Analysis of predictive models for correlation of irradiation effects on pressure vessel steels

Abstract

The reactor pressure vessel steel of a nuclear power plant is subjected to fast neutron induced embrittlement. More speciffcally, hardening nanofeatures are produced within the materials by the neutron bombardment, which is of importance for the toughness of the material. The decrease in toughness is characterized by the change in ductile to brittle transition temperature. A number of predictive semi-empirical models, for evaluation of the transition temperature shift, ΔT, have been compared, considering the Ringhals units 3 and 4 base metal data. The US Regulatory Guide 1.99 Revision 1, followed by the French FIM and Miannay formulas, is found to best agree with the results for all Ringhals units 3 and 4 base materials. The transition temperature shift is found to follow a high exponent (& 0:5)fluence behavior, and lacks a significant saturation effect. This conclusion is drawn both from the comparison of semi-empirical models, and from calculations by the least squares method. An effect of the copper content on ΔT is clear for the base metals. Also, next to Cu, analysis indicates strongest ΔT correlations with manganese, considering elemental contents of impurities and alloying elements in the materials. When also the weld metals are taken into account, this effect becomes more pronounced and is found to follow the same behavior as among base materials. While correlations between the data and manganese are strong, effects of nickel and other elements remain uncertain. Comparing data among Ringhals unit 2 materials, an effect of the neutron flux is noticeable. Here, half the samples are subjected to a twice as large neutron flux, and display higher temperature shifts, compared to the other half. It needs to be noted that the database is rather small, i.e. fourteen and nine discrete data points for the base- and weld metals, respectively. Hence, all results are of poor significant quality and additional effects may be recognized when further data is added.