Generalized self-similarity of intermittent plasma turbulence in the boundary layers of fusion devices (Intermittent and solitary transport events in the plasma edge)

• Viacheslav Budaev (NRC Kurchatov Institute, NRU 'MPEI')

Statistical characteristics of edge plasma turbulence in fusion devices are reviewed. The edge turbulence has a complicated multiscale structure. The anomalous transport of mass and momentum is intermittent and is carried by sporadic plasma flux bursts with non-Gaussian statistics, long-range correlation and multifractality. Fluctuations are characterized by the generalized (extended) self-similarity in an extended scale range. The generalized self-similarity (scale invariance) is described in frame of the log-Poisson turbulence model which allows to suggest different functional dependences of the scaling on the geometry of the particular dissipative structures. 1D (filamentary) and 2D (sheet-like) dissipative structures emerging in the cascade process were studied. Many of the scalings experimentally obtained in fusion as well as in space plasmas have been found to be quite close to the scaling predicted by the model for 1D dissipative structures. Turbulent transport scaling can be evaluated using the universal log-Poisson model parameters. Intermittent turbulent transport in such plasmas is characterized by superdiffusion with power law *x* ^ 2 ~ *t* ^ *a*, *a* is from 1.2 to 1.8 demonstrating the universal properties of intermittent turbulence. The work was supported by the RF Megagrant No. 14.Z50.31.0042. 1. V.P. Budaev, S.P. Savin, L.M. Zelenyi, Intermittency and generalized observations self-similarity of turbulent boundary layers in laboratory and magnetospheric plasmas: towards quantifying transfer characteristics, Physics Uspekhi, 181, №9, p. 905-952 (2011). 2. V. P. Budaev, L. M. Zelenyi and S. P. Savin, Generalized self-similarity of intermittent plasma turbulence in space and laboratory plasmas, Journal of Plasma Physics 81, 06 , 395810602 (53 pages) (2015).

Full-torus gyrokinetic simulations of limited FT-2 tokamak plasma with the ELMFIRE code (Role of the electric field in coupling divertor, SOL and edge plasma)

• Laurent Chôné (Aalto University)

The tokamak boundary is of particular importance for magnetic fusion research. Indeed, it gathers several critical technological and scientific challenges that need addressed to achieve and sustain burning plasma regimes [1-5]. This includes plasma surface interaction with high fluence during long pulses, heat and particle transport and exhaust, the physics of edge transport barriers (ETB) formation in advanced scenarios, mitigation of edge-localised modes (ELM) and so forth. Because the scrape-off layer (SOL) and confined plasma are interwoven through these phenomena, faithfully modelling the tokamak edge requires encompassing both regions. In this work, the 3D global full-f particle-in-cell gyrokinetic code ELMFIRE [6] is used to study the plasma edge of the FT-2 tokamak. Recent ELMFIRE developments have enabled full-torus simulations of FT-2 from the magnetic axis to the wall, including the SOL [7,8]. Here, we perform a validation of ELMFIRE SOL modelling and compare the numerical realisations to experimental measurements in FT-2. The simulation results are in turn analysed to provide further insight on the properties of the FT-2 SOL and the nature of the turbulence developing inside it. [1] E. J. Doyle et al., Nucl. Fusion 47 (2007) S18 [2] A. Loarte et al., Nucl. Fusion 47 (2007) S203 [3] P. C. Stangeby, The Plasma Boundary of Magnetic Fusion Devices, IOP, 2000 [4] F. Wagner, Plasma Phys. Control. Fusion 49 (2007) B1 [5] F. Ryter et al., Nucl. Fusion 54 (2014) 083003 [6] J. A. Heikkinen et al., J. Comput. Phys. 227 (2008) 5582‒5609 [7] T. Korpilo et al., Contrib. Plasma Phys. 56 (2016) 549‒554 [8] L. Chôné et al., Contrib. Plasma Phys. to be published

Core and edge turbulence experiments during the 2018 campaign of W7-X (New experiments and diagnostic developments)

• Gavin Weir (Max-Planck-Institute for Plasma Physics)

The optimized stellarator Wendelstein 7-X (W7-X) [1] is designed to have an approximately quasi-isodynamic magnetic configuration and reduced neoclassical transport across the plasma minor radius [2]. Two of the high-level goals of the 2018 experimental campaign of W7-X (OP1.2b) are to achieve high-performance stationary hydrogen discharges and to demonstrate the optimization of W7-X. Assessing the importance of turbulent transport, and experimentally measuring predicted changes in Ion Temperature Gradient (ITG) and Trapped Electron Mode (TEM) driven turbulent transport with quasi-isodynamic optimization are important aspects of the OP1.2b experimental campaign. The collisionless TEM is expected to be stable in a maximum-J, quasi-isodynamic magnetic configuration [3] for a large range of plasma density and electron temperature gradients, 0<η_e<2/3, and the drive for ITG modes is predicted to highly localized in W7-X [4]. W7-X is equipped with multiple plasma density and electron temperature fluctuation diagnostics, including three Doppler reflectometers, a poloidal correlation reflectometer, a radial correlation radiometer and a poloidal correlation radiometer. One of the Doppler reflectometers is equipped with a phased array antenna that can sweep +/-20 degree for density fluctuation k-spectra measurements, while the two other Doppler reflectometers are used for poloidal and zonal flow measurements. In addition to these microwave diagnostics, a soft x-ray tomography diagnostic and a phase contrast imaging diagnostic are also used to measure plasma density fluctuations and their characteristics. In this work, electron temperature and plasma density fluctuation measurements from correlation ECE and reflectometry diagnostics during the first island divertor campaign of W7-X will be summarized (OP1.2a), and an overview of plasma turbulence experiments planned for OP1.2b will be presented. [1] T. Klinger et al. Plasma Phys. Control. Fusion 59, 014018 (2017). [2] J. Nührenberg and R. Zille, Phys. Lett. A 129, 113 (1988). [3] J. H. E. Proll, P. Helander, J. W. Connor, G. G. Plunk, PRL. 108, 245002 (2012). [4] P. Xanthopoulos, PRL 113, 155001 (2014).

Dynamics of turbulent particle flux across the L-H transition in JFT-2M tokamak (Role of the electric field in coupling divertor, SOL and edge plasma)

• Tatsuya Kobayashi (National Institute for Fusion Science)

Importance of role of the radial electric field was widely known for turbulent transport suppression across the L-H transition. Focusing on the ExB shear decorrelation effect, a possible model explaining the high confinement regime was proposed [1], which is nowadays subject to experimental validation [2,3]. Recently, not only the ExB shear but also the curvature of the radial electric field is regarded as an important player to suppress the turbulence transport [3,4]. At the radius of the radial electric field well, confinement improvement owing to the radial electric field curvature, possibly via a modulational coupling between turbulence and flow, was confirmed [4]. How the turbulent transport is reduced is also important issue. There are several routes to reach the reduced turbulent transport state in the H-mode, e.g., via turbulence amplitude suppression or via reduction of phase difference between the density and potential fluctuations. Here, we analyze a turbulence data set obtained with a heavy ion beam probe (HIBP) to investigate how the radial electric field shear and curvature affect the turbulent particle transport across the L-H transition in JFT-2M tokamak. With the HIBP system, the electron density fluctuation and the electrostatic potential fluctuation are simultaneously and directly measured with a high time resolution. In the L-mode, turbulence amplitude is enhanced and the outward particle flux is observed, which results in the confinement degradation. After emergence of the radial electric field structure in the H-mode, the outward particle flux is significantly reduced. The particle flux is suppressed predominantly by reducing the density fluctuation amplitude and the cross phase between the density fluctuation and the potential fluctuation. Amplitude reduction of the potential fluctuation is only moderate. Both the shear and the curvature are found to play an important role to reduce the particle flux. Different time scale of changes in the density fluctuation amplitude and the cross phase between the density fluctuation and the potential fluctuation is also observed. [1] H. Biglari, P. H. Diamond, and P. W. Terry, Phys. Fluids B 2, 1 (1990). [2] J. A. Boedo et al., Nucl. Fusion 42, 117 (2002). [3] R. A. Moyer et al., Phys. Plasmas 2, 2397 (1995). [4] K. Kamiya et al., Sci. Rep. 6, 30585 (2016)

