摘要
针对新干法水泥预分解系统具有多输入变量、大滞后、大惯性和不确定的特征,文中提出了一种分层模糊建模方法,该方法采用増一型和二叉树型混合的分层结构,借助聚类方法和ANFIS获取各层模型变量的隶属度函数及推理规则。模型合理引入了历史信息以表达对象的动态特性。经系统模型的仿真结果与实际数据相比较,其拟合误差约为2%,表明所给模型具有良好的逼近性能,并且已经被应用于熟料生产过程仿真与操作培训系统。
To solve the problems such as more variables, large delay,large inertia and uncertainty characteristics, a kind of hierarchical fuzzy model was presented in this paper which combined the ‘one-plus' and ‘binary tree' structure. The cluster analysis and ANFIS were used to get the membership functions and the inference rules. The history information was considered in this model,which gave the model dynamic characteristics. With the contrast between the simulation model and the real data, the error of fitting is turned out to be about 2%, which indicate the capability of approximating to the plant,and has been used in clinker production process simulation and operator training systems.
出处
《自动化与仪表》
北大核心
2011年第9期1-5,共5页
Automation & Instrumentation
基金
河南省科技厅2010年河南省重点科技攻关计划项目(102102210376)
河南省教育厅自然科学基金项目(2010A120008)
关键词
分层模糊
建模
动态
新干法水泥
预分解
hierarchical fuzzy system
modeling
dynamic
new dry production
predecompose