Combustion optimization adjustment method and system for high-alkali coal boiler

Abstract

The invention relates to the field of combustion control, and discloses a combustion optimization adjustment method and system for a high-alkali coal boiler, and the method comprises the steps: collecting multi-source heterogeneous data through a monitoring module, and building a high-dimensional state vector matrix through an intelligent control module according to the multi-source heterogeneous data; an oxygen and temperature sliding mode controller is used for adjusting a novel oxygen and temperature wide-range adjustable combustor and a circulating clean flue gas multi-point injection device, and steady-state combustion optimization is executed; a heating surface cleaning factor is calculated, when the heating surface cleaning factor is lower than an ash deposition critical threshold value, combustion pre-compensation reinforcement is executed firstly through dynamic stabilization cooperative control logic, and a cleaning execution mechanism is driven to act after it is confirmed that the combustion pre-compensation reinforcement is stable; alkali metal phase change and dew point tracking logic are utilized to regulate and control a phase change conditioning device to induce alkali metal condensation, and a multi-target model prediction control model is utilized to drive an ammonia spraying device to perform denitration. The problems that high-alkali coal combustion contamination is serious and pollutant emission is difficult to cooperatively control are solved, and the operation stability of a unit is improved.

Description

Technical Field

[0001] This invention relates to the field of combustion control, specifically to a method and system for optimizing and adjusting the combustion of a high-alkali coal boiler. Background Technology

[0002] High-alkali coal, as an important fossil energy source, has attracted much attention due to its abundant reserves and low mining costs. However, the high content of alkali metal elements such as sodium and potassium in its fuel characteristics poses a severe challenge to the safe and stable operation of boilers. In practical industrial applications, the combustion process of high-alkali coal boilers involves complex gas-solid two-phase flow and multi-component chemical reactions, exhibiting nonlinear, strongly coupled, and large delay characteristics.

[0003] Existing combustion control systems for high-alkali coal boilers largely rely on single-loop feedback regulation using limited discrete measuring points, making it difficult to reconstruct and grasp the combustion state and temperature field distribution throughout the furnace in real time. Conventional PID control strategies struggle to achieve precise decoupling control under dynamic load conditions. Consequently, when attempting to suppress nitrogen oxide formation through low-oxygen combustion, they often fail to effectively control the flue gas temperature at the furnace outlet. This leads to alkali metal gasification induced in localized high-temperature zones within the furnace, resulting in supersaturated precipitation and adhesion on heating surfaces, exacerbating coking and fouling problems.

[0004] As the fouling level on the heated surfaces increases, frequent cleaning operations such as soot blowing are necessary to maintain heat exchange efficiency. However, existing cleaning strategies typically lack quantitative characterization of the fouling status of the heated surfaces, often employing open-loop control methods based on time or differential pressure. More critically, the actions of the cleaning actuators cause severe disturbances to the aerodynamic field inside the furnace, and existing control logic lacks mechanisms for predicting and compensating for combustion stability during cleaning operations. During this process, airflow impacts can easily cause significant fluctuations in furnace negative pressure and flickering flame intensity. Especially under low load or unstable combustion conditions, hastily executing cleaning operations poses a safety hazard of combustion instability or even boiler flameout.

[0005] Furthermore, in the treatment of pollutants in the tail-end flue gas, high concentrations of gaseous alkali metals flow with the flue gas into downstream equipment. These metals are highly susceptible to physical deposition or chemical reactions on the catalyst surface of the selective catalytic reduction (SCR) denitrification system, occupying active sites and leading to catalyst poisoning and deactivation, shortening catalyst life and reducing denitrification efficiency. Simultaneously, the denitrification reaction process itself has significant inertia and lag. Simply relying on feedback control based on the outlet nitrogen oxide concentration cannot respond promptly to inlet concentration fluctuations caused by changes in combustion-side operating conditions, resulting in delayed ammonia injection adjustment and causing problems such as excessive nitrogen oxide emissions or persistently high ammonia slip rates. Summary of the Invention

[0006] The first aspect of this invention provides a combustion optimization and adjustment method for a high-alkali coal boiler, which aims to solve the technical problems of severe fouling of the heating surface, unstable combustion during the cleaning process, and difficulty in coordinating the control of pollutant emissions during the combustion of high-alkali coal.

[0007] The method includes the following steps: First, a comprehensive perception and state reconstruction of multi-source heterogeneous data is performed. The monitoring module collects operational data from the high-temperature zone of the boiler furnace, the heating surfaces, and the tail flue. The intelligent control module processes the collected data using the data acquisition and preprocessing unit to establish a high-dimensional state vector matrix characterizing the boiler's operating state. Specifically, this process involves fusing data acquired by a laser diagnostic device and an auto-emission spectroscopy analysis device to reconstruct the gas phase sodium and potassium atom concentration distribution map of the furnace cross-section; reconstructing the temperature field distribution of the furnace outlet cross-section using an acoustic temperature measurement system; and simultaneously acquiring local heat flux density data of the heating surfaces and tube wall metal temperature data using a tubular heat flux meter and an N-type sheathed thermocouple. By performing spatiotemporal alignment processing on the above optical, acoustic, and thermal data, a digital mapping benchmark reflecting the combustion and fouling state inside the furnace is constructed.

[0008] Secondly, steady-state combustion optimization based on sliding mode is implemented. The intelligent control module utilizes an oxygen and temperature sliding mode controller to extract average furnace oxygen content and furnace outlet temperature data from a high-dimensional state vector matrix. Based on these two feedback variables, the system adjusts the novel oxygen and temperature wide-range adjustable burner and the circulating clean flue gas multi-point injection device in the combustion control module. The core of this step lies in changing the air distribution ratio and swirl intensity by adjusting the damper of the external secondary air duct and the axially moving swirl generator in the internal secondary air duct, thereby constructing a central recirculation zone to stabilize combustion; simultaneously, the circulating clean flue gas multi-point injection device is adjusted according to temperature feedback, injecting circulating clean flue gas into the upper part of the burnout zone and the wall-adhering protective layer, using physical dilution and heat absorption to lower the furnace outlet flue gas temperature, and forming a low-temperature gas insulation film on the water-cooled wall surface, thereby suppressing the generation of nitrogen oxides while preventing alkali metal gasification and adhesion.

[0009] Secondly, dynamic collaborative cleaning is performed under ash accumulation conditions. The intelligent control module, based on data from the high-dimensional state vector matrix, uses a quantitative characterization model of the heating surface's contamination characteristics to calculate the heating surface cleanliness factor in real time. This calculation process determines the actual heat absorption of the heating surface based on the working fluid mass flow rate and the specific enthalpy difference between the working fluid inlet and outlet, and determines the theoretical heat absorption based on the Stefan-Boltzmann law and the geometric radiation characteristics of the heating surface. The ratio of the actual heat absorption to the theoretical heat absorption is used as the heating surface cleanliness factor. When the heating surface cleanliness factor falls below the preset ash accumulation critical threshold for different heating surface areas, the dynamic mitigation collaborative control logic is triggered. This logic does not directly trigger the cleaning action, but first switches the control target to a stable combustion and anti-disturbance mode, increasing the swirling intensity by advancing the burner's swirling generator and increasing the primary air rigidity, implementing combustion pre-compensation reinforcement. Only after monitoring that the furnace negative pressure fluctuation rate and flame intensity flicker frequency meet the stability indicators is the cleaning actuator driven to perform cleaning operations on specific areas, and the low-oxygen combustion parameters are restored after cleaning.

