Combustion optimization control method for coal-fired unit of thermal power plant

Abstract

The invention discloses a combustion optimization control method for a coal-fired unit of a thermal power plant. The method comprises the following steps: acquiring a historical data set on a coal mill; according to the coal mill heat balance principle and in combination with the historical data set, the coal quality real-time moisture is obtained; establishing a coal calorific value regression model and a coal volatile component regression model by using the historical data set based on a BP neural network; calculating the coal ash content according to the output of the coal calorific value regression model and the moisture calculation result; coal volatile components are obtained through the coal volatile component regression model; the method comprises the following steps: acquiring a historical data set of a coal-fired unit, performing steady-state processing, and combining coal quality real-time moisture, coal quality calorific value, coal quality ash content and coal quality volatile components as fired coal quality information; establishing a coal quality prediction model, predicting each parameter of the coal-fired unit, calculating an evaluation function according to a prediction result of each parameter, performing genetic algorithm optimization by taking a historical optimal value of a current working condition as an initial population to obtain an optimal working condition of an optimal evaluation function result, and outputting the optimal working condition to perform real-time control. The control precision can be improved.

Description

Technical Field

[0001] The invention relates to the technical field of thermal power generation, and in particular to a combustion optimization control method for a coal-fired unit in a thermal power plant. Background Art

[0002] With the rapid development of large-capacity, high-parameter coal-fired power generation units, the increasingly stringent standards for atmospheric pollutant emissions, and the rapid breakthroughs in the application of advanced measurement and control technologies, there is an urgent need for the in-depth development of power plant boiler combustion measurement and closed-loop optimization control technology. The in-furnace combustion control of coal-fired boilers is quite complex, especially in terms of high requirements for coal adaptability. There will be great differences in boiler characteristics during the combustion of different types of coal, such as the calorific value of coal, ignition distance, furnace temperature distribution characteristics, coking characteristics, etc., which will affect the rationality of the control outputs such as furnace air distribution, excess air coefficient, and cooling water distribution to varying degrees, thereby affecting the combustion control level of the unit, and further affecting the unit's operating economy and pollutant emission rate. Therefore, timely and accurate identification of the coal types used by each burner and real-time combustion optimization based on the identification results provide guarantees for the stable and economical combustion of coal-fired boilers under superimposed factors such as deep peak regulation and multiple coal types.

[0003] For intelligent combustion optimization control systems, the methods proposed in the prior art are as follows: the invention patent application with publication number CN116241905A discloses a combustion optimization control system based on multi-objective optimization, the principle of which is to establish a boiler combustion optimization model based on a large amount of test data, adopt an intelligent particle swarm optimization algorithm, and obtain the optimal input variable data in the boiler combustion model by calculating the objective function, thereby optimizing the combustion control system. Due to the complex and changeable types of boiler coal, this method cannot achieve refined control according to different coal quality information entering the furnace. The invention patent application with publication number CN116753538A discloses an intelligent combustion optimization control system and method for coal-fired power plants, including a preheating temperature determination system, a pulverized coal optimization power generation preheating system, and a boiler coal circulation system; the preheating temperature determination system is used to identify and distinguish the types of pulverized coal to be heated near the boiler of the coal-fired power plant, match the stored data, and determine the preheating temperature of the coal type; the pulverized coal optimization power generation preheating system, according to the temperature set by the preheating temperature determination system, the method controls the set temperature of the waste heat system according to the preheating temperature of the coal type to achieve combustion optimization control. The invention patent application with the publication number CN116123561A discloses a temperature optimization control method and system for boiler combustion, which uses an image acquisition device to acquire images of a target boiler, builds a three-dimensional fitting model based on the image acquisition results and the design size parameters, sets the temperature measurement points of the furnace wall through the three-dimensional fitting model to arrange the infrared acoustic temperature measurement device, and acquires temperature data through the infrared acoustic temperature measurement device to obtain a temperature data acquisition set and obtain real-time control parameters of the target boiler. This patent requires the addition of image acquisition equipment to obtain the real-time temperature field of the furnace, and the on-site equipment installation cost is high. In addition, due to the harsh working environment of the boiler furnace (high temperature and large airflow disturbance), the equipment reliability is not high. A real-time measurement method for the coal quality of a power station boiler with the publication number CN115753886A uses the combination of boiler positive balance and reverse balance to calculate the total heat, and realizes the measurement of the coal quality of the furnace based on the linear relationship between the received basic moisture content, ash content and calorific value of the coal, but the received basic moisture content, ash content and calorific value of the coal quality vary greatly with the type of coal, and it is impossible to obtain an accurate relationship between the components through a linear relationship. Summary of the invention

