多目标神经PID控制器及其在过热蒸汽温度控制中的应用研究  

STUDY ON MULTI-PURPOSE NEURAL PID CONTROLLER AND ITS APPLICATION IN SUPERHEATED STEAM TEMPERATURE CONTROL

作　　者：陈亮[1] 俞海斌[1] 金建祥[1] 褚健[1] 

机构地区：[1]浙江大学,浙江杭州310027

出　　处：《热力发电》2005年第6期19-22,33,共5页Thermal Power Generation

摘　　要：为了适用大滞后、大惯性复杂过程的高质量控制,对神经网络PID控制器进行了改进,提出了多目标优化算法。该算法具有自适应特点,克服了常规PID超调量大,神经PID响应速度慢的缺点,并容易实现实时控制。应用此算法对电站锅炉过热蒸汽汽温控制进行了仿真,取得了理想的效果。For implementation of high quality control of complex process with long time delay and large inertia, a multi-purpose optimal algorithm has been put forward on the basis of improving neural network proportional, integral and differential (PID) controller. The said algorithm boasts adaptive behavior, eliminating defects of large over shot in traditional PID controller, and low response speed in neural PID controller, as well as easily realizing the real-time control. Simulation study has been carried out by using the said algorithm in steam temperature control of superheated steam for utility boilers, obtaining ideal effectiveness.

关 键 词：神经网络 PID控制器 过热蒸汽 汽温控制 仿真 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]