Combustion efficiency control method of circulating fluidized bed boiler based on adaptive genetic algorithm

In order to improve the traditional boiler combustion efficiency control method, which has the problem of low efficiency after control, a new combustion efficiency control method of circulating fluidized bed boiler based on adaptive genetic algorithm is proposed in this paper. Firstly, the iterative principle of adaptive genetic algorithm is analyzed; Secondly, the combustion process model of circulating fluidized bed boiler is constructed by BP neural network method; Finally, based on the above model, the optimal control objective function of adaptive genetic algorithm is constructed, and the objective function is solved according to the iterative principle of adaptive genetic algorithm to complete the control of combustion efficiency of circulating fluidized bed boiler. The experimental results show that compared with the traditional control method, the combustion efficiency of the boiler controlled by this method is always more than 98%.


Introduction
Thanks to the development of production technology, industrial production technology has also developed rapidly, especially the indispensable combustion boiler in industrial production [1] .Circulating fluidized bed boiler is a clean combustion boiler with high efficiency combustion and low pollution.The combustion mode of this boiler is between coal layer combustion and powder combustion, which can greatly improve the efficiency of coal combustion, and the maximum combustion efficiency can reach about 97% [2][3][4] .Moreover, in the fluidized bed boiler, the combustion performance of coal is good, the efficient combustion can be completed without adding too many auxiliary materials, and has good desulfurization effect, which can produce better sulfur oxides and nitrogen oxides and reduce the pollution to the environment.Although the circulating fluidized bed boiler has good combustion performance, the control of combustion efficiency has become a problem to be solved in the use of circulating fluidized bed boiler.Only by controlling the combustion efficiency of circulating fluidized bed boiler, can the combustion performance of circulating fluidized bed boiler be improved comprehensively.Reference [5] proposes a boiler combustion efficiency control method based on optimized K-means clustering algorithm.The improved k-means clustering algorithm is used to cluster the boiler combustion parameters to obtain the best combustion parameters.Under the optimal combustion parameters, the optimal control function of combustion efficiency is constructed with the maximum combustion efficiency as the goal.Bayesian least square method and support vector machine are used to solve the objective function, so as to complete the control of boiler combustion efficiency.However, this method can not improve the combustion response of the boiler.Reference [6] proposes a boiler combustion efficiency control method based on fuzzy self optimization algorithm.To begin with, a mathematical model for combustion control is developed, and an adaptive optimization method is utilized to regulate the boiler combustion parameters with the aim of achieving efficient and effective combustion control.However, this method yields control results that deviate noticeably from the actual results, leading to unsatisfactory control precision.In contrast, reference [7] puts forth a boiler combustion efficiency control strategy that employs a combination of quantum genetic algorithm and neural network, which has been shown to offer superior performance.This method first collects the characteristic data of boiler combustion, constructs the boiler combustion model based on neural network, and uses quantum genetic algorithm to optimize the damper opening and burner yaw angle of boiler, so as to complete the control of boiler combustion efficiency.However, the control delay of this method is high.In order to address the issues associated with the aforementioned control methods, we propose a combustion efficiency control method for circulating fluidized bed boilers based on an adaptive genetic algorithm.

Combustion efficiency control method of circulating fluidized bed boiler based on adaptive genetic algorithm
From the perspective of combustion efficiency control of circulating fluidized bed boiler, the control problem can be transformed into an optimization problem, that is, given a certain input or constraint, the optimization algorithm is introduced to optimize the objects to be controlled.The control diagram of adaptive genetic algorithm is shown in Figure 1.  1, the constraint condition is the fitness function in the iterative process of the adaptive genetic algorithm, and the operation module in the adaptive genetic algorithm is selection, crossover and mutation.The optimization object is the combustion parameters of the circulating fluidized bed boiler, so as to achieve the purpose of efficiency control.Therefore, it is not difficult to find that the adaptive genetic algorithm is somewhat similar to the PID control idea, and can control the current and future state of the control object [8][9][10] .However, the essence of adaptive genetic algorithm is random algorithm.Compared with traditional classical algorithm, the coupling degree of adaptive genetic algorithm is very loose, which increases the instability of its optimization results.The connection of adaptive genetic algorithm to circulating fluidized bed boiler is mainly reflected in two aspects: fitness function constraint and combustion parameter control, so as to improve the effectiveness of boiler combustion efficiency control [11] .

