Coal-fired boiler control method directly facing NOx emission

Abstract

The invention discloses a coal-fired boiler control method directly oriented to NOx emission, and mainly relates to the technical field of intelligent environmental protection. Comprising the following steps: S1, establishing a boiler combustion control model for NOx emission; s2, designing an MPC controller, on the basis of a boiler combustion control model, calculating a Jacobian matrix at a current working point to realize local linearization of the model so as to construct a target function, and solving and updating an optimal control signal at each sampling moment; s3, designing a state observer, solving unknown state variables and interference in the system through the state observer, and improving the control performance of the system through error compensation; according to the method, the dynamic model directly oriented to the NOx concentration is established, the chemical reaction process of NOx is influenced by controlling the air inlet amount of the boiler system, the NOx emission concentration accurately tracks the target curve, and the corresponding emission standard is met.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent environmental protection, and in particular to a control method for a coal-fired boiler directly facing NOx emissions. Background Art

[0002] As an energy device widely used in thermal power generation and industrial production, a coal-fired boiler will generate a large amount of pollutants during the coal combustion process, causing air pollution problems. Therefore, it is necessary to analyze the operation mechanism of the boiler more deeply, design a stable and effective boiler combustion controller to make the boiler emissions meet the corresponding emission standards, and studying the low-nitrogen combustion control method of the boiler can also help improve the operation efficiency of the boiler.

[0003] A reasonable combustion model is the basis of boiler combustion control technology. Currently, many scholars conduct two-dimensional or three-dimensional simulation studies on the boiler combustion process through the numerical simulation method of Computational Fluid Dynamics (CFD). Based on software such as ANSYS and FLUENT, by using the actual physical parameters of the boiler, appropriate turbulence and chemical models are selected to simulate the internal combustion phenomenon, fluid flow and local heat transfer characteristics of the boiler. However, the CFD method requires building a huge numerical simulation model, with high computational costs, and it is difficult to embed control methods for simulation. It can only verify its effectiveness by changing its physical parameters or chemical coefficients, and cannot support the continued research of boiler combustion control theory.

[0004] Currently, data-driven modeling methods are also widely used in the modeling and prediction of industrial systems. However, under the complex combustion conditions of the boiler, sensors are difficult to effectively collect data due to the extremely high working environment temperature, interference from fly ash and slag, and during the data transmission process, not only is the transmission distance long, but also the noise is large, and its data cannot accurately reflect the working conditions, while effective actual data is a necessary condition for data-driven modeling. In addition, building an online update data model that can truly simulate complex working conditions on a model based on a large amount of labeled data will bring great computational pressure to the control system. Therefore, this method also has great deficiencies in terms of real-time performance and flexibility.

[0005] In the control research of boiler combustion, currently, a distributed control system is generally adopted, with the air intake as the main control variable, and the controlled variables are mostly steam pressure and drum water level related to the boiler output power, etc. Its control loop mostly adopts PID cascade control. In the design of a controller considering the NOx emissions of the boiler, only NOx is used as a constraint condition, and the generation and destruction processes of NOx are not directly considered in the design of the controlled object model. The current control method is difficult to achieve an ideal level in NOx emissions restrictions, and cannot adjust the NOx emission concentration of the boiler in real time and quickly, bringing great pressure to the denitration device of the boiler. Summary of the Invention

[0006] The object of the present invention is to solve the problems existing in the prior art, and provide a control method for a coal-fired boiler directly facing NOx emissions, establish a kinetic model directly facing NOx concentration, and obtain the chemical reaction process of NOx affected by the air intake through controlling the boiler system, so that the NOx emission concentration accurately tracks the target curve and meets the corresponding emission standards.

