Multiple Model Predictive Control of Component Content in Rare Earth Extraction Process

Abstract

Aiming at the complicated characteristic of rare earth extraction process and combining the material balan- ce model, a multiple models modeling and control method is proposed. Based on the data collected in an industrial field, an improved subtractive clustering algorithm is employed to obtain steady operation points for the process; the recuresi- ve least squares algorithm is adopted to identify submodel parameters and establish multiple linear models. According to the model switching index, an online optimal predictive model is obtained. And the efficiency of the model is verified by taking a certain rare earth company extraction as an ex- ample. In the end, generalized predictive controller of the corresponding sub-model is designed, so that component content is controlled in real time and accurately. Simulation results show the effectiveness of the method above.