Review on electrostatic SOL turbulence simulations based on probes experiments (Role of the electric field in coupling divertor, SOL and edge plasma)

• Jan Horacek (Institute of Plasma Physics)

I’m reviewing a large European team progress in 2003-2018, showing inter-connections between various tokamak SOL turbulence studies. It is based mostly on tokamak TCV reciprocating probe data. Thanks to relatively high data quality we found interesting fractal properties in analogy with a sandpile [Graves’05, Horacek’05, Labit’07]. We cross-checked them with a 2D fluid turbulence plasma model ESEL with transport coefficients based on fundamental principles [Fundamenski’07]. This describes plasma turbulence (split into blobs), driven by fundamental interchange instability through gradients of the magnetic field and plasma pressure. This yielded many interlinked statistical benchmarks [Garcia’06, Pitts’07, Garcia’07] with especially the plasma density spatial profiles and its fluctuations we found described very well by this simple 2D turbulence model also on tokamak ASDEX Upgrade [Horacek’10]. We concluded later [Ondac’15], however, quite bad model description of both the potential and temperature profiles and fluctuations on both ASDEX Upgrade and COMPASS obtained by ball-pen probes [Adamek’10]. Even though we attempted joining ESEL with another 1D model [Havlickova’11], we stopped using ESEL as being too simplistic. Later we continued with a more complex 3D model [Halpern’13, Halpern’16] capable of quantifying the Near-SOL feature studied experimentally using tokamak TCV wall probes [Tsui’17] and IR camera [Nespoli’17] with practical implications towards ITER first wall design [Horacek’15, Kocan’15, Horacek’16] observed on COMPASS and another 10 tokamaks. Using extremely long probe time-series of TCV edge plasma turbulence, we argue that the established paradigm of the "long-range correlations" is false; rather the observed correlations are a pure statistical coincidence of uncorrelated blobs passing by a probe [Garcia’15, Theodorsen’16]. Extremely high TCV statistics allowed scaling of blob velocities [Tsui18] in four established edge plasma regimes.

ELM electromagnetic fine structure in tokamak discharges (Intermittent and solitary transport events in the plasma edge)

• Monica Spolaore (Consorzio RFX, Padova)

The presence of filamentary structures widely characterizes the edge region of fusion devices, independently from their magnetic configuration. These structures are generally revealed by pressure peak locally emerging on the cross-field plane from the plasma background and the name filaments is due to their extended size along the magnetic field line. Filaments emerging from turbulence background share these general features with Edge Localized Modes (ELMs), which are responsible of a large fraction of transport towards the plasma wall and divertor plates. The filament electromagnetic features were experimentally studied in the recent years due to the relevance of this aspect on the transport, considering as an example the transition between closed and open magnetic field line topology, i.e X-points and advanced divertor magnetic configurations, up to the possible magnetic field line bending effect in case of enough high current associated to ELM filaments, which could enhance their interaction with the first wall. A further aspect to be considered is the enhancement of the electromagnetic effect is expected at high beta plasmas, in particular an increasing current density characterizing turbulent structures was experimentally observed to characterize turbulent filaments as the beta increases. Local magnetic fluctuations were experimentally detected in presence of ELMs in different tokamaks such as MAST, ASDEX-U and EAST and DII-D and JET. In this contribution a direct measurement of current density associated to ELM will be presented in the two tokamak experiments COMPASS and RFX-mod. Measurements were performed with specifically designed insertable probe heads, properly equipped with 2D arrays of electrostatic and magnetic sensors. The diagnostic set is completed in both experiments with arrays of probes located on the divertor or first wall. The plasmas analyzed are based on SN shaped tokamak discharges where the H-mode was achieved in ohmic or NBI heated plasmas, in COMPASS, or edge biasing stimulated in RFX-mod. The latter representing a newly explored scenario with this technique, where ELM electromagnetic composite filamentary structures are observed. They are characterized by clear vorticity and parallel current density patterns. On the other side, recent preliminary analysis revealed the presence of parallel current associated to the COMPASS ELMs and evidenced their fragmented structure within the Scrape-Off Layer (SOL) region. The comparative analysis of the experimental results will be based on the same method applied on the two devices, aiming to the identification of insight on possible common physics.

Two-dimensional structure of nonlinear wave in linear magnetized plasma (New experiments and diagnostic developments)

• Fumiyoshi Kin (Kyushu University)

Nonlinearly generated structure and its properties have been studied commonly for general physics. In magnetized plasma, pressure inhomogeneity generates drift waves to form several types of nonlinear waves. In particular, a streamer is known as a nonlinear wave [1] that is important for the plasma confinement. It is because that the streamer has radially elongated structure, thus enhances radial transport. The streamer was found in a linear plasma device, PANTA, for the first time, and to be produced through the couplings between drift waves and low frequency fluctuations, namely mediator [2]. Recently, we have succeeded in extracting the precise structure of nonlinear wave of streamer and mediator using the conditional averaging. It is found that the degree of localization increases with the amplitude of the streamer and mediator, which is in agreement with the nonlinear property of soliton. Moreover, as is shown in Fig. 1, the two-dimensional structure of the streamer envelope is reconstructed. This talk discusses the newly obtained nonlinear characteristics and two-dimensional spatial structure of streamer and mediator. Figure 1 : Reconstructed 2-dimensional structure of streamer envelope. References [1] S. Champeaux et al., Phys. Lett. A 288, 214 (2001) [2] T. Yamada et al., Nature Phys. 4, 721 (2008)

GAM and zonal flow damping for high collision numbers and by passing electrons in gyrokinetic tokamak simulations (Role of the electric field in coupling divertor, SOL and edge plasma)

• Klaus Hallatschek (Max-Planck-Institute for Plasma Physics)

Using the gyrokinetic code CGYRO [1] employing the sophisticated collision operator of Sugama [2] - essentially the gyro averaged lowest order Hirshman-Sigmar operator -- the transition between the fully kinetic and the fully fluid regime has been mapped for GAMs and zonal flows. The collision operator is sufficiently accurate to reproduce the two-fluid damping of the GAMs and residual zonal flows in the limit of large collision number. For GAMs for edge typical safety factors, surprisingly the damping by electrons is always important, even in collisionless cases without trapped electrons. The damping can be understood by an argument similar to Fermi's golden rule. Moreover with increasing collision numbers a maximum of the damping rate occurs at $$\nu_{ii}\sim\omega_{GAM}.$$ The maximum damping is relatively small (much smaller than the one of the zonal flows) so that the worst quality factor of the GAM resonance is still of the order ~100, which would allow, e.g., its external excitation. Finally at very high collision numbers the damping rate decreases again, while the frequency approaches the fluid value [3,4]. For zonal flows, while at small collision numbers the damping is of the order of an ion collision time, at sufficiently high collision numbers the collisional damping decreases again, while the residual approaches the fluid value, which is higher than the collisionless value. References [1] J. Candy, E.A. Belli, R.V. Bravenec, J. Comput. Phys. **324** 73 (2016) [2] H. Sugama et al., Phys. Plasmas **16**, 112503 (2009) [3] K. Hallatschek, A. Zeiler, Phys. Plasmas **7**, (2000) 2554 [4] K. Hallatschek, Plasma Phys. Control. Fusion **49**, B137-B148 (2007)

