PSI - Issue 42

Lucas Carneiro Araujo et al. / Procedia Structural Integrity 42 (2022) 163–171 Lucas Carneiro Araujo/ Structural Integrity Procedia 00 (2019) 000 – 000

164

2

1. Introduction Recently there have been failures in crankshafts of Brazilian thermoelectric generator sets, which are made of AISI 4140 (DIN 42CrMo4) or AISI 4340 (DIN 34CrNiMo6) steel. This type of material has impurities known as non metallic inclusions. These impurities contaminate the steel during its casting or forging process and can considerably reduce the fatigue strength of the components [1]. In fact, one can find many reports in the literature of fatigue failures caused by the presence of non-metallic inclusions or other types of small defects, involving gearboxes in wind turbines, railway wheels, pipelines, turbine blades and crankshafts for example [2 – 5]. Most of these components are expected to experience time-varying multiaxial stresses during their in-service lifetime. In view of this, some works have been developed addressing the fatigue problem in the presence of small defects with multiaxial loading conditions[6 – 12]. Most of these works propose new relationships to estimate the multiaxial fatigue limit in the presence of small defects. However, instead of creating new models, the authors recently proposed the use of classic multiaxial criteria combined with the √ parameter by Murakami and Endo (1983) [13 – 15] and Yanase and Endo (2014) [8] to address the problem [10 – 12]. The proposed adaptation consists of using the fatigue limits obtained from the √ parameter to calculate the material constants of the multiaxial fatigue criteria, that is, to calibrate them. The great advantages of this adaptation are that, in addition to inserting the effect of small defects in the prediction of the classic multiaxial models, the use of the √ parameter eliminates the need to carry out lengthy and expensive fatigue tests to experimentally obtain the fatigue limits of the material containing small defects. This approach has already been tested in previous works [10 – 12] and obtained good results, using different multiaxial fatigue models, when compared with experimental data of AISI 4140 steel, the same material of the of thermogenerator crankshafts. In these works, the fatigue tests were conducted with uniaxial and combined in-phase and out-of-phase loads, considering the non-metallic inclusions of the material. In this work, a summary of the results obtained so far will be presented, considering the analysis made for the material only with its natural defects, the non-metallic inclusions. And in addition, new experimental data from specimens where a micro hole was machined will be presented. The size of these micro holes is 550 mm in depth and diameter, in order to simulate possible defects much larger than the inclusions observed in the fatigue specimens and thus test the scope of the proposed adaptation, since the crankshafts are extremely larger than the specimens, with the possibility of having much greater natural defects. 2. Multiaxial fatigue models To evaluate the proposed methodology, two critical plane models will be used, the Findley model and the Susmel and Lazzarin model, which are well-established models in the literature and which differ in terms of the definition of a critical plane. The Findley model is defined from the linear relationship between the shear stress amplitude, , and the maximum normal stress, , . For him, the critical plane is the one where the greatest linear combination between these two parameters occurs [16], according to Eq. (1). ( , ∅ ) : ,∅ { ( , ∅) + , ( , ∅)} (1) Where and ∅ are the spherical coordinates of the critical plane and is a factor that accounts for the material's sensitivity to normal stresses. Thus, it is considered that fatigue failure will occur if Eq. (2) is not met, where is the multiaxial fatigue limit. ,∅ ( + , ) ≤ (2)