Adaptive generic model control scheme for an optimized external heat integrated air separation column using unscented kalman filter

Abstract

An External Heat-Integrated Air Separation Column (E-HIASC) process is a promising air separation technology. This study focuses on the operational stability of the optimized E-HIASC process for separating nitrogen, oxygen, and argon mixtures. The operation stability of process is achieved through an Adaptive Generic Model Control (AGMC) scheme which is designed by incorporating the identified E-HIASC state-space dynamic model into the controller algorithm. The controller synthesizes the Generic Model Control (GMC) algorithm, decoupled ARX model, and Unscented Kalman Filter (UKF) algorithm to enable the auto-regression and exogenous (ARX) for model identification and the UKF algorithm to estimate time-varying parameters and compute unmeasured E-HIASC state parameters required in the GMC algorithm. A Generic Model Control (GMC) and Multivariable PID (M-PID) control schemes were also designed for benchmarking study. Simulation results show that an AGMC scheme performs better than the GMC and M-PID schemes in tracking the product concentration set point and disturbances rejection.