摘要
将神经网络与预测控制相结合 ,采用递阶遗传算法对神经网络的权值和结构同时进行训练 ,实现了非线性、大时滞系统模型的精确预测 ;然后将聚类算法和模糊控制相结合 ,设计了一种新型的聚类自适应模糊预测控制器 ,实现了对非线性和大时滞系统的自适应控制。将该控制器应用于锅炉的单元机组负荷控制系统中 ,仿真表明该方案的适应性、实时性和鲁棒性都很强 ,具有工程实用价值。
Considering nonlinear and big-lagged characteristics of system,such as drum boiler,This paper,first,combines neural network with forecast control,at the time,optimizes the neural network by genetic algorithm and realizes accurate forecast to nonlinear big-lagged system;then,combines fuzzy control with clustering algorithm and presents a new clustering adaptive fuzzy forecast controller.The controller is applied to the unit load control system of drum boiler,and the simulation results show that the project has good adaptability,strong robustness and real prospect for further applications.
出处
《电工技术学报》
EI
CSCD
北大核心
2003年第3期77-80,共4页
Transactions of China Electrotechnical Society
关键词
神经网络
预测模型
模糊控制
遗传算法
聚类算法
鲁棒性
Neural network,forecast model,fuzzy control,genetic algorithm,clustering algorithm,robustness
