In this study, crack initiation and propagation under cyclic loading are first experimentally studied by testing specimens fabricated from a rotor material containing manufacturing defects. The latter represent clusters of non-metallic inclusion which size and location are examined by both ultrasonic testing (UT) and fractographic analyses. Three test specimens were extracted from a material block in such a way that the UT indications were located in the middle part of the cross-section. The specimens were then subjected to cyclic loading, applying tensile stresses with different magnitude and stress ratio. This test procedure produced beach marks on specimen fracture surfaces from which the crack initiation and propagation were backtraced. The experimental results suggest that a considerable number of cycles are required for a crack with a size corresponding to the UT indication to be formed. A numerical approach was then adopted for modeling damage accumulation and crack formation starting from a defect cluster. A material model was first calibrated to describe both the cyclic hardening behavior of the defect-free material and strain controlled low cycle fatigue tests. Subsequently, the model was applied to predict the damage evolution and crack formation at a defect group representative of that in one of the specimens studied.

Press-fitted railway axles and wheels are subjected to fretting fatigue loading with a potential hazard of crack initiation in press fits. Typically, the resistance against crack initiation and propagation in press fits is investigated in full-scale tests, which procedure is both costly and time consuming. In this context, combined experimental and numerical approaches are of increasing practical importance, as these may reduce the experimental effort and, moreover, provide a basis for the transferability of experimental results to different axle geometries and materials. This study aims at evaluating stress–strain conditions under which fretting fatigue crack initiation is likely to occur. Experiments on small-scale specimens under varying fretting fatigue load parameters and their finite-element modelling to characterize the resulting stress–strain fields are performed. Subsequently, different multiaxial fatigue parameters are applied to predict crack initiation under fretting fatigue conditions.

Fatigue behaviour of the austenitic steel 1.4550 (X6CrNiNb18-10) under low-cycle fatigue and high-cycle thermal fatigue was investigated with in two research projects supported by the Federal Ministry of Economic Affairs and Energy and the Ministry of Education and Research. The objectives of the projects were the gain of deep understanding of the damage mechanisms under mechanical and thermal cyclic loading and the development of material models and simulation procedures for an improved lifetime assessment. In comparison to the advanced mechanism based material models engineering computational procedures were proven with respect to their applicability and conservatisms. For thermal cyclic loading, test equipment and technique were developed which allow for cyclic thermal loading with temperature ranges between 1 00 C and 300 C and frequencies between 0.1 and 1 Hz. As a result, tests with a temperature range of 150 C and lower showed no crack formation up to 300,000 cycles. For temperature ranges of 200 C and higher multiple crack patterns were observed with the deepest crack of about 1.3 mm after 1,000,000 cycles, whereas the difference in crack depth between 300,000 and 1,000,000 cycles was negligibly small. To model the fatigue lifetime, the D{sub TMF} damage parameter was applied to the low-cycle fatigue and the thermal, high frequent fatigue tests. For thermal fatigue, the analyses predicted in agreement with the tests crack initiation followed by crack propagation, subsequent retardation and arrest. This behaviour can be explained qualitatively and quantitatively using the methods of linear-elastic fracture mechanics, whereas the consideration of the interaction of multiple cracks is essential to describe the experimentally observed crack retardation. The results for thermal fatigue are in the scatterband of the mechanical p and thermo-mechanical fatigue results and the cycles to failure are 10 times higher than those estimated according to the KTA fatigue curve. As a rule, the assessment using simplified engineering methods based on elastic stresses is very conservative. Using the RCC-M concept and a modification developed from AREVA, realistic and conservative results were obtained for all cases. An improved assessment is mainly achieved by the separate treatment of mechanical and thermal stresses.

This paper focuses on the evaluation of predictive methods for the analysis of fatigue crack growth (FCG) under mixed mode I/II loading. The experimental part of the study consists of a series of fatigue crack growth tests performed on bend and tension specimens with geometries similar to the standard SE(B) and SE(T) ones. For the above specimen types, both mode I and mixed mode FCG tests are first performed. To facilitate an accurate evaluation of the mixed mode test results, finite element analyses of stress intensity factors for crack geometries following the experimentally measured trajectories are carried out. Additionally, the XFEM based algorithm available in the finite element code ABAQUS is explored with respect to its performance in predicting crack growth paths. The subsequent test evaluation focuses on examining a correlation between FCG rates for mixed mode loading conditions with the mode I baseline curve. The results suggest that, using mode I experimental data along with a mode I specimen analysis, both conservative and non-conservative prediction of mixed mode fatigue crack growth is possible. In this context recommendations of failure assessment procedures regarding the flaw re-characterisation and projection onto principal stress planes, as well rules for the transferability of mode I FCG curves to mixed mode conditions are discussed.

The paper presents results of numerical modelling of elastic-plastic stressand strain fields at the tip of a propagating crack under cyclic loading. A particularmotivation is to investigate the difference in fatigue crack growth rates previouslyobserved in tests on M(T) and C(T) specimens made of 25CrMo4 (EA4T) steel. Thestress field triaxiality (constraint) is considered as a factor influencing the deformationand, accordingly, closure behaviour at the crack tip. Among numerical issues studied inthe paper are the strain hardening behaviour, consideration of the crack face contact,definition of the onset of crack opening, possible simplifications of numerical modellingby using the boundary layer formulation. The numerical results suggest that, using theeffective stress intensity factor range, a reasonable explanation to the experimentalfindings can be provided.

This paper focuses on the evaluation of methods for estimating fatigue crack growth (FCG) under mixed mode I/II loading. The experimental part of the study consists of a series of fatigue crack growth tests on bend and tension specimens with geometries similar to the standard SE(B) and SE(T) ones. For the above specimen types, both Mode I and mixed mode FCG tests are first performed. To facilitate an accurate evaluation of the test results, finite element analyses of stress intensity factors for crack geometries following the experimentally measured crack trajectories are carried out. Additionally, the XFEM based algorithm available in the finite element code ABAQUS is explored with respect to its performance in estimating crack growth paths.The subsequent test evaluation focuses on examining a correlation between FCG rates for mixed mode loading conditions with the Mode I baseline curve. The results suggest that, using Mode I experimental data, both conservative and non-conservative estimation of mixed mode fatigue crack growth is possible. In this context, recommendations of failure assessment procedures regarding the flaw re-characterisation and projection onto principal stress planes, as well as rules for the transferability of Mode I FCG curves to mixed mode conditions, are discussed.

This paper presents some experimental results demonstrating the geometry dependence of fatigue crack growth (FCG) curves for the steel EA4T (25CrMo4). The experimental results exhibit considerable differences in FCG rates measured on M(T) and C(T) standard specimens, as well as specimens with surface cracks. To explore the possibility of an analytical description of these effects, a numerical analysis is applied to simulate crack growth behaviour for the M(T) and C(T) geometries with special emphasis put on modelling plasticity induced crack closure. The analysis results provide an adequate qualitative description of the crack propagation in different specimens and suggest an explanation for the difference in respective FCG data. On the other hand, a re-evaluation of the experimental results is undertaken to explore a correlation between crack growth rates and the amount of crack tip yielding, with the latter being quantified in terms of the plasticity parameter Lr of the failure assessment diagram. Such an analysis demonstrates that, even though small scale yielding conditions prevail at the crack tip, the FCG curves for individual specimens depend upon and can be arranged according to the plasticity level. Thus, the results suggest that engineering calculations of fatigue crack propagation can be facilitated by incorporating the Lr factor as an additional influencing parameter to take into account the geometry and load effects on FCG rates.