[0004] The technical problem to be solved by the present invention is to solve the problem of low combustion control accuracy of coal quality entering the furnace of a coal-fired unit.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A combustion optimization control method for a coal-fired unit in a thermal power plant, comprising:

[0007] S10, obtaining a historical data set including coal records on a coal mill and corresponding coal quality information and historical operation data of the coal mill;

[0008] S20, based on the thermal balance principle of coal mill and combined with historical data sets, obtains the real-time moisture content of coal;

[0009] S30, based on a multi-layer feedforward neural network, using historical data sets, establishing a coal calorific value regression model to obtain the coal calorific value;

[0010] S40, obtaining coal ash content according to the calorific value of the coal and the real-time moisture content of the coal;

[0011] S50, based on a multi-layer feedforward neural network, uses historical data sets to establish a coal volatile matter regression model to obtain coal volatile matter;

[0012] S60, performing steady-state processing on the historical data set of the coal-fired unit, and combining the steady-state processed data with the real-time moisture content, calorific value, ash content and volatile matter of the coal as the coal quality information for entering the furnace;

[0013] S70, using the coal quality information entering the furnace, based on a multi-layer feedforward neural network, establish a coal quality prediction model for predicting the NOx concentration at the SCR inlet, the desuperheating water at the A and B sides of the reheater, the exhaust gas temperature at the air preheater inlet, the water-cooled wall, the highest wall temperature of the final reheater, and the steam temperature at the final superheater and reheater outlet;

[0014] S80, using the NOx concentration at the SCR inlet, the desuperheating water at the A and B sides of the reheater, the exhaust gas temperature at the air preheater inlet, the water-cooled wall, the final superheater, the maximum wall temperature of the final reheater, and the steam temperature at the reheater outlet output by the coal quality prediction model as the evaluation function, setting the weights of each parameter;

[0015] S90, compare the predicted data of each parameter with the historical optimal data of each parameter, determine the comprehensive evaluation index of the current operating condition of each parameter, use the genetic optimization algorithm to optimize the operating oxygen content, the opening size of each secondary air and the burnout damper, and when optimizing the genetic algorithm, select the historical optimal operating conditions of the current load range and coal loading method from the historical data of each parameter as the initial population of the genetic algorithm, perform optimization operations, obtain the optimal optimized operating parameters of the current operating condition, and output them for control.

[0016] In one embodiment of the present invention, the real-time moisture content of coal is obtained by the following formula:

[0017]

[0018] in,

[0019] Where M m is the coal feeding capacity of the coal feeder, t / h; M fis the primary air volume at the coal mill inlet, t / h; t1 is the primary air temperature at the coal mill inlet, ℃; t2 is the air-powder mixture temperature at the coal mill outlet, ℃; K is the grindability coefficient; W is the coal mill power, KW; C1 is the specific heat of the desiccant at the coal mill inlet, KJ / kg*℃; C2 is the specific heat of the desiccant at the coal mill outlet, KJ / kg*℃; C lk is the specific heat of cold air, KJ / kg*℃; M ar is the coal received basis moisture, C rd is the dryer specific heat of coal, KJ / kg*℃; t lk is the temperature of the leaking cold air, ℃; t r is the temperature of coal entering the system, °C; ΔM is the heat consumed by evaporating water, q z and the amount of water evaporated per kg of raw coal in the coal mill, K lf is the air leakage coefficient of the coal mill; R 90 is the coal powder fineness, %; Q5 is the heat dissipation loss of the subsystem, KW; K nm M is the coefficient of coal mill input power converted into heat; ar The received base moisture; M mf It is the moisture content of coal powder.

