Pressurized water reactor environments are known to reduce the fatigue life of austenitic stainless steel components when compared to air environments. Laboratory testing has provided a means of quantifying this, allowing conservative plant assessments to be made. The majority of this testing has been isothermal and carried out on membrane loaded hollow or solid specimens. The geometry and loading of laboratory test specimens is significantly different to that experienced on plant, where complex strain waveforms are generally out of phase with temperature changes, and significant through wall strain gradients may be present. To address the issue of realistic loading, a test facility has been developed which can simulate thermal shock loading on a tubular specimen.

The capability of the test facility was presented at the PVP2016 conference [PVP2016-63161]. Since then the facility has evolved, with modifications made to the rig configuration and specimen geometry in order to maximize the strain amplitude from the thermal shock, including the adoption of an annular flow geometry. These modifications were designed to optimize both the heat transfer coefficient and the speed of cycling between hot and cold water in order to induce a thermal strain that can cause mechanical failure within practicable test durations. In order to calculate the magnitude of the thermal strain, detailed calculations were required both in terms of thermal hydraulics as well as stress analyses. The latest stress analysis has been combined with state of the art life prediction models to estimate the time for crack initiation.

This paper presents the results of the latest stress analysis and life prediction, including the derivation of the heat transfer coefficient for an annular flow region. The life prediction method uses best estimate strain-temperature histories from elastic-plastic finite element analysis (FEA). Heat-specific material properties have been developed during accompanying tests within the same experimental programme, and have been applied to enable cyclic hardening to be taken into account. The comparison of the prediction to an on-going test is also discussed.

This content is only available via PDF.
You do not currently have access to this content.