METHODOLOGY BASED ON SVD FOR CONTROL STRUCTURE DESIGN

摘要

This work presents a methodology for control structure design based on the Singular Value Decomposition (SVD) of the Hankel matrix, constructed from the observability and controllability matrices. A weight index is proposed as a measure of the impact of each input and output variable and, based on it, an input-output pairing selection is performed. The methodology is validated using the Tennessee Eastman process and the result is compared to the relative gain method in frequency.