[0007] To achieve the above object, the present invention is realized through the following technical solutions:

[0008] A control method for a coal-fired boiler directly facing NOx emissions, comprising the steps of:

[0009] S1. Establish a boiler combustion control model facing NOx emissions,

[0010]

[0011] wherein, the state variable x1 represents the NOx concentration, with the unit of mg / m 3 , which is the main output index of the system; x2 represents the coal feeding amount, with the unit of kg / s, reflecting the fuel flow rate entering the furnace; x3 represents the oxygen concentration, with the unit of kg / m 3 , used to describe the oxygen content in the combustion process; x4 represents the nitrogen concentration, with the unit of kg / m 3 , used to describe the nitrogen component in the combustion air; the control quantity u(t) is the control input signal, which represents the air intake; the parameter ρ represents the mass fraction of nitrogen element in the coal, with the unit of kg / kg, reflecting the nitrogen content characteristics of the coal; φ represents the molar ratio of hydrogen cyanide (HCN) in the volatile components during the pulverized coal pyrolysis process, with the unit of dimensionless, used to describe the potential ratio of fuel nitrogen migrating to NOx; the constant k is the control gain; T is used to describe the sensitivity of the combustion temperature unit of K (Kelvin) to the NOx generation rate; in addition, δ(t) describes the dynamic disturbance or change rate of the coal feeding amount, with the unit of kg / s 2 , used to model the influence of external disturbances on the combustion process; through the above model, the generation mechanism of NOx during the coal combustion process of the boiler can be completely described, and then a mathematical basis can be provided for the model-based control method to achieve precise control of the NOx emission concentration and ensure that the boiler operation meets the environmental emission standards.

[0012] S2. Design an MPC controller, linearize it by solving the Jacobian matrix on the basis of the boiler combustion control model, construct an objective function, and solve and update the optimal control signal at each sampling moment;

[0013] S3. Design a state observer to solve the unknown state variables and disturbances in the system through the state observer, and then improve the control performance of the system through error compensation.

[0014] Preferably, in step S1, the construction of the boiler combustion control model includes: based on the NOx generation mechanism in boiler combustion, taking the chemical reaction process of fuel NOx as the basis, selecting state variables, control variables, and output variables to establish a control-oriented mechanism model; the main chemical reaction equations of fuel NOx are:

[0015] HCN / NH3 + O2 → NO +…

[0016] NO + HCN / NH3 → N2 +…

[0017] During the combustion process, the nitrogen element (N) in the fuel will first be converted into hydrogen cyanide (HCN) or ammonia (NH3). These nitrogen-containing substances are oxidized by oxygen (O2) to form NOx, and at the same time, part of the NOx will further react with HCN or NH3 to form nitrogen; the kinetic rate equations of the above two chemical reaction processes are described as follows:

[0018]

[0019] Among them, X represents the molar concentration of the corresponding component (unit: mol / m 3 ), X HCN is the molar concentration of hydrogen cyanide (mol / m 3 ), is the molar concentration of oxygen (mol / m 3 ), X NO is the molar concentration of nitric oxide (mol / m 3 ), and b is the empirical reaction order of oxygen participating in the reaction (generally determined according to experimental data, dimensionless).

[0020] In order to further incorporate the fuel characteristics into the model, assume that the coal storage in the boiler is M (unit: kg), the fuel used is standard coal, and its nitrogen content is defined as ρ (unit: kg / kg). During the pyrolysis process of pulverized coal, the proportion of hydrogen cyanide (HCN) in the volatile components is defined as φ (dimensionless). Then the above reaction rate equations can be rewritten as:

[0021]

[0022] The source of O2 in the boiler is the O2 in the air supply. Based on the principle of mass balance, the change rate of O2 is:

[0023]

[0024] Among them, V represents the volume of the boiler combustion space, unit: m3 ; A in is the air intake of the boiler, unit: m 3 / s; k in is the molar conversion coefficient of oxygen in the incoming air, unit: mol / m 3 ; k vo is the molar coefficient of oxygen consumption during the release of volatile matter from pulverized coal pyrolysis, unit: mol / kg; k br is the molar coefficient of oxygen consumption during the combustion process, unit: mol / kg; k ou is the molar conversion coefficient of oxygen in the outgoing air, unit: mol / m 3 ; In this model, the NOx concentration coal feeding rate (M), oxygen concentration and nitrogen concentration are selected as the state variables of the system, and the intake air volume (A in ) is set as the control variable to construct the controlled object model of the coal-fired boiler, thereby providing a mathematical basis for the subsequent design of the controller.

