Assessment of Current High-Temperature Design Methodology Based on Structural Failure Tests

A mature design methodology, consisting of inelastic analysis methods provided in U.S. Department of Energy guidelines and failure criteria contained in ASME Code Case N-47, exists in the United States for high-temperature reactor components. The objective of this paper is to assess the adequacy of that overall methodology by comparing predicted inelastic deformations and lifetimes with observed results from structural failure tests and from an actual service failure. Comparisons are presented for three structural cases: 1) nozzle-to-spherical shell specimens, emphasizing stresses at structural discontinuities; 2) welded structures, emphasizing metallurgical discontinuities; and 3) thermally loaded cylinders and pipes, emphasizing thermal discontinuities. The comparisons between predicted and measured inelastic responses are generally reasonable; quantities are sometimes overpredicted somewhat and sometimes underpredicted. However, even seemingly small discrepancies in predicted stresses and strains can have a significant effect on life, which is thus not always as closely predicted. For a few cases, the lifetimes are substantially overpredicted, which raises questions regarding the methodology and/or the adequacy of the current design margins.