Reliability of Fluence-Embrittlement Projections for Pressure Vessel Surveillance Analyses

Abstract

A simulated vessel wall environment was constructed that provided for measurements of steel embrittlement increase and neutron flux detector activation at two typical surveillance program locations, as well as at five additional locations, through the thickness of an 8-in.-thick steel “vessel wall.” Neutron spectra for these locations were calculated using one transport and two diffusion theory reactor physics spectrum codes plus a multiple-foil spectrum analysis code. The measured increases in steel transition temperature from the experimental locations revealed the expected gradient of highest embrittlement near the core to least embrittlement at the outer edge of the simulated vessel. Good agreement with published trends was observed. Comparisons of the code calculations versus measurements of the decrease in fluence level between locations, however, were favorable only over the shortest distances. Neutron fluences for the critically important region between fuel core and pressure vessel inner edge were significantly higher from a multiple-foil spectrum analysis code than from transport and diffusion codes when based on comparisons of spectrally adjusted iron activation measurements. This evidence of the possibility that the real fluence values in the pressure vessel wall environment are higher than those produced by common measurement techniques suggests the need for reevaluation of current surveillance data, improvements in reactor physics codes, and continuing assessments of measured versus calculated surveillance fluence data.