Modeling and optimization of burner combustion efficiency and low NOx emission

　　Burner control system is the core of the burner to achieve the goal of high efficiency and low pollution, but due to the characteristics of the burner control object such as pure delay, large lag and multi-distribution parameters [1], the control process is difficult.It is an effective, simple and quick measure to solve this problem and achieve the goal of high efficiency and low pollution by using burner system combustion optimization control.At present, there are two kinds of combustion optimization control for burner optimization: numerical simulation and combustion modeling.The numerical simulation mainly adopts the computational fluid dynamics method, which shows a good effect in the simulated combustion process [2-3], but the calculation time of this method is long and it is difficult to apply it to the online optimization.Combustion modeling is the basis of combustion optimization control, and it has a good effect on real-time modeling and optimization [4]. Common modeling methods include mechanism modeling, modeling based on state estimation, regression analysis modeling, artificial neural network modeling, fuzzy modeling and hybrid modeling [5].Among them, artificial neural network modeling is the most widely used, but it has the following problems [6]:1) the training time of multi-layer forward neural network is long, which limits the linearity of combustion optimization system;2) the multi-layer forward neural network is sensitive to the incompleteness and errors of data samples;3) the multi-layer forward neural network has the problem of poor ability of overfitting and generalization.In recent years, the support vector regression machine (SVM) with excellent performance in nonlinear regression estimation has become a research hotspot at home and abroad.In this paper, based on the previous work, the least square support vector machine (LSSVM) is used to study the combustion modeling and optimization of burners.The author explored the amount of O2, CO in flue gas quantity feedback control scheme based on the combustion efficiency of give attention to two or morethings and nitrogen oxide (NOx) emissions two indicators, using LSSVM for burner nonlinear relation between input and output of the burner for regression modeling, and by using the genetic algorithm (ga) feedback gas amount of O2, CO quantity value optimization, realize the combustion optimizing control of burner.This paper intends to solve the following problems :1) to determine the combustion optimization control system and scheme of the burner, and to analyze the feedback control principle of the amount of O2 and amount of CO in the flue gas;2) LSSVM was used for burner combustion modeling, and LSSVM model parameters were optimized by particle swarm optimization algorithm;3) genetic algorithm is used to optimize the set values of feedbacks O2 and CO amount of flue gas;4) analyze the combustion modeling and optimization simulation results of the burner.Burner combustion control system and optimization program burner combustion control system is mainly divided into two categories: air volume control system and fuel control system.The two control systems cooperate with each other to ensure the full combustion of fuel entering the furnace, so that the combustion system achieves higher combustion efficiency and lower pollution emissions.The traditional control system only USES O2 as the adjustment and correction element of air volume. However, because O2 is easily affected by air leakage and fuel performance, it cannot directly reflect the quality of mixed working conditions in the system and the existence of local severe hypoxia areas, so O2 feedback alone cannot truly reflect the combustion state.The 1% CO component can represent the reducing atmosphere in the local area of the furnace. The CO content increases with the increase of the flame temperature and is not easily affected by air leakage.Therefore, CO feedback and O2 feedback are introduced to adjust and correct the air volume.The schematic diagram of burner combustion optimization control system is shown in figure 1.As shown in FIG. 1, air volume is more fully combined with fuel under the feedback regulation of O2 and CO volumes of flue gas, resulting in higher combustion efficiency and lower nox emissions.FIG. 1 schematic diagram of burner combustion optimization control system on the basis of FIG. 1, the lssvm-based combustion optimization control scheme designed in this paper is shown in FIG. 2.The scheme consists of two parts: combustion model and optimization algorithm.Among them, the combustion model to determine the combustion efficiency and NOx emissions n(n3) main factors Xi(I =1,2,...,n) as the input, with combustion efficiency and NOx emissions as the output, used to describe the combustion process and provide the prediction of combustion efficiency and NOx emissions in the optimization process.The input of the optimization algorithm is the output of the combustion model (combustion efficiency, NOx emission) and other input of the combustion model Xi(I =1,2...,n-2), the output is the optimized flue gas O2 quantity O2, CO quantity CO.The O2 and CO cyclic feedback was used for modeling and optimization until the O2 and CO quantities of flue gas reached the optimal output values :O2b and COb, as the given values of closed-loop control of the burner, participated in FIG. 1