IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 46, NO. 7, JULY 2018 2691 A First Analysis of JET Plasma Proﬁle-Based Indicators for Disruption Prediction and Avoidance A. Pau , A. Fanni, B. Cannas, S. Carcangiu, G. Pisano, G. Sias, P. Sparapani, M. Baruzzo, A. Murari, F. Rimini, M. Tsalas, P. C. de Vries, and the JET Contributors * Abstract—Reliable algorithms for disruption avoidance and prediction are foreseen to play a fundamental role in the JET control system for the successful operation of the machine in the upcoming deuterium-tritium campaigns. The integration of such algorithms is expected to be a key part also in the implementation of the ITER plasma control system. So far, most of the effort has been devoted to the prediction of disruptions, which is required to mitigate the effects of these transient events, protecting the integrity of in-vessel components. Nevertheless, in order to put in place recover strategies or to have the possibility of a soft landing for the plasma current, the paradigm must be shifted to avoiding disruptions. In this paper, plasma proﬁle-based indicators will be statistically analyzed showing their potential in such a perspective, where warning times and reliability of detection are crucial. Index Terms— Disruption avoidance, disruption prediction, operational space mapping, plasma proﬁles. I. I NTRODUCTION D ISRUPTIVE events still pose a serious problem for the operation of large size tokamak devices, representing, therefore, a key aspect to be considered for the design and operational strategies of next step fusion devices such as ITER and DEMO [1]. If an efﬁcient mitigation is required to avoid damage to the machine, efﬁcient avoidance schemes are needed to possibly bring the plasma back to a nondis- ruptive operating condition. The better the avoidance schemes (i.e., lower risk of a disruption), the lower is the performance Manuscript received June 30, 2017; revised February 12, 2018; accepted April 6, 2018. Date of publication June 12, 2018; date of current version July 9, 2018. This work was supported by the Euratom Research and Training Program 2014–2018 under Grant 633053. The review of this paper was arranged by Senior Editor E. Surrey. * See the author list of “Overview of the JET results in support to ITER” by X. Litaudon et al. to be published in Nuclear Fusion Special issue: overview and summary reports from the 26th Fusion Energy Conference (Kyoto, Japan, 17–22 October 2016). (Corre- sponding author: A. Pau.) A. Pau, A. Fanni, B. Cannas, S. Carcangiu, G. Pisano, G. Sias, and P. Sparapani are with the Electrical and Electronic Engineering Department, University of Cagliari, 09123 Cagliari, Italy (e-mail: alessandro.pau@ diee.unica.it). M. Baruzzo and A. Murari are with the Consorzio RFX (CNR, ENEA, INFN, Universitá di Padova, Acciaierie Venete SpA), Corso Stati Uniti 4, 35127 Padova, Italy. F. Rimini is with the CCFE, Culham Science Center, OX14 3DB Abingdon, U.K. M. Tsalas and P. C. de Vries are with ITER Organization, 13067 Saint- Paul-lès-Durance, France. EUROfusion Consortium, JET, Culham Science Centre, Abingdon, OX14 3DB, U.K. Color versions of one or more of the ﬁgures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identiﬁer 10.1109/TPS.2018.2841394 requirements for the mitigation schemes [2]. In this frame- work, disruption prediction plays a key role and in the last few years, a substantial effort has been devoted to developing more sophisticated prediction systems and improving their perfor- mance both in terms of success rate and warning time [3]–[6]. Many of the presently developed disruption predictors mainly rely on magnetohydrodynamic (MHD) markers related to still rotating modes and, especially, to locked modes, which are basically the ﬁnal precursor of most of the disruptions. Never- theless, in many cases, the warning time is still unsatisfactory with respect to avoidance requirements, and a signiﬁcant step forward needs to be taken. This paper deals with the development of “plasma proﬁle- based indicators” for disruption prediction and avoidance in JET, where parameterized peaking factors (PFs) have been implemented for electron temperature, density, and plasma radiation proﬁles. Proﬁle-based indicators are impor- tant because of their close connection with the plasma stability and the destabilization of MHD modes that even- tually cause the disruption. The basic interplay of the time evolution of different proﬁles will be described in relation to the phenomenology characterizing speciﬁc disruption types together with the relevant time scales. Furthermore, a statis- tical analysis aiming to describe differences and boundaries between the nondisruptive and the disruptive space as well as among speciﬁc types of disruptions will be presented, discussing the implications in terms of disruption prediction and avoidance. II. PEAKING FACTORS In [3]–[7], several physics and engineering parameters have been considered both for disruption prediction and classiﬁ- cation purposes, but many of them depend signiﬁcantly on speciﬁc machine conﬁgurations or characteristics, and on the scientiﬁc program carried out throughout the experimental campaigns. As known, the performance of machine learning algorithms deteriorates outside the training domain and this poses serious concerns in terms of extrapolation to the next step fusion devices. That is why it is very important to focus the attention on something more invariant with respect to the operational domain than the engineering parameters traditionally used. For this purpose, some physics-based indi- cators have been synthesized, which contain information on the time evolution of the main plasma proﬁles such as the electron temperature, the radiated power, and the electron density. In fact, in many cases, the information contained in the proﬁles is strongly related to the phenomenology that 0093-3813 © 2018 EU