火电机组SCR喷氨量的智能优化控制
发布时间：2018-09-11 14:45
【摘要】：随着我国经济的快速增长,火电厂排放的氮氧化物(NO_x)也逐年增多,对人类健康和自然环境造成一定威胁。近年来,国家对燃煤电厂NO_x排放做出更严格的规定,因此,研究安全、环保、经济的氮氧化物脱除技术非常重要。选择性催化还原(Selective Catalytic Reduction,SCR)技术具有设计简单、控制方便、脱硝效率高等优点,广泛应用在燃煤电厂中。长期以来,对SCR脱硝系统的研究主要关注脱硝原理、催化剂、反应器流场等方面,SCR系统控制方法的研究没有得到重视。但脱硝系统控制精度不仅决定烟气排放是否达标,也影响着电厂运行成本。本文分析研究了SCR脱硝系统的动力学过程,基于吸附脱附及化学反应,构建脱硝系统机理模型,同时将前馈控制与动态矩阵控制方法相结合,设计出基于智能前馈的SCR系统喷氨量预测控制方法,提高喷氨量的控制精度。论文从以下三个部分进行研究:阐述了动态矩阵控制算法(Dynamic Matrix Control,DMC),同时考虑到SCR脱硝系统实际运行时存在的约束条件,设计了DMC串级控制回路,并通过仿真分析了动态矩阵控制参数对控制效果的影响。其次介绍了最小二乘支持向量机(Least Squares Support Vector Machine,LS-SVM)的基本原理及特点。介绍了SCR系统布置、工艺流程和反应原理两种喷氨量控制方式:固定摩尔比控制方式和脱硝反应器出口NO_x定值控制方式。设计了基于出口NO_x定值控制方式的PID控制方案,并进行仿真,根据电厂实际运行特点,分析总结了SCR脱硝系统特点及现有控制方案存在的不足。分析SCR反应过程的动力学原理,建立SCR系统机理模型,作为被控对象。脱硝系统的输入参数受锅炉燃烧状态影响很大,而且,SCR反应器中的化学反应、出口NO_x检测以及喷氨阀门的调整都存在惯性和迟延,在工况变化时仅依靠反馈控制难以实现出口NO_x浓度的准确快速控制。因此,本文依据电厂历史数据,采用锅炉侧可调参数作为输入,以锅炉出口NO_x浓度作为输出,利用最小二乘支持向量机算法构建锅炉出口NO_x浓度模型,该模型作为智能前馈控制器,将出口NO_x浓度转变成阀门开度信号,根据锅炉侧参数变化实时输出前馈控制信号,来快速响应锅炉侧工况的变化,将前馈控制与动态矩阵控制方法相结合,设计出基于机理模型的喷氨量最优控制系统。仿真结果表明,本模型实现NO_x浓度的全工况快速准确控制,减小了氨逃逸。
[Abstract]:With the rapid economic growth in China, the emission of nitrogen oxides (NO_x) from thermal power plants is increasing year by year, which poses a certain threat to human health and natural environment. In recent years, the state has made more stringent regulations on NO_x emissions from coal-fired power plants, so it is very important to study the safe, environmentally friendly and economical nitrogen oxides removal technology. Selective catalytic reduction (Selective Catalytic Reduction,SCR) technology is widely used in coal-fired power plants because of its simple design, convenient control and high denitrification efficiency. For a long time, the research of SCR denitrification system has been paid little attention to, such as the principle of denitrification, catalyst, reactor flow field and so on. But the control precision of denitrification system not only determines whether the flue gas emission is up to standard, but also affects the operation cost of power plant. In this paper, the kinetic process of SCR denitrification system is analyzed and studied. Based on adsorption desorption and chemical reaction, the mechanism model of denitrification system is constructed, and the feedforward control is combined with the dynamic matrix control method. A predictive control method for ammonia injection in SCR system based on intelligent feedforward is designed to improve the precision of ammonia injection. In this paper, the following three parts are studied: the dynamic matrix control algorithm (Dynamic Matrix Control,DMC) is introduced, and the DMC cascade control loop is designed considering the constraints of SCR denitrification system. The effect of dynamic matrix control parameters on the control effect is analyzed by simulation. Secondly, the basic principle and characteristics of least squares support vector machine (Least Squares Support Vector Machine,LS-SVM) are introduced. The SCR system layout, process flow and reaction principle are introduced in this paper. Two kinds of ammonia injection control methods, namely, fixed molar ratio control and NO_x fixed value control at the outlet of denitrification reactor, are introduced. The PID control scheme based on export NO_x fixed value control is designed and simulated. According to the actual operation characteristics of power plant, the characteristics of SCR denitrification system and the shortcomings of existing control schemes are analyzed and summarized. The kinetic principle of SCR reaction process is analyzed and the mechanism model of SCR system is established as the controlled object. The input parameters of the denitrification system are greatly affected by the combustion state of the boiler, and there are inertia and delay in the chemical reaction in the SCR reactor, the detection of the outlet NO_x and the adjustment of the ammonia injection valve. It is difficult to realize the accurate and fast control of NO_x concentration at outlet only by feedback control when the working condition changes. Therefore, according to the historical data of the power plant, the boiler side adjustable parameters are used as input, the boiler outlet NO_x concentration is taken as the output, and the NO_x concentration model of boiler outlet is constructed by using the least square support vector machine algorithm. As an intelligent feedforward controller, the model converts the outlet NO_x concentration into the valve opening signal, and outputs the feedforward control signal according to the boiler side parameter change in real time to quickly respond to the boiler side working condition change. Combining feedforward control with dynamic matrix control, an optimal control system for ammonia injection is designed based on mechanism model. The simulation results show that the model can quickly and accurately control the concentration of NO_x and reduce the ammonia escape.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM621.8;TP273
