Adaptive Soft-Sensor Modeling of SMB Chromatographic Separation Process Based on Dynamic Fuzzy Neural Network and Moving Window Strategy

Abstract

Simulated moving bed (SMB) chromatographic separation technology is a new type of separation technology developed on the basis of traditional fixed bed adsorption operation and real moving bed (TMB) chromatographic separation technology. The component purity of the extract and raffinate during the SMB chromatographic separation was used as the prediction object. An adaptive soft-sensing modeling method for SMB chromatographic separation process based on dynamic fuzzy neural network (D-FNN) and moving window strategy. Dynamic fuzzy neural network soft-sensing models based on Kalman filter (KF) algorithm, linear least squares (LLS) method, and extended Kalman filter (EKF) method. The moving window strategy is then adopted to realize the adaptive revision on the soft-sensing model, and the prediction performances are with compared with the soft-sensing model established by the generalized dynamic fuzzy neural network (GD-FNN). The simulation results show that the proposed soft-sensing model can not only effectively achieve accurate prediction of key economic and technical indicators of the SMB chromatographic separation process, but also meat the real-time, efficient, and robust operation of the SMB chromatographic separation process.