Comparative edge‐SOL density profile and turbulence measurements during the I‐phase in JET, ASDEX, COMPASS and EAST (Intermittent and solitary transport events in the plasma edge)

• Dániel Réfy

An intermediate phase (M-mode [1], I-phase [2][3] or LCO [4]) close to the L-H transition has been identified at various magnetic confinement fusion devices, sharing the same feature of low frequency oscillation (LFO, few kHz), quasi-periodic changes of the plasma edge pressure, turbulence amplitude and flow velocity. This contribution aims at a comparative analysis of this phase in terms of the modulation of edge and SOL density and temperature pedestal properties and the high-frequency oscillations at the JET, AUG, COMPASS and EAST tokamaks. The main tool used in the investigation is the Lithium Beam Emission Spectroscopy (Li-BES) diagnostic whose capabilities depend on the observation geometry and the amount of light collected. Consequently, the information from different devices is of different nature or resolution. However, the common observation is that the edge density gradient and turbulence amplitude is modulated at the LFO frequency. At AUG and JET the analysis of the ECE signals, while at COMPASS the probe measurements revealed that the electron temperature is also modulated in phase with the density. Further, a high frequency oscillation (HFO, 10-150 kHz) of the magnetics and the density was characterized, its amplitude is modulated with the LFO frequency. The relation between these and both the density and the temperature pedestal modulation will be shown. The virtual 2D measurement at COMPASS allows investigation of the poloidal flow velocity modulation, while the real 2D Li-BES system at EAST revealed the poloidal and radial wave structure of the HFO activity. Our results indicate that these phenomena at the various experiments are likely related. [1] E. R. Solano et al 2017 Nucl. Fusion 57 022021 [2] G. Birkenmeier et al 2016 Nucl. Fusion 56 086009 [3] G.S. Xu et al 2014 Nuclear Fusion 54 013007 [4] O. Grover et al accepted for publication in Nuclear Fusion

GAM evolution in L-mode approaching the L-H transition on JET (Role of the electric field in coupling divertor, SOL and edge plasma)

• Carlos Silva (IST)

The interaction between zonal flows (ZFs) and turbulence is a self-regulating mechanism. Understanding this interaction is crucial to control plasma confinement. The shearing due to ZFs is thought to dominate in regimes when the mean shear flow is modest as before and during the L-H transition [e.g. 1]. This was corroborated by findings in different devices demonstrating the importance of both the oscillating and mean flow shear and their interaction in triggering the transitions [1-4]. While on AUG the sheared flow below the L-H threshold is dominated by Geodesic Acoustic Modes (GAMs) [1], on devices such as DIII-D [2], EAST [3] and HL-2A [4] GAMs do not appear to be important on the way to H-mode. The reported results reveal that no clear picture exists on the relevance of GAMs in the turbulence collapse required for the formation of steep pressure gradients at the transition. This contribution focuses on the characterization of GAMs in JET L-mode plasmas when approaching the L-H transition aiming at understanding its possible role in triggering the transition. Doppler backscattering has been used to investigate GAMs by measuring oscillations in the E × B flow velocity. Experiments were performed in NBI heated discharges for different values of plasma current (2.2 < Ip < 3.2 MA) and line-averaged density (1.6 < n < 3.1×1019 m-3). The dataset also includes variations in toroidal magnetic field, magnetic configuration and hydrogen isotopes. Results demonstrate that parameters such as plasma current and density have a strong effect on the GAM amplitude. By assessing the importance of critical parameters such as safety factor and collisionality, experimental evidence is found for the different mechanisms determining the GAM amplitude: turbulence drive, collisional and collisionless damping. GAMs have been studied along the power ramp used to induce the L–H transition, taking advantage of the unique JET dataset. As the heating power increases, the GAM amplitude first increases but then is reduced as the L-H transition is approached. GAMs are either suppressed or have a modest amplitude at the transition. The cause of this reduction is however unclear as the GAM damping rate is expected be reduced along the heating power ramp and density fluctuations levels and the E × B shear flow display modest changes. Experimental investigation of isotope effects in hydrogen and deuterium plasmas was also performed. Stronger GAMs were found in D than in H plasmas at low heating power (PNBI < 4 MW) associated with larger edge density fluctuations in H. However, above PNBI  4 MW, the GAM amplitude is reduced in D plasmas while increases for H plasmas. Unfortunately, the L-H transition could not be achieved in H plasma due to heating power limitations.

The impact of the radial electric fields on turbulence spreading in the edge and Scrape-Off Layer of TJ-II (Role of the electric field in coupling divertor, SOL and edge plasma)

• Gustavo Grenfell (Consorzio RFX)

Turbulence spreading is the transfer of free turbulent energy from strongly (i.e., unstable regions) to weakly driven locations \cite{garbet:1994}. The net effect of this phenomenon is the radial distribution of energy, which couples different turbulent regions and changes local plasma features. It has been pointed out that spreading may be important in setting the Scrape-Off Layer (SOL) width. Since the peak divertor heat load is intimately related to the SOL width, the comprehension of the mechanisms setting the SOL properties is essential in the framework of the prediction of the SOL width in ITER. In this work, we report measurements of the turbulence drive and turbulence spreading profile, as defined by Manz, P. et al \cite{Manz:2015}, from the near edge to the far SOL region of the TJ-II stellarator, for ECRH and NBI heated plasmas. A 2-D Langmuir probe array \cite{Alonso:2015} was used to measure both parameters as well as the profiles of floating potential, plasma density, radial turbulent particle flux, effective radial velocity, turbulence correlation time and phase velocity of the fluctuations. In addition, the plasma edge was modified by means of a biasing electrode, inserted into the edge of the plasma ($\rho \approx 0.85$) and delivering a voltage $\pm$ 350 V (with respect to the wall), in a square waveform, with a frequency of 40 Hz. The transition from ion to electron root (when the heating is changed from ECRH to NBI) deeply impacts the plasma parameters, due to the establishment of a shear layer close to the Last Close Flux Surface (LCFS). The local turbulence drive is mainly reduced in the shear layer, while the turbulence spreading decreases in the SOL, where its value is higher than in the edge for both regimes. On the other hand, in the NBI at -350 V, the velocity shear reached its maximum, resulting in a strong suppression of turbulent transport and the effective fluctuating radial velocity, not only in the shear layer, but also in the far SOL. Moreover, the ion saturation profile got steeper near the shear layer and more truncated in the SOL. Finally, both the local turbulence drive and turbulence spreading are impacted by the biasing. The driving term is strongly reduced in the shear layer, and only slightly reduced in the SOL. However, turbulence spreading is mainly modified in the SOL in the vicinity of the LCFS, when the $E_r\times B$ shear reaches values close to the inverse of the turbulence correlation time. In short, biasing was found to reduce edge-SOL coupling by decreasing turbulence spreading, thus affecting the ion saturation current profile, which may have an impact in the SOL width.

