COMPUTATION LOSS MODELING AND SUPPRESSION

A COMPREHENSIVE COMPUTATIONAL STUDY OF MULTI-DIMENSIONAL (GAS/CURRENT) FLOW EFFECTS IN INHERENTLY TRANSIENT MHD SYSTEMS IS PROPOSED. TWO CLASSES OF FLOWFIELD NON-UNIFORMITY SHALL BE ADDRESSED: 1) ONE CLASS WITH INHERENT STREAMWISE VORTICITY, AND 2) ANOTHER CLASS WITH STREAMWISE VORTICITY ARISING FROM THE E-M INTERACTIONS. BOTH THREEDIMENSIONAL AND TWO DIMENSIONAL TRANSIENT FLOWS SHALL BE STUDIED, WITH THE TWO-DIMENSIONAL INCLUDING BOTH INFINITE SPAN-WISE EXTENT AND SPAN-WISE AVERAGED INCLUDING 3-D EFFECTS. THE TEMPORAL RESOLUTION SHALL INCLUDE: 1) FULL NUMERICAL SIMULATION OF TURBULENT AND CO HERENT FLUCTUATIONS AT LOW RE, 2) LARGE-EDDY SIMULATION OF THE TURBULENCE (WITH APPROPRIATE SGS PROCESS MODELING) AND FULL RESOLUTION OF COHERENT FLUCTUATIONS, AND 3) MODELED TURBULENCE WITH FULL RESOLUTION OF COHERENT FLUCTUATIONS. IN ADDITION, TEMPRALLY-PERIODIC REPETITIVE OPERATION SHALL BE COMPARED TO A SINGLE BURST IN ORDER TO EXPOSE THE 'MEMORY' EFFECT OF FLOWFIELD COMPONENTS WITH LONGER TIME SCALES. FINALLY, THE EFFECT OF SPATIALLY-PERIODIC ELECTRODE/ INSULATOR GEOMETRY SHALL BE COMPARED TO A LOCALLY SPATIALLY-AVERAGED APPROACH WHICH IS THE FORMALLY CORRECT EQUIVALENT TO THE AD HOC INFINITE SEGEMENTATION MODEL. THE COMPUTER DATA SHALL BE RETAINED FOR USE BY THE GOVERNMENT AND ITS CONTRACTORS. THE DATA SHALL ALSO BE PRESENTED IN COLOR '3-D' PLOTS AND VIDEO-TAPES FOR USE IN INTERPRETING EXPERIMENTAL DATA, QUANTIFYING 'LOSS' MODELS FOR SIMPLIFIED CODES, AND SUPPRESSING 3-D EFFECTS VIA SPECIFIC SYSTEM DESIGN MODIFICATIONS.