RNGA for non-square multivariable control systems: properties and application

TL;DR

This paper introduces a new RNGA method for non-square multivariable control systems that considers both steady-state and transient data, improving loop pairing and control performance over traditional RGA.

Contribution

The paper extends RNGA to non-square systems with a new column-major approach and demonstrates its effectiveness through a practical laboratory application.

Findings

RNGA provides better loop pairing than RGA for non-square systems.

RNGA reduces control interactions and improves system performance.

Application to a laboratory setup confirms practical benefits.

Abstract

The Relative Gain Array (RGA) and Relative Normalized Gain Array (RNGA) have received considerable attention for square systems. In this paper RNGA with the column-major, for non-square multivariable systems is introduced. RNGA of the paper has a row-column inequality, i.e. the number of rows is less than the number of columns. Unlike the conventional RGA, the RNGA loop pairing criteria of the paper considers both steady-state as well as transient information for the assessment of control-loop interactions. The RNGA for square systems is extended for non-square multivariable systems by thoroughly deriving its supporting properties. The RNGA method is applied to a non-square multivariable radiator laboratory test setup for loop pairing. Closed-loop results arising from the RNGA-based loop pairing are depicted in the paper. The lacuna of the conventional RGA loop pairing has been overcome by the application of the developed RNGA of this paper. The results unfold the effectiveness of RNGA over RGA for non-square multivariable systems to have minimum interactions and better control.