Impact of the near SOL radial electric field on the H-mode power threshold and turbulence suppression using EDGE2D-EIRENE simulations (Role of the electric field in coupling divertor, SOL and edge plasma)

• Alex Chankin (Max-Placnk-Institute for Plasma Physics)

Scrape-off layer (SOL) and divertor make a direct impact on the outer core region via neutral penetration and ionization. Experiments on several tokamaks show however that even when neutrals’ ionization doesn’t significantly alter plasma parameter profiles in the core, changing divertor geometry can still make a strong impact on the H-mode power threshold, PLH, (factor 4 variation in the case of JET [1]), suggesting that there may be other, more subtle mechanisms for the influence of the SOL and divertor on the core plasma. EDGE2D-EIRENE simulations of JET plasmas revealed that the near SOL radial electric field, Er, is the parameter most affected by the divertor geometry change, suggesting that the strong local ExB shear may suppress turbulence [1,2]. It may serve as an additional mechanism for the confinement improvement, with the main one located in the outer core (2D edge fluid code simulations show the largest drop in transport coefficients occurring inside of the separatrix). Further code modelling of JET plasmas by changing hydrogen isotopes (H-D-T), toroidal field direction, and by replacing carbon with the ITER-like wall, showed that the magnitude of the localized near SOL Er anti-correlates with experimentally observed PLH variations [2]. The isotope effect on the near SOL Er, in particular, was due to the increase in the convective power flux through the separatrix for lighter neutrals, which, for the given total input power, reduced electron temperature, Te, at the separatrix, causing a strongly amplified Te reduction at the strike point on the divertor target. Code simulations of isotope exchange (D vs. H) experiments in JET showed that a large reduction of anomalous diffusion coefficients, D_purp, in the outer core and SOL is necessary to match experimental density profiles [3]. This applies to both L- and H-mode plasmas, and it is the most obvious explanation for poorer particle confinement in H. The exact mechanism responsible for poorer particle (and energy) confinement in H is unknown. If SOL physics was to contribute to this, it would likely be via the influence of Er on turbulence suppression. Earlier code modelling didn’t contradict this, as it showed progressively larger near SOL Er in the H-D-T series of runs (see above). In that modelling, however, D_purp was assumed the same for all isotopes. Recent modelling with the reduction of D_purp, from H to T cases, showed that a simultaneous reduction in electron and ion perpendicular heat conduction coefficients, Chi_e,i, is required to achieve an amplification of the isotope influence on the near SOL Er. Whether a positive feedback loop exists, with larger SOL Er in heavier isotope plasmas causing turbulence reduction and a drop in transport coefficients, leading to sharper Te gradients and an increasing local Er, can only be established in further modelling to be carried out soon.

Enhancement and suppression of turbulence by energetic-particle-driven geodesic acoustic modes (Role of the electric field in coupling divertor, SOL and edge plasma)

• Makoto Sasaki (Kyushu University)

We investigate the phase-space dynamics of spatially inhomogeneous turbulence with a transport barrier in the presence of EGAMs, based on the wave-kinetic equation. It is found that the trapped turbulence clumps leak across the transport barrier, and they can propagate in the turbulence stable region. As a result, turbulence is enhanced by EGAMs in the stable region, while turbulence suppression is obtained in the unstable region. The propagation of the turbulence is ballistic, with the phase velocity of the EGAM. Hence, there appear a new global characteristic velocity for turbulence dynamics, in addition to the local group velocity and that of the turbulence spreading. The propagation of trapped turbulence is different from processes such as turbulence spreading and avalanches.

Experimental observation of Edge Coherent Mode during the L-H transition (Role of the electric field in coupling divertor, SOL and edge plasma)

• Frederic Clairet (CEA)

The foreseen plasma operation in future tokamak devices to reach fusion would be in H-mode. It is desirable to understand the underlined physic, and particularly the turbulence, that provides the improvement of the plasma confinement. Major modification occurs close the plasma edge into the pedestal region where the turbulence not only reduces dramatically but also changes in nature. In this work, experimental observation of these changes have been performed using reflectometry diagnostic which provides high radial (in the mm range) and temporal resolution (in the µs range). During the L to H transition the density fluctuation level (n/n) drops significantly at the edge and the frequency spectra turn from broadband to coherent modes in a narrow region of the pedestal. This change occurs at the position of the Er well position (calculated from the diamagnetic contribution) along with its growth. Moreover, a detailed analysis of the complex reflected signal displays a single side band feature at low frequency around few kHz in the spectra which reverses of sign radially. This feature is discussed in terms of eddy tilting in changing ExB poloidal shear flow. This shear decorrelation mechanism of the turbulence is suspected to be at the origin of the observed particle transport reduction.

GAM-controlled TEM turbulence self-organization and particle transport isotope effect in FT-2 tokamak experiment and gyrokinetic modeling (Role of the electric field in coupling divertor, SOL and edge plasma)

• Evgeniy Gusakov (Ioffe Institute)

In this paper the trapped electron mode turbulence and ExB flows are investigated in hydrogen and deuterium Ohmic FT-2 tokamak plasmas experimentally by a set of microwave Doppler backscattering diagnostics and numerically by the global gyrokinetic ELMFIRE code. The plasma poloidal velocity associated with GAMs is shown to be larger than the mean flow velocity, thus possessing a strong impact on the ExB flow shear. The effective shearing manifests itself as a modulation of the turbulence level and particle flux by the space-time GAM pattern. GAMs are modulating turbulence due to rotation shear and, in its turn, the modulated turbulence is pumping the GAM due to the Reynolds stress shear (RSS). The numerical analysis demonstrates the self-organized turbulent dynamics, when the pumping of the flow by the turbulence is found to be balanced by collisional damping. The GAM activity is shown, both in experiment and computation, to be more intensive in deuterium because of smaller damping in the case of heavier isotope. The RSS, the ExB flow and the particle flux oscillate close to in-phase in time, while the flow shear induced by GAMs follows with a phase shift of pi/2. This behavior is consistent with experimentally observed high coherency of turbulence level and plasma poloidal velocity dynamics at the GAM frequency.

Phase Contrast Imaging: A Versatile Diagnostic to Measure Turbulence, MHD Modes and RF Waves in Magnetically Confined Fusion Plasmas* (New experiments and diagnostic developments)

• Miklos Porkolab (MIT)

