(t) based on the change in injection volume in each injection sub-zone within the sliding historical window and the change in NOx concentration after hysteresis compensation; it calculates the ammonia slip risk parameter R_p(t) based on the ammonia slip monitoring value, low temperature duration, cumulative injection volume, and NOx decreasing trend; and it calculates the injection execution deviation parameter D_p(t) based on the target flow rate, actual flow rate, valve opening, and reducing agent main pipe pressure of each injection nozzle path.
[0294] The state parameters of each injection sub-region during this control cycle are shown in the table below:
[0295]
[0296] Among them, the estimated base value of the temperature in sub-region Z6 is higher than the upper limit of the preset effective reaction temperature window for SNCR, its temperature window matching parameter is low, and its ammonia slip risk parameter is high. Therefore, it is marked as a restricted injection sub-region. Although Z4 does not exceed the effective reaction temperature window, its temperature window matching parameter is low and its ammonia slip risk parameter is high. Therefore, its injection allocation weight is reduced in the partitioned collaborative optimization process.
[0297] During this control cycle, the outlet NOx concentration after hysteresis compensation is higher than the NOx control target value, and the control unit determines that the reducing agent injection rate needs to be increased. The control unit determines the injection priority based on the temperature window matching parameters, denitrification response parameters, ammonia slip risk parameters, and injection execution deviation parameters for each injection sub-zone. Specifically, Z2 and Z5 have higher injection priority, followed by Z1, then Z3 and Z4, while Z6 does not participate in the injection increment allocation.
[0298] The final reducing agent injection setpoints for each injection sub-region in the previous control cycle are shown in the table below:
[0299]
[0300] Based on the NOx control deviation, the upper limit of the total injection quantity, and the injection priority of each injection sub-zone, the control unit obtains the recommended reducing agent injection quantity for the t-th control cycle as follows:
[0301]
[0302] Subsequently, it was recommended that the reducing agent injection volume enter the safe mixing integration step. After single injection sub-zone limiting, total volume limiting, adjacent control cycle change rate limiting, reducing agent main pipe pressure verification, atomizing air pressure verification, and DCS / PLC interlock status verification, no abnormal backoff conditions were triggered. The control unit outputs the corresponding final reducing agent injection setpoint to the DCS / PLC control system, which then controls the corresponding spray gun path to execute the injection action.
[0303] In this embodiment, Z2 and Z5 are given priority for injection increments due to their high temperature window matching and denitrification response capabilities; Z6 is not allocated injection increments because its temperature deviates from the effective reaction temperature window and its ammonia escape risk is high. This demonstrates that the present invention can perform differentiated injection control based on the actual state of each injection sub-region, rather than uniformly adjusting the total ammonia injection amount solely based on the outlet NOx concentration.
[0304] Example 2: In this example, the injection sub-region division method is the same as in Example 1. During the (t+k)th control cycle, the local temperature of the furnace decreases due to the increased moisture content of the incoming waste. After preprocessing and injection sub-region mapping, the estimated base values of the temperature of each injection sub-region are shown in the table below:
[0305]
[0306] Among them, Z1 and Z4 are below the lower limit of the preset effective reaction temperature window for SNCR. The state parameters of each injection sub-region calculated by the control unit are as follows:
[0307]
[0308] During this control cycle, Z1 and Z4 were marked as restricted injection sub-regions because their temperatures were below the effective reaction temperature window, and further increases in the reducing agent injection rate were not permitted. Although Z2's temperature was within the effective reaction temperature window, the target flow rate of its corresponding spray gun circuit R2 was 35 L / h, while the actual flow rate was 27 L / h, the valve opening had reached over 90%, and the reducing agent main pipe pressure was within the normal range. Based on this, the control unit determined that R2 had a local execution limitation and marked Z2 as an execution-limited sub-region.
[0309] The final reducing agent injection setpoints for each injection sub-region in the previous control cycle are shown in the table below:
[0310]
[0311] During this control cycle, the outlet NOx concentration after hysteresis compensation is higher than the NOx control target value, and the control unit determines that the reducing agent injection rate needs to be increased. If only the temperature window and NOx control deviation are considered, Z2 may continue to receive injection increments; however, since Z2 has been marked as an execution-restricted sub-region, continuing to increase its target injection rate cannot guarantee a synchronous increase in actual flow. Therefore, the control unit limits the increase in Z2's injection rate and transfers some of the injection demand to Z3 and Z5, which are adjacent to its flue gas flow influence area and are in normal condition.