[0010] Finally, deep, tiered purification of tail-end pollutants is performed. The intelligent control module utilizes an alkali metal phase change and dew point tracking logic unit to calculate the critical temperature for saturation precipitation of gaseous alkali metals based on the gaseous alkali metal concentration and flue gas component partial pressure in a high-dimensional state vector matrix, and sets a phase change conditioning target temperature below this critical temperature. The system then adjusts the cooling medium flow rate of the tubular cryogenic heat exchanger array in the phase change conditioning device to reduce the flue gas temperature to the target range, inducing heterogeneous condensation of gaseous alkali metals. Subsequently, the flue gas enters a particle agglomeration device. Under the synergistic effect of a non-uniform high-voltage electric field and high-intensity sound waves, the fine particles carrying alkali metal condensates become charged and collide, forming large-diameter agglomerates that are captured by an electrostatic precipitator. In the denitrification process, a multi-objective model predictive control model is used, with the average oxygen content in the furnace and the furnace outlet temperature as feedforward disturbance variables, to predict the dynamic response trajectory of nitrogen oxide concentration in the future time domain. The optimal ammonia injection flow control increment sequence is solved by a rolling optimization algorithm under the conditions of satisfying the upper limit constraint of ammonia escape and the valve opening constraint, and the partitioned grid-type injection array of the ammonia injection device is driven to perform on-demand ammonia injection.

[0011] A second aspect of the present invention provides a combustion optimization and adjustment system for a high-alkali coal boiler, the system being configured to perform the method described in the first aspect above.

[0012] The system includes a monitoring module, a combustion regulation module, a pollutant control module, and an intelligent control module.

[0013] The monitoring module is configured to collect furnace and flue parameters, and specifically includes a laser diagnostic device, an auto-emission spectroscopy analysis device, an acoustic temperature measurement system, a tubular heat flow meter, and an N-type armored thermocouple.

[0014] The combustion control module is configured to regulate the combustion conditions inside the furnace, and includes a new type of oxygen and temperature wide-range adjustable burner and a circulating clean flue gas multi-point injection device. The new type of oxygen and temperature wide-range adjustable burner has an adjustable internal secondary air swirl generator and an external secondary air damper; the circulating clean flue gas multi-point injection device has a multi-stage nozzle group connected to the upper part of the burnout zone and the wall-mounted protective layer.

[0015] The pollutant control module is configured to treat flue gas pollutants, and it includes, in sequence along the flue gas flow direction, a phase change conditioning device, a particle agglomeration device, an electrostatic precipitator, an ammonia injection device, and a selective catalytic reduction denitrification system.

[0016] The intelligent control module includes a data acquisition and preprocessing unit and a core computing and processing unit. The core computing and processing unit internally runs an aerobic and temperature sliding mode controller, a quantitative characterization model of the fouling characteristics of the heated surface, a multi-dimensional ash and dirt removal decision model, a dynamic suppression and collaborative control logic unit, an alkali metal phase change and dew point tracking logic unit, and a multi-objective model predictive control model, which are used to achieve closed-loop optimization control of the entire process.

[0017] This invention provides a method and system for optimizing combustion in a high-alkali coal boiler. It offers the following advantages: 1. This invention achieves full-domain state perception by constructing a high-dimensional state vector matrix and utilizes an oxygen and temperature sliding mode controller to adjust a novel wide-range adjustable burner and a multi-point injection device for circulating clean flue gas. This method overcomes the nonlinearity and large delay characteristics of high-alkali coal combustion by using a sliding mode variable structure control law. While maintaining a low excess air coefficient to suppress nitrogen oxide generation, it lowers the furnace outlet flue gas temperature through the physical dilution and endothermic effect of circulating flue gas, thereby suppressing the supersaturation precipitation of gaseous alkali metals and fouling of heating surfaces at the source, achieving a synergistic improvement in combustion stability and anti-coking performance.

[0018] 2. This invention utilizes a quantitative characterization model of the fouling characteristics of heated surfaces to calculate the cleanliness factor in real time and establishes a cleaning mechanism based on dynamic suppression and collaborative control logic. Before performing cleaning operations, the system prioritizes the implementation of a combustion pre-compensation reinforcement strategy. By increasing the swirl intensity and primary air rigidity to construct a stable combustion core, it solves the problem of combustion instability or flameout caused by load disturbances and airflow impacts during ash removal in high-alkali coal units. This ensures that while maintaining the cleanliness of the heated surfaces, the furnace negative pressure fluctuation rate and flame intensity flicker frequency remain within the safe operating range.

[0019] 3. This invention combines alkali metal phase change and dew point tracking logic with multi-objective model predictive control technology to achieve tiered deep purification of tail-end pollutants. Phase change conditioning induces heterogeneous condensation of gas-phase alkali metals, and a particle agglomeration device promotes the growth of fine particles, significantly reducing the alkali metal load entering the denitrification system and preventing catalyst poisoning and deactivation. Simultaneously, model predictive control utilizes combustion-side feedforward signals to compensate for the large inertial lag of the denitrification system in advance, achieving ultra-low nitrogen oxide emissions and low ammonia escape control through precise on-demand ammonia injection. Attached Figure Description

[0020] Figure 1 This is a schematic diagram illustrating the overall architecture and core algorithm interaction of the high-alkali coal boiler combustion optimization and adjustment system of the present invention; Figure 2 This is a flowchart illustrating the overall implementation of the combustion optimization and adjustment method of the present invention. Figure 3 This is a detailed diagram of the dynamic suppression and collaborative control logic of the present invention. Detailed Implementation

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see the appendix Figure 1 This invention provides a combustion optimization and adjustment method and system for a high-alkali coal boiler. The system mainly includes a monitoring module, a combustion regulation module, a pollutant control module, and an intelligent control module.

[0023] The monitoring module is physically divided into a high-temperature zone monitoring subset in the furnace, a heat exchange monitoring subset on the heating surface, and a tail flue monitoring subset.

[0024] The high-temperature zone monitoring subset of the furnace includes a laser diagnostic device and an autoemission spectroscopy analysis device. The laser diagnostic device's transmitting and receiving units are installed in pairs on the boiler furnace sidewalls, forming a multi-path intersecting optical network, configured to monitor the concentrations of gaseous sodium and potassium atoms based on tunable semiconductor laser absorption spectroscopy. The autoemission spectroscopy analysis device acquires the flame's spontaneous emission spectrum through a high-temperature resistant fiber optic probe with a coaxial sealed air jacket. The data from both are fused to reconstruct a two-dimensional distribution field of gaseous alkali metal concentrations within the furnace. The heat transfer monitoring subset for the heating surfaces includes tubular heat flow meters and type N-sheathed thermocouples. The tubular heat flow meters and type N-sheathed thermocouples are arranged in pairs at the same physical coordinates on the water-cooled walls, the screen-type superheater, and the high-temperature reheater, configured to simultaneously acquire local heat load data and metal wall temperature data. The tail flue monitoring subset includes a multi-point matrix zirconia oxygen analyzer located at the economizer outlet, an acoustic temperature measurement system located at the furnace outlet plane, and flow rate monitors, nitrogen oxide concentration monitors, and particulate matter quantitative monitoring devices located before and after the denitrification system.

[0025] The combustion control module includes a novel oxygen and temperature wide-range adjustable burner and a circulating clean flue gas multi-point injection device. The novel oxygen and temperature wide-range adjustable burner adopts a coaxial sleeve structure, including a central primary air duct, an inner secondary air channel, an outer secondary air channel, and a sliding expansion cone adjustment mechanism. A corrugated stabilizing blister is installed at the end of the central primary air duct. An axially moving swirl generator driven by a mechanical linkage is installed in the inner secondary air channel, configured to continuously adjust the swirl intensity. The novel oxygen and temperature wide-range adjustable burner is configured to achieve wide-range adjustment of the flame pattern by independently adjusting the air ratio and swirl intensity of the inner and outer secondary air channels. The circulating clean flue gas multi-point injection device includes a high-temperature recirculation booster fan, a main delivery header, and a furnace multi-stage injection nozzle assembly. The furnace multi-stage injection nozzle assembly is spatially divided into: a burner mixing injection layer connected to the burner air distribution box, a wall-mounted protective injection layer arranged tangentially along the water-cooled wall, and an upper injection layer in the burnout zone located above the main burner. The circulating clean flue gas multi-point injection device is configured to independently control the flow rate of circulating clean flue gas at each level through electric regulating valves in each branch, thereby creating an aerodynamic field that includes a low-temperature oxidation atmosphere near the wall and a high-temperature combustion atmosphere in the center.

