Safety Assessment of Reactor Pressure Vessel Integrity for Loss of Coolant Accident Conditions

One of the critical issues for Reactor Pressure Vessel (RPV) structural integrity is related to the Pressurized Thermal Shock (PTS) event. Therefore, within the framework of safety assessments special emphasis is given to the effect of PTS-loadings caused by the non-uniform azimuthal temperature distribution due to cold water plumes or stripes during emergency coolant injection. The paper describes the method used to predict the thermal mechanic boundary conditions (system pressure, wall temperature). Using a system code the pressure and global temperature distributions were calculated, systematically varying the leak size and the location of the coolant water injection. Local and temporal temperature distributions in the main circulation pipes and at the RPV wall were predicted by mixing analyses with a Computational Fluid Dynamics (CFD) code. The model used for these calculations was validated by post-test calculations of a UPTF (Upper Plenum Test Facility) experiment simulating cold leg injection during a small break Loss of Coolant Accident (LOCA). Comparison with measured temperatures showed that the modelling used is suitable to obtain bounding results. Fracture mechanics analyses were carried out for circumferential flaw sizes in the weld joint near the core region and between the RPV shell and the flange, as well as for axial flaws in the nozzle corner. Stress intensity factors K_I were calculated numerically using the finite element program ANSYS and analytically on the basis of weight and polynomial influence functions using stresses obtained from elastic finite element analyses. Benchmark tests revealed good agreement between the results from numerical and analytical calculations. In order to determine the worst case loading conditions a wide spectrum of thermal-hydraulic transients was considered. Since the resulting load paths decrease with lower temperatures after a maximum, the warm prestress (WPS) effect was employed. The fracture toughness curve determined by deeply notched specimens with high constraint is not representative of the nozzle corner due to the considerable loss of constraint at LOCA conditions. Hence the influence of constraint on fracture toughness was accounted applying the constraint modified master curve concept and the relationship between the T-stress and the reference temperature T₀. According to ASME Code Cases N-629 and N-631 the reference temperatures RT_NDT and RT_T0 can be used alternatively for the adjustment of the K_IC-curve. Therefore both the RT_NDT- and the RT_T0-concept were considered. For all regions of the RPV investigated and the most severe transients it was demonstrated that a large safety margin against crack initiation exists and brittle fracture of the RPV can be excluded.