基于DBFNN的非线性自适应控制在CSTR中的应用  

Application of Nonlinear Adaptive Control Based on DBFNN for CSTR

作　　者：时海涛[1] 王心刚[1] 郭曙光[1] 

机构地区：[1]中国石油大学(华东)信息与控制工程学院

出　　处：《科学技术与工程》2010年第23期5776-5778,5781,共4页Science Technology and Engineering

基　　金：山东省自然科学基金(ZR2009FL027)资助

摘　　要：讨论了一类放射非线性系统的自适应控制问题。首先对于利用方向基神经网络（DBFNN）对系统的不确定性进行建模。所得到的系统模型作为对象的数学模型用来设计控制器。首先设计了反馈线性化控制器,为了克服建模误差的影响又引入了内模控制机制。证明了只要网络的学习精度足够高,所设计的闭环系统是最终一致有界的。把所设计的控制策略用于CSTR的控制中,仿真结果表明了所设计的控制器有效性。Adaptive control using DBFNN（Direction Basis Function Neural Network）of a class of nonlinear systems is discussed.First,DBFNN is used to model the uncertainties of nonlinear systems.Then the acquired NN model is used as system model to and a feedback linearization controller is designed based the trained well network.In order to overcome the model mismatch the internal model control（IMC）mechanism is introduced in.It is proved the closed-loop system is uniformly ultimately（UUB）.The proposed control strategy is application one CSTR（continuous stirred tank reactor）system.Simulations show the proposed strategy is effective.

关 键 词：连续搅拌反应器 自适应控制 DBF神经网络 非线性系统 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]