[0026] The pollutant control module, arranged sequentially along the flue gas flow direction, includes a phase change conditioning unit, a particle agglomeration unit, an electrostatic precipitator, an ammonia injection unit, and a selective catalytic reduction (SCR) denitrification system. The phase change conditioning unit employs a tubular low-temperature heat exchanger array, configured to cool the flue gas temperature below the dew point temperature corresponding to the saturated vapor pressure of alkali metals. The particle agglomeration unit integrates a high-pressure electrostatic agglomeration unit and an acoustic excitation unit, configured to promote the agglomeration of fine particles and condensed alkali metals. The electrostatic precipitator is configured to capture large-diameter agglomerates rich in alkali metals. The ammonia injection unit uses a zoned grid-type injection array, configured to inject ammonia gas as needed according to the flow field distribution. The SCR denitrification system uses an alkali-resistant catalyst, configured to efficiently remove nitrogen oxides under low-alkali load conditions.

[0027] The intelligent control module includes a data acquisition and preprocessing unit, a core computing and processing unit, a storage unit, and an instruction output interface unit.

[0028] The core computing and processing unit runs a quantitative characterization model of the fouling characteristics of the heated surface, a multi-dimensional ash and dirt removal decision model, a dynamic suppression and collaborative control logic unit, an oxygen and temperature sliding mode controller, and a multi-objective model predictive control model.

[0029] The quantitative characterization model for the fouling characteristics of the heated surface is configured to calculate the cleanliness factor of the heated surface in real time based on the working fluid mass flow rate, inlet and outlet specific enthalpy of the working fluid, Stefan-Boltzmann constant, heat transfer area of ​​the heated surface, comprehensive correction coefficient, average absolute temperature on the flue gas side, and average absolute temperature on the outer surface of the tube wall metal. The multidimensional ash removal decision model is configured to generate a cleaning request signal when the cleanliness factor of the heated surface is lower than the critical threshold for ash accumulation. The dynamic suppression and collaborative control logic unit is configured to respond to the cleaning request signal by first sending a combustion pre-compensation reinforcement command to the oxygen and temperature sliding mode controller. After confirming that the combustion stability index meets the requirements, the cleaning actuator is triggered through the command output interface unit. The oxygen and temperature sliding mode controller is configured to calculate the control command for the new oxygen and temperature wide-range adjustable burner and the circulating clean flue gas multi-point injection device based on the furnace average oxygen deviation and furnace outlet temperature deviation. The multi-objective model predictive control model is configured to calculate the optimal ammonia injection flow rate control increment for the ammonia injection device based on the nitrogen oxide generation and reference trajectory in the prediction time domain.

[0030] A quantitative characterization model for the fouling characteristics of the heated surface is embedded in the core computing unit of the intelligent control module. This model is configured based on the principle of solving the inverse problem of heat transfer, converting multi-source heterogeneous data collected by the monitoring module into a dimensionless index characterizing the physical cleanliness of the heated surface. The data input terminal of the model is connected to the data acquisition and preprocessing unit, configured to read working fluid-side parameters and flue gas-side parameters at a fixed sampling period. The working fluid-side parameters include the working fluid mass flow rate, inlet specific enthalpy, and outlet specific enthalpy of the working fluid flowing through a specific heated surface tube assembly. The flue gas-side parameters include the average absolute temperature of the flue gas in the heated surface region and the average absolute temperature of the outer surface of the tube wall metal.

[0031] The quantitative characterization model for the fouling characteristics of the heating surface first calculates the actual heat absorption of the heating surface. The actual heat absorption is obtained based on the law of conservation of energy by calculating the enthalpy increase of the working fluid inside the heating surface tubes. The model then uses the real-time monitored working fluid mass flow rate multiplied by the difference between the working fluid outlet specific enthalpy and the working fluid inlet specific enthalpy to obtain the actual heat power absorbed by the heating surface from the flue gas under the current operating conditions.

[0032] The quantitative characterization model of the fouling characteristics of the heating surface was then used to calculate the theoretical heat absorption of the heating surface under ideal clean conditions. Given that heat transfer in the furnace and high-temperature heating surface area of ​​a high-alkali coal boiler is primarily radiative, the Stefan-Boltzmann law was used as the basis for the theoretical heat transfer calculation in the quantitative characterization model of the fouling characteristics of the heating surface. The theoretical heat absorption represents the maximum radiative heat transfer capacity achievable by the heating surface under the assumptions of no fouling on the tube wall surface, inherent metallic emissivity, and a constant geometric angle coefficient.

[0033] The quantitative characterization model for the fouling characteristics of heated surfaces defines a cleanliness factor by constructing the ratio of actual heat absorption to theoretical heat absorption. The cleanliness factor directly reflects the degree to which the thermal resistance of fouling affects heat transfer efficiency. The specific calculation formulas used in the quantitative characterization model for the fouling characteristics of heated surfaces are as follows: ; in, Defined as the heat-receiving surface cleanliness factor; Defined as the mass flow rate of the working fluid flowing through the heated surface; Defined as the specific enthalpy of the working fluid outlet; Defined as the specific enthalpy of the working fluid inlet; Defined as the Stefan Boltzmann constant; Defined as the heat transfer area of ​​the heated surface; Defined as the comprehensive correction factor, the comprehensive correction factor is used to characterize the influence of geometry and radiation properties; Defined as the average absolute temperature on the flue gas side; Defined as the average absolute temperature of the outer surface of the pipe wall metal. The intelligent control module is configured to generate a cleaning request signal when the cleanliness factor of the heated surface is less than 0.35.

[0034] The quantitative characterization model for the fouling characteristics of heated surfaces is configured to output a real-time numerical sequence of the cleanliness factor of the heated surfaces. When alkali metal ash from the combustion of high-alkali coal deposits on the surface of the heated surface tubes, the thermal conductivity of the ash layer is much lower than that of the metal tube wall, leading to an increase in additional thermal resistance. This results in a reduction in the actual heat transferred to the working fluid under the same flue gas temperature and wall temperature difference. The quantitative characterization model for the fouling characteristics of heated surfaces can sensitively capture this decrease in heat flux density, which is reflected in the monotonically decreasing trend of the calculated cleanliness factor value. The quantitative characterization model for the fouling characteristics of heated surfaces sends the calculation results to the multidimensional ash removal decision model in real time, serving as the sole quantitative basis for triggering the cleaning mechanism.

[0035] The oxygen and temperature sliding mode controller, as a key algorithm component in the intelligent control module, is configured to address the inherent strong nonlinearity, large delay, and model uncertainty issues in the combustion process of high-alkali coal. By constructing a control law with variable structure characteristics, the oxygen and temperature sliding mode controller forces the state trajectory of the combustion system to reach and maintain a preset sliding surface within a finite time, thereby achieving robust control of the furnace oxygen content and furnace outlet temperature.

[0036] The input of the oxygen and temperature sliding mode controller is connected to the data acquisition and preprocessing unit, configured to read the furnace average oxygen feedback value and furnace outlet temperature feedback value in real time. The oxygen and temperature sliding mode controller has preset furnace average oxygen setpoints and furnace outlet temperature setpoints corresponding to the current unit load command. The oxygen and temperature sliding mode controller first calculates the state error vector. The state error vector includes an oxygen error component and a temperature error component. The oxygen error component is defined as the furnace average oxygen setpoint minus the furnace average oxygen feedback value. The temperature error component is defined as the furnace outlet temperature setpoint minus the furnace outlet temperature feedback value.