[0025] Preferably, in step S2, the NOx concentration (unit: mg / m 3 ) emitted during boiler combustion is selected as the controlled variable, and the total incoming air volume of the boiler system (unit: m 3 / s) is selected as the control variable. By adjusting the air intake volume, the combustion process of pulverized coal is affected, so that the NOx concentration dynamically tracks the set target value.

[0026] To achieve the above control objectives, the model predictive control (MPC) method is adopted to design the controller and construct the optimization performance index as follows:

[0027]

[0028] Among them, J(k) is the optimization performance index; y(k) is the output value of the NOx concentration at the current sampling moment (mg / m 3 ); is the set target value of the NOx concentration (mg / m 3 ).

[0029] Preferably, in step S2, the boiler combustion control model is discretized by the Euler method, and then through the mechanism of "rolling optimization - repeated execution" of predictive control, the Jacobian matrix of the state space equation is solved at each sampling point to obtain the linearized system equation at each moment. Based on the optimization performance index, the optimal control rate of the actuator is solved to perform tracking control on it.

[0030] Preferably, in step S2, an MPC controller with error compensation is constructed, specifically as follows:

[0031] First, the unknown state variables and disturbances at the (k + 1)-th moment are obtained through a state observer based on the system state and output at the k-th moment. Then, based on the error compensation principle, according to Solve the control compensation amount u ′* (k); where: C0 represents the output matrix of the system, reflecting the mapping relationship from the system state to the output; B represents the system control matrix, describing the influence of the control input on the state change; i ′* (k) is the control correction amount under error compensation, and the unit is the same as that of the control amount; is the external disturbance term estimated by the state observer, and the unit is the same as that of the output quantity; e(k) represents the observation error at the moment k, defined as the deviation between the actual output value and the predicted value, and the unit is mg / m 3 .

[0032] Subsequently, substitute the disturbance value estimated by the state observer into the optimization problem minJ(k), and solve the corresponding quadratic programming problem to obtain the optimal control sequence u(k + 1) = u ′* (k) + u * (k + 1|k);

[0033] Through the above steps, combined with the state observer and the rolling optimization strategy, the real-time and precise control of NOx emissions from coal-fired boilers can be achieved. At each sampling moment, the processes of state estimation, error compensation, and optimization solution are repeated to ensure that the control strategy can adapt to the working condition changes in real time, and improve the dynamic response ability and robustness of the control system.

[0034] Preferably, in step S3, an error feedback compensation mechanism is constructed to compensate the prediction error of the previous moment into the boiler combustion control model of the next moment, where the prediction error is

[0035] where, represents the prediction error at the sampling moment k, and the unit is mg / m 3 , reflecting the deviation between the actual output value and the model predicted output value; y(k) represents the actual measured output value of the system at the moment k, usually the NOx emission concentration, and the unit is mg / m 33 ; represents the predicted value of the NOx emission concentration by the system model at the moment k, and the unit is mg / m 3 .

[0036] By using the prediction error The control model introduced and compensated for the next moment can effectively reduce the prediction deviation caused by model uncertainty, system disturbances or operating condition fluctuations, thereby improving the stability and control effect of the boiler combustion control system and achieving accurate tracking of the NOx concentration and effective compliance with emission standards.

[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0038] The main purpose of the present invention is to analyze the NOx generation mechanism during the boiler combustion process, establish a non-linear dynamic model directly facing the NOx concentration, design an MPC controller based on this model, use the air intake as the control variable, solve the optimal control law of the air intake according to the optimization objective function at each sampling moment, and minimize the NOx emission concentration under the premise of meeting the constraints; due to the existence of unknown state variables and disturbances in the system, a state observer is established to estimate the target value and input it into the prediction model to improve the accuracy of the prediction model. In order to compensate for system errors, an error feedback mechanism is constructed, and based on the prediction error and disturbance, the compensation control law is solved to control the NOx concentration to track the target curve more quickly and accurately, enabling the NOx emission concentration of the boiler system to meet the corresponding standards and improving the response speed and tracking accuracy of the boiler system. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is the structural diagram of the boiler combustion system;

[0040] Figure 2 is the structural diagram of the state observer;

[0041] Figure 3 is the control structure diagram of the boiler low-nitrogen combustion system;

[0042] Figure 4 is the simulation result diagram of the control system output;

[0043] Figure 5 is the tracking effect diagram of the NOx emission concentration of the boiler combustion system. DETAILED DESCRIPTION OF THE INVENTION

[0044] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and not to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by this application.