Phase contrast imaging (PCI) is an internal reference beam interferometric technique which provides a direct image of line integrated electron density fluctuations in a plasma. The method has been used with great success to measure waves and turbulence in magnetically confined high temperature plasmas over the past three decades. The principle of PCI was developed in optics in the 1930s by the Dutch physicist Zernike, leading to the development of phase-contrast microscopy for which he received the Nobel Prize. The technique allows one to detect the variation of the index of refraction of a dielectric medium (such as a plasma) due to the presence of waves or turbulent density fluctuations. The concept relies on the introduction of a phase plate in the light beam path, causing a 90 phase shift in the unscattered portion of the beam relative to the scattered one, and subsequently the expanded laser beam is imaged onto a cryogenically cooled detector array that combines the two beams, thus allowing one to measure wavelengths and correlation lengths of fluctuations. The first time such method was used to measure waves in fusion plasmas is attributed to Henry Weisen who detected Alfvén waves in the TCA tokamak in Lausanne. Other experiments followed in rapid succession by Porkolab and coworkers, measuring not only turbulence in the edge and the core plasmas in Alcator C-Mod and DIII-D but also detecting coherent waves in the ICRF regime with heterodyne techniques that use Acousto-Optic Modulators. Further work included studies of MHD modes, for example Alfvén wave cascades in the presence of reversed shear and energetic ions in ICRF heated plasmas in C-Mod. More recently, the PCI diagnostic has been implemented in the Wendelstein 7-X stellarator at the Max Planck Institute in Greifswald and initial results have been presented very recently. In this talk, the principle of PCI is discussed and examples of measurements will be presented from a wide range of recent and past experiments, demonstrating the versatility of this diagnostic. From practical considerations, typical wave numbers that can be measured with PCI using a 10-50 Watt cw CO2 laser at 10.6 um wavelength are in the range of 1.5  k(cm-1)  30 and frequencies 10 kHz to 2.0 MHz. This covers most of the wave-numbers and frequencies of interest in turbulence studies, including ITG, TEM and some ETG modes. It also includes typical wave numbers of ICRF waves. Finally, we note the recent extension of the PCI diagnostic on the DIII-D tokamak where it was combined with an interferometer method to measure long (0  k(cm-1)  4.0) wavelength fluctuations at frequencies up to 1 MHz. * Work supported by the US Department of Energy

Importance and interplay of neoclassical and turbulent transport in modern stellarators (New experiments and diagnostic developments)

• Felix Warmer (Max Planck Institute for Plasma Physics)

With the recent start of the optimized modern stellarator Wendelstein 7-X, the upgrade of the Large Helical Device and the advent of gyrokinetic simulations in stellarator geometry, for the first time it has now become possible to study and disentangle neoclassical and turbulent transport effects in stellarator geometry. ECRH experiments conducted in the latest campaigns of W7-X and LHD give first indications towards a better understanding of the interplay of neoclassical and turbulent transport in stellarators. The strong ECRH heating gives rise to electron-root conditions in both devices with strongly peaked electron temperatures and flat ion temperatures as well as a strong positive radial electric field in the region where Te is peaked. In both devices it has been observed that the measured radial electric field is in excellent agreement with neoclassical theory despite the fact that the neoclassical energy and particle fluxes are insufficient to explain the experimental fluxes over a wide range of parameters. Thus, it seems that the radial electric field in stellarators is well described by enforcing the ambipolarity constraint on the neoclassical particle fluxes as provided by standard neoclassical theory. This means that the remaining turbulent transport must be intrinsically ambipolar which is consistent with gyrokinetic theory for electromagnetic micro-turbulence. The fraction of turbulent energy flux, however, seems to vary with plasma parameters. In standard performance scenarios, the neoclassical energy flux can explain up to half the energy flux and becomes more dominant in high-performance, high-temperature scenarios due its strong temperature dependence. First gyrokinetic simulations indicate that W7-X electron-root plasmas are dominated in the plasma centre by ETG while in similar LHD plasma scenarios, TEM seems to be the main drive in the plasma centre but not at the plasma edge. A general statement about the micro-turbulence characteristics cannot be easily given due the strong dependence on the magnetic geometry. While the scales of neoclassical and turbulent transport are well separated, they still condition each other due to their interdependency resulting from plasma profile and gradient modifications. A specific influence has the radial electric field and its shear, which is determined by the ambipolarity condition (but is connected to the plasma profiles via the neoclassical particle fluxes) and directly acts as an external constraint on micro-turbulence. But also the neoclassical transport coefficients themselves are strongly affected by the radial electric field. An often neglected effect is the total particle transport. The combination of the neoclassical and turbulent particle fluxes as well as particle sources determine the density profile and its gradient, which has a strong impact on e.g. stabilization of ITG or the drive of TEM. Consequently, a first-principle understanding will always result in a connection of particle and energy transport.

Method of Bursts for Investigation of Plasma Turbulent Transport (Intermittent and solitary transport events in the plasma edge)

• Irakli Nanobashvili (Andronikashvili Institute of Physics of the Ivane Javakhishvili Tbilisi State University)

As it is well known transport in magnetised plasma is highly bursty, intermittent and has strong convective character. Coherent turbulent structures bring important contribution to such transport. They are formed intermittently on diffusive background and propagate radially outwards at a speed which is a fraction of ion sound speed. Different methods are used to study turbulent transport in magnetised plasma. We proposed method – hereafter “method of bursts” – which is based on selection of large amplitude fluctuations – bursts of plasma density (the method can be applied to any other fluctuating signal – floating potential etc.). After the selection of bursts from fluctuating signal we calculate their temporal characteristics - burst rate, inter-burst time and burst duration. Radial dependence of these characteristics together with statistical properties of fluctuations allow better understanding of plasma turbulent transport. Method of bursts has been applied to the analysis of plasma turbulent fluctuations measured at the edge of Tore Supra tokamak. Physical picture, which is based on dynamics of coherent turbulent structures, for the explanation of experimental observations has been proposed. The method has been also used for the analysis of data generated by one of the two-dimensional fluid codes. Good agreement has been found between experiment and modeling and all this will be presented in detail. Comparative analysis of plasma turbulent fluctuations measured on Tore Supra and CASTOR tokamaks has been performed. This analysis by means of the method of bursts allowed to conclude that blob-like structures should be responsible for plasma turbulent transport at the edge of Tore Supra tokamak, while at the edge of CASTOR we have elongated streamer-like structures. Method of bursts has been also used to analyse the experimental data obtained during electrode biasing discharges on the CASTOR tokamak. Method of bursts allowed for the first time to demonstrate experimentally, that biasing splits coherent turbulent structures into larger number of smaller structures thereby reducing plasma transport and improving the confinement. The main reason is that biasing modifies the radial electric field and imposes strongly sheared poloidal rotation on plasma through Er × Bt electric drift.

Planned Physics Investigations with an Imaging Heavy Ion Beam Probe at ASDEX Upgrade (New experiments and diagnostic developments)

• Gregor Birkenmeier (Max Planck Institute for Plasma Physics)

We present a conceptual study and the potential physics investigations of an imaging heavy ion beam probe (i-HIBP), which is planned to be developed for the ASDEX Upgrade tokamak. The main feature of this new type of a heavy ion beam probe (HIBP) is the imaging of the secondary beams by means of an in-vessel scintillator screen in combination with a high speed camera. The spatio-temporal pattern on the scintillator contains two-dimensional information about the plasma density, the plasma potential and the poloidal magnetic field at the points of ionization of the primary beam, where the secondaries are created. Due to the use of a neutral beam as primary beam and the in-vessel imaging of the secondaries, the i-HIBP system is much more compact than a classical HIBP. A numerical study for a neutral 80 keV cesium beam as primary beam has shown, that information about the density, the plasma potential and the poloidal magnetic field from 10 cm inside the last closed flux surface up to the far scrape-off layer can be obtained for a typical low-density ASDEX Upgrade plasma. On large time scales (several milliseconds) the i-HIBP data could be used to estimate the electron density profile by means of a forward modelling approach as already standard for beam emission spectroscopy applications. On the same time scale, a more accurate estimation of the equilibrium poloidal magnetic field can be used to study the edge current density evolution during an ELM cycle. On smaller time scales, the expected resolution allows for radially resolved signatures of geodesic acoustic modes and zonal flows. In general, the expected radial resolution and the possibility to get 2D information indicates that the i-HIBP could become an excellent diagnostics for edge physics investigations.

