Investigation of Conditions for Activation of Rupture Disk in BWR Containment Filtering System

• Main Author: Beltran Arroyos, Guillem
• Format: Others
• Language: English
• Published: KTH, Kärnkraftsäkerhet 2011
• Subjects: GOTHIC software; rupture disk; large LOCA
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-45667

Due to the Three Mile Island accident in 1979 the Swedish government took the decision in 1986 to impose a pressure relief system for Swedish BWR’s which prevents containment overpressure in case of LOCA. This pressure relief system consists of a rupture disks in two different systems, non-filtered system 361 and filtered system 362. During a steam line break it is not clear if an unjustified activation of rupture disk 361 or 362 could possibly occur. If significant amount of nitrogen will leak out from the containment then, there is a risk of low pressure in the containment (e.g. due to activation of containment spray) with leaking rupture disks, which might cause air inflow to the containment and burning of hydrogen, so conditions of activation of rupture disk must be studied. The main objective of this master thesis is the investigation of conditions of activation of rupture disk in BWR containment filtering system. In order to find out these conditions specific software called GOTHIC has been used. The methodology of this master thesis has been modeling different containments with GOTHIC software; this thesis work will go from a simple GOTHIC model, that consist in nine lumped control volumes connected by flow paths, until a more complex GOTHIC model that consist in a combination of lumped and 3D control volumes, connected among them by flow paths and 3D connectors. A large LOCA in the upper part of the reactor vessel will be considerate, due to this severe accident; conditions for the activation of the rupture disk will be complying. It has to be mentioned that pressure in the lumped modeling will be lower than pressure in the 3D volumes. Activation time for the lumped modeling will be 8,5 seconds after the steam break for system 362 and activation time for 3D modeling will be 2,8 seconds for system 362 as well. In neither case 361 system will be activated. Considering this is a nuclear safety study and accuracy must be a key point, for further investigations it might be more than advisable using 3D control volumes instead of lumped control volumes. It has to be mentioned also that due to there is no experimental data, uncertainty regarding to the results exist, and if a further safety analysis want to be done, sensitive study of the parameters implemented on GOTHIC software should be performed in the future.