Although drift-ordered fluid models are widely applied in tokamak edge turbulence simulations, the models used are acknowledged not to conserve energy or even electrical charge. The present paper aims to remove many of the existing pitfalls in drift-fluid models, however, with the objective of finding a solution simple enough to be implemented in numerical applications. Our main result is an improved version of the drift-Braginskii equations involving a generalized vorticity function. In the new drift-Braginskii system, the quasi-neutrality condition translates into a transport equation for a generalized vorticity, expressed in conservation form, and related to the total mass-weighted circulation. It is found that kinetic energy conservation can be achieved if the polarization flow is defined recursively. The resulting model conserves the kinetic energy associated with E×B and diamagnetic flows and retains the associated perpendicular kinetic energy flux.

We describe results obtained from Gkeyll, a full-F continuum gyrokinetic code, designed to study turbulence in the edge region of fusion devices. The edge region is computationally very challenging, requiring robust algorithms that can handle large amplitude fluctuations and stable interactions with sheath boundary conditions. Results of turbulence in a scrape-off layer (SOL) for NSTX-type parameters with a model magnetic geometry have been obtained. Key physics of SOL turbulence, such as drive by toroidal bad curvature and steep gradients and interactions with a model sheath boundary condition are included. This allows us to perform parameter scans and physics studies, such as the physics of heat flux width on the divertor plate, and the amplitude and intermittency of SOL turbulence. Initial results find that the heat flux narrows as the connection length is made shorter (the poloidal field becomes stronger). To validate the code, we have studied turbulence in the straight-field LAPD device at UCLA and the Helimak device at the University of Texas. We will also describe the extension of the GENE gyrokinetic code to be full-F, and initial GENE simulations for LAPD.