TURBULENT NATURAL CONVECTION IN REACTOR CORE MELT

PROGRESSION OF MOLTEN POOL, FORMED IN SEVERE ACCIDENT SCENARIOS, IS STRONGLY DEPENDENT ON ITS HEAT TRANSFER CHARACTERISTICS. THE PURPOSE OF THE PROPOSED RESEARCH IS TO PERFORM A COMPUTATIONAL ANALYSIS OF THE TURBULENT NATURAL CONVECTION IN INTERNALLY HEATED POOLS. TWO VERSIONS OF THE LOW-REYNOLDS-NUMBER MODELS WILL BE FIRST VALIDATED FOR THE PREDICTION OF TURBULENT NATURAL CONVECTION FLOWS OF INTEREST BY COMPARING THEIR PREDICTIONS WITH THE AVAILABLE EXPERIMENTAL DATA; THE REYNOLDS STRESS MODEL WILL BE EMPLOYED, IF NEEDED, IN THE SUBSEQUENT PHASES OF THE WORK. PREDICTIONS FOR TURBULENT FLOW AND HEAT TRANSFER IN HEMISPHERICAL POOLS WILL BE CARRIED OUT TO REMOVE THE UNCERTAINTIES IN THE FREQUENTLY EMPLOYED APPROACH OF DEDUCING THE HEAT TRANSFER CHARACTERISTICS FROM AVAILABLE CORRELATIONS FOR SEMICIRCULAR GEOMETRIES. THE VALIDATED TURBULENCE MODEL WILL THEN BE INCORPORATED IN A COMPUTATIONAL METHOD FOR BOUNDARY FITTED COORDINATES TO ENABLE PREDICTIONS OF FLOWS IN COMPLEX POOL GEOMETRIES. THE OVERALL OBJECTIVE OF THIS RESEARCH PROJECT IS TO DEVLOP A MODEL FOR THE PREDICTION OF TURBULENT NATURAL CONVECTION IN POOLS OF VARIOUS SHAPES CONTAINING INTERNALLY HEATED FLUID AND TO GENERATE COMPREHENSIVE DATA FOR THE TYPICAL POOL GEOMETRIES ANTICIPATED IN THE LATE PHASES OF SEVERE REACTOR ACCIDENTS.