Bilinear model of rare earth cascade extraction process and its parameter identification
稀土串级萃取分离过程的双线性模型及其参数辨识

In the process of rare earth cascade extraction, the online measurement of element component content is extremely important to improve the direct recovery of metals. Because of the difficulty of implementing online measurement, research on the dynamical model becomes the key problem in soft-sensor technology. In this paper, based on the material balance equations in the process of rare earth cascade extraction, a bilinear model with time-delay is proposed. The established model is also simplified by reducing its order. The recursive least-squares identification method is then adopted to identify the model parameters by using the data selected in an industrial field. The optimal time delay constant is simultaneously selected according to the error index. The mathematical model representing the rare earth cascade extraction process is thus obtained. Finally, simulations are conducted by using the practical data from the rare earth cascade extraction process, and the results show the effectiveness of the modeling method and the validity of the model.