CONSISTENCY PRINCIPLES FOR STABILITY IN DECENTRALIZED CONTROL SYSTEMS

Abstract The Niederlinski Index (NI) is generalized to systematically address closed loop stability from steady state information of both the process and the controller. A series of stability conditions, referred to as ‘consistency principles for stability’, which allows for joint consideration of controller design and variable pairing, are established. It is shown that systems paired on negative NI can still be stabilized by properly setting directions of individual controllers, and that positive NI is not only desirable to facilitate controller tuning and failure tolerance, but is also always achievable for any decentralized control system.