[0045] The present invention establishes a control model for NOx concentration, enabling the control system to use the NOx concentration generated during the combustion process as the state variable and directly adjust its emission concentration through a controller; by designing a state observer to estimate the unknown states and disturbances in the system, adopting the mechanism of "rolling optimization - repeated incoming line", at each sampling moment, linearize the model, solve the output of the prediction model, and based on the optimized performance index function, calculate the optimal control law, and design an error compensation structure, so that the control system is a repeated decision-making process based on compensating for prediction errors and control errors.

[0046] The present invention provides a control method for a coal-fired boiler directly facing NOx emissions. By analyzing the NOx generation mechanism during the boiler combustion process, a control model directly facing NOx concentration is established based on physical and chemical laws; for the established nonlinear model of the boiler combustion system, a combustion control system is designed using the model predictive control (MPC) method. Based on the optimized performance index function, the mechanism of rolling optimization and repeated incoming line is adopted to solve the optimal control law to make the NOx emission concentration of the boiler system meet the national standards; by designing a state observer and an error feedback system, the response speed and tracking accuracy of the boiler system are improved. The method includes the following steps:

[0047] Step 1: Based on the boiler combustion mechanism, according to the laws of mass and energy balance and thermodynamics principles, and based on the NOx chemical reaction process, establish a boiler combustion control model facing NOx emissions.

[0048] Step 2: Design an MPC controller, linearize it by solving the Jacobian matrix based on the mechanism model, construct an objective function, and solve and update the optimal control signal at each sampling moment, so that the control system can meet the NOx emission index.

[0049] Step 3: Design a state observer to solve the unknown state variables and disturbances in the system through the state observer. Then, through error compensation, improve the control performance of the system.

[0050] Specifically:

[0051] In Step 1, based on the NOx generation mechanism during boiler combustion, and based on the chemical reaction process of fuel-type NOx, select state variables, control variables, and output variables to establish a mechanism model for control.

[0052] The main chemical reaction equations of fuel-type NOx are:

[0053]

[0054] As can be seen from Equation (1), the generation of fuel-type NOx is a chemical reaction process in which oxidation and reduction compete simultaneously. The nitrogen element in the fuel is converted into HCN / NH3 during combustion. While being oxidized into NOx, it will also react with NOx to form N2. The reaction rates of these two processes are given below:

[0055]

[0056] Among them, X represents the molar concentration of the substance, b is the reaction coefficient of O2, and T represents the temperature of the boiler.

[0057] Let the coal storage in the boiler be M. In a pulverized coal boiler, the combustion rate of coal is equal to its coal feeding rate. Assume the coal being burned is standard coal with a nitrogen content of ρ. During the pyrolysis of pulverized coal, assume the proportion of HCN in the volatile components is Then its reaction rate is:

[0058]

[0059] The source of O2 in the boiler is the O2 in the air supply. Based on the principle of mass balance, the change rate of O2 is:

[0060]

[0061] Among them, V is the volume of the boiler, k in is the molar coefficient conversion ratio of O2 in the air supply, A in is the air intake, k vo and k br are the molar coefficient conversion ratios of the O2 amounts for volatile release and combustion, and k ou is the molar coefficient conversion ratio of O2 in the air outlet.

[0062] Here, select the NOx concentration, coal feeding amount, O2 concentration, and N2 concentration as state variables, and take the intake air volume as the control variable to establish a controlled object model.

