Fatigue Testing and Assessment of Test Data

Abstract

Planning of Testing Fatigue testing may be planned for different purposes, such as testing for documentation of a general design S-N curve for a considered detail or for qualification of a detail in a project. More time is usually available for planning of the testing for the first case than for the second case. The first case also normally involves tests of more specimens than time allows for in the second case. When planning the testing, the purpose of the test must be clearly defined as early as possible. When the purpose is defined, the number of tests and the testing time required can be planned. During this initial planning phase, how the test data will be assessed and transferred into a recommended design methodology for the considered project should also be considered. Some examples of fatigue testing of different details are included in Sections 2.2–2.6 for butt welds in plated structures and piles, small-scale specimens for simulation of fatigue strength in sailing ship structures, large-scale specimens from sailing ships and floating production ships, fillet welded connections, and cover plates or doubling plates. In Section 2.7 approaches to how fatigue test data can be used for assessment of a recommended design procedure, where the principal stress direction during load cycling is not normal or parallel with the weld direction, are provided. Constant Amplitude versus Variable Amplitude Testing Constant amplitude testing is normally performed for derivation of test data representative for the left section of the high cycle part of the S-N curve (or the part of the S-N curve between the low cycle region and the fatigue limit). The fatigue limit in air is understood to refer to the position of the transition of slope in the S-N curve at 10 7 cycles in Figure 1.1. However, in order to obtain relevant test data for the high cycle region to the right of the fatigue limit, a variable amplitude loading, also known as a spectrum loading, must be used. The reason for this is that some load cycles with a stress range larger than the fatigue limit are needed to initiate a crack, such that crack growth will also occur for stress cycles that are below the fatigue limit. This is further illustrated in Sections 3.2.2 and 3.2.3.