RELAP5 post-test analyses of OECD/NEA ROSA project experiments on steam generator depressurization with or without non-condensable gas inflow

Two LSTF experiments were conducted for OECD/NEA ROSA Project simulating PWR 0.5% cold leg small break LOCA. Steam generator (SG) secondary-side depressurization was performed by fully opening the relief valves at 10 minutes after a safety injection signal with or without non-condensable gas (air) inflow from accumulator tanks with total failure of high pressure injection system. Further assumptions were made to conduct enhanced SG depressurization by fully opening the safety valves when the primary pressure decreased to 2 MPa and no actuation of low pressure injection system, both to well observe natural circulation (NC) phenomena at low pressures. The primary depressurization rate decreased when non-condensable gas started to enter primary loops because of degradation in the condensation heat transfer in SG U-tubes, while two-phase flow NC has continued even after non-condensable gas inflow. Asymmetric NC behaviors appeared between two loops due probably to different number of forward flow SG U-tubes which would have been under influences of non-condensable gas. Post-test analyses by using JAEA-modified RELAP5/MOD3.2.1.2 code indicated that the code has remaining problems in propor prediction of the primary loop flow rate and SG U-tube liquid level behaviors especially after non-condensable gas inflow. The improvement of the condensation heat transfer model under non-condensable gas mixture condition and the SG U-tube model may be necessary for correct analysis of the LSTF SG depressurization transient. (author)