[0312] After zoned collaborative optimization, the recommended reducing agent injection amount for each injection sub-region is as follows:
[0313]
[0314] Among them, Z1 and Z4 do not increase the injection volume, Z2 maintains the set value of the previous control cycle, Z3 and Z5 undertake the main injection increment, and Z6 increases slightly.
[0315] In this embodiment, multiple spray gun paths of the SNCR system share a common reducing agent manifold. To reduce the impact of inter-branch fluid coupling on the actual spray volume, the control unit further performs branch fluid coupling compensation. Based on the branch coupling characteristic matrix C established during the commissioning phase, the control unit converts the suggested reducing agent spray volume for each spray sub-region into the suggested target flow rate for the corresponding spray gun path, and calculates the change in the branch target flow rate relative to the previous control cycle.
[0316] Assume the following changes in the target flow rate of some spray gun routes during this control period:
[0317]
[0318] Based on the branch coupling characteristic matrix C, the control unit predicts that increasing the injection volume of R3 and R5 may cause a slight decrease in the actual flow rates of R1 and R2, with a predicted disturbance of -0.3 L / h for R1 and -0.4 L / h for R2. Since R2 is already in a restricted execution state, the control unit does not perform directional compensation for R2. For R1, assuming its branch state is normal and does not exceed the rate of change limit, reverse compensation is performed with a compensation coefficient of 0.7, resulting in a compensation of approximately +0.21 L / h for R1.
[0319] After branch fluid coupling compensation, the compensated branch target flow rate enters the safe mixing integration step. The safe mixing integration module performs single-branch limiting, total reducing agent limit, adjacent control cycle change rate limitation, reducing agent main pipe pressure verification, atomizing air pressure verification, and DCS / PLC interlock status verification on the compensation results. After verification, if the set values of each branch do not exceed the safety boundary, the control unit generates the final reducing agent injection set value and outputs it to the DCS / PLC control system.
[0320] If the temperatures of Z1 and Z4 recover to within the preset effective SNCR reaction temperature window in subsequent control cycles, and the ammonia escape risk parameter decreases, the control unit will release its restricted injection state; if the actual flow rate of the R2 branch recovers to normal and the injection execution deviation parameter is lower than the preset deviation threshold, the execution restriction state of Z2 will be released, allowing it to re-participate in normal injection quantity allocation.
[0321] This embodiment illustrates that, under the simultaneous presence of local low temperature, spray gun path execution deviation, and main pipe branch coupling, the present invention can identify restricted spray sub-regions and execution-limited sub-regions based on sub-region state vectors, transfer the spray demand to adjacent spray sub-regions with better state, and reduce the impact of inter-branch disturbances on the actual spray volume through branch fluid coupling compensation, thereby improving the feasibility of zonal spray control at the physical execution end.
[0322] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any ordinary changes and substitutions made by those skilled in the art within the scope of the technical solutions of the present invention should be included within the protection scope of the present invention.
Claims
1. A household garbage incinerator SNCR denitration control method, characterized by, Includes the following steps: S1. Obtain multi-source operational data of the SNCR denitrification process in the municipal solid waste incinerator. The multi-source operational data includes furnace temperature-related data, flue gas emission monitoring data, boiler operation process data, and reducing agent injection execution data. S2. Based on the arrangement of the SNCR spray gun in the furnace, the SNCR spray area is divided into multiple spray sub-areas, and a mapping relationship between multi-source operating data and each spray sub-area is established. S3. Based on the mapping relationship, the multi-source operation data is fused to obtain the sub-region state vector of each injection sub-region. The sub-region state vector includes at least the temperature window matching parameter, denitrification response parameter, ammonia escape risk parameter and injection execution deviation parameter. S4. Based on the state vector of each injection sub-region, determine the injection priority of each injection sub-region, and construct a zoned collaborative optimization process with the objectives of NOx emission control, ammonia slip risk control and reducing agent consumption control, to obtain the recommended reducing agent injection amount for each injection sub-region. S5. Apply safety constraints to the recommended reducing agent injection amount for each injection sub-zone to generate the final reducing agent injection setting value that meets the execution requirements of the DCS / PLC control system. S6. Output the final reducing agent injection setpoint to the DCS / PLC control system, which then controls the corresponding spray gun paths of each injection sub-zone to perform reducing agent injection.
