Closed-loop identification and controller reconfiguration for EXTRAP T2R reversed-field pinch

Abstract

We briefly summarize a supposedly efficient novel method for measuring the external plasma response as applied to the inherently unstable reversed-field pinch EXTRAP T2R. The set of parameters estimated with this particular method is then harvested and fed as input to a discrete-time fixed-order fast fourier transform decoupled multi-input-multi-output controller synthesis. The thus reconfigured feedback system is implemented and experimentally tested on the real plant T2R. A particular and intended behaviour is observed, it seems, from these firstdeployment results. I. INTRODUCTORYOVERVIEW Robust controller synthesis and simulations have been worked out for e.g. the DIII-D Tokamak [1] in the case of RWMs. Apparently, to properly do control system synthesis availability of some plant model is needed. Minimal requirements include (vague) bounds on important parameters in a plausible model structure. There are several studies of external plasma response measurements in the literature [2], [3], [4], [5], [6], and also computational results on the effect of 3D external conducting structures on growth-rates, eddies and eigenmode geometry [7], [8]. Significant efforts and interest is spent on the topic of realistic experimental-conditions MHD dynamics and the associated control issues. Hosted by the Alfvén Laboratory at the Royal Institute of Technology (KTH) in Stockholm, Sweden, the reversed-field pinch (RFP) machine EXTRAP T2R is armed with dense arrays of saddle-shaped coils, both sensors and actuators. Typical parameters for T2R [9] are I p = 100 kA, R = 124 cm and a = 18 cm. A speciality of T2R is feedback control of magnetohydrodynamic (MHD) instabilities including the resistive-wall mode (RWM). Theoretically, given the geometry and dimensions of T2R and the RFP equilibria, the normalmodes spectrum of magnetic perturbations express ∼ 15 unstable solutions [10], [11], [3]. Experimentally, they have all been shown to be simultaneously suppressable [12], [13]. More recently, experimental verification of stabilized sustained nonaxisymmetrical perturbations to the plasma column has been produced [14], in essence by merely rehashing and refurbishing the classic active shell algorithm of [15], [16].The role of this particular work is to carefully develop a fourier mode decoupled plasma control system, yet another recycling of the mode-controller (MC) concept, in a partially automated manner.MHD control research at T2R strives to generate results not only relevant to RFPs, but also to macroscopic plasma dynamics in general with particular focus on Tokamaks. For example, RWMs in Tokamaks are typically more involved than the nonresonant RWM perturbation in RFPs, which for the latter is current-driven and present for all β. Tokamak RWMs in contrast, are resonant and grows while rotating with respect to the vessel, as well as being pressure-driven, i.e. triggered by high-β. Although different in detail RFP and Tokamak RWMs are basically different instances of the same underlying physical principles and T2R is a highly versa...