[0037] The oxygen and temperature sliding mode controller designs a sliding surface function based on the dynamic characteristics of the error. The sliding surface function defines the dynamic trajectory of system error convergence. When the system state lies on the sliding surface, the system error will decay to zero according to a predetermined exponential law, and this process is invariant to external disturbances such as fluctuations in the quality of high-alkali coal. The formula for the sliding surface function used by the oxygen and temperature sliding mode controller is as follows: ; in, Defined as the sliding surface function value; Defined as the system tracking error vector, the system tracking error vector is the difference between the set value and the actual feedback value; Defined as the first derivative of the error vector with respect to time; Defined as a positive definite gain matrix. The intelligent control module is configured to calculate control commands based on the sliding mode surface function value, and maintain the system state on the sliding mode surface by adjusting the new type of oxygen and temperature wide-range adjustable burner and the circulating clean flue gas multi-point injection device.

[0038] The oxygen and temperature sliding mode controller is configured to calculate control outputs based on the calculated sliding surface function value vector. These control outputs include damper opening correction commands for the new wide-range adjustable oxygen and temperature burner and flow valve opening correction commands for the circulating clean flue gas multi-point injection device.

[0039] The oxygen and temperature sliding mode controller employs an exponential reaching law algorithm to design the control law, ensuring that the system state can rapidly approach and traverse the sliding surface from any initial position. When the sliding surface function value vector is not zero, the oxygen and temperature sliding mode controller outputs a control signal with switching characteristics. The sign of this control signal is opposite to the sign of the sliding surface function value vector, thereby generating a restoring force that drives the system state to move towards the sliding surface.

[0040] Specifically, when the oxygen content error component causes the corresponding sliding mode variable to deviate from zero, the oxygen and temperature sliding mode controller adjusts the air distribution ratio of the inner and outer secondary air channels of the new oxygen and temperature wide-range adjustable burner, as well as the position of the sliding expansion cone adjustment mechanism. By changing the swirl intensity and local oxygen concentration distribution, the oxygen and temperature sliding mode controller corrects the combustion chemical reaction rate within the furnace.

[0041] When the temperature error component causes the corresponding sliding mode variable to deviate from zero, the oxygen and temperature sliding mode controller adjusts the opening of the electrically controlled regulating valves in each branch of the circulating clean flue gas multi-point injection device. By changing the flow rate of the low-temperature inert gas injected into the furnace, the oxygen and temperature sliding mode controller adjusts the heat capacity and temperature gradient distribution within the furnace.

[0042] The oxygen and temperature sliding mode controller, through the above-mentioned high-frequency variable structure adjustment, can effectively suppress the combustion pulsation caused by the rapid precipitation of volatiles in high-alkali coal, strictly lock the flue gas temperature at the furnace outlet within the low-temperature window that suppresses alkali metal gasification, and at the same time maintain the excess air coefficient in the optimal range for low nitrogen oxide generation, ensuring that the control deviation is always less than 5%.

[0043] The multi-objective model predictive control operates within the core computational processing unit of the intelligent control module. This model is configured to overcome the control lag problem of the selective catalytic reduction denitrification system under conditions of high inertia and large delay, and to collaboratively handle sudden changes in nitrogen oxide (NOx) generation caused by fluctuations in high-alkali coal combustion. The multi-objective model predictive control includes a NOx generation prediction sub-model and a rolling optimization control sub-model.

[0044] The data input terminal of the nitrogen oxide (NOx) generation prediction sub-model is connected to the data acquisition and preprocessing unit. The NOx generation prediction sub-model is configured to read operating parameters from the combustion control module and status parameters from the pollutant control module. Operating parameters include the average oxygen content in the furnace, furnace outlet temperature, coal feed rate, and total air volume. Status parameters include the NOx concentration at the inlet of the selective catalytic reduction (SCR) denitrification system, the NOx concentration at the outlet, and the flue gas flow rate.

[0045] The nitrogen oxide (NOx) generation prediction sub-model is constructed based on a controlled autoregressive integral moving average algorithm to create a dynamic mathematical model of the controlled object. The NOx generation prediction sub-model is configured to use the changing trends of the furnace average oxygen content and furnace outlet temperature as feedforward disturbance variables, and the ammonia flow rate of the ammonia injection unit as the control input variable, to predict the dynamic response trajectory of the NOx concentration at the outlet of the selective catalytic reduction (SCR) denitrification system within a finite time domain. Due to the instability of high-alkali coal combustion leading to drastic fluctuations in furnace outlet NOx concentration, the NOx generation prediction sub-model, by introducing combustion-side parameters as feedforward signals, can calculate the impending NOx load change due to changes in combustion conditions before the inlet NOx concentration analyzer detects the change.

[0046] The rolling optimization control sub-model is configured based on the prediction results of the nitrogen oxide generation prediction sub-model. It determines the optimal sequence of control actions by solving a constrained quadratic programming problem. The objective of the rolling optimization control sub-model is to ensure that the system output in the future time domain closely tracks the desired ultra-low emission setpoint trajectory, while penalizing the severity of control actions to prevent actuator wear and system oscillations. The optimization objective function formula based on the rolling optimization control sub-model is as follows: in, Defined as a control performance index function; Defined as the prediction time domain length; Defined as the length of the control time domain; Defined as the predicted value of nitrogen oxide concentration at the current moment for future moments; Defined as a set reference trajectory for nitrogen oxide emission concentration; Defined as the control increment of ammonia injection flow rate at a future time; Defined as the output error weighting matrix; Defined as a control increment weighted matrix. The intelligent control module is configured to solve for the control increment sequence that minimizes the control performance index function and send this sequence as a command to the ammonia injection unit.

[0047] The rolling optimization control sub-model is configured to, within each control cycle, solve online for the minimum value of the aforementioned objective function while satisfying the upper limit constraint of ammonia escape and the ammonia injection valve opening constraint, based on the latest measured and predicted values, thereby obtaining a future optimal ammonia injection flow control increment sequence. The rolling optimization control sub-model sends only the first element of this sequence as the actual control command to the command output interface unit, driving the high-precision flow regulating valve in the ammonia injection unit. At the next sampling time, the multi-objective model predictive control model repeats the above prediction and optimization process based on the new state measurements, achieving rolling time-domain control.

[0048] The alkali metal phase transition and dew point tracking logic unit is embedded in the core computing processing unit of the intelligent control module. The alkali metal phase transition and dew point tracking logic unit is configured to dynamically calculate the saturation precipitation temperature of gaseous alkali metal compounds and adjust the flue gas cooling depth according to the saturation precipitation temperature, so as to use the phase transition pressure difference to drive the directional migration of gaseous alkali metals to the surface of solid particles.

[0049] The data input of the alkali metal phase transition and dew point tracking logic unit is connected to the data acquisition and preprocessing unit. The alkali metal phase transition and dew point tracking logic unit is configured to read gaseous sodium atom concentration data and gaseous potassium atom concentration data provided by the furnace high-temperature zone monitoring subset, as well as flue gas static pressure data provided by the tail flue monitoring subset, at millisecond-level cycles.

[0050] The alkali metal phase transition and dew point tracking logic unit internally stores a thermodynamic vapor-liquid equilibrium database of alkali metal compounds. This database includes saturated vapor pressure characteristic curves for sodium sulfate, potassium sulfate, sodium chloride, and potassium chloride at different temperatures. The alkali metal phase transition and dew point tracking logic unit first derives the theoretical partial pressure values ​​of the main gaseous alkali metal compounds under the current flue gas conditions through chemical equilibrium calculations, based on real-time acquired gaseous sodium and potassium atom concentration data, combined with the sulfur oxide and chlorine content in the flue gas components.

