Data from top injection ECCS tests conducted at Columbia University have been analyzed as part of an effort to qualify the FLASH-6 computer program for performing post-blowdown heat transfer calculations for the LWBR Safety Analysis. These experiments, which employed a full-scale fuel assembly with electrical heater rods to simulate an inlet rupture for a pressurized water reactor, provided test conditions and rod cooling mechanisms quite similar to those encountered in the postulated LWBR cold leg break loss-of-coolant accident. Clad temperature predictions were obtained using both the modified Dittus-Boelter and the Dougall-Rohsenow correlations to evaluate beyond CHF heat transfer coefficients. Overall comparisons using the FLASH calculated flow rates indicated that the rod temperature calculations were conservative using either of the heat transfer correlations because virtually none of the coolant was calculated to penetrate the heated test assembly. Heat transfer model comparisons were also performed by adjusting the calculation so that coolant was injected directly into the top of the rod bundle to simulate the experimentally observed flow conditions. Once this downflow was forced, conservative temperature predictions were obtained using the Dougall-Rohsenow correlation, whereas the modified Dittus-Boelter beyond CHF option yielded non-conservative results.

A model was developed to estimate the radiation dose commitments received by people in the vicinity of a facility that releases radionuclides into the atmosphere. This model considers dose commitments resulting from immersion in the plume, ingestion of contaminated food, inhalation of gaseous and suspended radioactivity, and exposure to ground deposits. The dose commitments from each of these pathways is explicitly considered for each radionuclide released into the atmosphere and for each daughter of each released nuclide. Using the release rate of only the parent radionuclide, the air and ground concentrations of each daughter are calculated for each position of interest. This is considered to be a significant improvement over other models in which the concentrations of daughter radionuclides must be approximated by separate releases.

METER is a Monte Carlo computer program which can be used to simulate the interaction between independent random variables and their effects on one or more dependent random variables. The program is easy to use for simple simulations but is capable of accommodating complex simulations. METER processes input, generates random numbers from several common frequency distributions under user control, performs the simulation which the user has coded in FORTRAN, and displays results.

Pressure pulse tests were conducted with both solid and flexible test sections installed in a test vessel filled with room temperature water. A surge tank whose volume was approximately equal to that of the test vessel with the test section installed was connected to the test vessel by a /sup 1///sub 8/ inch I.D., 8 inch long surge line. Pressure pulses of magnitude up to 1275 psid and durations from 4.6 to 55.8 msec were generated in the test vessel with a drop hammer and piston pulse generator. FLASH-34 calculations show good agreement with the test data. In particular, FLASH-34 accurately predicts (a) the decrease in peak pressure and the increase in pulse duration due to the presence of a flexible test section, (b) the time delay between the occurrence of the pressure pulse in the test vessel and its arrival in the surge tank and (c) the magnitudes of the transient pressure differences between the test vessel and surge tank caused by the time delay. All of the structural responses were in the elastic range and were approximately quasi-static for the pulss tested. The test data versus calculation comparisons presented here provide preliminary qualification for FLASH-34 calculations of transient …