Tuning of controller parameters by FCbT with stability constraints for non-minimum phase plants

Abstract

We propose the controller tuning method considering closed-loop stability after tuning without the identified plant modeL The methods are formulated as a model reference control problem in which the reference model is chosen by the user. In the case of non-minimum phase systems, the unstable zeros of the plant need to be included in the reference model to avoid unstable pole-zero cancellation. This paper proposes the model-free controller tuning for non-minimum phase plants estimating its unstable zeros by the flexible reference model in the process of parameter updates. The effectiveness of the proposed method is confirmed by simulation examples.

Novel data-driven control approaches : application in essential oil extraction processes

“…Regarding to CbT, it is worth to mention that exist two correlation based approaches for data-driven controller design, one of them is iterative (iCbT) (Karimi et al, 2002), another is non-iterative (CbT-GS) (van Heusden et al, 2011), as above mentioned, in this work our interest is limited to non-iterative techniques, that is why principal characteristics of CbT-GS are listed: (i) The problem is solved via model reference criterion, but it doesn't need a virtual reference signal, (ii) In the same way that in VRFT, the controller is linearly parametrized, (iii) It is one-shot method and is better suited to deal with measurement noise (iv) In CbT-GS a sufficient conditions is proposed to guarantee closed-loop stability, (v) This method has not been extend to non linear framework. Few extensions on non-iterative CbT have been found in literature, one of them is about the tuning of controller parameters taking into account stability constraints for non-minimum phase plants (Matsuo et al, 2013), meanwhile in (Yubai et al, 2009) and (Usami, 2010) was addressed the MIMO case.…”