[0020] In one embodiment of the present invention, a method for constructing a coal calorific value regression model comprises the following steps:

[0021] S31, create input matrix x [1] =[N,M], where N is the number of historical data sets and M is the number of input variables;

[0022] S32, random initialization: weight matrix w [1] =[M,q], error b [1] =[1,q],a [1] =x [1] w [1] +b [1] , where q is the number of nodes in the first hidden layer; a [1] is the calculation result matrix of the first layer of neural network;

[0023] S33, obtain the output z of the first hidden layer [1] =sigmod(a [1] );

[0024] S34, the output z of the first hidden layer [1] As the input parameter of the second hidden layer, and so on, get the output z of the fth hidden layer [4] , the output variable is 1, and the output z of the fth hidden layer [4] As the input variable, repeat S33 to calculate the predicted value of the output layer output target variable The predicted value is calculated using the L2 norm loss function The loss with the actual value y of the target variable;

[0025] S35, update each hidden layer w according to the L2 norm loss function S [i] , b [i] ;

[0026] S36, after completing the update, repeat S33 to S35, iterate n times, and calculate the predicted value The root mean square error (MSE) between the target variable and the actual value y.

[0027] In one embodiment of the present invention, the method for building the coal volatile matter regression model and the coal quality prediction model is the same as the method for building the coal calorific value regression model.

[0028] In one embodiment of the present invention, the coal ash content is obtained by the following formula:

[0029]

[0030] Among them, A ar Q is the basic ash content of the coal fed into the furnace, %, net,ar The minimum calorific value of the coal fed into the furnace, KJ / Kg, M ar It is the basic moisture content of the coal entering the furnace, %.

[0031] In one embodiment of the present invention, steady-state processing is performed on the historical data set of the coal-fired unit, including: judging whether the absolute value of the difference in coal feed rate of each pulverizer of two groups of operating conditions within one minute is greater than 10t / h, or whether the change in the opening of each secondary air is greater than 20, or whether the change in the opening of each electric adjustment damper at the inlet of each secondary air box is greater than 10, or the change in the opening of each SOFA burnout air is greater than 20, or the deviation of the speed set value of each flue gas recirculation fan inverter is greater than 10; when one of the above conditions is met, it is considered that the operating condition is in a fluctuating state, and the data within 15 minutes of the operating condition is eliminated.

[0032] In one embodiment of the present invention, a single parameter evaluation value = |parameter prediction value - historical optimal value| / historical optimal value is used as the single parameter evaluation value.

[0033] In one embodiment of the present invention, full assist is set for each evaluation parameter. Among them, p j is the single parameter evaluation value, h j is the weight of each parameter, j is the number of parameters, and J is the last parameter.

[0034] In one embodiment of the present invention, obtaining the optimal optimization operating parameters of the current working condition and outputting them for control includes:

[0035] S91, numerically encode the operating oxygen quantity, each secondary air, and each burnout air door opening;

[0036] S92, establishing a fitness function according to the total evaluation value;

[0037] S93, establishing an initial population: selecting historical optimal operating conditions of different load ranges and different coal mill combinations from the coal quality information entering the furnace as the initial population of the genetic algorithm;

[0038] S94, select individuals for replication: replicate and evolve the individuals ranked at the top according to a certain ratio in a single working condition of the initial working condition, and discard the individuals ranked at the bottom;

[0039] S95, crossover mutation: chromosome A and chromosome B cross at point c, and obtain new individuals A1 and B1 through genetic algorithm;

[0040] S96, individual evaluation: calculate the fitness of the new individual to determine whether the new working condition generated after genetic iteration is better than the previous working condition. If the set fitness is met, proceed to the next step. If not, repeat step S94 and perform the second round of iteration until the design requirements are met or the maximum number of iterations is reached.

[0041] S97, output results: output the optimized operating oxygen content, secondary air damper, and burnout air damper opening; realize real-time boiler combustion control.

[0042] In one embodiment of the present invention, the fitness function is obtained by the following formula:

[0043] In the formula, F is the fitness function and E is the total evaluation value.

[0044] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention proposes a combustion optimization control method for a coal-fired unit based on real-time monitoring of the quality of coal entering the furnace without adding measuring points or making technical modifications. Real-time monitoring of the quality of coal entering the furnace of each coal mill can guide combustion optimization control. The combustion optimization search method based on historical data is closer to actual operation and saves a lot of calculation time. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 The present invention is a flow chart of a method for optimizing combustion control of a coal-fired unit in a thermal power plant according to an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to facilitate those skilled in the art to understand the technical solution of the present invention, the technical solution of the present invention is further described in conjunction with the accompanying drawings of the specification.

