A comparative analysis of multiaxial fatigue models under random loading. (1st March 2019)

• Record Type: Journal Article
• Authors: Portugal, Ibai; Olave, Mireia; Urresti, Iker; Zurutuza, Aitor; López, Arkaitz; Muñiz-Calvente, Miguel; Fernández-Canteli, Alfonso
• Is Part Of: Engineering structures. Volume 182(2019)
• Page Start: 112
• Page End: 122
• Publication Date: 2019-03-01
• Subjects: Component design -- Critical plane -- Multiaxial fatigue -- Non-proportional loading; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2018.12.035
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: The different multiaxial models applied are demonstrated as reasonable candidates. McD is the most conservative. Findley, FS and BM criteria are the least in order. Orthogonal shear proves to underestimate the damage compared to the other models. The influence of the initial ball position proves to exert big influence on damage. The advances in bearings design could be extended to other industrial components. Abstract: A number of structural components, such as those related to wind turbines, are permanently subject to randomly varying load conditions during their service life and are therefore exposed to fatigue failure hazards. For this reason, the use of models for lifetime prediction is indispensable to ensure the structural integrity of at component and the correct operation conditions during its service life. This article compares different multiaxial fatigue criteria, in order to promote a secure and optimal design. Orthogonal shear stress and critical plane models based on stresses (e.g., McDiarmid and Findley), strains (e.g., Brown-Miller) and energy (e.g., Fatemi-Socie and Smith-Watson-Topper) are therefore considered. With this goal, the effect of randomly distributed load histories is analysed using a novel methodology based on the calculation of the stress tensor as a function of time, by interpolating loads with those obtained from response surfaces using Finite Element Method (FEM) models. The critical values of the selected parameters involved in the failure criterion, once estimated, are considered as references and used to determine the fatigue damage based on the Wöhler curves of the material. The methodology proposed allows the most suitable multiaxial fatigue criterion to be recognised by comparing predicted and experimental lab fatigue lives and levels of safety reserve. In this way, a methodology is provided to advance in the optimal test fitting and lifetime prediction of components under real fatigue conditions.