Power Threshold Studies on ASDEX Upgrade and the Influence of the SOL on the H-mode Onset (Role of the electric field in coupling divertor, SOL and edge plasma)

• Ulrike Plank (Max-Planck-Insitut für Plasmaphysik)

The characteristic feature of the H-mode in a tokamak plasma is the edge transport barrier (ETB). It is believed to be caused by the edge radial electric field $E_r,$ which leads via the $\mathbf{E} \times \mathbf{B}$ velocity shear to the suppression of turbulent transport and thus, to the formation of an ETB. Due to its important role in H-mode physics, the edge $E_r$ is considered to be a key player also in the transition from L- to H-mode. On the macroscopic scale it is seen that the heating power has to exceed a certain threshold $P_{\rm thr}$ to transit from L- to H-mode. According to an inter-machine scaling, $P_{\rm thr}$ scales with the line-averaged density $\langle n_\mathrm{e} \rangle $, toroidal magnetic field $B_\phi$ and plasma surface $S$ [1]. The scaling is, however, only valid for high densities, at which the electron and ion channels are coupled [2] and for magnetic configurations with the ion $\nabla B$ drift pointing towards the X-point. Furthermore, it does not include other dependencies, such as the divertor condition or the wall material. Recent experimental studies from ASDEX Upgrade indicate that a critical ${\mathbf{E} \times \mathbf{B}}$ shear is needed to enter H-mode [3]. They could explain the $B_\phi$-dependence of $P_\mathrm{thr},$ as the $E_r$ shear is observed to increase with increasing magnetic field. For a more complete understanding of the L-H transition, the exact location of the $E_r$ shear is crucial. Findings from several machines suggest that the scrape-off layer (SOL) physics influences $P_\mathrm{thr}.$ These are supported by modelling results showing that the divertor and wall conditions change $E_r$ in the SOL [4]. This in turn would impact the ${\mathbf{E} \times \mathbf{B}}$ shear across the separatrix and, therefore, could modify the condition for the L-H transition. To experimentally address these relations, a new gas puff-based charge exchange recombination spectroscopy diagnostics was installed at ASDEX Upgrade in order to measure $E_r$ across the separatrix with high radial resolution at the outer mid-plane. **References** [1] Y. R. Martin et al., *J. Phys.: Conf. Ser.* **123** 012033 (2008). [2] F. Ryter et al., *Nucl. Fusion* **54** 083003 (2014). [3] M. Cavedon et al., *EPS Conf. Proc.*, 2017. [4] A.V. Chankin et al., *Plasma Phys. Control. Fusion* **59** 045012 (2017).

Role of the X-point height in the access to H-mode in COMPASS (Role of the electric field in coupling divertor, SOL and edge plasma)

• Renaud Dejarnac

The access to H-mode is a crucial step for ITER operation but is still not completely understood and the prediction of its L- to H-mode power threshold (PLH) relies on a multi-machine scaling law based on macroscopic parameters such as density, magnetic field and machine size [1]. However, it has been demonstrated that local effects, e.g., divertor geometry or X-point height, can affect strongly but also differently PLH from one machine to another. The reliability of the empirical scaling law is therefore questioned, making the predictions for ITER uncertain. In the tokamak COMPASS, the PLH strongly increases with the height of the X-point (up to 50% after a fresh boronization). COMPASS has the particularity to access the H-mode in ohmic when the power crossing the separatrix reaches a certain threshold. In this study, the L-H transition is triggered with a plasma current ramp-up and mechanisms at play are investigated for a low and a high X-point configuration using an extensive set of diagnostics. In L-mode just before the L-H transition, a larger radial electric field (Er) is measured close to the separatrix by the horizontal reciprocating probe for the low X-point configuration when PLH is the lowest, thus confirming the role of Er in stabilizing the turbulent transport. Considering the gradient of the electron temperature (dTe/dr) being proportional to Er, a similar result is observed by the high resolution Thomson scattering system around the separatrix. Prior the L-H transition, dTe/dr is two times higher for the low X-point, whilst exhibiting the same value in H-mode for both the high and low configurations. In the divertor, the electron temperature at the outer strike-point measured by Langmuir probes is 50% higher in the low X-point configuration. In the frequency spectra, a well localized mode at 6 kHz is observed only for the high X-point configuration prior the L-H transition and disappears when the H-mode is reached. This mode runs in the ion diamagnetic drift direction, therefore it is not a dithering mode nor a low frequency zonal flow and is seen on electric probe, Da and magnetic coil signals. [1] Y. R. Martin et al., J. of Physics: Conf. Series 123 (2008) 012033

Basic characteristics of the multiscale interaction between drift tearing modes and ITG modes (Role of the electric field in coupling divertor, SOL and edge plasma)

• Lai Wei (Dalian University of Technology)

Basic features of the multiscale interaction between drift tearing modes (TM) and ion temperature gradient (ITG) modes are studied by using an electromagnetic Landau fluid model in a slab geometry. It is observed that the energy spectra with respect to wavenumbers become broader during the transition phase from ITG-dominated stage to TM-dominated stage. Accompanied with the fast growth of magnetic island, the frequency of TM/ITG with long/short wavelength fluctuations in the electron/ion diamagnetic direction decreases/increases respectively. The decrease of the TM frequency is identified to result from the effect of the profile flattening in the vicinity of magnetic island, while the increase of the frequencies of ITG fluctuations is mainly due to the eigenmode transition of ITG induced by the large scale zonal flow and zonal current related to the drift TM. Roles of zonal current induced by the ITG fluctuations in the instability of drift TM are also analyzed. Finally, the electromagnetic transport features in the vicinity of magnetic island are discussed.

Multi-field/-scale Interactions of Turbulence with Magnetic Islands in the DIII-D Tokamak and Comparison to Gyrokinetic Simulations (New experiments and diagnostic developments)

• Laszlo Bardoczi (UCLA, ORAU)

We present the first localized measurements of ITG scale temperature and density fluctuations and TEM scale density fluctuations modified by m=2, n=1 magnetic islands. These islands are formed by a Neoclassical Tearing Mode (NTM) deep in the core plasma at the q=2 surface. NTMs are important as they often degrade confinement and lead to plasma termination. This is the first experimental confirmation of a long-standing theory prediction [1] of decreased local small-scale turbulence levels across large-scale magnetic islands. Our measurements capture a mean reduction of turbulence inside (and enhancement just outside) the island region during island evolution. Additionally, in the island saturated state the fluctuations at the O-point are reduced compared to the X-point [2]. A novel, non-perturbative measurement finds reduced cross-field electron thermal diffusivity (by 1-2 orders of magnitude) at the O-point, consistent with the local turbulence reduction. Initial comparisons to the GENE non-linear gyrokinetic code are promising with GENE predicting the observed turbulence reduction inside the island and increase just outside the island and replicating the observed scaling with island size [3]. These results are significant as they allow the validation of gyrokinetic simulations modeling the interaction of these multi-scale phenomena [4]. Improved understanding of NTM physics and interaction with turbulence can lead to improved control of NTMs and plasma performance, therefore has potential implications for future fusion devices such as ITER. Supported by USDOE under DEFG02-08ER54984, DE-FG02-08ER54999 and DE-FC02-04ER54698. [1] C. J. McDevitt and P. H. Diamond, PoP, 13 032302 (2006); [2] L. Bardoczi et al, PRL 116 215001 (2016); [3] A. B. Navarro et al, 59 034004 PPCF (2017); [4] L. Bardoczi et al, PoP 24 056106 (2017).

