过热汽温智能控制系统研究

Intelligent Control for Boiler Superheated Steam Temperature

经过分析某超临界600MW直流锅炉高温过热器动态特性,过热器动态特性随机组的运行工况变化而变化,可以认为用蒸汽流量就能近似表示运行工况。在此基础上,提出了串级过热汽温智能控制系统。该智能控制系统中的副调节器仍然采用比例调节器;设计该主调节器时,把整个副回路和主对象看成广义被控对象,主调节器是一个自整定PID控制器,需要用到被控对象的Jacobian信息。随后构造了一个径向基函数(RBF)神经网络对被控对象进行辨识,从而求出了被控对象的Jacobian信息。最后将文中提出的智能控制系统用到了某超临界600MW直流锅炉的过热汽温控制中,仿真结果表明,用这种方法建立的过热汽温控制系统在机组运行工况发生变化时,具有较好的控制品质和较强的自适应能力。

This paper presents an intelligent control system for superheated steam temperature. Firstly, this paper analyzes the dynamic characteristics of a supercritical oncethrough 600MW boiler. It is found that while the operating condition is changing, the dynamic characteristics of the superheated steam plant are also varying. So the steam flow can be used to represent the operating condition of boiler. And then, a design of an intelligent control system for superheated steam temperature is presented. Cascade control structure is used in the proposed control scheme, where the secondary controller is a proportional controller and the main controller is a self-tuning PID controller. The whole inner loop and main object are regarded as a general object. The self-tuning PID controller is designed with the, aid of Jacobian information of the controlled plant, which is obtained from a RBF neural networks by which the general object is identified. Finally, the proposed control system is used to control the superheated steam temperature of a oncethrough 600MW boiler. Simulation results show that the control system has better performances and stronger adaptive abilities even if the operating condition is changing.