Estimating inclusion distributions of hard metal using fatigue tests

The maximum inclusion size in clean steels influences fatigue behaviour and other mechanical properties. In this paper three different approaches to estimating the distribution of the inclusion sizes which cause the failure are compared. The data used are from a staircase test. The load levels on the broken and unbroken specimens as well as the size of the inclusions which cause the failures are recorded. One model uses only the load levels and the other two models use the recorded inclusion size on the broken specimens and the load level for the unbroken specimens. The difference between the models using the inclusion size is that one is based on the assumption that the specimen breaks in the largest inclusion, and the other one on the assumption that the specimen breaks in some inclusion over a critical size, not necessarily the largest one. It is shown that from a staircase test with measured inclusion size on the broken specimens it is better to use both load levels and inclusion sizes to estimate the inclusion size distribution instead of using just the different load levels. The two models using the inclusion sizes almost coincide for big inclusions. These models are also used to verify or reject possible fatigue limit models. (C) 2002 Elsevier Science Ltd. All rights reserved.