Search

In this study, various experimental methods are employed to determine the anisotropic fatigue behavior of a 25MnCrSiVB6 forged steel (Metasco MC). This material has a bainitic microstructure and contains many elongated non-metallic inclusions in the rolled direction, which are grouped into clusters. Specimens with different orientations relative to the rolling direction have been extracted from a hot rolled bar and the ability of certain experimental techniques to capture the fatigue anisotropy has been tested. Results obtained from monotonic tensile tests and Charpy impact tests show that the material has isotropic fracture strength and anisotropic ductility. The influence of the inclusion clusters is clearly demonstrated via observation of the fracture surfaces. Concerning the fatigue behavior, results from a classical staircase experimental procedure are compared to results from self-heating fatigue tests. For specimens orientated at 0° relative to the rolled direction, microcrack initiation is controlled by the material matrix and the prediction of the fatigue strength with the self-heating method has been observed to be correct. For specimens orientated at 45° and 90°, the elongated manganese sulfide inclusion clusters are the origin of crack initiation and the fatigue strength drops significantly. For this case, it appears that the self-heating method has difficulty predicting the fatigue behavior.

This paper presents a new model that accounts, on a local scale, for the coupling between plasticity due to gliding in shear bands and damage occurring when the accumulated plastic strain has reached a threshold value. The irreversible thermodynamics with internal state variables is employed to keep a middle way between extensive description of plastic and damage flow and application of accessibility requirements. Plasticity and damage are governed by their proper complementary rules (yield functions and potentials). At the same time, a coupling occurs between the damage variable and the hardening parameters. A large experimental database relative to the fatigue behavior of a mild steel C36 submitted to different loading modes (tension, torsion, combined proportional tension and torsion) proves the efficiency of such a model. The prediction of Wöhler curves for cyclic complex stress states can be readily done, but the main feature of this approach is to ensure a clear link between mesoscopic parameters like the hardening behavior of individual grains and the subsequent local damage.

The stress gradient and the size of a component are known to influence the fatigue strength of metallic components. Indeed, in high-cycle fatigue, experiments prove that the stress distribution as well as the size of the loaded specimen can be responsible for changes in the fatigue limit (for instance, the fatigue limits in tension and bending are different, and decrease with the size of the specimen). When dealing with multiaxial load conditions, those effects still act but a relevant criterion must be used to account for the complex state of stress. The weakest-link concept together with a multiaxial endurance criterion based on a microplasticity analysis are then combined to describe the fatigue limit distribution of different metallic materials. Several load conditions are analysed: tension-compression, torsion, rotating bending and plane bending. By means of the proposed model, all the known effects on fatigue strength can be reflected. First, the endurance probability can be adequately predicted for any complex load conditions knowing some reference data from uniaxial fatigue tests. It can be linked to the probability of finding a defect with a critical size. The weakest-link theory also accounts for the decrease of multiaxial fatigue limit with the stressed volume. For the same load condition (i.e. for the same stress distribution in the volume), the probability of finding a critical defect increases with the component size and then according to the weakest-link theory the fatigue strength drops. A second model, based only on the damage developed at the surface, is also proposed. While the original Weibull theory makes no distinction between potential initiation sites at the free surface and in the volume and can lead to unsatisfactory predictions when applied to materials containing defects such as nodular cast iron, the new surface approach distinguishes between surface and volume effects.

The multiaxial fatigue loading in the high-cycle regime leads to localized mesoscopic plastic strain that occurs in some preferential directions of individual grains for most metallic materials. Crack initiation modeling is difficult in this fatigue regime because the scale where the mechanisms operate is not the engineering scale (macroscopic scale), and local plasticity and damage act simultaneously. This article describes a damage model based on the interaction between mesoplasticity and local damage for the infinite and the finite fatigue life regimes. Several salient effects are accounted for via a simple localization rule, which connects the macroscopic scale with the mesoscopic one, and by the model presented here, which describes the coupled effects of mesoplasticity and damage growth. Irreversible thermodynamics concepts with internal state variables are used to maintain a balance between extensive descriptions of plastic flow and damage events. Cyclic hardening behavior is described by a combined isotropic and kinematic hardening rule while the damage evolution is governed notably by the accumulated plastic mesoscopic strain. In this study, predictions are compared to fatigue tests performed on a mild steel (C36) under different loading modes. All the experiments are carried out under in-phase loading conditions: reversed tension, torsion, and combined tension—torsion. The mean stress effect is also studied through tests conducted under tension. The predicted Wöhler curves under any loading mode can be readily obtained with this model, but the main feature of this approach is to ensure a clear link between the mesoscopic parameters like the hardening behavior of individual grains and the subsequent local damage.

Lien vers la version éditeur: http://www.sciencedirect.com/science/article/pii/S0142112306003240; International audience; The aim of this paper is to present new modelling dedicated to multiaxial high cycle fatigue (HCF), and applied to polycrystalline metals. The model presented is based on the experimental characterization of damage during HCF tests, under pure tension and torsion modes. The origin of this approach is a mesoscopic model considering three plastic behaviour stages (hardening, saturation and softening) suggested by Papadopoulos [Papadopoulos I.V. Fatigue limit of metals under multiaxial stress conditions: the microscopic approach. Technical Note No. I.93.101, Commission of the European Communities, Joint Research Centre; 1993. [ISEI/IE 2495/93].], and used by Morel [Morel F. A critical plane approach for life prediction of high cycle fatigue under multiaxial variable loading. Int J Fatigue 2000;22:101-119.]. The principal evolution brought in by this study is a competition description during all the sample lifetime of the plasticity and damage effects. The plasticity mechanisms induce a hardening saturating effects (resulting from movement and accumulation of dislocations), especially significant at the beginning of the crystal life. Damage, present at the end of crystal lifetime, is considered as a degradation process inducing a strong reduction of the crystal ductility, leading to its failure (decohesion). The coexistence and the competition between these two effect (hardening and damage-induced softening) describe cyclic crystal behavior, including shakedown phase. The model is formulated in the framework of the continuum damage mechanics, according to the identified physical mechanisms during the tests. The second purpose is to compare the model predictions with experimental data, after identification of the parameter for a ferritic-pearlitic steel. The case of in-phase loading is merely studied here. In particular, the evolution of a few internal variables is discussed and correlated with the available physical features. It is shown that this model provides a complementary insight into a crystal with respect to the endurance criterion of Dang Van. The model predicts, in a particular stress amplitude range, the damage growth arrest.