Successive Linearization-based Stochastic Repetitive Control Technique and its Applications to SMB and CATOFIN Processes

摘要

A repetitive control technique which is based on a time-varying stochastic state space model obtained by cycle-to-cycle linearization of a nonlinear fundamental process model around the operation trajectories in the previous cycle is introduced. The control technique was applied to two repetitive chemical processes: simulated moving bed, a continuous chromatograph separation process, and CATOFIN process, a catalytic propane dehydrogenation reactor where dehydrogenation and coke oxidation alternate over short periods. Numerical studies have shown that the proposed RC techniques are very effective in improving the operation of both processes.