作　　者： ; ;

机构地区： 茂名学院机电工程学院自动化系

出　　处： 《计算机仿真》 2009年第6期292-295,共4页

摘　　要： 针对模糊控制表达能力强,但学习能力、适应性差的特点,提出在聚合反应过程的温度控制系统中使用模糊控制与神经网络相结合的控制方案,控制方案给出了基于Takagi-Sugeno模型的神经模糊网络结构以及优化模糊规则和隶属函数的学习算法;仿真研究表明,神经模糊网络可以根据输入数据的分布情况,自动修改模糊规则的加权系数和隶属函数,克服了单纯的模糊控制其控制规则和隶属函数一经得到就固定不变的缺陷,具有较强的适应能力和学习能力,较单纯的模糊控制有更高的控制精度。 In view of the characteristics of fuzzy control, such as high expression capability, low learning capability and adaptability, a control method of combining fuzzy control with neural network for polymerizing - kettle temperature control is proposed. The constitution of neural fuzzy network based on T - S model, and an algorithm for optimizing fuzzy rules and membership function are given by the method. The MATLAB＇s simulation experiments proved that neural fuzzy network could automatically amend weighing parameters of fuzzy rules and membership function according to the distribution of inputs, thus overcoming the disadvantage of the pure fuzzy control in which the fuzzy rules and membership function remaining to be the same during control. This method has very good adaptability and learning capability, and has higher control precision than simple fuzzy control.

关 键 词： 神经模糊网络 仿真 隶属函数 模糊规则

领　　域： [自动化与计算机技术] [自动化与计算机技术]