神经网络在火电厂过热蒸汽温度控制系统中的应用  

Neural networks control in the superheated steam temperature system

作　　者：周光明[1] 马永光[1] 

机构地区：[1]华北电力大学工业过程仿真与控制实验室,保定071003

出　　处：《仪器仪表用户》2005年第5期67-68,共2页Instrumentation

摘　　要：本文针对惰性区具有多个参数及迟延摄动的火电厂锅炉过热蒸汽温度控制问题及其对汽温响应和抗扰动特性的要求,设计了内回路采用神经网络控制、外回路采用PID控制的串级控制系统。仿真结果表明,在过热蒸汽温度控制系统中,采用复合串级控制,显著的改善了汽温响应特性,取得了令人满意的效果。This paper presents the design of a system of cascade control that its inner loop adopted neural networks control and its outer loop adopted PID control, aiming at the super-heated steam temperature cascade control system with many parameter and timedelay perturbations and its special requires of step response and anti-jamming characters. The simulation results indicate that the complex cascade control system have better performances and robusticity than PID control in the super-heated steam temperature system.

关 键 词：过热汽温 串级控制 神经网络控制 PID控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]