Multi-target boiler combustion optimization control method under variable loads
By collecting and analyzing boiler combustion data, combining particle swarm algorithm and neural network control model, the combustion parameters of large coal-fired boilers are automatically adjusted, which solves the problem that traditional control means are difficult to achieve precise regulation, and improves the safety and operation efficiency of the unit. And optimized combustion efficiency and pollutant emissions.
Patent Information
- Application Number
- CN202410953728.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-16
- Publication Date
- 2024-11-08
AI Technical Summary
Traditional control methods are difficult to achieve precise regulation of the combustion state of large coal-fired boilers, resulting in problems such as instability in combustion, coking and over-temperature burning, affecting the safe and stable operation of the unit.
The data acquisition module collects temperature, pressure, oxygen content and combustion efficiency data during the boiler combustion process, establishes a multi-objective optimization model, and combines particle swarm algorithm and neural network control model to realize adaptive adjustment of the combustion process under variable loads and optimize, automatically adjust secondary air and other parameters.
It improves the safe and stable operation efficiency of the unit, reduces the impact of load changes on the system, and realizes the optimization of combustion efficiency and the control of pollutant emissions.
Smart Images

Figure CN118915441A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of boiler combustion control, and in particular to a multi-objective boiler combustion optimization control method under variable loads. Background Art
[0002] Multi-objective boiler combustion under variable load means that in the boiler system, the fuel combustion process needs to adapt to a variety of load changes and needs to optimize multiple performance indicators at the same time. In the face of changes under different loads, the boiler system needs to maintain a stable working state. Neural network control can better adapt to load changes and improve the stability and robustness of the system. In actual operation, in addition to meeting basic heat production needs, other performance indicators such as NOx, CO and other pollutant emissions, fuel flexibility, etc. need to be considered at the same time. Neural network control can achieve multi-objective optimization and achieve a balance between these performance indicators. Through neural network control, boiler combustion requirements can be adjusted more quickly, the response speed can be improved, and the impact of load changes on the entire system can be reduced.
[0003] The current market has put forward more stringent requirements on boiler combustion, such as high flexibility and ultra-low load operation. The boiler combustion process is characterized by strong coupling of multiple variables, nonlinearity, time variation and large delay. Traditional control methods are difficult to achieve precise regulation of the combustion state. At present, the adjustment of parameters such as secondary air of large coal-fired boilers mostly relies on the experience of on-site operators, which can easily cause problems such as unstable combustion, coking, and over-temperature, affecting the safe and stable operation of the unit. Summary of the invention
[0004] 1. Technical issues to be resolved
[0005] In view of the shortcomings of the prior art, the present invention provides a multi-objective boiler combustion optimization control method under variable loads, which has the characteristics of collecting temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data Rsxl in the boiler combustion process through a data acquisition module, and the data analysis module performs characteristic analysis based on the numerical values generated in the boiler combustion process, calculates the boiler combustion efficiency Glrs, emission concentration Pfnd and fuel consumption rate Rlxh, and establishes a multi-objective optimization model according to the above data, and the algorithm model module combines the established multi-objective optimization model with the particle swarm algorithm to achieve optimal combustion control under variable loads, and the control strategy module establishes a neural network control model, and establishes a neural network model by taking the temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data in the boiler combustion process as input variables, so as to achieve adaptive adjustment and optimization of the combustion process under variable loads, and automatically adjusts the parameters such as the secondary air of a large coal-fired boiler by establishing a boiler combustion control system, thereby improving the safety and stability of the unit. The operation efficiency and other advantages solve the above problems.