[0051] The alkali metal phase transition and dew point tracking logic unit then performs inverse dew point optimization calculations. It substitutes the calculated theoretical partial pressures of each component into the thermodynamic vapor-liquid equilibrium database and retrieves the corresponding saturation condensation temperature for each component. The alkali metal phase transition and dew point tracking logic unit is configured to compare the saturation condensation temperatures of all components, select the highest value, and define this highest value as the current critical phase transition dew point temperature.

[0052] The alkali metal phase transition and dew point tracking logic unit is configured to set the phase transition conditioning target temperature based on the critical phase transition dew point temperature. The phase transition conditioning target temperature is set to a temperature range of 15°C to 30°C below the critical phase transition dew point temperature, which is used to create a supersaturated environment in the flue gas. The thermodynamic potential energy driving force generated by this supersaturated environment is configured to force gaseous alkali metal compounds to undergo heterogeneous nucleation using fly ash particles present in the flue gas as condensation nuclei, or to directly adsorb and deposit on the surface of existing fly ash particles.

[0053] The alkali metal phase change and dew point tracking logic unit compares the target temperature of the phase change tempering with the actual flue gas temperature at the outlet of the phase change tempering device. When the actual flue gas temperature is higher than the target temperature, the alkali metal phase change and dew point tracking logic unit generates a cooling load increase command. The cooling load increase command is transmitted through the command output interface unit and converted into a physical control signal that drives the opening of the cooling medium flow regulating valve in the phase change tempering device to increase. By continuously tracking the fluctuation of the alkali metal concentration in the furnace and adjusting the tail flue temperature in real time, the alkali metal phase change and dew point tracking logic unit ensures that the gaseous alkali metal has completed the phase change transformation from gaseous to solid or liquid aerosol state before the flue gas enters the particle agglomeration device.

[0054] Please see the appendix Figure 2 The present invention provides a combustion optimization and adjustment method for a high-alkali coal boiler, comprising the following steps: S1: Real-time monitoring and reconstruction of multiple field parameters specifically includes the following processes: The laser diagnostic device and the spontaneous emission spectroscopy analyzer in the monitoring module initiate a synchronous acquisition program to optically scan the gaseous alkali metal concentration field inside the furnace. The transmitting unit of the laser diagnostic device emits a modulated laser beam with wavelengths covering the characteristic absorption lines of sodium (589.0 nm) and potassium (766.5 nm) towards the core combustion region of the furnace. The receiving unit of the laser diagnostic device receives the transmission spectral signal after passing through the combustion flame and records the intensity attenuation data of the transmission spectral signal. Simultaneously, the spontaneous emission spectroscopy analyzer captures the spontaneous emission spectrum of the flame at a specific monitoring point in the furnace using a high-temperature resistant fiber optic probe.

[0055] The data acquisition and preprocessing unit in the intelligent control module reads the aforementioned transmission spectrum signal and spontaneous emission spectrum data, and executes a tomographic inversion algorithm. Using multipath integral concentration data provided by the laser diagnostic device as constraints and local point emission intensity data provided by the spontaneous emission spectroscopy analysis device as prior weights, the data acquisition and preprocessing unit reconstructs the gaseous sodium and potassium atom concentration distribution maps on the furnace cross-section using an algebraic iterative reconstruction algorithm. The spatial resolution of these concentration distribution maps is set to a grid accuracy of no less than 100 mm x 100 mm, and the refresh rate is set to no less than 10 Hz.

[0056] Simultaneously, the heat exchange monitoring subset of the monitoring module performs contact-type measurements of the thermal state of the heating surfaces. Tubular heat flux meters, located in the water-cooled walls, screen-type superheaters, and high-temperature reheater areas, output local heat flux density electrical signals in real time. N-type sheathed thermocouples, located at corresponding positions, output tube wall metal temperature electrical signals in real time. The data acquisition and preprocessing unit synchronously acquires the working fluid-side operating parameters from the boiler's distributed control system. These parameters include feedwater flow rate, main steam flow rate, and the inlet and outlet working fluid temperatures and pressures at each stage of the heating surface.

[0057] The tail flue monitoring subset in the monitoring module performs measurements of combustion products and flow field conditions. The acoustic temperature measurement system controls the acoustic transceiver array to transmit acoustic signals according to a preset timing sequence and measures the transit time of the acoustic signals across the furnace outlet section. Based on the differences in the propagation speed of acoustic waves in different temperature media, it calculates and outputs two-dimensional distribution data of the flue gas temperature field at the furnace outlet. A multi-point matrix zirconia oxygen analyzer collects oxygen content data of the flue gas at different grid locations on the economizer outlet section.

[0058] The data acquisition and preprocessing unit performs time-axis alignment and outlier removal on the heterogeneous data from optical non-contact sensors, thermal contact sensors, and acoustic sensors. The unit employs a moving average filtering algorithm to remove high-frequency random noise from the signals and maps all measurement data to the same standard timestamp sequence, forming a high-dimensional state vector matrix containing the alkali metal concentration field, temperature field, oxygen field, and the metal temperature field of the heated surface within the entire furnace. This high-dimensional state vector matrix is ​​written in real-time to the shared memory area of ​​the intelligent control module, serving as the real-time input data source for the heated surface fouling characteristic quantitative characterization model, the oxygen and temperature sliding mode controller, and the multi-objective model predictive control model.

[0059] S2: The steady-state combustion optimization based on sliding mode specifically includes the following processes: The oxygen and temperature sliding mode controller in the intelligent control module periodically reads the average oxygen content feedback value and the furnace outlet temperature feedback value from the data acquisition and preprocessing unit. The oxygen and temperature sliding mode controller compares the average oxygen content feedback value with a preset low-oxygen combustion setpoint to calculate the oxygen state error. It also compares the furnace outlet temperature feedback value with a preset alkali metal precipitation critical temperature setpoint to calculate the temperature state error. The preset low-oxygen combustion setpoint is typically set to an oxygen volume fraction range of 2.0% to 3.0%, and the preset alkali metal precipitation critical temperature setpoint is set below the ash melting point of high-alkali coal and the gasification temperature of alkali metals, typically between 1050°C and 1150°C.

[0060] The oxygen and temperature sliding mode controller substitutes the oxygen and temperature state errors into a preset sliding surface function formula to calculate the sliding surface function value vector at the current moment. Using an exponential reaching law algorithm, the oxygen and temperature sliding mode controller calculates the control quantity correction value for the combustion control module based on the sign and magnitude of the sliding surface function value vector. This control quantity correction value aims to drive the system state trajectory to converge towards the sliding surface on the phase plane and maintain it on the sliding surface, thereby achieving decoupled control of the combustion process.

[0061] For the oxygen control loop, when the oxygen and temperature sliding mode controller calculates a control command requiring a reduction in the furnace oxygen concentration, it sends an action signal to the new type of wide-range adjustable oxygen and temperature burner. Responding to this signal, the burner drives the damper in the external secondary air duct to close, reducing the flow rate of the high-speed direct-flow air jet. Simultaneously, the burner drives the axially moving swirl generator within the internal secondary air duct to move outwards from the furnace. This movement alters the tangential velocity component of the internal secondary air, enhancing the swirl intensity. This enhanced swirl intensity forms a central recirculation zone at the burner outlet, which entrains high-temperature flue gas to maintain the ignition stability of the pulverized coal flow, thus ensuring combustion without flameout or instability under conditions of overall low excess air coefficient.

[0062] For the temperature control loop, when the oxygen and temperature sliding mode controller calculates a control command requiring a reduction in the furnace outlet temperature, it sends a flow increase signal to the circulating clean flue gas multi-point injection device. The circulating clean flue gas multi-point injection device responds to this flow increase signal by adjusting the electrically operated regulating valves on the main delivery pipe and each branch pipeline.