Modeling of combustion process in circulating fluidized bed boiler
In the process of using adaptive genetic algorithm to control the combustion efficiency of circulating fluidized bed boiler, it is necessary to calculate the combustion efficiency and NOx emission of the boiler as the control parameters.The combustion process of circulating fluidized bed boiler is a nonlinear and strong coupling process, and the traditional ordinary modeling method is difficult to meet the requirements of combustion process analysis.Therefore, in this combustion process modeling, BP neural network is used to construct the combustion process model of circulating fluidized bed boiler [12] .In this study, the input data of BP neural network model is the oxygen and coal feed of boiler combustion, and the output of the model is the sum of heat loss and NOx emission, so as to build the completed BP neural network model, as shown in Figure 2.After completing the construction of BP neural network model, the strict ASME standard is used to calculate the combustion efficiency of circulating fluidized bed boiler, so as to obtain accurate combustion efficiency calculation results.When BP neural network is used to build the boiler combustion process model, it is not necessary to directly calculate the boiler combustion efficiency, but only to calculate the parameters that can reflect the boiler combustion efficiency.Therefore, the sum of dry flue gas loss and unburned carbon heat loss of circulating fluidized bed boiler is selected as the relevant parameters [13][14][15] .
In the formula, l y represents the amount of oxygen, py t represents the temperature of flue gas discharged from the boiler, and jf t represents the temperature of boiler inlet air.
The calculation formula of the sum loss of unburned carbon heat is as follows: 2 0.9 0.1 4.1868 7850 100 100 In the formula, f h represents the ash content in the boiler, r d represents the low heating capacity of the boiler, t f represents the carbon content in the boiler fly ash, and z h represents the carbon content in the bottom slag of the boiler.Step 1: set the parameters of the adaptive genetic algorithm.The main parameters include the number of individuals, the number of iterations and the length of individuals.In the process of parameter setting, it should be noted that once the individual length increases, although the accuracy of the calculation results is improved, the overall calculation time will also increase relatively.Therefore, it is necessary to select the individual length appropriately.In this calculation process, considering the calculation requirements and the control accuracy requirements of boiler combustion efficiency, the individual length is set to 10, the number of iterations is set to 20, and the number of groups is also set to 20, which can meet the requirements of population information calculation.

Combustion efficiency control based on adaptive genetic algorithm
Step 2: using binary coding method, convert the objective function solution data into the coding string of adaptive genetic algorithm, and generate the initial calculation population.
In the formula, loss P represents the controllable combustion loss of circulating fluidized bed boiler and max c represents the maximum combustion efficiency.After completing the construction of fitness function, judge whether the fitness results of individuals in the population meet the standard of calculation requirements.If they meet the requirements, the optimal control of boiler combustion efficiency can be carried out, otherwise genetic operation can be carried out.
Step 4: sort the individual fitness of the population, and calculate the fitness selection probability of all individuals by random traversal: In the formula, represents the fitness calculation result, and n represents the number of individuals in all populations.Through the calculation of formula ( 5), the cumulative probability of each individual can be obtained, so as to sort the individual fitness at an equal distance.
Step 5: select the single point crossing method to cross the individuals in the population.The cross probability is 0.85.After cross processing, new individuals can be obtained.
Step 6: Based on the new individuals obtained by the above cross operation, exchange 0 and 1 in binary, and the mutation probability is 0.05.After mutation, new population individuals will also be generated for subsequent calculation.
Step 7: determine whether the new population generated by the above cross meets the iteration termination conditions.If the fitness calculation results meet the requirements and meet the termination conditions, complete the control of combustion efficiency.

Experimental verification
In order to verify the practical application effect of the proposed circulating fluidized bed combustion efficiency control method based on adaptive genetic algorithm, simulation and comparative verification experiments are carried out.The fluidized bed boiler selected in this experiment is 1025t/h boiler.The parameters of this circulating fluidized bed boiler are shown in Table 1.300MW According to the selected circulating fluidized bed boiler and the relevant parameters of the boiler, under the normal operation state of the boiler, take 24 hours as an operation cycle and sample the combustion parameter data at an interval of 1 hour for subsequent experiments.In this experimental study, we compare the efficacy of the method proposed in this paper with that of the methods presented in reference [6] and reference [7] , using the response results of boiler steam, bed temperature, and combustion efficiency control as the basis for comparison.

Boiler steam and bed temperature response results
To evaluate the control performance of this method in terms of boiler combustion efficiency, a comparison is made between the responses of boiler steam and bed temperature before and after implementing the optimal control using adaptive genetic algorithm.The comparison results, depicted in Figure 5 and Figure 6, illustrate the changes in boiler steam pressure and bed temperature response as a result of the optimization control.Comparison results of boiler bed temperature response before and after optimization From the comparison results of Fig. 5 and Fig. 6, it can be seen that the overall pressure and bed temperature response of circulating fluidized bed boiler can quickly reach the standard level after the optimized control of this method, which fully shows that this method has strong boiler combustion control ability.

Effectiveness of boiler combustion efficiency control
The comparison results of the combustion efficiency control effectiveness of this method with those of references [6] and [7] are shown in Table 2. Reference [6] method Reference [7]  Based on the comparison results of boiler combustion efficiency control effectiveness presented in Table 2, it is evident that after the optimization control using this method, the combustion efficiency of the circulating fluidized bed boiler is consistently maintained above 98%.In contrast, the maximum combustion efficiency of the boiler under the control of two traditional comparison methods does not exceed 90%.Therefore, this fully demonstrates that this method can effectively control the combustion efficiency of the circulating fluidized bed boiler.

