Multiaxial fatigue strength of common structural steel and the response of some estimation methods. (November 2017)

• Record Type: Journal Article
• Authors: Papuga, Jan; Fojtík, František
• Is Part Of: International journal of fatigue. Volume 104(2017)
• Page Start: 27
• Page End: 42
• Publication Date: 2017-11
• Subjects: Multiaxial fatigue limit -- Fatigue strength -- Multiaxial loading -- Manson-McKnight method; Materials -- Fatigue -- Periodicals; Materials -- Fatigue; Periodicals; 620.1122
• DOI: 10.1016/j.ijfatigue.2017.07.001
• Languages: English
• ISSNs: 0142-1123

Abstract

Highlights: A set of fatigue strengths close to the fatigue limit condition is provided for CSN 411523 steel. The new simply calculated method MMP is derived from the Manson-McKnight method. Its results are compared with 4 multiaxial methods and 3 variants of the Manson-McKnight method. MMP provides results comparable with the multiaxial models. Simplicity of MMP makes it a perfect candidate for quick estimates. Abstract: This paper describes the multiaxial fatigue strengths obtained from a series of load-controlled fatigue experiments in the high-cycle fatigue domain. The experiments were carried out on unnotched specimens manufactured from common structural steel CSN 41 1523. Various load combinations were induced, including some rare combinations, e.g. a combination of pressurizing and plane bending. The stress states on the inner surface and also on the outer surface of hollow specimens are therefore evaluated in the subsequent fatigue strength analyses. The goal of the paper is to derive a new solution for estimating multiaxial fatigue strength that will better conform to the described experiments than the Manson-McKnight method, but which will keep its simplicity of calculation. Our final proposal – the MMP criterion – is compared with various versions of the Manson-McKnight method, and also with four multiaxial methods (Dang Van, Liu-Zenner, Crossland and Papuga PCr). The results for the MMP criterion are quite motivating, thanks to the simplicity of the computational algorithm and the quality of the prediction output, which is comparable with the output of the evaluated multiaxial methods. However, the set of experiments described here is not large enough either for more thorough statistical analyses of prediction quality, or for more enhanced improvement of the MMP criterion, which would include e.g. the effect of non-proportional loading.