[0063] Specifically, the multi-point injection device for circulating clean flue gas increases the flow rate of circulating clean flue gas leading to the wall-mounted protective injection layer. The high-momentum circulating clean flue gas cuts tangentially into the water-cooled wall, forming a low-temperature gas insulation film on the inner surface of the water-cooled wall. This low-temperature gas insulation film blocks direct radiative heat exchange between the high-temperature flame in the furnace and the metal surface of the water-cooled wall, reducing the temperature level in the near-wall region and thus inhibiting the adhesion of molten slag to the water-cooled wall.

[0064] Simultaneously, the multi-point injection device for circulating clean flue gas increases the flow rate of circulating clean flue gas leading to the upper injection layer of the burnout zone. Low-temperature circulating clean flue gas is injected into the upper space of the furnace, mixing with the rising high-temperature flue gas flow. Through physical heat absorption and dilution, the circulating clean flue gas lowers the overall flue gas temperature at the furnace outlet plane, ensuring that the gaseous alkali metals carried by the flue gas begin to undergo supersaturation precipitation before reaching the convective heating surface, thus preventing direct contact between the gaseous alkali metals and the high-temperature superheater tube walls.

[0065] Throughout the steady-state combustion optimization process, the oxygen and temperature sliding mode controller continuously monitors the system's response to control actions. If the monitoring module's feedback data shows that the deviation of the average oxygen content in the furnace or the deviation of the furnace outlet temperature exceeds the preset steady-state dead zone range, the oxygen and temperature sliding mode controller will adjust the control gain through high-frequency switching according to the sliding mode variable structure control law, outputting a large correction command to overcome the nonlinear disturbances caused by the fluctuation of high-alkali coal quality, forcing the system state to return to the preset sliding mode surface. This control method ensures that the boiler always operates under the dual constraints of low oxygen inhibiting nitrogen oxide generation and low temperature inhibiting alkali metal gasification under normal operating conditions.

[0066] Please see the appendix Figure 2 -Appendix Figure 3 S3: Monitoring, cleaning, and dynamic control of combustion specifically include the following processes: The quantitative characterization model for heating surface fouling characteristics transmits the real-time calculated heating surface cleanliness factor data stream to the multidimensional ash removal decision model within each calculation cycle. The multidimensional ash removal decision model is configured to preset independent ash accumulation critical thresholds for different heating surface regions of the boiler. For the water-cooled wall region, the ash accumulation critical threshold is set to 0.35; for the high-temperature superheater region, the ash accumulation critical threshold is set to 0.45. The multidimensional ash removal decision model compares the received heating surface cleanliness factor with the corresponding ash accumulation critical threshold.

[0067] When the cleanliness factor value of a specific heated surface area falls below the set critical threshold for ash accumulation, the multidimensional ash removal decision model determines that there is a serious risk of ash accumulation or slagging in that area and generates a cleaning request signal for that specific area. The multidimensional ash removal decision model sends this cleaning request signal to the dynamic mitigation and collaborative control logic unit, rather than directly to the cleaning execution mechanism.

[0068] Upon receiving a cleaning request signal, the dynamic suppression and coordination control logic unit immediately initiates the combustion pre-compensation and reinforcement program. Because soot blowing or hydraulic decoking operations introduce low-temperature media into the furnace or cause drastic fluctuations in localized heat load, they can easily lead to combustion instability or even flameout under high-alkali coal combustion conditions. Therefore, the dynamic suppression and coordination control logic unit sends a combustion state switching command to the oxygen and temperature sliding mode controller, temporarily switching the control target from a low-oxygen, low-nitrogen mode to a stable combustion and disturbance-resistant mode.

[0069] In response to the combustion state switching command, the oxygen and temperature sliding mode controller advances the swirl generator in the inner secondary air channel towards the furnace interior by adjusting the sliding expansion cone adjustment mechanism of the new oxygen and temperature wide-range adjustable burner. This mechanical action increases the tangential momentum of the inner secondary air, raising the swirl intensity number to the range of 1.2 to 1.5. The high swirl intensity creates a strong aerodynamic recirculation zone at the burner outlet, which forcibly entrains high-temperature flue gas to the ignition root of the pulverized coal, thereby improving the temperature level at the flame root and its resistance to flameout. Simultaneously, the oxygen and temperature sliding mode controller fine-tunes the momentum ratio of primary air to secondary air, increasing the rigidity of the central primary air and preventing airflow disturbances during cleaning operations from causing flame deflection or disruption of the tangential firing condition.

[0070] The dynamic suppression and coordination control logic unit continuously monitors the furnace negative pressure fluctuation rate and flame intensity flicker frequency fed back by the monitoring module. When the furnace negative pressure fluctuation rate is less than 50 Pascals and the flame intensity flicker frequency is stable within ±5% of the normal baseline value, the dynamic suppression and coordination control logic unit determines that the combustion state inside the furnace has established sufficient anti-disturbance redundancy. At this time, the dynamic suppression and coordination control logic unit sends a permission action command to the field cleaning actuator through the command output interface unit.

[0071] The on-site cleaning actuators include acoustic soot blowers, steam soot blowers, or hydraulic soot blowers. The command output interface unit addresses and activates the corresponding cleaning actuator based on the specific heated surface area identified in the cleaning request signal. For example, if the cleaning request signal originates from the water-cooled wall area, the command output interface unit drives the corresponding hydraulic soot blower to perform a diagonal spraying operation. During the operation of the cleaning actuators, the heated surface contamination characteristic quantification model continues to calculate the heated surface cleanliness factor in real time.

[0072] When the heat transfer surface cleanliness factor calculated by the quantitative characterization model of heat transfer surface fouling characteristics rises above 0.85, or after the cleaning actuator completes the preset operation time, the multidimensional ash removal decision model cancels the cleaning request signal. Upon detecting the disappearance of the cleaning request signal, the dynamic suppression and collaborative control logic unit sends a recovery command to the oxygen and temperature sliding mode controller. The oxygen and temperature sliding mode controller then switches the control strategy back to the sliding mode-based steady-state combustion optimization state of step two, reducing the swirl intensity number and restoring low-oxygen combustion parameters to re-enter the low-NOx emission and alkali metal gasification suppression operation mode. This dynamic suppression control process ensures that the boiler maintains the continuity and stability of the combustion process throughout the cyclical changes of ash accumulation-cleaning-re-ash accumulation, avoiding the risk of unplanned boiler shutdowns due to forced cleaning.

[0073] S4: The coordinated defense strategy against pollutants specifically includes the following processes: The alkali metal phase change and dew point tracking logic unit first dynamically calculates the saturation precipitation critical temperature of the gaseous alkali metal based on the real-time feedback of the gaseous alkali metal concentration and flue gas component partial pressure from the monitoring module. The logic unit then sends a cooling medium flow rate adjustment command to the phase change conditioning device in the pollutant control module. In response, the phase change conditioning device increases the cooling medium flow rate within the tubular cryogenic heat exchanger array, forcibly reducing the flue gas temperature flowing through this area below the calculated saturation precipitation critical temperature. This subcooling operation creates a thermodynamically non-equilibrium environment in the flue gas flow field, forcing gaseous sodium and potassium atoms to undergo heterogeneous nucleation using fine fly ash particles in the flue gas as condensation nuclei, or directly sublimate and adsorb onto the surface of the fly ash particles. This transforms the difficult-to-physically capture gaseous alkali metal into easily captureable solid particulate alkali metal.

[0074] Subsequently, the flue gas enters the particle agglomeration device. The high-voltage electrostatic agglomeration unit and the acoustic excitation unit within the device are activated simultaneously. The high-voltage electrostatic agglomeration unit applies a non-uniform high-voltage electric field to the fine particles carrying alkali metal condensates, causing the particles to become charged and migrate directionally. Simultaneously, the acoustic excitation unit emits high-intensity sound waves into the flue, inducing high-frequency vibrations in the airflow medium. Under the synergistic effect of Coulomb force and acoustic turbulence, intense orthogonal collisions and friction occur between the fine particles, causing the alkali metal-carrying particles to rapidly adhere and agglomerate into large-diameter agglomerates. Immediately afterwards, the electrostatic precipitator utilizes the strong electrostatic attraction generated by the high-voltage electric field to efficiently capture these large-diameter agglomerates that have undergone phase change growth and physical agglomeration. This physically separates and removes most of the fly ash particles containing solidified alkali metals before the flue gas enters the denitrification system, cutting off the poisoning path of alkali metals to subsequent catalysts at the source.

