Residual Stresses in Clad Nuclear Reactor Pressure Vessel Steels: Prediction, Measurement and Reconstruction

Most residual stress measurement methods are limited in terms of their stress and spatial resolution, number of stress tensor components measured and measurement uncertainty. In contrast, finite element simulations of welding processes provide full field distributions of residual stresses, with results dependent on the quality of the input conditions. Measurements and predictions are often not the same, and the true residual stress state is difficult to determine. In this paper both measurements and predictions of residual stresses, created in clad nuclear reactor pressure vessel steels, are made. The measurements are then used as input to a residual stress mapping technique provided within a finite element analysis. The technique is applied iteratively to converge to a balanced solution which is not necessarily unique. However, the technique aids the identification of locations for additional measurements. This is illustrated in the paper. The outcomes from the additional measurements permit more realistic and reliable estimates of the true residual state to be made. The outcomes are compared with the finite element simulations of the welding process and used to determine whether there is a need for additional input to the simulations.

Download Full-text

Assessment of a Stress-Free Temperature Model for Residual Stresses in Surface Cladding of a Reactor Pressure Vessel

Volume 6B: Materials and Fabrication ◽

10.1115/pvp2017-65255 ◽

2017 ◽

Author(s):

B. Richard Bass ◽

Paul T. Williams ◽

Terry L. Dickson ◽

Hilda B. Klasky

Keyword(s):

Residual Stress ◽

Finite Element ◽

Thermal Expansion ◽

Residual Stresses ◽

Base Metal ◽

Pressure Vessel ◽

Reactor Pressure Vessel ◽

Normal Operation ◽

Inner Surface ◽

Reactor Pressure

This paper describes further results from an ongoing study of a simplified engineering model that is intended to account for effects of clad residual stresses on the propensity for initiation of preexisting inner-surface flaws in a commercial nuclear reactor pressure vessel (RPV). The deposition of stainless steel cladding during fabrication of an RPV generates residual stresses in the cladding and the heat affected zone of the under-lying base metal. In addition to residual stress, thermal strains are generated by the differential thermal expansion (DTE) of the cladding and base material due to temperature changes during normal operation. A simplified model used in the ORNL-developed FAVOR probabilistic fracture mechanics (PFM) code accounts for the clad residual stress by incorporating a stress-free temperature (SFT) approach. At the stress-free temperature (Ts-free), the model assumes there is no thermal strain, i.e., the thermal expansion stresses and clad residual stresses offset each other. For normal cool-down transients applied to the RPV, interactions of the latter stresses generate additional crack driving forces on shallow, internal surface-breaking flaws near the clad/base metal interface; those flaws tend to dominate the RPV failure probability computed by FAVOR. In a previous report from this study (PVP2015-45086), finite element analysis was used to compare the stresses and stress-intensity factors (SIF) during a cool-down transient for two cases: (1) the existing SFT model of FAVOR, and (2) directly applied RPV clad residual stress (CRS) distribution obtained from empirical (hole-drilling) measurements made at room temperature on an RPV that was never put into service. However, those analyses were limited in scope and focused on a single flaw orientation. In this updated study, effects of CRS on the SIF histories computed for both circumferential and axial flaw orientations subjected to a cool-down transient were determined from an extended set of finite element analyses. Specifically, comparisons were made between results from applying CRS experimental data to ABAQUS two-dimensional, inner-surface flaw models and those generated by the FAVOR SFT model. It is demonstrated that the FAVOR-recommended SFT value of 488 °F produces conservatively high values of SIF relative to the use of CRS profiles in the ABAQUS models. For the vessel and flaw geometry and transient under study, the circumferential flaw (360° continuous) required a decrease of SFT down to 390 °F to match the CRS SIF histories. For the infinite axial flaw model, a decrease down to 300 °F matched the CRS SIF histories. Future plans are described to develop more general conclusions regarding the FAVOR model.

Download Full-text