The Research On AC LF Electrode Positioning Intelligent Decoupling Controller

The 110ton Lade Furnace (LF) of Tianjin Iron & steel CO.LTD have been researched in this paper. Fore expect to have already completed the foundation of the intelligence controlled up, study further the intelligent decoupling problem of three-phase AC LF.Three-phase LF is a kind of metallurgical system which include the characteristic of three-phase unbalance,strong coupling,time lag,time variation and strong nonlinearity. There is the coupling between the three-phase electrodes, and two electrodes could be willfully affected by one's positioning, and it was very difficult to be stably adjusted for the three-phase electrodes, besides, electrode loss increased and massive energy was wasted. As there is the ability to approach any nonlinear mapping and to reject the interference in the diagonal recurrent neuronal network (DRNN), besides, good robustness was included in the internal model control, so the intelligent controller which including the two ones was adopted. By the state of the DRNN network system's online learning, Jacobian information was solved and sent to the internal model controller which decouple and improve the system's performance of control by simulating the contrary model of object. The online internal model control based on DRNN network discrimination and RBF network controller were mainly researched in this paper. The control algorithm is successfully achieved by computer programming, and the hardwired connection simulation platform is built which based on industrial computer,data acquisition board and hardware circuit. Then the research on testing the ability of the control algorithm's performance is conducted by the simulation platform.