电厂主汽温串级控制系统的PID控制器算法改进  

Improved Algorithm of PID Controller in Cascade Control System of Power Plant Main Stream Temperature

作　　者：赵为光[1] 杨莹[1] 

机构地区：[1]黑龙江科技学院电气与信息工程学院

出　　处：《电站系统工程》2008年第6期61-63,共3页Power System Engineering

摘　　要：针对火电厂主汽温控制系统常规固定参数PID串级控制性能欠佳的问题,提出一种改进算法整定PID参数。先用RBF网络在线辨识温度采样离散信息,得到控制器所需的Jacobian信息,然后用Levenberg-Marquardt算法代替传统的梯度法整定PID参数,得到改进的PID算法。通过对电厂主汽温串级系统仿真分析,验证了算法的快速性、抗扰性、鲁棒性。To the problem that satisfactory control effects can hardly be obtained with fixed-parameter PID cascade control in the main steam temperature control system of thermal power plants, a betterment PID controller based on an improved algorithm is proposed. RBF network is built to construct processing model through on-line identification and the Jacobian information of the system controller is obtained. Levenberg-Marquardt algorithm is used to adjust the parameter of PID controller instead of traditional gradient descent tuning method. Simulation of main steam temperature control system is carded out to prove that the improved algorithm is faster, stronger anti-jamming and stronger robustness.

关 键 词：主汽温 RBF神经网络 PID控制 LEVENBERG-MARQUARDT算法 

分 类 号：TK311[动力工程及工程热物理—热能工程]