蒸汽发生器水位SDRNN优化自抗扰控制  

Steam Generator Water Level Control Based on SDRNN-ADRC Optimization

作　　者：周洪煜[1] 汪正海[1,2] 黄建平 赵乾[3] 

机构地区：[1]重庆大学动力工程学院,重庆400030 [2]华电电力科学研究院,杭州310030 [3]中国核动力研究设计院,成都610041

出　　处：《动力工程学报》2013年第10期789-794,807,共7页Journal of Chinese Society of Power Engineering

基　　金：重庆市科委重大科技攻关项目(CSTC 2009AB1008)

摘　　要：针对蒸汽发生器(SG)水位控制过程存在的主要问题,引入水位自抗扰控制(ADRC)方案.通过扩张状态观测器实时估计系统内、外扰动,并采用前馈方式予以动态补偿,同时依据状态误差矢量进行非线性反馈调节,缓解控制系统快速与超调之间的矛盾.并引入二阶对角递归神经网络(SDRNN)动态辨识SG Jacobian信息,实时优化自抗扰控制器参数.分别在水位、蒸汽和给水扰动下进行SG水位仿真实验,并对比了前馈串级PI控制与SDRNN-ADRC控制的响应曲线.结果表明:在扰动工况及控制对象参数时变下,此SG水位控制系统的控制响应迅速、超调小且稳态误差小,具有优良的动、静态性能.To solve the problems existing in steam generator （SG） water level control system, an active disturbance rejection control （ADRC） scheme was proposed, in which an extended state observer was used to estimate and dynamically compensate the internal and external disturbance in feed forward way, and simultaneously, the contradiction between rapidity and overshoot of the control system was alleviated through nonlinear feedback regulation based on state error vector. Meanwhile, an identification method based on second-order diagonal recurrent neural network （SDRNN） was proposed to dynamically identify SG Jacobian information for real time optimization of the ADRC parameters. Besides, simulation tests of SG water level were carried out with the controller under different disturbances, such as water level dis- turbance, steam disturbance and feed water disturbance, of which the results were compared with that of feedforward-cascade PI controller. Results show that the SG water level control system based on SDRNN- ADRC optimization has excellent dynamic and static control performance, with small overshoot and small steady-state error.

关 键 词：蒸汽发生器 水位 自抗扰控制 二阶对角递归神经网络 

分 类 号：TK172[动力工程及工程热物理—热能工程]