[0063] In Step 2, take the NOx concentration generated during boiler combustion as the controlled variable, take the total air intake of the boiler system as the control variable, and affect the combustion process of pulverized coal by adjusting the air intake volume, so as to make the generated NOx concentration track the set value. Design an MPC controller and construct an optimization performance index as shown in Equation (5):

[0064]

[0065] Since the controlled object model is a continuous input-affine nonlinear system, which is not applicable to discrete sampling and control systems in industrial scenarios, it is discretized by the Euler method. Then, through the mechanism of "rolling optimization - repeated execution" of predictive control, the Jacobian matrix of the state-space equation is solved at each sampling point to obtain the linearized system equation at each moment. Based on the optimized performance index, the optimal control rate of the actuator is solved to perform tracking control on it.

[0066] In step 3, a state observer is designed to estimate the unknown state variables in the first step. To cancel the error caused by linearization and improve the response speed, an error feedback mechanism is constructed to improve the accuracy of the system.

[0067] The principle of the present invention is as follows: By analyzing the main generation mechanism of NOx in the boiler, the substances (HCN, O2, and N2) participating in its chemical reaction are obtained. Based on physical principles and thermodynamic laws, the relationship between the reaction substances and the coal feeding amount and air supply amount is obtained, and a controlled object model of the boiler system with NOx concentration as the output is established. Then, an MPC control system capable of accurately and quickly tracking the NOx target concentration curve is designed, and its optimized objective function is set as Equation (5); for the nonlinear boiler combustion control system, it is first discretized by the Euler method, and then through the mechanism of "rolling optimization - repeated execution" of predictive control, the Jacobian matrix of the state-space equation is solved at each sampling point to obtain the linearized system equation at each moment. Based on the optimized performance index, the optimal control rate of the actuator is solved to perform tracking control on it. Subsequently, to compensate for the unknown state variables and disturbance quantities in the system, a state observer is constructed to obtain the complete state information of the system. To cancel the errors caused by linearization and disturbance variables, an error feedback system is designed to improve the tracking accuracy of the boiler control system.

[0068] The present invention adopts a strategy of model predictive control + setpoint tracking control for NOx concentration. The NOx concentration, coal feeding amount, O2 concentration, and N2 concentration are selected as state variables, the intake air volume is used as the control quantity, and the NOx concentration is used as the output quantity. The continuous nonlinear boiler combustion system is discretized. By solving the Jacobian matrix of the state-space equation, the linearized equation is obtained. Since there are unknown variables and disturbances in the system, the optimal control sequence cannot be solved based on the set objective function. First, a state observer is designed. Through the NOx emission concentration and air inlet control quantity at the previous moment, the unknown state variables and disturbances in the system are estimated and input into the prediction model to solve the system output and optimal control quantity at the next moment. At the same time, based on the estimated state variables and disturbance quantities, the compensation control quantity u is solved through an error compensation mechanism. ′* The state observation value and disturbance estimation value are substituted into the optimization problem Equation (5), and the optimal control sequence u is solved by minimizing the objective function. *, select the optimal control signal at the current moment, apply it to the air inlet controller, and then repeat the above process at the next moment to continuously adjust the NOx output concentration of the system to track the target curve.

[0069] Embodiment: The present invention relates to a control method for a coal-fired boiler directly facing NOx emissions, which is specifically as follows:

[0070] First step: First, analyze the physical structure of the boiler. Figure 1 A simple structure diagram of a boiler combustion system is given. Then, based on the generation mechanism of NOx during the boiler combustion process, according to the laws of mass, energy conservation, and thermodynamics principles, with the chemical process of NOx generation as the main research object, a boiler combustion control model facing NOx emissions is established.

[0071] Equation (1) gives the competitive oxidation-reduction reaction mechanism of NOx during the combustion process of a coal-fired boiler. Determine its related variables according to the reaction rate of Equation (2): HCN concentration, O2 concentration, and N2 concentration. Select the NO x concentration, coal feeding amount M, O2 concentration, and N2 concentration as state variables x1, x2, x3, and x4. Set the air inlet volume of the boiler as the control variable u, and the output variable y of the system as the NO x concentration.

