Identification of an Ill-Conditioned Distillation Column Process using Rotated Signals as Input

Abstract

The standard uncorrelated test signals estimate the low-gain direction of ill-conditioned multivariable systems poorly. Therefore, the low-gain information needs to be excited more. In this paper, identification of an ill-conditioned distillation column process using rotated signals is proposed. Rotated input signals allow more excitation to be applied in the weak gain direction of the process and less excitation in the strong gain direction. In this approach, the singular value decomposition (SVD) of the steady state gain matrix is used to rotate the input signals along the directions of the right singular vectors. Simulation results show good accuracy of the proposed method in identifying low and high gain directions.