A 2-LEVEL ACCELERATION SCHEME APPLIED TO FULL-POTENTIAL FLOW COMPUTATIONS

The TRANAIR aerodynamics code is used as a testbed for developing adva nced iterative methods that lend themselves to parallel or distributed computing. In the case of aerodynamics design optimization, a signifi cant part of the computational cost is in the solution of large sparse nonsymmetric linear systems. We have implemented a two-level method f or solving such linear systems and present preliminary computational r esults for problems in two space dimensions. We observe a lime-space t rade-off mediated by parameters that govern the amount of fill-in in i ncomplete factorizations on each level. In both subsonic and transonic flow regimes, parameter choices exist that lead to substantial improv ements in runtime for comparable memory requirements, relative to a si ngle grid method.