摘要
针对循环流化床(CFB)锅炉床温的非线性、大惯性和大延迟等特性,提出了1种基于粗糙集的自适应模糊神经网络的床温控制方法,并且通过大量已知数据的学习得到模糊规则及其隶属度函数。为了减少规则的数目,提高数据的学习效率,引入了粗糙集,从采集数据中提取最小规则集,从而解决了自适应模糊神经网络中的规则爆炸问题。以CFB锅炉床温为控制对象,对基于粗糙集的自适应模糊神经网络控制器进行仿真比较。结果表明,该控制器控制效果优于常规PID控制器,但稳态误差较常规PID控制器大,其稳态误差小于1.7%,在允许范围内。
Directing against the features of non-linearity,large inertia,and long time delay of the bed temperature in CFB boilers,a control method of bed temperature using adaptive fuzzy neural network based on rough set has been put forward.This method boasts learning ability of neural network,and can obtain fuzzy rules and membership functions through learning a large number of available data.In order to reduce the number of fuzzy rules for improving the learning efficiency of data,a rough set has been introduced to draw minimal set of rules,thereby solving the problem of 'explosion of rules'.Taking the bed temperature in CFB boiler as the control object,an emulation and comparison of the adaptive neural network controller based on rough set have been made.Results obtained from emulation
出处
《热力发电》
CAS
北大核心
2012年第7期101-104,共4页
Thermal Power Generation
基金
国家自然科学基金资助项目(60774028)
河北省自然科学基金资助项目(F2010001318)