Conclusion
In order to improve the combustion performance of a circulating fluidized bed boiler, a combustion efficiency control method based on adaptive ant colony algorithm is proposed.The performance of this method has been verified theoretically and experimentally.When controlling the combustion efficiency of the circulating fluidized bed boiler, this method exhibits high combustion efficiency.In comparison to the control method utilizing optimized K-means clustering algorithm and fuzzy self-optimization algorithm, the combustion efficiency of the circulating fluidized bed boiler consistently exceeds 98% when under the influence of this proposed method.This clearly demonstrates that the control method based on adaptive genetic algorithm is better suited to fulfill the requirements for controlling the combustion efficiency of the circulating fluidized bed boiler.

Figure 1 .
Figure 1.Control chart of adaptive genetic algorithm In the control diagram shown in Figure1, the constraint condition is the fitness function in the iterative process of the adaptive genetic algorithm, and the operation module in the adaptive genetic algorithm is selection, crossover and mutation.The optimization object is the combustion parameters of the circulating fluidized bed boiler, so as to achieve the purpose of efficiency control.Therefore, it is not difficult to find that the adaptive genetic algorithm is somewhat similar to the PID control idea, and can control the current and future state of the control object[8][9][10] .However, the essence of adaptive genetic algorithm is random algorithm.Compared with traditional classical algorithm, the coupling degree of adaptive genetic algorithm is very loose, which increases the instability of its optimization results.The connection of adaptive genetic algorithm to circulating fluidized bed boiler is mainly reflected in two aspects: fitness function constraint and combustion parameter control, so as to improve the effectiveness of boiler combustion efficiency control[11] .

Figure 2 .
Figure 2. Neural network model of fluidized bed boiler combustion process The calculation formula of dry flue gas loss is as follows: 2 1 100 21 0.4 3.55 (0.1 ) ( ) 21 10000 Circulating fluidized bed boiler is a clean combustion boiler vigorously developed in China.With the gradual increase of application scope, the research on combustion control of circulating fluidized bed boiler is increasing, and the research on combustion chemical reaction, combustion efficiency and combustion gas flow is deepening.Reasonably controlling the combustion efficiency of circulating fluidized bed boiler can greatly improve the performance of circulating fluidized bed boiler.The structure of circulating fluidized bed boiler is shown in Figure 3.

Figure 3 .
Figure 3. Structure of circulating fluidized bed boiler

Figure 4 .
Figure 4. Combustion efficiency control flow of adaptive genetic algorithm The specific steps of using adaptive genetic algorithm to optimize and control the combustion efficiency of circulating fluidized bed boiler are as follows:Step 1: set the parameters of the adaptive genetic algorithm.The main parameters include the number of individuals, the number of iterations and the length of individuals.In the process of parameter setting, it should be noted that once the individual length increases, although the accuracy of the calculation results is improved, the overall calculation time will also increase relatively.Therefore, it is necessary to select the individual length appropriately.In this calculation process, considering the calculation requirements and the control accuracy requirements of boiler combustion efficiency, the individual length is set to 10, the number of iterations is set to 20, and the number of groups is also set to 20, which can meet the requirements of population information calculation.Step 2: using binary coding method, convert the objective function solution data into the coding string of adaptive genetic algorithm, and generate the initial calculation population.Step 3: Based on the above constructed population, calculate the fitness function of a single individual in the population, and transform the problem of solving the objective function into a problem of fitness calculation.The objective function of boiler combustion efficiency control is the maximum combustion efficiency, so the fitness function can be expressed as:

Step 3 :
Figure 4. Combustion efficiency control flow of adaptive genetic algorithmThe specific steps of using adaptive genetic algorithm to optimize and control the combustion efficiency of circulating fluidized bed boiler are as follows:Step 1: set the parameters of the adaptive genetic algorithm.The main parameters include the number of individuals, the number of iterations and the length of individuals.In the process of parameter setting, it should be noted that once the individual length increases, although the accuracy of the calculation results is improved, the overall calculation time will also increase relatively.Therefore, it is necessary to select the individual length appropriately.In this calculation process, considering the calculation requirements and the control accuracy requirements of boiler combustion efficiency, the individual length is set to 10, the number of iterations is set to 20, and the number of groups is also set to 20, which can meet the requirements of population information calculation.Step 2: using binary coding method, convert the objective function solution data into the coding string of adaptive genetic algorithm, and generate the initial calculation population.Step 3: Based on the above constructed population, calculate the fitness function of a single individual in the population, and transform the problem of solving the objective function into a problem of fitness calculation.The objective function of boiler combustion efficiency control is the maximum combustion efficiency, so the fitness function can be expressed as: max max , ( ) 0, loss loss loss c P P c fit P else

Figure 5 .
Figure 5.Comparison results of boiler steam pressure before and after optimization control

Figure 6 .
Figure 6.Comparison results of boiler bed temperature response before and after optimization From the comparison results of Fig.5and Fig.6, it can be seen that the overall pressure and bed temperature response of circulating fluidized bed boiler can quickly reach the standard level after the optimized control of this method, which fully shows that this method has strong boiler combustion control ability.

Table 1 .
Parameters of circulating fluidized bed boiler

Table 2 .
Effectiveness of boiler combustion efficiency control