(EX/P3-12) Inter-Machine Comparison of Spontaneous Toroidal Rotation

J.E. Rice1), A. Ince-Cushman1), J. S. deGrassie2), L.-G. Eriksson3), Y. Sakamoto4), A. Scarabosio5), A. Bortolon5), K. H. Burrell2), C. Fenzi-Bonizec3), M. J. Greenwald1), R. J. Groebner2), G. T. Hoang3), Y. Koide4), E. S. Marmar1), A. Pochelon5), Y. Podpaly1)
1) MIT/PSFC, Cambridge, United States of America
2) General Atomics, La Jolla, United States of America
3) EURATOM-CEA, Cadarache, France
4) Japan Atomic Energy Agency, Naka, Japan
5) CRPP EPFL, Lausanne, Switzerland

Abstract.  Plasma rotation plays an important role in the L-H transition, ITB formation and in the stabilization of resistive wall modes. In the current generation of tokamaks, this rotation is usually provided by the external momentum input from neutral beams. In future reactor-grade devices, this may not be available. The spontaneous rotation observed in many tokamaks without external momentum input may provide the solution. Since the mechanism driving spontaneous rotation is not well understood, and in order to anticipate the level of rotation expected in ITER and other reactor devices, a database of observations on several current devices has been constructed. In H-mode or in other enhanced confinement regimes, spontaneous toroidal rotation is generally observed to be in the co-current direction. Co-current spontaneous rotation has been seen on many devices and produced with a wide variety of techniques, demonstrating its fundamental nature. Substantial rotation velocities have been generated with ICRF heating on JET, Alcator C-Mod and Tore Supra. Co-current rotation has been seen in Ohmic H-mode plasmas in COMPASS-D, C-Mod and DIII-D. Co-current rotation has been observed at the edge of ECH plasmas on DIII-D and on JT-60U with a combination of LH waves and ECH. Velocities up to 150 km/s have been measured, without external momentum input. A common feature of all of these observations is a correlation between the toroidal rotation velocity and the plasma pressure or stored energy. On devices which can operate with a large range of plasma current, the rotation velocity is found to be inversely proportional to I. The coefficient of this scaling is different on different devices, and probably includes some machine size scaling information. In order to unify the rotation measurements on all of these devices, the approach of utilizing dimensionless variables has been followed. For the rotation velocity, a Mach number using the background ion sound speed has been chosen. The plasma stored energy or pressure normalized to the plasma current is contained in the normalized plasma β. All of these devices follow a scaling of M increasing with βN. Based on this simple scaling, a Mach number of 0.3 may be expected for an ITER discharge with βN = 1.8, and for T = 15 keV, this corresponds to a spontaneous rotation velocity of 250 km/s or 40 kRad/s.

Full paper available (PDF)