[0006] (II) Technical solution
[0007] To achieve the above object, the present invention provides the following technical solution: a multi-objective boiler combustion optimization control method under variable load, specifically comprising the following steps:
[0008] S1, collecting temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data Rsxl during boiler combustion through a data acquisition module, and transmitting the collected data to a data analysis module;
[0009] S2. Perform characteristic analysis based on the numerical values generated during boiler combustion through the data analysis module, calculate the boiler combustion efficiency Glrs, emission concentration Pfnd and fuel consumption rate Rlxh, and establish a multi-objective optimization model based on the above data;
[0010] S3. The established multi-objective optimization model is combined with the particle swarm algorithm through the algorithm model module to achieve optimal combustion control under variable loads;
[0011] S4. A neural network control model is established through the control strategy module. The temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data in the boiler combustion process are used as input variables to establish a neural network model to achieve adaptive adjustment and optimization of the combustion process under variable loads.
[0012] Preferably, the data acquisition module compiles the collected temperature data into a data set, and the data set is {Rswd n-9 , Rswd n-8 , Rswd n-7 , ..., Rswd n}.
[0013] Preferably, the data acquisition module compiles the collected pressure data into a data set, and the data set is {Rsyl n-9 , Rsyl n-8 , Rsyl n-7 , ...., Rsyl n}.
[0014] Preferably, the data acquisition module compiles the collected oxygen content data into a data set, and the data set is {Rshy n-9 , Rshy n-8 , Rshy n-7 , ..., Rshy n}.
[0015] Preferably, the data acquisition module compiles the collected combustion efficiency data into a data set, and the data set is {Rsxl n-9 , Rsxl n-8 , Rsxl n-7 , ..., Rsxl n}.
[0016] Preferably, the data analysis module performs characteristic analysis based on the values generated during the boiler combustion process and calculates the boiler combustion efficiency Glrs. The calculation formula is as follows:
[0017]
[0018] In the formula, Glrs represents boiler combustion efficiency, glsr represents boiler output heat, rlrz represents fuel calorific value, and pyss represents exhaust heat loss. The above numerical parameters are automatically obtained by the system.
[0019] Preferably, the data analysis module performs characteristic analysis based on the values generated during the boiler combustion process and calculates the emission concentration Pfnd using the following formula:
[0020]
[0021] In the formula, Pfnd represents the emission concentration, pfzl represents the emission mass, qttj represents the gas volume, and qtls represents the gas flow rate. The above values are automatically measured by the system.
[0022] Preferably, the data analysis module performs characteristic analysis based on the values generated during the boiler combustion process and calculates the fuel consumption rate Rlxh. The calculation formula is as follows:
[0023]
[0024] In the formula, Rlxh represents the fuel consumption rate, lzxh represents the fuel consumption, which is obtained through the fuel metering instrument, and time represents the time of fuel consumption, which is obtained through the system timestamp.
[0025] Preferably, the particle swarm algorithm used by the algorithm model module is explained as follows:
[0026] Assume that there are N particles in the group, the position of each particle i is represented by {x_i=x_i1,x_i2,x_i3,...,x_id}, where d represents the dimension of the problem, and the velocity is recorded as {v_i=v_i1,v_i2,v_i3,...,v_id}. Each particle has a fitness value, which represents the quality of the solution of the particle. The fitness value of particle i at time t is recorded as f_i(t). The particle position and velocity update rules are as follows:
[0027] Update speed: v_id(t+1)=w*v_id(t)+c1*rand()*(p_id-x_id+c2*rand()*(p_gd-x_id));
[0028] Where w represents the inertia weight, c1 and c2 are learning factors, rand() is a random number uniformly distributed between [0,1], p_id represents the optimal position of particle i, and p_gd represents the global optimal position of all particles;
[0029] Update position: x_id(t+1)=x_id(t)+v_id(t+1)
[0030] After the update, the new fitness value f_i(t+1) of particle i is calculated, and the individual optimal value p_id and the global optimal value p_gd in the group are updated.
