【作者】 张志君；

【Author】 Zhijun Zhang,,(Department of Automatic ,Dalian Technology University ,Dalian 116024 )

【机构】 大连理工大学自动化系；

【摘要】 针对复杂、未知、非线性及不确定过程,提出一种新的参考模型自适应控制方案。采用两个多层模糊神经网络分别作为过程的辨识器与控制器,控制器根据辨识器提供的信息通过在线修正隶属度函数及权值,自适应地学习调节控制率。网络训练采用拟牛顿法(BFGS)与最小二乘混合算法,极大的提高了网络收敛速度,将该方案用于复杂的非线性化学反应器(CSTR)控制中,仿真结果表明了该方法的有效性.更多还原

【Abstract】 A new adaptive control of model reference is presented for chemical engineering process having non-linear , complex , unknown, uncertain systems. In the proposed scheme, two multi-layer fuzzy-neural networks are used for identification and control of process systems ,the controller is able to learn to control a process adaptively by updating the fuzzy membership functions, the BFGS is incorporated into the least squared algorithm for training neural network on-line, The precision of model is raised and the control performance is improved significantly, So it can overcome the effect of model-mismatching and time-varying. The method is applied to control the reactor (CSTR), which has a complex dynamic behavior and nonlinear, The results illustrate effectiveness of the proposed approach.更多还原