[0047] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0048] See also Figure 1 As shown, the present invention provides a method for optimizing combustion control of a coal-fired unit in a thermal power plant, comprising:

[0049] S10, obtaining a historical data set including coal records on a coal mill and corresponding coal quality information and historical operation data of the coal mill.

[0050] In this embodiment, historical operation data of the direct-blowing coal mill are collected, including: current A, coal feed rate t / h, inlet wind pressure KPa, coal mill differential pressure KPa, inlet wind temperature ℃, outlet wind temperature ℃, inlet air volume t / h, load MW, 1st powder pipe wind speed m / s, 2nd powder pipe wind speed m / s, 3rd powder pipe wind speed m / s, 4th powder pipe wind speed m / s, and the coal feeding record and coal quality information of the coal mill, that is, the coal quality industrial analysis data, are added to the historical operation data of the coal mill at the corresponding historical time to establish a historical data set.

[0051] S20, based on the heat balance principle of direct-blowing coal mill and combined with historical data sets, obtains the real-time moisture content of coal.

[0052] In this embodiment, according to the principle of conservation of energy:

[0053] q gz +q lf +q nm +q r =q z +q jc +q2+q5;

[0054] In the formula,

[0055]

[0056] q z =ΔM(2491+1884t2-419t r );

[0057]

[0058] In the formula, q gz ,q lf ,q nm ,q r ,q z ,q jc, q2, q5 can be understood as different variables to simplify the formula. m is the coal feeding capacity of the coal feeder, t / h; M f is the primary air volume at the coal mill inlet, t / h; t1 is the primary air temperature at the coal mill inlet, ℃; t2 is the air-powder mixture temperature at the coal mill outlet, ℃; K is the grindability coefficient, W is the coal mill power, KW; C1 is the specific heat of the desiccant at the coal mill inlet, KJ / kg*℃; C2 is the specific heat of the desiccant at the coal mill outlet, KJ / kg*℃; C lk is the specific heat of cold air, KJ / kg*℃; M ar To receive the base moisture, C rd is the dryer specific heat of coal, KJ / kg*℃; t lk is the temperature of the leaking cold air, ℃; t r is the temperature of coal entering the system, °C; ΔM is the heat consumed by evaporating water, q z and the amount of water evaporated per kg of raw coal in the coal mill, K lf is the air leakage coefficient of the coal mill; R 90 is the coal powder fineness, %; Q5 is the heat dissipation loss of the subsystem, KW.

[0059] in,

[0060] Where: M ar The received base moisture; M mf It is the moisture content of coal powder.

[0061]

[0062] Coal powder moisture M mf It is generally about equal to the moisture content of the air dryer, and is usually lower, at around 15%.

[0063] The real-time moisture content of coal can be obtained through the following formula:

[0064]

[0065] In the formula, K nm It is the coefficient of converting the input power of coal mill into heat.

[0066] S30, based on a multi-layer feedforward neural network and using historical data sets, establishes a coal calorific value regression model to obtain the coal calorific value.

[0067] In one embodiment of the present invention, the historical data set is used as the input of a multi-layer feedforward neural network, and the multi-layer feedforward neural network is used as the prediction output.

[0068] S31, create input matrix x [1] =[N,M], where N is the number of historical data sets and M is the number of input variables;

[0069] S32, random initialization: weight matrix w [1] =[M,q], error b [1] =[1,q],a [1] =x [1] w [1] +b [1] , where q is the number of nodes in the first hidden layer; a [1] is the calculation result matrix of the first layer of neural network;

[0070] S33, obtain the output z of the first hidden layer [1] =sigmod(a [1] ); and, the sigmod function is as follows:

[0071]

[0072] Among them, there are f hidden layers in total.

[0073] S34, the output z of the first hidden layer [1] As the input parameter of the second hidden layer, and so on, get the output z of the fth hidden layer [4] , the output variable is 1, and the output z of the fth hidden layer [4] As the input variable, repeat S33 to calculate the predicted value of the output layer output target variable The predicted value is calculated using the L2 norm loss function The loss of the actual value y of the target variable; the calculation formula is:

[0074]

[0075] Where S is the L2 norm loss function, n is the number of iterations, and i is the i-th iteration.