[0031] Preferably, the control strategy module establishes a neural network control model, and establishes a neural network model by taking the temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data in the boiler combustion process as input variables, so as to realize adaptive adjustment and optimization of the combustion process under variable loads. The neural network control model designs a neural network controller to control a dynamic system with input m and output n, and the output of the neural network controller can be expressed as:
[0032]
[0033] Where m is the output of the neural network controller, f θ is a neural network model, θ is a parameter of the neural network, and the input of the neural network model includes the output n of the system and the derivative of n Control input m, derivative of m
[0034] Compared with the prior art, the present invention provides a multi-objective boiler combustion optimization control method under variable load, which has the following beneficial effects:
[0035] The present invention collects temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data Rsxl in the boiler combustion process through a data acquisition module. The data analysis module performs characteristic analysis based on the numerical values generated in the boiler combustion process, calculates the boiler combustion efficiency Glrs, emission concentration Pfnd and fuel consumption rate Rlxh, and establishes a multi-objective optimization model based on the above data. The algorithm model module combines the established multi-objective optimization model with the particle swarm algorithm to achieve optimal combustion control under variable loads. The control strategy module establishes a neural network control model, and establishes a neural network model by taking the temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data in the boiler combustion process as input variables, thereby achieving adaptive adjustment and optimization of the combustion process under variable loads. By establishing a boiler combustion control system, parameters such as secondary air of a large coal-fired boiler are automatically adjusted, thereby improving the safe and stable operation efficiency of the unit. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION
[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0038] At present, the adjustment of parameters such as secondary air of large coal-fired boilers mostly relies on the experience of on-site operators, which is easy to cause problems such as unstable combustion, coking, and overheating. Therefore, a multi-objective boiler combustion optimization control method under variable load is proposed. Please refer to Figure 1 , the method comprises the following steps:
[0039] S1, collecting temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data Rsxl during boiler combustion through a data acquisition module, and transmitting the collected data to a data analysis module;
[0040] The data acquisition module compiles the collected temperature data into a data set, which is {Rswd n-9 ,Rswd n-8 ,Rswd n-7 ,...,Rswd n}, compile the pressure data into a data set, which is {Rsyl n-9 ,Rsyl n-8,Rsyl n-7 ,...,Rsyl n}, the collected oxygen content data is compiled into a data set, which is {Rshy n-9 ,Rshy n-8 ,Rshy n-7 ,...,Rshy n}, compile the collected combustion efficiency data into a data set, which is {Rsxl n-9 ,Rsxl n-8 ,Rsxl n-7 ,...,Rsxl n}, each set of the above data contains 10 values;
[0041] S2. Perform characteristic analysis based on the numerical values generated during boiler combustion through the data analysis module, calculate the boiler combustion efficiency Glrs, emission concentration Pfnd and fuel consumption rate Rlxh, and establish a multi-objective optimization model based on the above data;
[0042] The boiler combustion efficiency calculation formula is:
[0043]
[0044] Calculating the combustion efficiency can help evaluate the energy utilization of the boiler system and understand the energy conversion efficiency of the fuel. By optimizing the combustion efficiency, it can reduce energy waste, reduce fuel consumption, achieve the goal of energy conservation and emission reduction, and reduce the emission of emissions such as carbon dioxide. In the formula, Glrs represents the boiler combustion efficiency, glsr represents the boiler output heat, rlrz represents the fuel calorific value, and pyss represents the exhaust heat loss. The above numerical parameters are automatically obtained by the system. High combustion efficiency helps to reduce the occurrence of problems such as vicious combustion or carbon deposition, thereby reducing damage to boiler equipment, extending the service life of the equipment, and reducing maintenance and replacement costs;
[0045] Emission concentration calculation formula:
[0046]
[0047] By calculating the emission concentration, we can understand the concentration of the emission substances in the production process, help monitor and evaluate the emission situation of the enterprise, find problems in time, take corresponding measures to ensure that the emission meets the environmental protection standards and protect the environmental quality. In the formula, Pfnd represents the emission concentration, pfzl represents the mass of the emission, qttj represents the gas volume, and qtls represents the gas flow rate. The above values are automatically measured by the system. By calculating the emission concentration, we can understand the source and composition of the emission, adjust the production process and equipment operation in a targeted manner, optimize the production process, and reduce the emission concentration;