[Abstract]:With the rapid economic growth in China, the emission of nitrogen oxides (NO_x) from thermal power plants is increasing year by year, which poses a certain threat to human health and natural environment. In recent years, the state has made more stringent regulations on NO_x emissions from coal-fired power plants, so it is very important to study the safe, environmentally friendly and economical nitrogen oxides removal technology. Selective catalytic reduction (Selective Catalytic Reduction,SCR) technology is widely used in coal-fired power plants because of its simple design, convenient control and high denitrification efficiency. For a long time, the research of SCR denitrification system has been paid little attention to, such as the principle of denitrification, catalyst, reactor flow field and so on. But the control precision of denitrification system not only determines whether the flue gas emission is up to standard, but also affects the operation cost of power plant. In this paper, the kinetic process of SCR denitrification system is analyzed and studied. Based on adsorption desorption and chemical reaction, the mechanism model of denitrification system is constructed, and the feedforward control is combined with the dynamic matrix control method. A predictive control method for ammonia injection in SCR system based on intelligent feedforward is designed to improve the precision of ammonia injection. In this paper, the following three parts are studied: the dynamic matrix control algorithm (Dynamic Matrix Control,DMC) is introduced, and the DMC cascade control loop is designed considering the constraints of SCR denitrification system. The effect of dynamic matrix control parameters on the control effect is analyzed by simulation. Secondly, the basic principle and characteristics of least squares support vector machine (Least Squares Support Vector Machine,LS-SVM) are introduced. The SCR system layout, process flow and reaction principle are introduced in this paper. Two kinds of ammonia injection control methods, namely, fixed molar ratio control and NO_x fixed value control at the outlet of denitrification reactor, are introduced. The PID control scheme based on export NO_x fixed value control is designed and simulated. According to the actual operation characteristics of power plant, the characteristics of SCR denitrification system and the shortcomings of existing control schemes are analyzed and summarized. The kinetic principle of SCR reaction process is analyzed and the mechanism model of SCR system is established as the controlled object. The input parameters of the denitrification system are greatly affected by the combustion state of the boiler, and there are inertia and delay in the chemical reaction in the SCR reactor, the detection of the outlet NO_x and the adjustment of the ammonia injection valve. It is difficult to realize the accurate and fast control of NO_x concentration at outlet only by feedback control when the working condition changes. Therefore, according to the historical data of the power plant, the boiler side adjustable parameters are used as input, the boiler outlet NO_x concentration is taken as the output, and the NO_x concentration model of boiler outlet is constructed by using the least square support vector machine algorithm. As an intelligent feedforward controller, the model converts the outlet NO_x concentration into the valve opening signal, and outputs the feedforward control signal according to the boiler side parameter change in real time to quickly respond to the boiler side working condition change. Combining feedforward control with dynamic matrix control, an optimal control system for ammonia injection is designed based on mechanism model. The simulation results show that the model can quickly and accurately control the concentration of NO_x and reduce the ammonia escape.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM621.8;TP273
【参考文献】
相关期刊论文 前10条
1 杨耀权;张新胜;;LSSVM动态软测量模型在磨煤机一次风量预测方面的应用[J];动力工程学报;2016年03期
2 周鑫;吴佳;;火电厂SCR烟气脱硝系统建模与喷氨量最优控制[J];浙江电力;2015年11期
3 侯玉婷;薛建中;王林;王U，
本文编号：2236993
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2236993.html