[0072] Since the corresponding substance concentrations cannot be directly controlled in the boiler control system, it is necessary to obtain the relationship between them and the controllable parameters of the boiler based on the physical laws and relevant thermodynamic laws. In the industrial coal-fired boiler system, the current main operable variables are the coal feeding amount and the air supply volume. In a pulverized coal boiler, the combustion rate of coal is equal to its coal feeding rate. Therefore, select the coal storage amount M in the boiler instead of the HCN concentration as the state variable. Since the coal feeding rate and the coal quality will fluctuate, set the disturbance amount as δ, and set the nitrogen content of standard coal as ρ. During the pulverized coal pyrolysis process, set the proportion of HCN in its volatile components as φ. The relationship between HCN and coal is:

[0073]

[0074] During the combustion process of a coal-fired boiler, the control system controls its NOx emission concentration by adjusting the total air inlet volume u. Set the volume of the boiler as V. During the generation process of NOx, O2 and N2 participate in its chemical reaction process. Here, set the proportionality coefficient of O2 at the air inlet as k in , and the proportionality coefficient at the air outlet as k ou . Since a part of O2 will be generated and consumed during the pulverized coal pyrolysis process, set their coefficients as k vo and k br respectively. According to the law of conservation of matter, its rate of change equation is given by Equation (4).

[0075] During the combustion process of a coal-fired boiler, the change in N2 concentration is caused by the reduction of NOx to N2. Therefore, only the N2 concentration rate equation in Equation (3) is considered.

[0076] Considering that there is a proportional relationship between the air supply volume and the control signal, and setting the control gain as k here, the state-space equation of the boiler combustion system is shown in Equation (7) as follows:

[0077]

[0078] Since in an actual industrial system, sensors and actuators are sampled and adjusted at a certain frequency, and digital signal processing technology is widely used in modern control systems, in order to conform to the actual application, the state-space model given in Equation (7) is discretized. The state-space equation after discretization by the Euler method is shown in Equation (8):

[0079]

[0080] Then, using the mechanism of "rolling optimization - repeated execution" of predictive control, a nonlinear quadratic dynamic matrix control algorithm is proposed. Among them, the nonlinear mechanism model will be linearized at each sampling moment and used to predict the future dynamics of the system. In this way, the problem is transformed into a quadratic programming problem that is repeatedly solved at each sampling moment. The Jacobian matrix is solved for the state-space equation at each sampling point to obtain the linearized system equation at each moment. Regarding the nonlinear function in Equation (8) as f, the goal of linearization is to transform the nonlinear system state-space equation into the linear state equation and output equation shown in Equation (9):

[0081]

[0082] Among them, A, B, and C are coefficient matrices, and D is the disturbance state vector.

[0083] The specific process is as follows:

[0084]

[0085] Since D(k) is an unknown state variable and disturbance in the system, the optimal control rate cannot be obtained by solving the objective function during the rolling optimization process. In order to estimate the value of D(k), a state observer is designed:

[0086]

[0087] Among them, are the state estimate value, disturbance estimate value, system output estimate value, and estimation error respectively. P0, L0, and L1 are constant matrices, and the disturbance estimation error can be made by The uniformly bounded convergence yields the values of three matrices, and the structure of the state observer is as shown in Figure 2 the following figure

[0088] Combined with the state observer, the present invention designs an MPC controller with error compensation, and its control structure diagram is as shown in Figure 3 the following figure, where y r (k + 1) is the tracking trajectory of the NOx concentration

[0089] First, the unknown state variables and disturbances at the (k + 1)-th moment are obtained through the state observer according to the system state and output at the k-th moment. Then, based on the error compensation principle, according to the control compensation amount u′ * (k) is solved

[0090] Then, the observed value is substituted into the optimization problem minJ(k) to solve the quadratic programming problem, and the optimal control sequence u * (k|k) is obtained. Then, the optimal control rate u(k) = u′ * (k) + u * (k|k)

[0091] In the next moment, the above steps are repeated to continuously and real-time perform rolling optimization to calculate the optimal control rate at each sampling moment

