Modeling of ITER-like Edge Plasmas by an Improved Adaptive Finite Element Method

An unstructured adaptive Finite-Element mesh generating software for plasma edge physics numerical analysis is developed. It combines the requirements introduced by plasma physics and FE-techniques to handle the extreme gradients of the plasma parameters as well as the complex geometry and magnetic field structure of e.g. ITER-like magnetic fusion devices [1]. The Equilibrium Fitting software output, used for grid generation, is especially suitable to represent magnetic flux behavior, separatrix strike points and the limiter outline. Studies demonstrate strong coupling of edge and core plasma, therefore the entire cross section of the poloidal plasma is treated. The initial representation of the grid is generated automatically at large and can be adjusted interactively. It follows the equipotential and slant lines of the magnetic field. Quadrilateral elements are used. Two edges are in parallel and two edges approximately perpendicular to the field. This takes care of the significant anisotropy, of the plasma transport coefficients along and across the magnetic field. Commonly used triangular elements do not satisfy this condition the same way. The simulation prediction of tokamak, scrape-off layer plasmas in a detached regime requires the resolution of sharp fronts. These are a consequence of essential radiation, ionization, and recombination effects. For this reason structured mesh restrictions app13, because of local refinement limitations. Earlier such FE-applications are known [2]. All this gives reason that unstructured adaptive grid algorithms are ideally, suited to model edge plasmas [3]. A mesh refinement and -coarsening procedure is used. Refinement element's edges are readjusted to the equipotential lines of the magnetic field. Convergence is improved and as a consequence the stability of the numerical solution. This approach emphasizes the application of appropriate FE-techniques to improve performance, convergence and stability of the numerical solutions, in between other essential, when solving problems like bifurcation of the solution.