Residual Zonal Flows in Reactor Tokamaks Plasmas (Role of the electric field in coupling divertor, SOL and edge plasma)

• T.S. Hahm (Seoul National Univeristy)

\documentclass{epsconf} \usepackage{graphicx} %\usepackage{epsfig} % use this package to include EPS format figures \usepackage{wrapfig} \usepackage{amsmath} \title{Residual Zonal Flows in Reactor Tokamaks Plasmas} \author{\underline{T.S. Hahm}$^{*}$ and Y.W. Cho} \institute{Department of Nuclear Engineering, Seoul National University, Seoul, 08826, Republic of Korea\\ $^*$ tshahm@snu.ac.kr} \begin{document} \maketitle %\begin{wrapfigure}{r}{70mm}\centering %\vspace{0cm} % Adjust vertical figure placement %\includegraphics[width=70mm]{epslogo} %\caption{\it \small EPS logo} %\label{fig:flow} %\vspace{0cm} % Adjust vertical figure spacing %\end{wrapfigure} Zonal flow plays an important role in turbulence regulation via its flow shearing\cite{Hahm1999}. It is mainly in poloidal direction and therefore gets reduced by collisionless process in toroidal geometry. But it asymptotes in time to a non-zero value called the residual zonal flow level \cite{RH1998}. In this study, we investigate the residual zonal flow in reactor plasmas using modern gyro-kinetic approach\cite{Hahm1988}. We find that the residual zonal flow can be enhanced for the intermediate radial wavelength range on the order of the ion banana width in the presence of slowing down fusion products. \begin{thebibliography}{99} \bibitem{Hahm1999} T.S. Hahm, M.A. Beer. Z. Lin, G.W. Hammett, W.W. Lee, and W.M. Tang, Phys. Plasmas \textbf{6}, 922 (1999) \bibitem{RH1998} M.N.Rosenbluth and F.L. Hinton Phys. Rev. Lett. \textbf{80}, 724 (1998) \bibitem{Hahm1988} T.S. Hahm, Phys. Fluids {\bf 31}, 2670 (1988) \end{thebibliography} \end{document} \endinput %% %% End of file `sample.tex'.

Spectral gyrokinetic implementation of sonic rotation (Role of the electric field in coupling divertor, SOL and edge plasma)

• Jeff Candy (General Atomics)

Sonic toroidal plasma flow, on the order of the ion sound speed, arises in tokamaks due to external torque driven by neutral beam injection. This flow has a profound effect on drift-wave turbulence and corresponding radial transport fluxes. Historically, gyrokinetic theory and simulation operate (almost) exclusively in the weak rotation limit, retaining only the E$\times$B flow, Coriolis drift and toroidal rotation shear. However, correct treatment of the sonic rotation regime requires the inclusion of centrifugal effects, which are quadratic in the Mach number. In 1998, Sugama formulated a comprehensive and rigorous gyrokinetic system that describes sonic rotation and associated centrifugal terms, and is valid for general electromagnetic perturbations [1]. This formulation, importantly, includes the corresponding particle, energy, momentum and exchange transport coefficients which are required to obtain the correct equations for profile evolution. We show that the most general implementation is critically important for the study of heavy impurity transport. In particular, using the more accurate theory, nonlinear turbulent fluxes for tungsten are radically different than in the weak-rotation regime. In this presentation we give particular emphasis to a discussion of a new approach for the implementation of shear in the E$\times$B flow. This shear is different than the previous rotation terms in that it cannot be treated simply or directly in a flux-tube. In the past, E$\times$B shear has been treated using either *non-periodic boundary conditions*, or in the case of flux-tube codes, using a discontinuous *wavenumber shift* method [2]. We report on the development of a new discrete *wavenumber advection* algorithm that treats the shear with spectral accuracy without spurious boundary effects or a discontinuous time-history. Because the new algorithm may also be used to treat profile shear, it is well-suited to treat multiscale gyrokinetic simulations in the steep-gradient pedestal region.

Edge-scrape-off layer coupling and turbulence spreading (Intermittent and solitary transport events in the plasma edge)

• Peter Manz (Max-Planck-Institut für Plasmaphysik)

Transport codes such as SOLPS, EDGE2D, EMC3, SOLedge2D and UEDGE are still the main workhorses for scrape-off layer (SOL) investigations. They are based on a diffusive description of the turbulent transport, where the transport is proportional to the mean gradient. In particular in the far-SOL or at high densities where a density shoulder is formed the mean gradients are flattened out and strong events cannot be locally excited. However, in particular in the far-SOL or at high densities strong fluctuation levels are observed. It is well known that these strong events called plasma blobs or filaments are generated in the plasma edge around the separatrix and propagate into the far SOL. Therefore, the plasma edge as the driving region has to be coupled to the SOL. This makes SOL transport intrinsically nonlocal. Usually, filamentary transport can be included by a convective term in the transport model. This in principle allows for transport across the flat gradient region. It will be shown that such a description faces problems once the perpendicular transport competes with the parallel transport. The transport of fluctuation amplitude is called turbulence spreading. In magnetically confined plasma physics turbulence spreading and nonlocal transport has been mainly investigated in the core in the context of zonal flow and pinch physics as well as heat transport phenomena. In the scape-off layer exhibiting low gradients and high fluctuation amplitudes this phenomenon has been rarely considered so far. In the present contribution it will be shown how nonlocal transport and turbulence spreading can be quantified and modeled with application to scrape-off-layer plasmas.

Poloidal asymmetry of the perpendicular plasma flows in Tore Supra tokamak (New experiments and diagnostic developments)

• laure vermare (LPP)

Simultaneous measurements of binormal velocity of density fluctuations using two separate Doppler backscattering systems at the low field side and at the top of the plasma show significant poloidal asymmetry. The measurements are performed in the core region between the radii (0.7<r/a<0.95), over a limited number L-mode discharges covering a wide range of plasma conditions in the Tore Supra tokamak. Ripple effect is an option to explain this poloidal asymmetry since non-ambipolar particle flux induced by ripple losses have been shown to be the dominant mechanism that sets the radial electric field in the plasma core on Tore Supra. In this previous work, only the flux surface average of the radial electric field was derived and compared to measurements. In reality, since the magnetic ripple decreases significantly from low field side toward high field side, a possible poloidal dependence of the radial electric field induced by ripple losses is conceivable. In addition, a possible generation mechanism by the ballooned structure of the underlying turbulence, especially in the form of convective cells, has proposed for explaining the observation of these poloidally asymmetric mean flows. In the present contribution, the both possible origins of this unexpected poloidal asymmetry are discussed and investigated through comparisons with numerical simulations. For this purpose, simulations are performed using a full-f, semi-Lagrangian, global gyrokinetic simulations obtained using the GYSELA code and as much as possible, the experimental radial profiles as well as a simplified synthetic Doppler backscattering diagnostic to reconstruct the perpendicular velocity of density fluctuations.