[0075] After the initial alkali removal pretreatment, the multi-objective model predictive control takes over the deep purification control of nitrogen oxides. The nitrogen oxide generation prediction sub-model within the multi-objective model predicts the nitrogen oxide concentration curve entering the selective catalytic reduction (SCR) denitrification system in the future time domain based on the current combustion parameters. The rolling optimization control sub-model, based on this prediction curve and combined with the ammonia slip rate constraint, calculates the optimal ammonia injection flux distribution matrix for the ammonia injection unit.

[0076] The command output interface unit parses the optimal ammonia injection flux distribution matrix into control signals, which drive the high-precision flow regulating valves on each independent branch of the partitioned grid-type injection array in the ammonia injection unit. Based on the non-uniformity of nitrogen oxide concentration distribution on the flue gas cross-section, the ammonia injection unit performs differentiated, on-demand ammonia injection operations, ensuring a precise molar ratio match between ammonia reducing agent and nitrogen oxides in spatial distribution.

[0077] Flue gas mixed with ammonia enters the selective catalytic reduction (SCR) denitrification system. Under the action of the alkali-resistant catalyst, nitrogen oxides are reduced to harmless nitrogen and water. Because the upstream process has significantly reduced the alkali metal content in the flue gas, the alkali-resistant catalyst avoids the risk of active sites being covered by alkali metal sulfates or pores being blocked, thus maintaining high denitrification efficiency at full load.

[0078] Finally, the nitrogen oxide concentration monitor and particulate matter quantitative monitoring device in the tail flue gas monitoring subset detect pollutant indicators in the final emission flue gas in real time. If the detection data shows a trend of exceeding the standard, the multi-objective model predictive control model immediately activates the feedback correction mechanism to adjust the control strategy weights for the next moment online, ensuring that the pollutant emissions of the high-alkali coal boiler always meet the ultra-low emission standards through closed-loop iteration.

[0079] This embodiment selects a 660 MW ultra-supercritical variable-pressure once-through boiler as the implementation object. The boiler body adopts a single-furnace inverted U-shaped arrangement, balanced ventilation, and front and rear wall opposed combustion. The new oxygen and temperature wide-range adjustable burners are arranged in layers on the front and rear water-cooled walls. The wall-mounted protective injection layer of the circulating clean flue gas multi-point injection device is arranged along the side wall water-cooled walls, together constructing an aerodynamic field containing a low-temperature oxidizing atmosphere near the wall and a high-temperature combustion atmosphere in the center.

[0080] The target fuel is high-alkali metal content bituminous coal from Zhundong, Xinjiang, with a net calorific value of 22 MJ / kg to 26 MJ / kg, a dry ash-free volatile matter content of more than 30%, a sodium oxide content of more than 2.0% by mass, an ash fusion softening temperature of less than 1150 degrees Celsius, and characteristics of easy fouling and strong slagging.

[0081] The operating conditions cover a range of 30% to 100% of the rated load: Under the condition of 100% boiler maximum continuous evaporation, the intelligent control module implements a collaborative pollutant defense strategy, maintaining the excess air coefficient at approximately 1.15 through the oxygen and temperature sliding mode controller, and controlling the circulating clean flue gas multi-point injection device to inject circulating flue gas at a high flow rate in the upper part of the burnout zone, controlling the furnace outlet flue gas temperature below 1050 degrees Celsius. Under the condition of 30% to 50% deep peak shaving, the intelligent control module implements steady-state combustion optimization based on sliding mode, driving the sliding expansion cone adjustment mechanism of the new oxygen and temperature wide-range adjustable burner to enhance the intensity of the internal secondary air swirl and construct a stable combustion recirculation zone. Under the condition of dynamic disturbance during heating surface cleaning, after the multi-dimensional ash and dirt removal decision model issues a cleaning request signal, the intelligent control module allows furnace negative pressure fluctuations of ±100 Pascals and a 10% increase in nitrogen oxide generation concentration, in exchange for combustion anti-interference robustness during the operation of the cleaning actuator.

[0082] The implementation process is divided into four stages: Phase 1: Global State Awareness and Baseline Establishment The laser diagnostic device and the self-emission spectroscopy analysis device in the monitoring module output the gas phase sodium and potassium atom concentration distribution map of the furnace cross section at a frequency of 10 Hz; the acoustic temperature measurement system reconstructs the temperature field of the furnace outlet cross section; and the heat transfer monitoring subset of the heating surface uploads heat flux density and tube wall metal temperature data. The data acquisition and preprocessing unit performs spatiotemporal alignment on the above data to establish a high-dimensional state vector matrix characterizing the operating status.

[0083] Phase 2: Steady-state combustion optimization operation based on sliding mode: The oxygen and temperature sliding mode controller performs adjustments based on the average oxygen content in the furnace and the feedback value of the outlet temperature. When the furnace outlet temperature approaches 1050 degrees Celsius, the multi-point injection device for circulating clean flue gas is driven to increase the flow rate of circulating clean flue gas in the upper part of the combustion zone, thereby reducing the flue gas temperature through dilution and heat absorption.

[0084] When there is a deviation in the average oxygen content in the furnace, adjust the opening of the external secondary air damper of the new type of oxygen and temperature wide-range adjustable burner to maintain the excess air coefficient between 1.1 and 1.2; if unstable combustion occurs, immediately fine-tune the intensity of the internal secondary air swirl to stabilize the flame root.

[0085] Phase 3: Dynamic collaborative cleaning under dust accumulation conditions: A quantitative characterization model for the fouling characteristics of the heated surface calculates the cleanliness factor of the heated surface in real time. When the cleanliness factor of the heated surface is lower than the critical threshold of ash accumulation of 0.45, the multidimensional ash removal decision model sends a request to the dynamic suppression and collaborative control logic unit. The dynamic suppression and collaborative control logic unit instructs the oxygen and temperature sliding mode controller to increase the swirl intensity number of the new oxygen and temperature wide-range adjustable burner and increase the primary air rigidity. After confirming that the flame intensity flicker frequency and furnace negative pressure fluctuation rate meet the stability indicators, the instruction output interface unit activates the corresponding steam soot blower or acoustic soot blower.

[0086] Phase 4: Deep, tiered purification of tail-end pollutants: The alkali metal phase transition and dew point tracking logic unit calculates the critical temperature for saturated precipitation, driving the phase transition conditioning device to cool the flue gas temperature below the dew point, inducing heterogeneous condensation of gaseous alkali metals. The flue gas containing condensates passes through a particle agglomeration device, where it agglomerates under the influence of an electric and acoustic field and is captured by an electrostatic precipitator. The dealkali-treated flue gas then enters the selective catalytic reduction (SCR) denitrification system. A multi-objective model predictive control model uses a rolling optimization algorithm to calculate the optimal ammonia injection increment, driving the ammonia injection device to implement on-demand ammonia injection, achieving ultra-low nitrogen oxide emissions under the action of an alkali-resistant catalyst.

