An efficient method to compute a steady-state solution suitable for computationally intensive turbulent confinement models such as TGLF was developed and incorporated into the ONETWO transport code. It is based on a hybrid approach where the Newton type linearization of the turbulent flux is used only on the predictor step of the computation. A modification of the iteration method that uses an additional stabilization contribution to the implicit flux while at the same time removing it in the explicit part of the iteration scheme was incorporated and appears essential. A significant modification of the turbulent flux calculations was made by smoothing some of the input parameters. Some of these quantities such as elongation and triangularity can propagate persistent noise into the calculations and lead to structure in the solutions if they are not smoothed. Other useful ONETWO improvements include allowance for time-dependent input for gyrotron ECH powers. These new features are available in the new ONETWO version 5.3 that will be released shortly.

A NIMROD study of runaway electron (RE) confinement during disruptions in limited and diverted DIII-D equilibria finds much better RE confinement for the limited, low-elongation shape, consistent with the experimental observation of more reliable appearance of REs for that shape. Despite global similarities in MHD mode amplitudes for the two shapes, the limited plasma displays both greater spatial localization of MHD activity, and wider separation of toroidal mode amplitudes than the diverted plasma. As a result, the minimal RE losses found in the limited plasma are not due to transport to the wall along stochastic magnetic fields (as in the diverted plasma), but instead are associated with a global n=1 shift of the plasma into the center column, which produces a tell-tale toroidally asymmetric RE striking pattern near the inner-wall limiting point. Similar asymmetric RE losses in limited DIII-D RE experiments are suggested by hard x-ray data, but further analysis is required for a detailed comparison of NIMROD results to DIII-D data.

Two major improvements have been made for DIII-D and EAST collaborations during the last EAST experimental campaign. The first was deploying the efficient EAST data replication infrastructure. This system automatically transfers EAST data to the dedicated MDSplus server at DIII-D (gateway.gat.com) immediately after the completion of each EAST pulse. The automation was done via MDSplus events, and the efficient data transfer was implemented with compression and by utilizing the GridFTP parallel data transfer code. As the result, the EAST data for the newest pulse is made available at DIII-D in less than a minute after the pulse, compared to more than 10 minutes previously. Another major improvement was the deployment of EAST web portal (http://webportal.ipp.ac.cn). The DIII-D web portal framework was customized for the EAST environment with three new applications: automatic signal plot generation, PCS status monitoring, and data replication/transfer status monitoring. Both the EAST data replication system and EAST web portal were in production during the last EAST experimental campaign and served the experimental data access and status monitoring needs of both DIII-D and PPPL researchers.

Analytic solution of a reduced model for the parallel velocity shear driven Kelvin Helmholtz (KH) mode in gyro-fluid theory showed that the maximum growth rate occurs at a parallel wave number that increases linearly with the parallel velocity shear. This scaling is very different than that for drift-waves, which have a parallel wave number that is not very sensitive to their temperature gradient drive. The analytic solution was used to refine the algorithm used in the TGLF code to find the parallel wave number that maximizes the growth rate. This analytic result also indicates that GYRO must use a higher poloidal resolution as the parallel velocity shear is increased. With these adjustments, the linear KH mode growth rates in TGLF and GYRO agree well.