Severe accident natural circulation flows have been investigated at the Idaho National Engineering Laboratory to better understand these flows and their potential impacts on the progression of a pressurized water reactor severe accident. Parameters affecting natural circulation in the reactor vessel and hot legs were identified and ranked based on their perceived importance. Reviews of the scaling of the 1/7-scale experiments performed by Westinghouse were undertaken. RELAP5/MOD3 calculations of two of the experiments showed generally good agreement between the calculated and observed behavior. Analyses of hydrogen behavior in the reactor vessel showed that hydrogen stratification is not likely to occur, and that an initially stratified layer of hydrogen would quickly mix with a recirculating steam flow. An analysis of the upper plenum behavior in the Three Mile Island, Unit 2 reactor concluded that vapor temperatures could have been significantly higher than the temperatures seen by the control rod drive lead screws, supporting the premise that a strong natural circulation flow was likely present during the accident. SCDAP/RELAP5 calculations of a commercial pressurized water reactor severe accident without operator actions showed that the natural circulation flows enhance the likelihood of ex-vessel piping failures long before failure of the reactor vessel lower head.

This report evaluates both the positive and negative consequences of adding water to a degraded reactor core during a severe accident. The evaluation discusses the earliest possible stage at which an accident can be terminated and how plant personnel can best respond to undesired results. Specifically discussed are (a) the potential for plant personnel to add water for a range of severe accidents, (b) the time available for plant personnel to act, (c) possible plant responses to water added during the various stages of core degradation, (d) plant instrumentation available to understand the core condition and (e) the expected response of the instrumentation during the various stages of severe accidents.

A conceptual model for the vapor and aerosol transport and deposition in the in situ vitrification large-field test off-gas system (OGS) has been developed. This model can be used to predict the emissions from the OGS under normal and off-normal conditions. Results generated by the model can be used to evaluate design and/or procedural modifications, define tests, and predict results. The OGS vapor and aerosol transport and deposition is modeled using the PULSE/MOD-ISV/VER 1.0.0 developmental computer code. Input data requirements for this code include the specific geometries of the OGS components; the composition, rate, and temperature of the vapors and aerosols entering the OGS; and the OGS component surface temperatures or heat fluxes. Currently, not all of these model inputs are available. Therefore, conceptual input parameters are developed. Using this input data, preliminary calculations with the code have been performed. These calculations include a demonstration that the code predicts convergent results, a comparison of predicted results with performance data for one of the OGS components, and a preliminary sensitivity study of the complete model.

This paper presents a simplified methodology for calculating the risk benefit associated with potential performance improvements for the dry pressurized water reactor (PWR) containment. The analysis is based on the June 1989 draft NUREG-1150 results for the Zion commercial nuclear reactor. Simplified containment event trees and the large accident progression event trees from draft NUREG-1150 are used to evaluate the effects of potential improvements on the response of the Zion containment to dominant severe accident sequences. Source terms are generated parametrically using the ZISOR code and offsite consequences are calculated with the MELCOR Accident Consequence Code System (MACCS). These results give point estimates of the risk reduction associated with each containment improvement identified by Brookhaven National Laboratory in their draft Issues Characterization Report.

An evaluation of a Pressurized Water Reactor (PWR) ice condenser containment was performed. In this evaluation, simplified containment event trees (SCETs) were developed that utilized the vast storehouse of information generated by the NRC's Draft NUREG-1150 effort. Specifically, the computer programs and data files produced by the NUREG-1150 analysis of Sequoyah were used to electronically generate SCETs, as opposed to the NUREG-1150 accident progression event trees (APETs). This simplification was performed to allow graphic depiction of the SCETs in typical event tree format, which facilitates their understanding and use. SCETs were developed for five of the seven plant damage state groups (PDSGs) identified by the NUREG-1150 analyses, which includes: both short- and long-term station blackout sequences (SBOs), transients, loss-of-coolant accidents (LOCAs), and anticipated transient without scram (ATWS). Steam generator tube rupture (SGTR) and event-V PDSGs were not analyzed because of their containment bypass nature. After being benchmarked with the APETs, in terms of containment failure mode and risk, the SCETs were used to evaluate a number of potential containment modifications. The modifications were examined for their potential to mitigate or prevent containment failure from hydrogen burns or direct impingement on the containment by the core, (both factors identified as significant contributors to risk in the NUREG-1150 Sequoyah analysis). However, because of the relatively low baseline risk postulated for Sequoyah (i.e., 12 person-rems per reactor year), none of the potential modifications appear to be cost effective. 15 refs., 10 figs. , 17 tabs.