| The powder metallurgy furnace playing an important role in industry is considered in this paper, it's control system is used as main content and analyzed, the flaw of control pattern is analyzed and introduced and put forward to using intelligent controller to replace traditional pure instrument + relay or PID control. Based on the current situation and effect in industry of intelligent control theory at present, fuzzy control and the advantage and flaw of neural network control are analyzed. The logical inference of fuzzy control and acquisition, remembrance of neural network are made full use, and the fuzzy neural network controller is constructed. Aimed to the flaw of fuzzy neural network by ordinary RLS algorithm which has the flaw of numerical stability, robustness and local minimum problem, we propose a control model of fuzzy neural network based on fast RLS algorithm. At last, by using of MATLAB simulation software for the model of powder metallurgy furnace's control system by fast RLS algorithm simulation, we analyzes and compare traditional PID control, fuzzy control, neural network control and fuzzy neural network by conventional RLS algorithm, and have the conclusion that the fuzzy neural network controller has a higher intelligent level. Using fuzzy neural network by fast RLS algorithm, steady-state overshot is zero. The system has no react with the fast signals of vibration, and is better than conventional RLS algorithm in training step, training time, and error precision. Convergent rate is faster and error curve is more stable. Thus fuzzy neural network controller based on fast RLS algorithm has a good prospect appliance in practice.
|
PDF Full Text Request