摘要
提出了一种新的神经 -模糊系统 ,结构上分为六层 ,包括一个输入层、一个输出层和四层隐层 .该系统规则的结论部分采用线性 ARX模型 ,训练方法结合最小二乘法与 BP算法 .将此神经 -模糊系统与广义预测控制算法相结合 ,形成一种神经 -模糊预测控制算法 .用一个具有三条规则的神经 -模糊系统建立某大型纸浆厂碱回收炉的模型 ,并应用神经 -模糊预测控制算法进行控制 ,结果显示神经 -模糊预测控制算法性能优良 。
A new kind of neural- fuzzy system is introduced.The neural- fuzzy network has six layers in- cluding input layer,output layer and four hidden layers.The conclusion part of its fuzzy rules is ARX model and the training method integrates LSand BP algorithm.A neural- fuzzy predictive controller which integrates Generalized Predictive Control algorithm with the neural- fuzzy system is also introduced.Three- rule neural- fuzzy system is used to model a recovery boiler in a pulp factory and the predictive control al- gorithms is applied to control the bed temperature and the oxygen content in the flu gas.The response curves show that the neural- fuzzy predictive control possesses good effectand has potential applications in industry processes.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2002年第2期123-126,共4页
Journal of Zhejiang University：Engineering Science
基金
浙江省教育厅资助项目 (2 0 0 10 5 3 9)
