A dynamic soft-sensor modeling method based on FC-GP for 4-CBA content

Abstract

Because the static soft-sensor modeling method can not reflect the dynamic information of the industrial process, a dynamic soft-sensor modeling method based on fuzzy curve and Gaussian process is proposed in this paper to overcome the problem of poor precision and robust of static model. Fuzzy curve is firstly used to manage the dynamic input data, which is corresponding to the output data of each sample time. This step will transform the input data matrix to a vector. These vectors and the corresponding output data are then buildup a new matrix. Last, Gaussian process is used to build soft sensor model based on the new data matrix. This dynamic soft sensor model is used to build 4-CBA soft sensor model in PTA oxidation process. Simulation results indicate that the proposed model is precise and efficient