[0076] S35, update each hidden layer w according to the L2 norm loss function S [i] , b [i] ;

[0077]

[0078] Where α is the hyperparameter learning rate, which is manually defined.

[0079] S36, after completing the update, repeat S33 to S35, iterate n times, and calculate the predicted value The root mean square error MSE with the actual value y of the target variable;

[0080]

[0081] S37, predicting the calorific value of coal based on the trained coal calorific value regression model.

[0082] In this embodiment, the network architecture of the multi-layer feedforward neural network has not been improved and is conventional. Only the L2 norm loss function calculation and the prediction value calculation are added to improve the prediction accuracy. The root mean square error (MSE) between the target variable and the actual value y.

[0083] S40, obtaining the ash content of the coal according to the calorific value of the coal and the real-time moisture content of the coal.

[0084] In this embodiment, according to the linear relationship between the received ash content of coal and the low calorific value and moisture content proposed in the paper "Research on the Relationship between Calorific Value, Moisture and Ash Content of Coal":

[0085]

[0086] Among them, A ar Q is the basic ash content of the coal fed into the furnace, %, net,ar The minimum calorific value of the coal fed into the furnace, KJ / Kg, M ar It is the basic moisture content of the coal entering the furnace, %.

[0087] S50, based on a multi-layer feedforward neural network, uses historical data sets to establish a coal volatile matter regression model to obtain coal volatile matter.

[0088] In this embodiment, step S30 is repeated, the steps of obtaining the model are the same, and the input data of the coal volatile matter regression model is also a historical data set. The difference is that the output data set is different, and the coal volatile matter regression model predicts the output of the coal volatile matter.

[0089] S60, obtaining a steady-state processing of the historical data set of the coal-fired unit, and combining the steady-state processed data with the real-time moisture content, calorific value, ash content and volatile matter of the coal as the coal quality information entering the furnace.

[0090] In this embodiment, the data in the historical data set of the coal-fired unit include data on parameters such as NOx concentration at the SCR inlet, cooling water on the A and B sides of the reheater, exhaust gas temperature at the air preheater inlet, water-cooled wall, final superheater, maximum wall temperature of the final reheater, and steam temperature at the reheater outlet.

[0091] The steady-state processing method is as follows: determine whether the absolute value of the difference in coal feed of each coal mill between the two groups of working conditions within one minute is greater than 10t / h, or whether the change in the opening of each secondary air is greater than 20, or whether the change in the opening of each electric adjustment damper at the inlet of each secondary air box is greater than 10, or whether the change in the opening of each SOFA burnout air is greater than 20, or whether the deviation of the speed setting value of each flue gas recirculation fan inverter is greater than 10; when one of the above conditions is met, it is considered that the working condition is in a fluctuating state, and the data within 15 minutes of the working condition are eliminated. The data within 15 minutes of the working condition is considered to be non-steady-state data and cannot be used for training.

[0092] S70, using the coal quality information entering the furnace, based on a multi-layer feedforward neural network, establishes a coal quality prediction model for predicting the NOx concentration at the SCR inlet, the desuperheating water at the A and B sides of the reheater, the exhaust gas temperature at the air preheater inlet, the water-cooled wall, the final superheater, the maximum wall temperature of the final reheater, and the steam temperature at the reheater outlet.

[0093] In this embodiment, the process of building the coal quality prediction model is the same as the step of obtaining the model in step S30, with the difference that the input data is the coal quality information entering the furnace, and the coal quality prediction model predicts the output of the NOx concentration at the SCR inlet, the cooling water on the A and B sides of the reheater, the exhaust gas temperature at the air preheater inlet, the water-cooled wall, the last stage superheater, the maximum wall temperature of the last stage reheater, and the steam temperature at the reheater outlet.

[0094] S80, with the NOx concentration at the SCR inlet, the desuperheating water at the A and B sides of the reheater, the exhaust gas temperature at the air preheater inlet, the water-cooled wall, the final superheater, the maximum wall temperature of the final reheater, and the steam temperature at the reheater outlet output by the coal quality prediction model as the evaluation function, the weights of each parameter are set.

[0095] In this embodiment, the current optimization condition is predicted, and each prediction result is compared with the historical optimal value. The specific calculation method is: single parameter evaluation value = | parameter prediction value - historical optimal value | / historical optimal value is used as a single parameter evaluation value, and full assist is set for each evaluation parameter.

