A closed-loop frequency domain PI retuning technique for multivariable systems

Aguiar, Anna Paula V. de A.; Júnior, George Acioli; Barros, Péricles Rezende; Perkusich, Angelo

doi:10.1016/j.ifacol.2021.08.285

Cited by 4 publications

(1 citation statement)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The method was presented for non-square systems based on the pseudoinverse evaluation of the process transfer function matrix at a lowfrequency point. In [37], the frequency response was used to retune a PID MIMO controller. Other works [2,38] proposed the design of centralized PID control based on the frequency response of the process as a convex optimization problem with LMIs.…”

Section: Introductionmentioning

confidence: 99%

Design of Multivariable PID Control Using Iterative Linear Programming and Decoupling

Garrido,

Garrido-Jurado,

Vázquez

et al. 2024

Electronics

View full text Add to dashboard Cite

The design of multivariable process control systems is specially complicated when there are strong interactions between the different control loops, and even more with multiple time delays. This paper proposes an iterative design method of centralized proportional-integral-derivative (PID) controllers for stable linear systems. The methodology is based on the linear parameterization of equivalent loop transfer functions (ELTFs) for centralized control. These functions capture the dynamics of the other loops and, from a prior design, allow solving the design problem at each iteration with linear programming that shapes the Nyquist plot of the ELTFs in the frequency domain, which also avoids the need for approximations. Two optimizations are proposed: (I) maximizing integral gains by fulfilling linear robustness margins in each ELTF and (II) maximizing linear robustness margins by fulfilling minimum bandwidths in each loop. In both optimizations, static decoupling and decoupling at a frequency close to the bandwidth of each loop are included as constraints, which improves the decoupling performance and the procedure convergence. The effectiveness of the method is verified in three simulation examples (square and non-square) and a lab experimental process. The proposed designs achieve a similar or better response when compared to that achieved by other authors.

show abstract

Section: Introductionmentioning

confidence: 99%