Deep radial electric field and Reynolds stress profiles measured with Langmuir and ball-pen probes on COMPASS (New experiments and diagnostic developments)

• Ondřej Grover (Institute of Plasma Physics, The Czech Academy of Sciences, Prague, Czech Republic)

The Reynolds stress and is gradient is a key quantity for the experimental validation of theories of turbulence suppression by zonal flows and other sheared flows. A common diagnostic for its measurement is an array of appropriately positioned Langmuir probes. However, simultaneous Reynolds stress profile measurements with Langmuir and ball-pen probes show a strong influence of electron temperature fluctuations on the final measured value. The diagnostic used on COMPASS is a complex probe head mounted on a horizontal reciprocating manipulator which allows measurements deep in the edge plasma. This enables a direct measurement of $E_r$ as well as the radial-poloidal component of the Reynolds stress tensor $R_{rp}$. While the Reynolds stress gradients calculated from either probe type are found to be comparable in the close vicinity of the LCFS, they are significantly different even in terms of polarity deeper inside the LCFS. Specifically, the $R_{rp}$ calculated from Langmuir probes is found to be lower than from ball-pen probes due to high-frequency $T_e$ fluctuations. The radial profile of the differences in the measured values is examined in detail and a correction for the Langmuir probe measurement in terms of local plasma parameters measurable by Langmuir probes is proposed.

Intermittency in ASDEX Upgrade I-mode edge plasmas (Intermittent and solitary transport events in the plasma edge)

• Tim Happel (Max-Planck-Institut für Plasmaphysik)

The I-mode is an improved confinement regime characterized by the presence of a temperature pedestal, while the density profile remains comparable to that in L-mode [1]. This means that in I-mode a transport barrier is formed in the heat transport channel, while the particle transport channel is not affected. This has several benefits for the operation of a tokamak: both impurity accumulation and ELMs are absent in I-mode, resulting in a comparably benevolent plasma. After substantial I-mode research by the fusion community in the last years, the mechanism which creates a transport barrier in only one of the transport channels is still not understood. In ASDEX Upgrade (AUG) I-modes, the edge fluctuation amplitude is reduced substantially, which leads to the dominance of the so-called weakly coherent mode (WCM) in the density turbulence spectrum [2,3]. It will be shown that both poloidal correlation reflectometry [4] and correlation electron cyclotron emission [5] diagnostics detect the WCM in the L-mode phase long before I-mode starts, showing that the WCM is not exclusive to the I-mode. A newly installed thermal helium beam diagnostic on AUG [6] confirms these results and allows a precise radial determination of maximum impact of the WCM. It will be reported that the reduction of fluctuation amplitudes at frequencies different from the WCM is accompanied by the emergence of solitary and intermittent structures with large amplitudes, even larger than in L-mode [7]. Hence the PDF of density fluctuations in I-mode exhibits a strong tail towards large amplitudes, which is observed for all wavenumbers investigated ($k_\perp = 5$-$12~{\rm cm}^{-1}$, with $k_\perp$ the perpendicular wavenumber of the density flucutations) [8]. These density turbulence bursts are linked to the WCM and their amplitudes scale with the I-mode confinement quality. The increased dominance of the WCM during improving I-mode, connected to the increased intensity of turbulence bursts could be an indication that the WCM is responsible for the generation of the turbulence bursts. [1] D. G. Whyte et al., Nucl. Fusion **50**, 105005 (2010). [2] A. E. Hubbard et al., Phys. Plasmas **18**, 056115 (2011). [3] P. Manz et al., Nucl. Fusion **55**, 083004 (2015). [4] D. Prisiazhniuk et al., Plasma Phys. Control. Fusion **59**, 025013 (2017). [5] S. J. Freethy et al., Rev. Sci. Instrum. **87**, 11E102 (2016). [6] M. Griener et al., Plasma Phys. Control. Fusion **60**, 025008 (2018). [7] T. Happel et al., Nucl. Fusion **56**, 064004 (2016). [8] T. Happel et al., Plasma Phys. Control. Fusion **59**, 014004 (2017).

Impact of flows on avalanche dynamics in a basic transport experiment (Intermittent and solitary transport events in the plasma edge)

• Bart Van Compernolle (UCLA)

Results of a basic heat transport experiment$^{1,2}$ involving an off-axis heating source are presented. Experiments are performed in the Large Plasma Device (LAPD) at UCLA. A ring-shaped electron beam source injects low energy electrons (below ionization energy) along a strong magnetic field into a pre-existing, large and cold plasma. The injected electrons provide an off-axis heating source that results in a long, hollow, cylindrical region of elevated plasma pressure embedded in a colder plasma, and far from the machine walls. The source is active for a period long compared to the density decay time, i.e. as time progresses the heating power per particle increases. Two distinct regimes are observed: 1) an initial regime dominated by avalanches, identified as sudden intermittent rearrangements of the pressure profile; 2) a state dominated by sustained drift-Alfven wave activity following a global collapse of the density profile. The avalanches are triggered by the rapid growth of drift-Alfven waves. The data suggest that ambient flows play a critical role in the dynamics, in particular in the onset of the avalanches through the interplay of the stabilizing flow shear and the destabilizing pressure gradient. The flows are a consequence of the boundary conditions at the ring-source. Recently a new configuration has been implemented that allows active control of the flows by changing the bias between the emitting ring and surrounding carbon masks. A parameter regime has been found in which avalanches are absent. The new source configuration also provides some control over the size and frequency of avalanches when present. This work is supported by the NSF grant PHY1619505, and is performed at the Basic Plasma Science Facility, sponsored jointly by DOE and NSF. **References:** $^{1}$ B. Van Compernolle et al. Phys Rev. E **91**, 031102 (2015) $^{2}$ B. Van Compernolle et al, Phys. Plasmas **24**, 112302 (2017)

3D modelling for transport and turbulence in the WEST divertor (Intermittent and solitary transport events in the plasma edge)

• Federico Nespoli (Aix-Marseille Universite)

First experimental results show the achievement of a variety of diverted configurations in WEST [1], which features a full tungsten divertor that will undergo heat fluxes of the same magnitude as for the ITER divertor [2]. The transport code SolEdge2D-EIRENE [3] has been used extensively in the past both in support of the WEST design [4] and campaign planning [5], and to model the first experimental results. The turbulence code TOKAM3X [6] is here used for the first time to model the plasma turbulence in the edge and scrape-off layer of WEST. TOKAM3X allows to model the 3D evolution of turbulence in a realistic magnetic geometry, featuring one or more X-points. We compare in particular two configurations, a standard single-null case featuring a secondary X-point at the top of the plasma, and a “shallow divertor” configuration, where the main X-point is located at the divertor plate. The influence of the divertor configuration on transport and the steady state profiles is discussed. The effect of the presence of the two X-points on turbulence is studied through the application of a blob detection and tracking technique, here presented, allowing to resolve the temporal evolution of the 3D shape of the filaments. The results from the blob tracking technique are compared with a conditional average sampling analysis, commonly used in experiments. The blob sizes and velocities obtained from both techniques are compared with existing scaling laws. This comparison, together with a cross-correlation analysis, indicates that the presence of the two X-points helps in disconnecting the low field side midplane blobs from both the divertor leg and the high field side of the tokamak. [1] M. Goniche et al, “First LHCD experiments in WEST”, 45th EPS Conference, Prague 2018 [2] C. Bourdelle et al., Nucl. Fusion 55 (2015) 063017 [3] H. Bufferand et al., Contrib. Plasma Phys. 56 (2016) 555 562 [4] H. Bufferand et al. Nucl. Fusion 55 (2015) 053025 [5] G. Ciraolo et al., Nuclear Material and Energy 12 (2017) 187-192 [6] P. Tamain et al., Journal of Computational Physics 321 (2016) 606623