摘要
针对具有实参数不确定性的热工过程,基于概率鲁棒方法,提出一种PID控制器设计方法。根据被控对象模型的参数摄动状态,计算闭环系统满足各条性能设计要求的概率,并对其进行综合作为优化算法的目标函数,利用遗传算法对PID控制器参数进行优化,用蒙特卡罗实验对控制系统进行鲁棒性检验。针对4种典型的热工过程进行了仿真试验,并与基于标称参数的设计方法以及传统的整定方法进行了比较。仿真结果表明,基于概率鲁棒的PID控制器设计方法对模型参数不确定性具有较好的鲁棒性,在被控对象存在一定的不确定性时,系统能以最大的概率满足设计要求,因此适用于具有参数不确定性的典型热力过程的控制。
A PID controller design method based on probabilistic robustness was presented for thermal process with varied parameters. Based on model uncertainties, the probability of closed system meeting every item of dynamic performance requirements was calculated and synthesized as the cost function of genetic algorithms which was used to optimize the parameters of PID controller. Monte-Carlo experiment was applied to test the control system robustness. Simulation for four kinds of typical thermal process was carried out. Comparison with the PID design method based on nominal conditions and some traditional tuning methods indicates that the method presented has better robustness, the systems could satisfy the design requirements in a maximal probability, so this method can be applied to typical thermal process with parameter uncertainty.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2007年第32期92-97,共6页
Proceedings of the CSEE
关键词
PID控制器
热工过程
概率鲁棒
蒙特卡罗实验
PID controller
thermal process
probabilistic robustness
Monte-Carlo experiment