[0096] Among them, p j is the single parameter evaluation value, h j is the weight of each parameter, j is the number of parameters, and J is the last parameter.

[0097] S90, compare the predicted data of each parameter with the historical optimal data of each parameter, determine the comprehensive evaluation index of the current operating condition of each parameter, use the genetic optimization algorithm to optimize the operating oxygen content, the opening size of each secondary air and the burnout damper, and when optimizing the genetic algorithm, select the historical optimal operating conditions of the current load range and coal loading method from the historical data of each parameter as the initial population of the genetic algorithm, perform optimization operations, obtain the optimal optimized operating parameters of the current operating condition, and output them for control.

[0098] In this embodiment, the following steps are specifically included:

[0099] S91, numerically encode the operating oxygen quantity, each secondary air, and each burnout air door opening;

[0100] S92, establish a fitness function according to the total evaluation value; the fitness function is derived from the total evaluation value and evaluates the quality of each individual (single working condition). Since the lower the total evaluation value, the better the working condition, the fitness function is obtained by the following formula:

[0101]

[0102] In the formula, F is the fitness function and E is the total evaluation value. The larger the fitness function value is, the better the degree of optimization of the working condition is considered to be.

[0103] S93, establishing an initial population: selecting the historical optimal operating conditions of different load ranges and different coal mill combinations from the coal quality information of the furnace as the initial population of the genetic algorithm; selecting the operating conditions in the same range as the current operating conditions according to the current operating conditions, and taking the operating oxygen content, secondary air door opening, and burnout air door opening of the optimal operating conditions as the initial population conditions, and performing replication and cross mutation;

[0104] S94, select individuals to replicate: replicate and evolve the individuals ranked at the top according to a certain ratio in the single working condition of the initial working condition, and discard the individuals ranked at the bottom; the present invention adopts the elite preservation method and the English preservation method as the optimal selection criterion to accelerate the convergence speed of the combustion optimization model. Among them, the probability of an individual being selected is the ratio of its fitness to the sum of the fitness of the population, which is obtained by the following formula:

[0105]

[0106] In the formula, f i is the intermediate variable for calculating the probability of an individual being selected in the population, F i is the fitness of the individual; P i is the probability of an individual being selected in the population, i is the number of individuals, and N is the population;

[0107] S95, crossover mutation: Chromosome A and chromosome B cross at point c, and obtain new individuals A1 and B1 through genetic algorithm; the genetic algorithm implementation process is as follows:

[0108] S951, set chromosome A and chromosome B;

[0109]

[0110] In the formula, a n is the nth gene in chromosome A, b n It is the nth gene in chromosome B.

[0111] S952, obtain the crossover of chromosome A and chromosome B at point c;

[0112]

[0113] In the formula, α′ c is the crossover result of chromosome A at point c, b′ c is the crossover result of chromosome B at point c, α is the hyperparameter learning rate, a c is the gene at c on chromosome A, b c It is the gene on chromosome B at position c.

[0114] S953, if the kth gene of chromosome A or chromosome B mutates, obtain new individuals A1 and B1 through genetic algorithm; specifically, taking the method of obtaining new individual A1 after chromosome A mutates as an example, the method of obtaining new individual B1 after chromosome B mutates is the same. After chromosome A mutates, the new individual A1 is obtained as shown in the following formula:

[0115]

[0116] In the formula, a′ k is the kth gene of chromosome A after mutation, a k is the kth gene of chromosome A, a max 、a min are the maximum and minimum values ​​of the genes of chromosome A respectively, f is the number of iterations of the genetic algorithm, F is the maximum number of iterations of the genetic algorithm, Q(f) is the intermediate variable for obtaining new individuals, and λ is a random number between 0 and 1.

[0117] In this embodiment, the crossover and mutation process of chromosomes in the genetic algorithm determines the diversity of the population during the iteration process. If the crossover probability is set too small, the probability of generating new individuals in the population will be reduced, while if it is set too large, the population diversity will be reduced. If the mutation probability is too small, the probability of generating new individuals in the group will also be reduced, while if the mutation probability is too large, the global optimization of the genetic algorithm will be transformed into a random search. Therefore, in order to ensure the diversity of the population and take into account the probability of new individuals, the crossover probability is usually set between 0.1 and 1, and the mutation probability is set between 0.01 and 0.1.

