一类典型工业过程的 RBF 神经网络控制  

Controlling a Typical Kind of Industrial Processwith the RBF Neural Network

作　　者：王旭东[1] 邵惠鹤[1] 

机构地区：[1]上海交通大学自动化系

出　　处：《上海交通大学学报》1998年第1期122-126,共5页Journal of Shanghai Jiaotong University

摘　　要：根据一类典型工业过程的特点，提出了一种神经网络控制方案．该方法基于ＲＢＦ（ＲａｄｉａｌＢａｓｉｓＦｕｎｃｔｉｏｎ）神经网络辨识过程模型，然后在此模型基础上设计内模控制．利用连续搅拌反应釜（ＣＳＴＲ）系统进行仿真设计，结果表明方案有效．According to the characteristic of a typical kind of industrial process, a new neural network control strategy is put forward. This strategy is based on the process model which is identified by the radial basis function neural network. With such a model, a nonlinear internal model control (IMC) strategy which has no offset is designed. The simulation with the CSTR shows that the CSTR is successfully identified and the IMC strategy is rather good.

关 键 词：RBF神经网络 连续搅拌反应釜 辨识 内模控制 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程] TQ052[自动化与计算机技术—控制科学与工程]