基于单神经元Smith预估的过热汽温控制  

Main Stream Temperature Control Based on Single Neuron with Smith Predictor

作　　者：赵静[1] 边立秀[1] 周俊霞[1] 

机构地区：[1]华北电力大学动力系,河北保定071003

出　　处：《电力情报》2002年第2期18-20,41,共4页Information on Electric Power

摘　　要：针对典型大时滞大惯性、参数变化的电站锅炉过热汽温受控对象 ,设计了基于单神经元Smith预估过热汽温控制系统。在仿真实验的基础上 ,对单神经元Smith预估控制和最优常规PID控制进行了比较和分析。仿真结果表明 ,单神经元Smith预估控制能够充分利用神经元自学习、自适应的能力以及Smith预估补偿控制的优点 ,使系统的控制品质提高 。Considering the controlled object of main stream temperature with large time delay and varying parameters in utility boilers, a single neuron controller with Smith predictor is presented. Based on the simulation, the comparison is made between single neuron controllers and optimal general PID controllers. The results indicate the proposed control system has advantages of the self-learning ability, the self adaptability and the compensation control so that the control quality, the robustness and the adaptability of the control system are superior to the general PID control.

关 键 词：汽温控制 锅炉 过热器 预估控制 单神经元控制器 

分 类 号：TK223.32[动力工程及工程热物理—动力机械及工程] TP183[自动化与计算机技术—控制理论与控制工程]