Skip to Main Content
The automatic pouring of molten metal from a ladle into a mold by a robot requires an high-precision tracking control of the flow rate. Due to the hydrodynamics of the liquid metal, the model of the pouring process is highly nonlinear which has to be considered for the model-based design of the tracking control. The two design approaches studied in the paper combine a feedforward and a feedback control in order to obtain a precise and robust tracking performance. Thereby, the property of differential flatness is used for the design of the tracking control, and the unmeasured states are reconstructed for the feedback control by an unscented Kalman filter. Two proposed control schemes are compared by simulations of the automatic pouring robot together with the pouring process.