[0048] Fuel consumption rate calculation formula:
[0049]
[0050] The fuel consumption rate reflects the combustion efficiency and operating status of the equipment. By calculating the fuel consumption rate, we can timely understand the equipment operation status, find abnormal equipment operation, perform maintenance in advance, extend the equipment life, and reduce maintenance costs. In the formula, Rlxh represents the fuel consumption rate, lzxh represents the fuel consumption, which is obtained through the fuel metering instrument, and time represents the time of fuel consumption, which is obtained through the system timestamp. By calculating the fuel consumption rate, enterprises can find energy waste or inefficiency problems in production, and then adjust the production process, optimize the process, and improve production efficiency and energy saving level;
[0051] S3. The established multi-objective optimization model is combined with the particle swarm algorithm through the algorithm model module to achieve optimal combustion control under variable loads;
[0052] The particle swarm algorithm used in the algorithm model module is explained as follows:
[0053] Assume that there are N particles in the group, the position of each particle i is represented by {x_i=x_i1,x_i2,x_i3,...,x_id}, where d represents the dimension of the problem, and the velocity is recorded as {v_i=v_i1,v_i2,v_i3,...,v_id}. Each particle has a fitness value, which represents the quality of the solution of the particle. The fitness value of particle i at time t is recorded as f_i(t). The particle position and velocity update rules are as follows:
[0054] Update speed: v_id(t+1)=w*v_id(t)+c1*rand()*(p_id-x_id+c2*rand()*(p_gd-x_id));
[0055] Where w represents the inertia weight, c1 and c2 are learning factors, rand() is a random number uniformly distributed between [0,1], p_id represents the optimal position of particle i, and p_gd represents the global optimal position of all particles;
[0056] Update position: x_id(t+1)=x_id(t)+v_id(t+1)
[0057] After the update, the new fitness value f_i(t+1) of particle i is calculated, and the individual optimal value p_id and the global optimal value p_gd in the group are updated;
[0058] S4. A neural network control model is established through a control strategy module. The temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data of the boiler combustion process are used as input variables to establish a neural network model, so as to realize adaptive adjustment and optimization of the combustion process under variable loads.
[0059] The neural network control model designs a neural network controller to control a dynamic system with input m and output n. The output of the neural network controller can be expressed as:
[0060]
[0061] Where m is the output of the neural network controller, f θ is a neural network model, θ is a parameter of the neural network, and the input of the neural network model includes the output n of the system and the derivative of n Control input m, derivative of m Neural network is a powerful nonlinear modeling tool used to deal with complex nonlinear systems. The design of neural network controller can handle systems that are difficult to describe with traditional control methods. It has better adaptability and flexibility. Neural network controller realizes control through parallel computing, has good real-time and rapid response capabilities, can make accurate decisions in a short time, and is suitable for systems that require high-speed control and high real-time requirements.
[0062] By building a boiler combustion control system, there is no need to rely on the experience of on-site operators when adjusting parameters such as the secondary air of large coal-fired boilers. Automatic optimization of control parameter adjustments can be performed, thereby improving the safe and stable operation efficiency of the unit.
[0063] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A multi-objective boiler combustion optimization control method under variable load, characterized by: The specific steps include: S1, collecting temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data Rsxl during boiler combustion through a data acquisition module, and transmitting the collected data to a data analysis module; S2. Perform characteristic analysis based on the numerical values generated during boiler combustion through the data analysis module, calculate the boiler combustion efficiency Glrs, emission concentration Pfnd and fuel consumption rate Rlxh, and establish a multi-objective optimization model based on the above data; S3. The established multi-objective optimization model is combined with the particle swarm algorithm through the algorithm model module to achieve optimal combustion control under variable loads; S4. A neural network control model is established through the control strategy module. The temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data in the boiler combustion process are used as input variables to establish a neural network model to achieve adaptive adjustment and optimization of the combustion process under variable loads.
2. The multi-objective boiler combustion optimization control method under variable load according to claim 1 is characterized in that: The data acquisition module compiles the collected temperature data into a data set, which is {Rswd n-9 , Rswd n-8 , Rswd n-7 , …, Rswd n }.
