Abstract
The paper presents a method which enables a probabilistic approach to fatigue life estimation under random loading. Based on a suitable segmentation of a sufficiently long random load process record, a set of life data is obtained which enables computation of a fatigue life distribution function (DF). Criteria for load process segmentation are determined providing the DF calculated represents the random nature of operational loading. The calculation method uses the fatigue damage accumulation hypothesis which considers the cyclic material properties, the rainflow method, and a new method for studying the mean stress effect. The DF of fatigue life obtained enables to compare different loading processes with respect to the required operational life.

Key words
fatigue life, random load process, fatigue life distribution function (DF)

Full text not available