[0092] Since there are errors when linearizing the prediction model, an error feedback compensation mechanism is designed to compensate the prediction error of the previous moment into the prediction model of the next moment. The prediction error is

[0093] Figure 3 The following figure shows the structure diagram of the boiler combustion control system. This control method is a model predictive control method based on a state observer and an error compensation mechanism. The control system directly faces the NOx concentration, and makes the emission concentration of NOx track the target curve by controlling the air intake of the combustion boiler. The simulation results are as shown in Figure 4 、 5 the following figures. Among them, through the control method proposed by the present invention, the boiler combustion control system obtains the optimal value of the air intake at each sampling moment Figure 4 The following figure shows the change curve of the optimal air intake Figure 5 The following figure shows the effect diagram of the boiler NOx emission concentration tracking the target concentration curve. It can be seen from the figure that this method can make the NOx concentration quickly and effectively track the target curve, so as to realize the low-nitrogen combustion control of the boiler combustion

Claims

1. A control method for a coal-fired boiler directly facing NOx emissions, characterized in that, Including the steps: S1. Establish a boiler combustion control model for NOx emissions, y = x1 Among them, the state variable x1 represents the NOx concentration, with the unit of mg / m 3 , which is the main output index of the system; x2 represents the coal feeding rate, with the unit of kg / s, reflecting the fuel flow rate into the furnace; x3 represents the oxygen concentration, with the unit of kg / m 3 , used to describe the oxygen content during the combustion process; x4 represents the nitrogen concentration, with the unit of kg / m 3 , used to describe the nitrogen component in the combustion air; the control variable u(t) is the control input signal, which represents the air intake; the parameter ρ represents the mass fraction of nitrogen element in the coal, with the unit of kg / kg, reflecting the nitrogen content characteristics of the coal; φ represents the molar ratio of hydrogen cyanide in the volatile components during the pulverized coal pyrolysis process, with the unit of dimensionless, used to describe the potential ratio of fuel nitrogen migrating to NOx; the constant k is the control gain; T is used to describe the sensitivity of the combustion temperature, with the unit of K, to the NOx generation rate; in addition, δ(t) describes the dynamic disturbance or change rate of the coal feeding rate, with the unit of kg / s 2 , used to model the influence of external disturbances on the combustion process; S2. Design an MPC controller, linearize it by solving the Jacobian matrix based on the boiler combustion control model, construct an objective function, and solve and update the optimal control signal at each sampling moment; S3. Design a state observer to solve the unknown state variables and disturbances in the system through the state observer, and then improve the control performance of the system through error compensation.

2. The control method for a coal-fired boiler directly facing NOx emissions according to claim 1, wherein In step S1, the construction of the boiler combustion control model includes: based on the NOx generation mechanism in boiler combustion, taking the chemical reaction process of fuel NOx as the basis, selecting state variables, control variables, and output variables to establish a mechanism model for control; the main chemical reaction equations of fuel NOx are: HCN / NH3 + O2 → NO + … NO + HCN / NH3 → N2 + … During the combustion process, the nitrogen element in the fuel will first be converted into hydrocyanic acid or ammonia, and these nitrogen-containing substances are oxidized to form NOx under the action of oxygen. At the same time, some NOx will further react with HCN or NH3 to form nitrogen; the kinetic rate equations of the above two chemical reaction processes are described as follows: Among them, X represents the molar concentration of the corresponding component, unit: mol / m 3 ; x HCN is the molar concentration of hydrocyanic acid, unit: mol / m 3 ; is the molar concentration of oxygen, unit: mol / m 3 ;,, X NO is the molar concentration of nitric oxide, unit: (mol / m 3 ); b is the empirical reaction order of oxygen participating in the reaction. To further incorporate fuel characteristics into the model, assume that the coal storage in the boiler is M, unit: kg; the fuel used is standard coal, and its nitrogen content is defined as ρ, unit: kg / kg; during the pulverized coal pyrolysis process, the proportion of hydrocyanic acid in the volatile components is defined as φ; then the above reaction rate equations can be rewritten as: The O2 source in the boiler is the O2 in the air supply. Based on the principle of mass balance, the change rate of O2 is: Among them, V represents the volume of the boiler combustion space, unit: m 3 ; A in is the air intake of the boiler, unit: m 3 / s; k in is the molar conversion coefficient of oxygen in the incoming air, unit: mol / m 3 ; k vo is the molar coefficient of oxygen consumption during the release of volatile matter from pulverized coal pyrolysis, unit: mol / kg; k br is the molar coefficient of oxygen consumption during the combustion process, unit: mol / kg; k ou is the molar conversion coefficient of oxygen in the outgoing air, unit: mol / m 3 ; In this model, the NOx concentration, coal feeding amount, oxygen concentration, and nitrogen concentration are selected as the state variables of the system, and the intake air volume is set as the control variable to construct the controlled object model of the coal-fired boiler, thereby providing a mathematical basis for the subsequent design of the controller.

