THE RESEARCH ON THE MODEL OF FLATNESS CONTROL BASED ON THE OPTIMIZED RBF FUZZY NEURAL NETWORK
As there are many non-linear elements influencing the flatness in the rolling process and its accurate mathematics model is too complex to build, a fuzzy neural network controller is proposed for the cold rolled flatness control Fuzzy neural controller does not require accurate model of plant and is able to learn to control adaptively.RBF network is adopted in the fuzzy neural network.To automatically acquire the fuzzy rule-base and the initial parameters of the RBF fuzzy model, the relationship clustering method is used in structure identification.Based on the clustering result, a fuzzy neural network is set up and then trained by genetic algorithm to obtain a precise flatness control model.The simulation result shows that it not only reduces the complexity of neural network, but also has faster convergence rate and less possibility to local minimum.The response is more favorable than that of conventional fuzzy controllers and that of fuzzy neural network based on BP network.
Fuzzy neural network Flatness control Relation clustering method
HAI-TAO HE LAN ZHANG
College of Information Science and Engineering, Yanshan University, Qinhuangdao, 066004, China
国际会议
2007 International Conference on Machine Learning and Cybernetics(IEEE第六届机器学习与控制论国际会议)
香港
英文
472-476
2007-08-19（万方平台首次上网日期，不代表论文的发表时间）