2. The refuse incinerator SNCR denitration control method according to claim 1, characterized by: In step S2, multiple injection sub-zones are divided according to the elevation of the SNCR spray gun, the arrangement of the spray gun on the cross-section of the furnace, and the direction of the main flue gas flow. Each injection sub-region corresponds to at least one injection gun path, and the injection gun paths within the same injection sub-region have the same or adjacent furnace height position, the same or adjacent flue gas flow influence area, or similar historical injection response characteristics. When the spray coverage of a spray gun path spans more than two spray sub-regions, the spray gun path is assigned a branch affiliation weight to the corresponding spray sub-region based on the spray coverage ratio of the spray gun path to each spray sub-region or its historical spray response contribution.
3. The refuse incinerator SNCR denitration control method according to claim 2, characterized by: Multi-source operational data is preprocessed before being used for fusion processing; Preprocessing includes: aligning furnace temperature-related data, flue gas emission monitoring data, boiler operation process data, and reducing agent injection execution data according to a unified control cycle; Based on the transmission time of flue gas from the injection area to the flue gas emission monitoring point, lag compensation is applied to NOx concentration data and ammonia slip monitoring data. Bad values such as abrupt changes, distorted values, and long-term unchanged values are removed from furnace temperature-related data. The dimensions of boiler load, flue gas oxygen content, furnace negative pressure, reducing agent main pipe pressure, nozzle valve opening and actual nozzle flow rate are standardized or normalized.
4. The refuse incinerator SNCR denitration control method according to claim 3, characterized by: The mapping relationships include temperature mapping weights, emission response mapping weights, and execution mapping weights; Temperature mapping weights are used to characterize the contribution of furnace temperature-related data to the temperature estimation of each injection sub-region; The emission response mapping weight is used to characterize the hysteretic response contribution of changes in the reducing agent injection rate in each injection sub-region to changes in flue gas emission monitoring data; The execution mapping weight is used to characterize the contribution of the set injection quantity and actual injection quantity of the spray gun path to the injection state of the corresponding injection sub-region; The temperature mapping weight, emission response mapping weight, and execution mapping weight are determined based on at least two of the following: the physical location of the spray gun, the direction of flue gas flow in the furnace, the location of the temperature measuring point, the location of the flue gas emission monitoring point, and historical injection response data.
5. The refuse incinerator SNCR denitration control method according to claim 4, characterized by: The temperature window matching parameters are determined based on the relationship between the estimated temperature of the injection sub-region and the preset effective SNCR reaction temperature window. When the estimated temperature of the sub-region is within the preset effective reaction temperature window of SNCR, the temperature window matching parameter increases as the estimated temperature of the sub-region approaches the preset optimal reaction temperature. When the estimated temperature of the sub-region is lower than the lower limit of the preset effective SNCR reaction temperature window, or higher than the upper limit of the preset effective SNCR reaction temperature window, the temperature window matching parameter is reduced. The denitrification response parameters are determined based on the change in the amount of reducing agent injected in the corresponding injection sub-zone during the historical control period, as well as the change in NOx concentration after hysteresis compensation, and are corrected according to the boiler load, flue gas oxygen content, and sub-zone temperature change trends.
6. The SNCR denitrification control method for municipal solid waste incinerators according to claim 5, characterized in that: The ammonia slip risk parameters are determined based on the outlet ammonia slip monitoring value, the duration of the temperature in the injection sub-zone being lower than the preset SNCR effective reaction temperature window, the cumulative amount of reducing agent injected in the corresponding injection sub-zone, the cumulative amount of reducing agent injected in the upstream adjacent injection sub-zone, and the NOx concentration change trend. When the temperature window matching parameter of the injection sub-region is lower than the preset matching threshold, or the ammonia escape risk parameter is higher than the preset risk threshold, the injection sub-region is marked as a restricted injection sub-region. For restricted spray sub-regions, reduce their spray allocation weight, limit the increase in their spray volume, or stop increasing the amount of reducing agent sprayed into them during the partitioned collaborative optimization process.
7. The SNCR denitrification control method for municipal solid waste incinerators according to claim 6, characterized in that: The injection execution deviation parameter is determined based on the deviation between the target flow rate, actual flow rate, valve opening, and reducing agent main pipe pressure in the spray gun path; When the deviation between the target flow rate and the actual flow rate of the spray gun path exceeds the preset deviation threshold, the spray allocation weight of the corresponding spray sub-region of the spray gun path in the next control cycle is reduced. When the valve opening of the spray gun circuit reaches the upper limit and the actual flow rate is lower than the target flow rate, or when the valve opening of the spray gun circuit reaches the lower limit and the actual flow rate is higher than the target flow rate, the spray sub-region corresponding to the spray gun circuit is marked as an execution restricted sub-region. For restricted sub-regions, during the partitioned collaborative optimization process, the increase of the recommended reducing agent injection amount is limited, or the corresponding injection demand is transferred to an injection sub-region adjacent to the area affected by its flue gas flow.
