基于双神经网络逆模型的屏式过热器汽温控制  

Steam Temperature Control of Panel Superheater Based on Dual Neural Network Inverse Model

作　　者：莫日格吉勒图 MORIGE Jiletu(Datang Thermal Power Technology Research Institute,Huhehaote 010020,China)

机构地区：[1]大唐火力发电技术研究院,内蒙古呼和浩特010020

出　　处：《电工技术》2021年第14期26-30,共5页Electric Engineering

摘　　要：过热汽温是燃煤锅炉运行中的关键参数,必须控制在规定范围内,过高或过低都会直接影响机组的安全经济性。由于大容量锅炉机组经常处于深度调峰大幅变工况运行,且过热汽温系统是一个典型的非线性、大惯性、大时延的被控对象,通常采用传统PID控制,往往很难获得满意的控制效果,因此智能控制的发展为过热汽温的控制提供了很好的研究方向。提出一种基于双神经网络逆模型的屏式过热器汽温控制方法,以克服屏式过热器蒸汽温度控制方法对机组大范围变工况适应性差,特别是当锅炉存在严重偏烧,导致喷水后汽温接近饱和蒸汽区时导前温度失灵的问题,从而增强工况适应性,有效应对喷水后蒸汽接近饱和区时温度对喷水量不敏感,引起屏过出口汽温控制效果变差的问题。The superheated steam temperature is a key parameter in the operation of coal-fired boilers and must be controlled within the specified range.Too high or too low will directly affect the safety and economy of the unit.Because the large capacity boiler unit is often in deep peak shaving and large variable condition operation,and the superheated steam temperature system is a typical nonlinear,large inertia and large time delay controlled object,it is often difficult to obtain satisfactory control effect by using traditional PID control.Therefore,the development of intelligent control provides a good research direction for superheated steam temperature control.A control method of panel superheater steam temperature based on double neural network inverse model is proposed to overcome the poor adaptability of platen superheater steam temperature control method to large-scale variable working conditions of the unit,especially when the boiler has serious partial combustion,which leads to the failure of leading temperature when the steam temperature is close to the saturated steam area after spraying water.So as to enhance the adaptability of working conditions,and effectively deal with the problem that the temperature is not sensitive to the amount of spray water when the steam is close to the saturation area after spray water,resulting in the poor control effect of the steam temperature at the outlet of platen superheater.

关 键 词：神经网络 逆模型 过热汽温 锅炉偏烧 PID算法 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]