3. The multi-objective boiler combustion optimization control method under variable load according to claim 2 is characterized in that: The data acquisition module compiles the collected pressure data into a data set, and the data set is {Rsyl n-9 , Rsyl n-8 , Rsyl n-7 , ..., Rsyl n }.
4. The multi-objective boiler combustion optimization control method under variable load according to claim 3 is characterized in that: The data acquisition module compiles the collected oxygen content data into a data set, which is {Rshy n-9 , Rshy n-8 , Rshy n-7 , ..., Rshy n }.
5. The multi-objective boiler combustion optimization control method under variable load according to claim 4 is characterized in that: The data acquisition module compiles the collected combustion efficiency data into a data set, which is {Rsxl n-9 , Rsxl n-8 , Rsxl n-7 , ..., Rsxl n }.
6. The multi-objective boiler combustion optimization control method under variable load according to claim 5 is characterized in that: The data analysis module performs characteristic analysis based on the values generated during the boiler combustion process and calculates the boiler combustion efficiency Glrs. The calculation formula is as follows: In the formula, Glrs represents boiler combustion efficiency, glsr represents boiler output heat, rlrz represents fuel calorific value, and pyss represents exhaust heat loss. The above numerical parameters are automatically obtained by the system.
7. The multi-objective boiler combustion optimization control method under variable load according to claim 6 is characterized in that: The data analysis module performs characteristic analysis based on the values generated during the boiler combustion process and calculates the emission concentration Pfnd. The calculation formula is as follows: In the formula, Pfnd represents the emission concentration, pfzl represents the emission mass, qttj represents the gas volume, and qtls represents the gas flow rate. The above values are automatically measured by the system.
8. The multi-objective boiler combustion optimization control method under variable load according to claim 7 is characterized in that: The data analysis module performs characteristic analysis based on the values generated during the boiler combustion process and calculates the fuel consumption rate Rlxh. The calculation formula is as follows: In the formula, Rlxh represents the fuel consumption rate, lzxh represents the fuel consumption, which is obtained through the fuel metering instrument, and time represents the time of fuel consumption, which is obtained through the system timestamp.
9. The multi-objective boiler combustion optimization control method under variable load according to claim 8, characterized in that: The particle swarm algorithm used in the algorithm model module is explained as follows: Assume that there are N particles in the group, the position of each particle i is represented by {x_i=x_i1,x_i2,x_i3,...,x_id}, where d represents the dimension of the problem, and the velocity is recorded as {v_i=v_i1,v_i2,v_i3,...,v_id}. Each particle has a fitness value, which represents the quality of the solution of the particle. The fitness value of particle i at time t is recorded as f_i(t). The particle position and velocity update rules are as follows: Update speed: v_id(t+1)=w*v_id(t)+c1*rand()*(p_id-x_id+c2*rand()*(p_gd-x_id)); Where w represents the inertia weight, c1 and c2 are learning factors, rand() is a random number uniformly distributed between [0, 1], p_id represents the optimal position of particle i, and p_gd represents the global optimal position of all particles; Update position: x_id(t+1)=x_id(t)+v_id(t+1) After the update, the new fitness value f_i(t+1) of particle i is calculated, and the individual optimal value p_id and the global optimal value p_gd in the group are updated.
10. The multi-objective boiler combustion optimization control method under variable load according to claim 9, characterized in that: The control strategy module establishes a neural network control model. By taking the temperature data Rswd, pressure data Rsyl, oxygen content data Rshy and combustion efficiency data in the boiler combustion process as input variables, the neural network model is established to realize adaptive adjustment and optimization of the combustion process under variable loads. The neural network control model designs a neural network controller to control a dynamic system with input m and output n. The output of the neural network controller can be expressed as: Where m is the output of the neural network controller, f θ is a neural network model, θ is a parameter of the neural network, and the input of the neural network model includes the output n of the system and the derivative of n Control input m, derivative of m
