JET is optimizing the plasma scenarios to exploit its capabilities at full power and plasma cur-rent. The goal is to match the best neutron yield achieved when equipped with a carbon wall and extend its stationary duration in the ITER-like Wall configuration, that is with Be tiles on the first wall and W tiles or W-coated tiles in the divertor. The presence of Be and W sets in fact a number of specific challenging conditions. Unlike other lighter elements W features a high radiation cooling rate that remains high well above 10 keV of electron temperature so that its presence in the plasma core should be minimized. Additionally, its high effective charge makes it very sensitive to neoclassical transport, leading to inward convection, rein-forced in presence of plasma rotation [1, 2], such that peaked electron density profiles easily drive W accumulation in the plasma center in absence of a sufficiently large logarithmic ion temperature gradient. Compared to carbon, beryllium radiates much less around the magnetic separatrix and does not offer the same cooling action that mitigates the sputtering on the tar-get walls. Main gas recycling reduces with metal walls.
In typical JET conditions, Edge Localized Modes (ELM) are the primary source of tungsten [3, 4] and in the same time contribute to the net abundance of W by flushing out to the Scrape Off Layer the impurities that enter during the inter ELM phase. Note that in ITER, instead, due to the dominance of the neoclassical temperature screening over the density gradient term , a positive gradient of W density builds up around the separatrix and the ELM is expected to bring W inside, according to JOREK simulations . Interestingly, in the JET edge barrier region it has been observed that the higher the input power the lower the neoclassical inward pinch of W, suggesting that at sufficiently high NBI power W could be screened at the edge by a positive (outward) velocity . However, the input power required to reverse the pinch raises if the plasma current is increased .
Active means such as pellet pacing or sudden vertical plasma displacements (kicks) are availa-ble in JET to artificially increase the ELM frequency  and sustain the expulsion of W when the natural frequency is too low. Increasing gas puffing at the edge also increases the ELM frequency for any given input power but at the cost of reducing the plasma temperature at the pedestal and affect the overall performance.
Tungsten transport in the core of JET is quite well understood . In absence of sawteeth that expel W from the plasma center, ion cyclotron frequency waves (ICRH) help to slow down the W penetration in the core predominantly by flattening the electron density profile and in-creasing the ion temperature gradient. However, the RF power (6 MW) that can be coupled to the plasma in JET is limited to a fraction of the available NBI power (34 MW), which strongly suggests that W source and edge penetration should be minimized. Finally MHD tearing modes can anticipate the drift of W to the core especially when the mode resonance is in the vicinity of the region where the centrifugal force pushes and accumulates tungsten.
The path towards the highest JET performances requires therefore a careful attention on and a deep knowledge of the impurity behavior.
Extending previous work [10,11,12], in this contribution we use the suite of codes available in COCONUT  to run predictive simulations of most of the above processes in the attempt to have an integrated description of the W behavior that encompasses both core and edge in the JET discharges with high input power. JETTO, covering the plasma core, is coupled to EDGE2D-EIRENE that describes the SOL up to plasma facing components. In JETTO and in this edge core-coupled configuration, SANCO describes the behavior of two impurities. Tur-bulent transport is computed by EDWM or GLF23 models, while the neoclassical one is given by NCLASS. In JETTO only cases, QuaLiKiz, TGLF and NEO can also be applied and com-pared to the other models. A sample result is given in Fig. 1 where the evolution of the density profile of injected neon (right) and W (left) is plotted versus time for a discharge with Bt= 2.8 T, Ip = 2.4 MA, 28 MW of NBI and 5 MW of RF deposited in the center (JPN 92419). In the simulation each ELM expels impurities, which enter mainly during the following inter-ELM phase.
The neon profile in particular fits the ELM resolved experimental Ne profile (not shown).
Several experimental conditions are studied and compared in terms of W behavior particularly at the edge transport barrier, including low vs high delta, different levels of gas puffing and input power, plasma current and toroidal magnetic field.
This work has been carried out within the framework of the EUROfusion Consortium and has received funding from the Euratom research and training programme 2014-2018 and 2019-2020 under grant agreement No 633053. The views and opinions expressed herein do not necessarily reflect those of the European Commission
1 C Angioni et al Phys. Plasmas 22, 102501 (2015)  F J Casson et al 2015 Plasma Phys. Control. Fusion 57 014031  N denHarder Nucl. Fus. 56 (2016) 026014  N FEDORCZAC et al. Journ. Nucl. Mat. 463 (2015),  R Dux et al IAEA Fusion Energy Conference St Peters-burg 2014 paper TH/P3-29,  van Vugt DC et al. Phys of Plasmas 26 042508 (2019) & GTA HUIJSMANS Phys. Plasmas 22, 021805 (2015).  M Valisa et al http://ocs.ciemat.es/EPS2018ABS/pdf/P2.1096.pdf
[8 ] E de la Luna et al. Nucl. Fusion 56 026001, 2016.  F J Casson et al . Submitted to Nucl Fusion.
 D.M. Harting et al./Journal of Nuclear Materials 463 (2015) 493–497  V Parail et al./Journal of Nuclear Materials 463 (2015) 611–614  F Koechl et al Pl. Phys. Contr. Fus. 60, (2018) 074008
 M Romanelli et al, Plasma and Fusion Res. Volume 9, 3403023 (2014)
|Country or International Organization||Italy|