Online Tuning Strategy for Multi-loop SISO PI Control Algorithms in Multivariable Interactive Systems

Abstract

Tuning of PI control algorithms for coupled multi input multi output (MIMO) systems is a challenging problem. This paper extends a previously developed model-based adaptive tuning method to handle the tuning problem of coupled multivariable systems. The performance of the proposed method is compared to those of existing methods such as Biggest Log Modulus (BLT), Sequential Loop Closing (SLC) and Multi-loop Modified Ziegler-Nichols (MMZN). The proposed method is not limited to any special form of process models and allows the use of a digital control law formulation. The tuning procedure is cast as a constrained least squares problem which can be easily solved enhancing its on-line implementation. Linear constraints are imposed on the least square problem to account for loop interactions. Application of the tuning algorithm to two simulated MIMO examples revealed its ability to perform as well as the other tested methods.