Magnus Holmgren, Klas Johansson, Erland Johnson, Thomas Svensson
May 26, 2013
This paper describes a fatigue testing inter-laboratory comparison between six Nordic laboratories, which performed fatigue tests on steel specimens. The specimens were tested with three different stress levels (maximum stress) and with a load ratio R = S min /S max = 0.1. Each laboratory tested four specimens at each stress level. The participating laboratories reported the results together with the uncertainty of measurement. The calculated results showed a significant difference between laboratories. However, when the differences in modelling were taken into consideration, no significant differences between the laboratories remained. Significant problems connected with calculations of uncertainty of measurement were detected. One of the major conclusions from the study is that it is of great importance how fatigue tests are reported, e.g. the method of modelling. Better knowledge of how to report fatigue tests, including uncertainty of measurement, as well as a more generally accepted way of reporting such tests – and especially how to treat run-outs – would be a great improvement for the customers of the laboratories.