| Nonlinear system is one of the hottest topics in present control field. In a manner of speaking, most of all the systems are nonlinear in real life. As the technology progressing and the industry development, all areas demand high quality control systems, and therefore the control systems become more and more complex. In addition, the control objects of modern industry always contain multiple characteristics, such as containing more and more variables, being serious nonlinear, being strong coupling, containing large delay, being time-varying, containing a wide range of interference and etc, and no other than these characteristics make a huge challenge to the traditional control theory. In order to solve these problems, intelligent control theory has been widely used to control nonlinear systems, further more the fuzzy neural networks become a hotspot yet. As a kind of excellent controller, fuzzy neural network has been used extensively in modeling and controlling. However, there are still lots of problems to been resolved.After introducing the history and present condition of nonlinear system and fuzzy neural network, dynamic fuzzy neural network(D-FNN) was described. Simultaneously D-FNN’s frame, arithmetic, generalization and the technology of networks trimming were studied. And then the validity and practicality of D-FNN was tested through approximating Hermite and identifying one nonlinear dynamical system. Ultimately, the thesis designed direct inverse control system of the combustion system of circulating fluidized bed boiler(CFBB) basing on the dynamic neural network. Subsequently simulation results of the inverse control system were compared with those of PID control system. Acting as a controller, D-FNN has many a advantages. Systems based on the D-FNN controller performs wonderfully in anti-jamming, fast and steadily responding, rapid convergence and small static error. The self-adaptation and robustness of these systems are perfect. |
PDF Full Text Request