[0118] S96, individual evaluation: calculate the fitness of the new individual to determine whether the new working condition generated after genetic iteration is better than the previous working condition. If the set fitness is met, proceed to the next step. If not, repeat step S94 and perform the second round of iteration until the design requirements are met or the maximum number of iterations is reached.

[0119] S97, output results: output the optimized operating oxygen content, secondary air damper, and burnout air damper opening; realize real-time boiler combustion control.

[0120] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting from any point of view, and the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention, and any reference numerals in the claims should not be regarded as limiting the claims involved.

[0121] The above-described embodiments merely represent implementation methods of the invention. The protection scope of the present invention is not limited to the above-described embodiments. For those skilled in the art, several modifications and improvements may be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for optimizing combustion control of a coal-fired unit in a thermal power plant, characterized in that: include: S10, obtaining a historical data set including coal records on a coal mill and corresponding coal quality information and historical operation data of the coal mill; S20, based on the thermal balance principle of coal mill and combined with historical data sets, obtains the real-time moisture content of coal; S30, based on a multi-layer feedforward neural network, using historical data sets, establishing a coal calorific value regression model to obtain the coal calorific value; S40, obtaining coal ash content according to the calorific value of the coal and the real-time moisture content of the coal; S50, based on a multi-layer feedforward neural network, uses historical data sets to establish a coal volatile matter regression model to obtain coal volatile matter; S60, performing steady-state processing on the historical data set of the coal-fired unit, and combining the steady-state processed data with the real-time moisture content, calorific value, ash content and volatile matter of the coal as the coal quality information for entering the furnace; S70, using the coal quality information entering the furnace, based on a multi-layer feedforward neural network, establish a coal quality prediction model for predicting the NOx concentration at the SCR inlet, the desuperheating water at the A and B sides of the reheater, the exhaust gas temperature at the air preheater inlet, the water-cooled wall, the final superheater, the maximum wall temperature of the final reheater, and the steam temperature at the reheater outlet; S80, using the NOx concentration at the SCR inlet, the desuperheating water at the A and B sides of the reheater, the exhaust gas temperature at the air preheater inlet, the water-cooled wall, the final superheater, the maximum wall temperature of the final reheater, and the steam temperature at the reheater outlet output by the coal quality prediction model as the evaluation function, setting the weights of each parameter; S90, compare the predicted data of each parameter with the historical optimal data of each parameter, determine the comprehensive evaluation index of the current operating condition of each parameter, use the genetic optimization algorithm to optimize the operating oxygen content, the opening size of each secondary air and the burnout damper, and when optimizing the genetic algorithm, select the historical optimal operating conditions of the current load range and coal loading method from the historical data of each parameter as the initial population of the genetic algorithm, perform optimization operations, obtain the optimal optimized operating parameters of the current operating condition, and output them for control.

2. The method for optimizing combustion control of coal-fired units in a thermal power plant according to claim 1, characterized in that: The real-time moisture content of coal can be obtained through the following formula: in, Where M m is the coal feeding capacity of the coal feeder, t / h; M f is the primary air volume at the coal mill inlet, t / h; t1 is the primary air temperature at the coal mill inlet, ℃; t2 is the air-powder mixture temperature at the coal mill outlet, ℃; K is the grindability coefficient, W is the coal mill power, KW; C1 is the specific heat of the desiccant at the coal mill inlet, KJ / kg*℃; C2 is the specific heat of the desiccant at the coal mill outlet, KJ / kg*℃; C lk is the specific heat of cold air, KJ / kg*℃; M ar is the coal received basis moisture, C rd is the dryer specific heat of coal, KJ / kg*℃; t lk is the temperature of the leaking cold air, ℃; t r is the temperature of coal entering the system, °C; ΔM is the heat consumed by evaporating water, q z and the amount of water evaporated per kg of raw coal in the coal mill, K lf is the air leakage coefficient of the coal mill; R 90 is the coal powder fineness, %; Q5 is the heat dissipation loss of the subsystem, KW; K nm M is the coefficient of coal mill input power converted into heat; ar The received base moisture; M mf It is the moisture content of coal powder.

