PSI - Issue 42

Robert Basan et al. / Procedia Structural Integrity 42 (2022) 655–662 R. Basan et al. / Structural Integrity Procedia 00 (2019) 000–000

656

2

Nomenclature b

fatigue strength exponent fatigue ductility exponent

c

Young’s modulus error criterion goodness of fit Brinell hardness

E

E f E a

HB 2N f

number of load reversals to failure 2N f,exp experimentally obtained number of load reversals to failure 2N f,est estimated number of load reversals to failure r coefficient of correlation R m ultimate strength Δ ε /2 total strain amplitude

Δ ε e /2 elastic strain amplitude Δ ε p /2 plastic strain amplitude ε f '

fatigue ductility coefficient fatigue strength coefficient

σ f '

H abbreviation for Hardness method UML abbreviation for Universal material law M abbreviation for Median method

The act of conducting tests is also fairly difficult because it entails avoiding blunders and errors as well as coping with expensive equipment availability. Testing numerous candidate materials is highly difficult, if not impossible, due to experiments' complexity and length (especially high-cycle fatigue tests). In an attempt to develop acceptably accurate alternative to complex, expensive and long-lasting experimental determination of fatigue data, numerous methods for estimation of strain-life fatigue parameters from materials’ monotonic properties have been proposed in the literature so far. First ones, namely, Four-point-correlation method and Universal slopes method were proposed by Manson (1965) soon after strain-based approach to material fatigue characterization was developed and proposed by Coffin and Manson. Afterwards, numerous others followed, such as Mitchell‘s method proposed by Socie, Mitchell and Caulfield (1979), Modified universal slopes by Muralidharan and Manson (1988), Uniform material law by Baumel and Seeger (1990), Modified four-point-correlation method by Ong (1993), Hardness method by Roessle and Fatemi (2000), Median method by Meggiolaro and Castro (2004), Uniform material law + by Hatscher, Seeger and Zenner (2007) and Uniform material law improvement by Wächter and Esderts (2018) to name some of them. With aim to further improve accuracy of estimations, with recent efforts shifting notably from empirical/analytical approaches and models to machine learning-based predictive models such as the ones based on artificial neural networks as given in Troschenko et al. (2011), Tomasella et al. (2011), Marohnić and Basan (2018), Linka et al. (2021). Significant advantages of such models, when they are developed properly, are the possibility to make use of more input variables (monotonic properties), and the ability to model much more complex relationships among monotonic and cyclic/fatigue parameters than is possible using conventional approaches. However, models with fewer input parameters are still frequently more beneficial for actual applications since some monotonic properties and related material data on which estimations are based, are not always available or simple to get. Another persuasive indicator that fatigue parameters estimation methods continue to be relevant and of interest are also recent developments and research on applicability of existing estimation methods and development of new ones for 3D printed metallic materials, Derrick and Fatemi (2022) as well as those for very high- and ultra high strength steels, Yadegari et al. (2022).