Hybrid fuzzy neural network control for complex industrial process

Fire exchange must be done for the large scale glass furnace in order to assure combustion equality. However, disturbance during fire exchange breaks the stability of temperature and pressure in furnace, and the automatic control precision is decreased greatly. A hybrid fuzzy neural network controller is employed to suppress these fluctuations in this paper. The algorithm and structure of hybrid FNNC are described in detail. During fire exchange, the optimal control coefficients are obtained using FNNC, at the same time the typical PID controller is closed. After fire exchange, the fuel and press output decrease or increase gradually according to these coefficients. The hybrid FNNC has already been implemented successfully. The results show that the hybrid FNNC algorithm improved the performance of automatic control greatly.