Hybrid intelligent optimal control of fused magnesium furnaces

Controlling the product quality and energy consumption within their desired ranges is the main objective for the operation of the smelting process of fused magnesia. With manual control, which is still widely used in practice, variations in the boundary conditions make it difficult for adjusting the setpoints for various control loops involved, thereby leading to compromised product quality and increased power consumption. In this paper, a hybrid intelligent optimal control method is proposed for the fused magnesium furnace so as to optimize the production indices by auto-adjusting online the setpoints of the currents for the three-phase electrodes in response to changes in boundary conditions, and to make the outputs of control loop track the setpoints. The techniques used consist of case based-reasoning (CBR), rule based-reasoning (RBR), iterative learning control (ILC) and PI control. Although each control element is well known, their innovative combination can generate better and more reliable performance. Industrial applications show the usefulness and effectiveness of the proposed hybrid intelligent optimal control method and its promises in other industrial processes with similar features.