The research on regenerative furnace combustion control based on multi-scale wavelet network

This article proposes a compound control strategy. It is based on the multi-resolution wavelet controller, the sub-controller and the PID-NNC master-controller and integrated with the WNN wavelet neural network identifier. This strategy can be used to overcome effectively bad influences caused by the system initiation, the control mode, the object change and the exterior disturbance to the regenerative furnace in the process of combustion temperature control. This strategy has such characteristics as quick response, small overshoot, high stable state precision, strong robustness and etc. It can not be realized easily by conventional PID control, which has been proved by the simulation result. It offers a new way for applying the wavelet transformation in the steel metallurgy controlling domain and a new direction for saving energy and reducing consumption.