Strain-controlled fatigue endurance testing in Pressurised Water Reactor (PWR) environments is a well-established testing method that is undertaken in many laboratories around the world. Jacobs have performed such testing over a number of years using both hollow specimens where the water sample flows through the specimen, and bar specimens where the water sample is provided by housing the specimen in an autoclave. Hollow specimen rigs are no longer used in the United Kingdom (UK) due to uncertainties over the impact of pressure loading on fatigue life. The method of strain control for bar specimens is a challenge in an autoclave, where space is limited, and high temperature and high pressure water directly impinges on the specimen surface. Shoulder controlled testing is currently the most reliable way of producing data at Jacobs. In order for shoulder controlled tests to be set up with the correct gauge length strain amplitude, a correlation is made between gauge length extension and the extension between specimen shoulders, i.e. a calibration curve. This has been derived at Jacobs by running a series of strain-controlled dual extensometer fatigue endurance tests in air. Variable Amplitude (VA) fatigue endurance testing in both air and simulated PWR conditions has recently been undertaken by Jacobs as it is considered to be an area where the results have a large potential to reduce conservatisms in fatigue assessments. It is also being investigated for INCEFA-SCALE, a collaborative European Commission (EC) funded Environmentally Assisted Fatigue (EAF) project involving 17 partners. VA testing typically involves running repeating blocks of cycles, where each block includes a small number of cycles at a large strain amplitude (i.e. Overload (OL) or Underload (UL) cycles) and a larger number of cycles at a small strain amplitude (i.e. baseline cycles). In an attempt to achieve the nominal test strains the calibration curve has been used to inform shoulder displacement for both the large and small cycles. There were concerns about using the calibration curve for VA loading, given that it was generated using data from single strain amplitude tests. It wasn't clear whether the shoulder to gauge correction factor for one strain amplitude might change once a specimen had been cycling at another strain amplitude, and whether complex softening and hardening responses in the specimen would affect the validity of the calibration curve. There was a need to understand how much variability there would be in gauge length strain, both throughout a test and within individual blocks, and whether some form of cycle-by-cycle correction might be necessary. Furthermore there was a need to understand the potential level of error in the maximum, minimum and mean strains throughout the test. Accordingly a series of dual extensometer VA tests were carried out in air. Detailed analysis of the data produced, and of VA test data produced from a high temperature water test suggested that it is possible for VA tests to viably be carried out under shoulder control using the calibration curve to inform shoulder displacement for both large and small cycles, at least for the loading conditions assessed in this paper. Supporting evidence, including calibration data, hysteresis loops and stress versus cycles plots, is presented and discussed.