Claims

1. A method for optimizing and adjusting the combustion of a high-alkali coal boiler, characterized in that, Includes the following steps: S1: The monitoring module collects multi-source heterogeneous data from the high-temperature zone of the boiler furnace, the heating surface, and the tail flue. The intelligent control module uses the multi-source heterogeneous data to establish a high-dimensional state vector matrix that characterizes the boiler's operating status. S2: The intelligent control module uses an oxygen and temperature sliding mode controller to adjust the new oxygen and temperature wide-range adjustable burner and circulating clean flue gas multi-point injection device in the combustion control module based on the furnace average oxygen content and furnace outlet temperature data extracted from the high-dimensional state vector matrix, and performs steady-state combustion optimization. S3: The intelligent control module calculates the cleanliness factor of the heated surface based on the data in the high-dimensional state vector matrix. When the cleanliness factor of the heated surface is lower than the ash accumulation critical threshold, the dynamic suppression and collaborative control logic unit executes the dynamic suppression strategy. First, combustion pre-compensation reinforcement is performed to ensure that the combustion state parameters in the high-dimensional state vector matrix meet the stability index, and then the cleaning actuator is driven to perform cleaning operations. S4: The intelligent control module uses the alkali metal phase change and dew point tracking logic unit to regulate the phase change conditioning device based on the gas phase alkali metal concentration data in the high-dimensional state vector matrix to induce alkali metal phase change, and uses the multi-objective model prediction control model to regulate the ammonia injection device based on the nitrogen oxide concentration data in the high-dimensional state vector matrix to perform denitrification.

2. The combustion optimization and adjustment method for a high-alkali coal boiler according to claim 1, characterized in that, The process of establishing the high-dimensional state vector matrix in step S1 specifically includes: The concentration distribution maps of gaseous sodium and potassium atoms in the cross-section of the furnace were reconstructed by fusing data obtained from laser diagnostic devices and auto-emission spectroscopy analysis devices. The temperature field distribution at the furnace outlet section was reconstructed using an acoustic temperature measurement system. Local heat flux density data of the heated surface and metal temperature data of the tube wall were collected simultaneously using a tubular heat flux meter and an N-type armored thermocouple. The gas phase sodium and potassium atom concentration distribution map, the temperature field distribution, the local heat flux density data, and the pipe wall metal temperature data are spatiotemporally aligned to form the high-dimensional state vector matrix.

3. The combustion optimization and adjustment method for a high-alkali coal boiler according to claim 1, characterized in that, The S2 step of adjusting the novel oxygen and temperature wide-range adjustable burner specifically includes: When the oxygen and temperature sliding mode controller calculates that the oxygen concentration in the furnace needs to be reduced, it drives the damper of the external secondary air duct to move in the closing direction. Simultaneously, it drives the axially moving swirl generator in the internal secondary air duct to move, changing the tangential velocity component of the internal secondary air to enhance the swirl intensity number, forming a central recirculation zone at the burner outlet.

4. The combustion optimization and adjustment method for a high-alkali coal boiler according to claim 1, characterized in that, The adjustment of the multi-point injection device for circulating clean flue gas in step S2 specifically includes: When the oxygen and temperature sliding mode controller calculates that the furnace outlet temperature needs to be reduced, the circulating net flue gas flow rate to the upper injection layer of the burnout zone is increased, and the furnace outlet plane flue gas temperature is reduced through physical dilution and heat absorption. Increase the flow rate of circulating clean flue gas to the wall-mounted protective spray layer, and cut tangentially into the water-cooled wall to form a low-temperature gas heat insulation film.

5. The combustion optimization and adjustment method for a high-alkali coal boiler according to claim 1, characterized in that, The calculation process of the cleanliness factor of the heated surface in step S3 includes: The actual heat absorption of the heating surface is calculated based on the mass flow rate of the working fluid flowing through the heating surface and the specific enthalpy difference between the inlet and outlet of the working fluid. Based on the Stefan-Boltzmann law, the theoretical heat absorption of the heat-receiving surface is calculated using the heat transfer area of ​​the heat-receiving surface, the comprehensive correction factor, the average absolute temperature of the flue gas side, and the average absolute temperature of the outer surface of the pipe wall metal. The ratio of the actual heat absorption to the theoretical heat absorption is calculated as the cleanliness factor of the heated surface.

6. The combustion optimization and adjustment method for a high-alkali coal boiler according to claim 1, characterized in that, The specific process of executing the dynamic suppression strategy using the dynamic suppression cooperative control logic unit in step S3 includes: After determining that the cleanliness factor of the heating surface is lower than the ash accumulation critical threshold, the dynamic suppression and collaborative control logic unit switches the control target to stable combustion and anti-disturbance mode. The ash accumulation critical threshold is an independent value preset for different heating surface areas of the boiler. Adjusting the sliding expansion cone adjustment mechanism of the novel oxygen and temperature wide-range adjustable burner will push the axially moving swirl generator in the inner secondary air channel toward the inside of the furnace to increase the swirl intensity and increase the rigidity of the central primary air. Monitor the furnace negative pressure fluctuation rate and flame intensity flicker frequency, and activate the cleaning actuator in the corresponding area when the set conditions are met; After cleaning, restore the low-oxygen combustion parameters and the original swirl intensity.

7. The combustion optimization and adjustment method for a high-alkali coal boiler according to claim 1, characterized in that, The process of controlling the phase change conditioning device using the alkali metal phase change and dew point tracking logic unit in step S4 includes: The critical temperature for saturated precipitation of gaseous alkali metals is calculated based on the concentration of gaseous alkali metals and the partial pressure of flue gas components. Set a phase change tempering target temperature below the saturation precipitation critical temperature; Adjusting the flow rate of the cooling medium in the tubular cryogenic heat exchanger array of the phase change conditioning device reduces the flue gas temperature to the target temperature of the phase change conditioning, inducing heterogeneous condensation of gaseous alkali metals.

8. The combustion optimization and adjustment method for a high-alkali coal boiler according to claim 7, characterized in that, Step S4, after the flue gas temperature decreases to the target temperature for phase change conditioning, also includes: Flue gas enters the particle agglomeration device, where a non-uniform high-voltage electric field is applied by the high-voltage electrostatic agglomeration unit to charge the fine particles, and a high-intensity sound wave is emitted by the acoustic excitation unit to induce the vibration of the airflow medium, which promotes the agglomeration of fine particles carrying alkali metal condensates into large-size agglomerates. The large-size agglomerates are captured using an electrostatic precipitator.

9. The combustion optimization and adjustment method for a high-alkali coal boiler according to claim 1, characterized in that, The process of using a multi-objective model predictive control model to regulate the ammonia injection device in step S4 includes: Using a nitrogen oxide generation prediction sub-model, with furnace average oxygen content and furnace outlet temperature as feedforward disturbance variables, the dynamic response trajectory of nitrogen oxide concentration at the outlet of the selective catalytic reduction denitrification system in the future time domain is predicted. Using a rolling optimization control sub-model, the optimal ammonia injection flow rate control increment sequence is solved to minimize the changes in control action while keeping the system output tracking the set trajectory, under the conditions of satisfying the upper limit constraint of ammonia escape and the valve opening constraint. The first element in the sequence is sent as a control command to the partitioned grid-type injection array of the ammonia injection device.

10. A combustion optimization and adjustment system for a high-alkali coal boiler, characterized in that, A combustion optimization and adjustment method for a high-alkali coal boiler according to any one of claims 1-9, comprising: The monitoring module is configured to collect parameters of the furnace and flue. Combustion control module, which includes a new type of oxygen and temperature wide-range adjustable burner and a circulating clean flue gas multi-point injection device; The pollutant control module includes a phase change conditioning device, a particle agglomeration device, an electrostatic precipitator, an ammonia injection device, and a selective catalytic reduction denitrification system; The intelligent control module includes a data acquisition and preprocessing unit and a core computing and processing unit; The core computing and processing unit operates an aerobic and temperature sliding mode controller, a quantitative characterization model of the contamination characteristics of heated surfaces, a multidimensional ash and dirt removal decision model, a dynamic suppression and collaborative control logic unit, an alkali metal phase transition and dew point tracking logic unit, and a multi-objective model predictive control model.

Images