Reduced Order Modeling of Reactor Pressure Vessels for Probabilistic Fracture Mechanics

In order to provide efficient estimates on the safe life of damaged nuclear Reactor Pressure Vessels (RPVs), simulation based models have been created for probabilistic analyses. As probabilistic analyses generally require millions of model evaluations, a model for rapid and accurate stress intensity factor calculations is necessary to obtain useful stochastic results. The Fracture Analysis of Vessels, FAVOR code, is a program designed to perform probabilistic assessments of RPVs, though its analysis is limited to axis aligned flaws for discrete flaw characteristic inputs. In an effort to extend this analysis to 3D off axis flaws for any continuous input, the weight function method utilized in FAVOR was implemented within a finite element framework more conducive to 3D flaw evaluations. Using the INL's fracture mechanics and material embrittlement code, Grizzly, RPV fracture models were utilized in the development of a reduced order model for approximating stress intensity factors to enable efficient probabilistic analyses. Additionally, surrogate modeling techniques were investigated and utilized within the reduced order model such that continuous flaw inputs could be evaluated.Thesis (M.S., Mechanical Engineering) -- University of Idaho, 201