MELCOR and GOTHIC analyses of a large dry pressurized water reactor containment to support resolution of GSI-191
Conference Paper
Overview
Identity
Additional Document Info
View All
Overview
abstract
The thermal-hydraulic response of large dry pressurized water reactor (PWR) containments under loss-of coolant accident (LOCA) conditions - particularly with respect to containment pressure and sump pool temperature - is crucial for risk informed decision making about GSI-191. Texas A&M University (TAMU) has developed models with several computer codes including MELCOR and GOTHIC. MELCOR is a best-estimate thermal-hydraulics and severe accident code created and actively maintained by Sandia National Laboratories (SNL) for the Nuclear Regulatory Commission (NRC). GOTHIC is a thermal-hydraulics software package meant for design, licensing, and safety calculations for, among other systems, nuclear power plant containments. It was developed and is maintained by Numerical Applications Inc. (NAI) for the Electronic Power Research Institute (EPRI). The overarching goal of analyses presented here is to produce a best-estimate time profile of sump pool temperature under double-end guillotine break (DEGB) accident conditions. Sump pool temperature has direct implications for sump pool chemistry, residual heat removal in the recirculation phase of a LOCA, and pressure drop across sump screens. For a best estimate of sump pool temperature response, the physics of pipe break effluent flashing, steam condensation and film drainage, and engineered safety features (fan coolers, sprays) must be modeled with good fidelity subject to code constraints. In this paper, aspects of the modeling strategies for MELCOR and GOTHIC are discussed and best estimates of containment thermal-hydraulic response are presented. Particularly for sump pool temperature, code predictions agree at certain transient times and significantly disagree at other times. Further exploration of observed differences via sensitivity calculations is the subject of a separate paper. However, possible explanations for said differences are hypothesized with due consideration given to code mechanics.