8. The SNCR denitrification control method for municipal solid waste incinerators according to claim 7, characterized in that: The partitioned collaborative optimization process includes: The injection priority of each injection sub-zone is determined based on the temperature window matching parameters, denitrification response parameters, ammonia slip risk parameters, and injection execution deviation parameters of each injection sub-zone. Prioritize meeting NOx emission target constraints, and select the injection quantity combination with the lowest overall risk of ammonia slip and reducing agent consumption from the feasible injection quantity set that meets NOx emission target constraints; Among them, the injection allocation weight is increased for injection sub-regions with higher temperature window matching parameters and higher denitrification response parameters; Reduce the injection allocation weight for injection sub-regions with high ammonia slip risk parameters, large injection execution deviation parameters, or those marked as restricted injection sub-regions or execution-limited sub-regions; The feasible injection quantity set satisfies the upper and lower limits of the injection quantity of a single injection sub-zone, the upper limit of the total injection quantity of reducing agent, the upper limit of the injection quantity change rate between adjacent control cycles, and the DCS / PLC interlock permission status.
9. The SNCR denitrification control method for a municipal solid waste incinerator according to any one of claims 1 to 8, characterized in that: Before generating the final reducing agent injection setpoint, a branch fluid coupling compensation step is also included, which includes: During the commissioning phase, the valve opening or target flow rate of a single spray gun circuit is changed sequentially while keeping the set values of other spray gun circuits unchanged. The changes in the actual flow rate of other spray gun circuits and the pressure of the reducing agent main pipe are recorded to establish a branch coupling characteristic matrix. During operation, when the recommended reducing agent injection quantity of any spray gun path changes, the disturbance of this change to the actual flow of other spray gun paths is predicted based on the branch coupling characteristic matrix. Reverse compensation is performed on the recommended reducing agent injection amount or the final reducing agent injection set value for other spray gun routes based on the disturbance amount, so as to reduce the impact of reducing agent main pipe pressure fluctuation on the actual injection amount of each spray sub-zone; The injection setpoint after reverse compensation still needs to be limited by single branch amplitude, total amplitude and adjacent control cycle change rate before it is output to the DCS / PLC control system as the final reducing agent injection setpoint.
10. A non-linear catalytic cracking (SNCR) denitrification control system for a municipal solid waste incinerator, characterized in that: include: The data acquisition interface module is used to acquire multi-source operational data of the SNCR denitrification process of the municipal solid waste incinerator. The multi-source operational data includes furnace temperature-related data, flue gas emission monitoring data, boiler operation process data, and reducing agent injection execution data. The injection sub-region mapping module is used to divide the SNCR injection area into multiple injection sub-regions according to the arrangement of the SNCR nozzles in the furnace, and to establish a mapping relationship between multi-source operating data and each injection sub-region. The multi-source fusion state characterization module is used to fuse multi-source operating data based on mapping relationships to obtain the sub-region state vector of each injection sub-region. The sub-region state vector includes at least temperature window matching parameters, denitrification response parameters, ammonia escape risk parameters, and injection execution deviation parameters. The partitioned collaborative optimization control module is used to determine the injection priority of each injection sub-region based on the sub-region state vector of each injection sub-region, and to construct a partitioned collaborative optimization process with the objectives of NOx emission control, ammonia slip risk control and reductant consumption control, so as to obtain the recommended reductant injection amount for each injection sub-region. The safety hybrid integration module is used to perform safety constraint processing on the recommended reducing agent injection amount of each injection sub-zone, generate the final reducing agent injection setting value that meets the execution requirements of the DCS / PLC control system, and output the final reducing agent injection setting value to the DCS / PLC control system. The branch fluid coupling compensation module is used to perform feedforward compensation on the suggested reducing agent injection amount or the final reducing agent injection set value of each injection sub-region based on the pre-established branch coupling characteristic matrix. The DCS / PLC control system is used to control the corresponding nozzle path of each injection sub-zone to perform reducing agent injection according to the final reducing agent injection set value, and to switch to the standby injection control curve based on boiler load and flue gas oxygen content when abnormal retreat conditions are triggered.
Citation Information
Patent Citations
Denitration system combining SNCR with low-temperature SCR and used for household garbage incinerator
CN104056541A
Dynamic collaborative optimization control method for denitrification in waste incinerator
CN120361695B