3. The method for optimizing combustion control of coal-fired units in a thermal power plant according to claim 1, characterized in that: The method for building a coal calorific value regression model includes the following steps: S31, create input matrix x [1] =[N,M], where N is the number of historical data sets and M is the number of input variables; S32, random initialization: weight matrix w [1] =[M,q], error b [1] =[1,q],a [1] =x [1] w [1] +b [1] , where q is the number of nodes in the first hidden layer; a [1] is the calculation result matrix of the first layer of neural network; S33, obtain the output z of the first hidden layer [1] =sigmod(a [1] ); S34, the output z of the first hidden layer [1] As the input parameter of the second hidden layer, and so on, get the output z of the fth hidden layer [4] , the output variable is 1, and the output z of the fth hidden layer [4] As the input variable, repeat S33 to calculate the predicted value of the output layer output target variable The predicted value is calculated using the L2 norm loss function The loss with the actual value y of the target variable; S35, update each hidden layer w according to the L2 norm loss function S [i] 、b [i] ; S36, after completing the update, repeat S33 to S35, iterate n times, and calculate the predicted value The root mean square error (MSE) between the target variable and the actual value y.

4. The method for optimizing combustion control of coal-fired units in a thermal power plant according to claim 2, characterized in that: The construction method of coal volatile matter regression model and coal quality prediction model is the same as the construction method of coal calorific value regression model.

5. The method for optimizing combustion control of coal-fired units in a thermal power plant according to claim 1, characterized in that: The coal ash content is obtained by the following formula: Among them, A ar Q is the basic ash content of the coal fed into the furnace, %, net,ar The minimum calorific value of the coal fed into the furnace, KJ / Kg, M ar It is the basic moisture content of the coal entering the furnace, %.

6. The method for optimizing combustion control of coal-fired units in a thermal power plant according to claim 1, characterized in that: Steady-state processing is performed on the historical data sets of coal-fired units, including: judging whether the absolute value of the difference in coal feed rate of each pulverizer between two groups of operating conditions within one minute is greater than 10t / h, or whether the change in the opening of each secondary air is greater than 20, or whether the change in the opening of each electric adjustment damper at the inlet of each secondary air box is greater than 10, or whether the change in the opening of each SOFA burnout air is greater than 20, or the deviation of the speed setting value of each flue gas recirculation fan inverter is greater than 10; when one of the above conditions is met, the operating condition is considered to be in a fluctuating state, and the data within 15 minutes of the operating condition are eliminated.

7. The method for optimizing combustion control of coal-fired units in a thermal power plant according to claim 1, characterized in that: Single parameter evaluation value = |parameter prediction value - historical optimal value| / historical optimal value is used as the single parameter evaluation value.

8. The method for optimizing combustion control of coal-fired units in a thermal power plant according to claim 7, characterized in that: Set full assists for each evaluation parameter. Among them, p j is the single parameter evaluation value, h j is the weight of each parameter, j is the number of parameters, and J is the last parameter.

9. The method for optimizing combustion control of coal-fired units in a thermal power plant according to claim 1, characterized in that: Obtain the optimal operating parameters for the current working conditions and output them for control, including: S91, numerically encode the operating oxygen quantity, each secondary air, and each burnout air door opening; S92, establishing a fitness function according to the total evaluation value; S93, establishing an initial population: selecting historical optimal operating conditions of different load ranges and different coal mill combinations from the coal quality information entering the furnace as the initial population of the genetic algorithm; S94, select individuals for replication: replicate and evolve the individuals ranked at the top according to a certain ratio in a single working condition of the initial working condition, and discard the individuals ranked at the bottom; S95, crossover mutation: chromosome A and chromosome B cross at point c, and obtain new individuals A1 and B1 through genetic algorithm; S96, individual evaluation: calculate the fitness of the new individual to determine whether the new working condition generated after genetic iteration is better than the previous working condition. If the set fitness is met, proceed to the next step. If not, repeat step S94 and perform the second round of iteration until the design requirements are met or the maximum number of iterations is reached. S97, output results: output the optimized operating oxygen content, secondary air damper, and burnout air damper opening; realize real-time boiler combustion control.

10. The method for optimizing combustion control of coal-fired units in a thermal power plant according to claim 9, characterized in that: The fitness function is obtained by the following formula: In the formula, F is the fitness function and E is the total evaluation value.

Images