COMPUTATIONAL METHODOLOGIES FOR CONVECTION-DIFFUSION PHASE CHANGE PROBLEMS

THE OBJECTIVE OF THIS PROJECT WAS TO DEVELOP NUMERICAL METHODOLOGIES FOR ANALYZING SOLID-LIQUID PHASE CHANGE PROBLEMS INVOLVING MULTI-COMPONENT MATERIALS SUCH AS BINARY MIXTURES AND ALLOYS. THE POTENTIAL NASA INTEREST FOR THIS WORK STEMS FROM MATERIALS SCIENCE APPLICATIONS IN SPACE, E.G. CRYSTAL GROWTH. THE AIM OF PHASE I WAS TO DEFINE A FRAMEWORK FOR DEVELOPMENTOF A GENERAL PURPOSE COMPUTER PROGRAM BY EXAMINING BOTH SINGLE-AND TWO-PHASE MODELS. THE SINGLE-PHASE MODEL EQUATIONS ARE RELATIVELY EASY TO SOLVE BUT REQUIRE A PRESCRIPTION OF THE SOLID PHASE-VELOCITY AND OTHER ASSUMPTIONS FOR COMPUTING THE LOCAL MASS FRACTIONS OF THE PHASES. THE TWO-PHASE MODEL DOES NOT HAVE THESE SHORTCOMINGS BUT NEEDS DATA ON INTERFACIAL EXCHANGE COEFFICIENTS WHICH REQUIRE FURTHER RESEARCH. THE PROBLEM OF FREEZING IN A SQUARE WAS STUDIED TO ILLUSTRATE THE USE OFTHE SINGLE- AND TWO-PHASE MODELS. THE CONCLUSION DRAWN FROMTHIS WORK IS THAT A GENERAL PURPOSE CODE FOR BINARY SOLID-LIQUID PHASE CHANGE PROBLEMS MUST INCLUDE BOTH SINGLE-AND TWO-PHASE OPTIONS.