Rapid Source Term Prediction for Swedish Pressurized Water Reactors: Development of a Generic Model
Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Development of tools for use in the fast, online diagnosis of an event or an accident, and in the subsequent radiological source term forecasting at nuclear power plants, is increasingly desired by off-site emergency planning and response personnel. In this thesis the framework for such a tool for Swedish Pressurized Water Reactors (PWRs) is developed. Bayesian belief networks (BBN) are used to model severe accident progression in a generic Swedish PWR. The advantage of using the BBN methodology is that meaningful results can be obtained despite missing information. By adding observable information, e.g. pressure, water level and activity measurements in certain plant compartments the most probable plant state, in case of a severe accident, will be predicted. The plant states will all have an associated environmental source term, i.e. the quantity, characteristics and timing of the release of radioactivity to the environment. The starting points of this thesis are two previously developed BBN models for accident progression in one specific Boiling Water Reactor (BWR) and one generic PWR. These two models provide significant input to the model developed in this master's thesis. Beyond examination of the previously developed models, investigation of plant specific information for the Swedish PWRs has been a key part when modelling the BBN. The model is presented in this report as well as in the BBN software Netica.
Place, publisher, year, edition, pages
2011. , 77 p.
IdentifiersURN: urn:nbn:se:ltu:diva-45456Local ID: 3244f7eb-b0e5-4b51-9c2f-ebe158b16beeOAI: oai:DiVA.org:ltu-45456DiVA: diva2:1018747
Subject / course
Student thesis, at least 30 credits
Fire Engineering, master's level
Validerat; 20110401 (anonymous)2016-10-042016-10-04Bibliographically approved