3. A coal-fired boiler control method directly facing NOx emissions according to claim 1, characterized in that, In step S2, select the NOx concentration emitted during boiler combustion as the controlled variable, and the total air intake of the boiler system as the control variable. Affect the combustion process of pulverized coal by adjusting the air intake, so that the NOx concentration dynamically tracks the set target value; To achieve the above control objective, adopt the model predictive control method, design the controller and construct the optimization performance index as follows: Among them, J(k) is the optimization performance index; y(k) is the NOx concentration output value at the current sampling moment; is the set NOx concentration target value.

4. A control method for a coal-fired boiler directly facing NOx emissions according to claim 1, characterized in that, In step S2, discretize the boiler combustion control model by the Euler method, and then through the "rolling optimization - repeated execution" mechanism of predictive control, solve the Jacobian matrix of the state space equation at each sampling point to obtain the linearized system equation at each moment. Based on the optimization performance index, solve the optimal control rate of the actuator to perform tracking control on it.

5. A control method for a coal-fired boiler directly facing NOx emissions according to claim 1, characterized in that, In step S2, construct an MPC controller with error compensation, specifically as follows: First, the unknown state variables and disturbances at the (k + 1)-th moment are obtained through a state observer based on the system state and output at the k-th moment. Then, based on the error compensation principle, according to Solve for the control compensation amount u′ * (k); where: C0 represents the output matrix of the system, reflecting the mapping relationship between the system state and the output; B represents the system control matrix, describing the influence of the control input on the state change; u′ * (k) is the control correction amount under error compensation, and its unit is the same as that of the control amount; is the external disturbance term estimated by the state observer, and its unit is the same as that of the output; e(e) represents the observation error at the k-th moment, defined as the deviation between the actual output value and the predicted value, and its unit is mg / m 3 ; Subsequently, the disturbance value estimated by the state observer is substituted into the optimization problem minJ(k), and the corresponding quadratic programming problem is solved to obtain the optimal control sequence u(k + 1) = u′ * (k) + u * (k + 1|k); Through the above steps, combined with the state observer and the rolling optimization strategy, the real-time and accurate control of NOx emissions from coal-fired boilers can be achieved; at each sampling moment, repeat the processes of state estimation, error compensation, and optimization solution to ensure that the control strategy can adapt to the working condition changes in real time, and improve the dynamic response ability and robustness of the control system.

6. A control method for a coal-fired boiler directly facing NOx emissions according to claim 1, characterized in that, In step S3, an error feedback compensation mechanism is constructed to compensate the prediction error at the previous moment into the boiler combustion control model at the next moment, where the prediction error is Among them, represents the prediction error at sampling time k, with the unit of mg / m 3 , reflecting the deviation between the actual output value and the model predicted output value; y(k) represents the actual measured output value of the system at time k, usually the NOx emission concentration, with the unit of mg / m 3 ; represents the predicted value of the NOx emission concentration by the system model at time k, with the unit of mg / m 3 ; By introducing and compensating the prediction error into the control model at the next moment, the prediction deviation caused by model uncertainty, system disturbance or operating condition fluctuation can be effectively reduced, thereby improving the stability and control effect of the boiler combustion control system, and achieving accurate tracking of NOx concentration and effective attainment of emission standards.

Images