Fretting fatigue under variable loading below fretting fatigue limit

The fatigue limit diagram provides the critical condition of non‐failure against fatigue under constant amplitude loading. The fatigue limit diagram is usually considered to give the allowable stress if every stress component is kept within the fatigue limit diagram. In the case of variable amplitude fretting fatigue, however, this study showed that fatigue failure could occur even when all stresses were within the fatigue limit diagram. An example of such a condition is a repeated two‐step loading such as when the first step stress is R=?1 and the second step stress has a high mean value. The reason why such a phenomenon occurs was investigated. A non‐propagating crack was formed by the first step stress even when well below the fatigue limit. The resultant non‐propagating crack functioned as a pre‐crack for the second step stress with a high mean value. Consequently, fatigue failure occurred even when every stress was within the fatigue limit diagram of constant amplitude loading. The fatigue limit diagram obtained in constant amplitude fatigue test does not necessarily guarantee safety in the case of variable amplitude loading in fretting fatigue.     

Deterministic Processing of Complex Multiaxial Fatigue Load Data on a Tubular Specimen with Hole

A study of the use of deterministic fatigue life prediction methods for a set of multiaxial experimental data on broad‐band random loading applied to a tubular specimen with a hole is presented. There is a discussion of the applicability of stress‐based predictions and strain‐based predictions based on the linear cumulative rule under current conditions. The only feasible criteria finally chosen for the application are strain‐based methods for analysing the fatigue life until fatigue crack initiation. Several methods differing in the damage parameter and application of the mean stress effect (MSE) are evaluated. The MSE is optimised so that a comparison of the predicted and experimental data have the least scatter and similar mean values. The results are presented, and the most promising method (by Erdogan and Roberts) has been selected. The positive properties of Walker mean stress correction and the failure of SWT and Landgraf methods are commented on.     

Fatigue crack growth in a structural steel under single and multiple periodic overload cycles

Preliminary results of a research program on fatigue crack growth in a low-carbon steel under a variable amplitude loading are presented. First, test results are reported on crack growth under simple loading sequences containing single and multiple tensile overloads applied periodically between smaller, constant amplitude cycles. Next, the observed crack growth behaviour is compared to predictions from a theoretical model developed by the authors.     

Fatigue life estimation under random loading using the energy criterion

A procedure for estimating the useful life of a component for a given (admissable) probability of fatigue fracture origination under random loading is presented. The method uses material constants obtained from the S/N and cyclic stress/strain curves, standard deviation and probability density distribution of the loading process and a macroblock of harmonic cycles obtained by applying the rainflow cycle counting method to the random loading process. Theoretical and experimental lives are found to exhibit good agreement.     

Fatigue strength of small-notched specimens under variable amplitude loading within the fatigue limit diagram

Although the fatigue limit diagram is defined in principle for constant stress amplitude, it is often considered that fatigue failure would not occur, even in varying loading, if applied stresses were kept within the fatigue limit diagram. However, it was shown in the case of small‐notched specimens that fatigue failure occurred in some special cases of variable amplitude loading, even when all stress amplitudes were kept within the fatigue limit diagram. The cause of this phenomenon was examined using two‐step stress and repeated two‐step stress patterns in which the first step stress was chosen to be equal to the fatigue limit with zero mean stress and a mean stress was superposed on the second step stress. A non‐propagating crack was formed by the first step stress. This crack functioned as a pre‐crack for the second step stress with high mean stress. Consequently, fatigue failure occurred even when all stress amplitudes were kept within the fatigue limit diagram. It was an unexpected fracture caused by the interference effect of a non‐propagating crack and a mean stress change.     

Cumulative fatigue damage and life prediction of elastomeric components

Elastomeric components are widely used in many applications due to their good damping and energy absorption characteristics. The type of loading normally encountered by these components in service is variable amplitude cyclic loading. Therefore, fatigue failure is a major consideration in their design. In this work capabilities of Rainflow cycle counting procedure, maximum principal strain as a damage criterion, and Miner's linear cumulative damage rule are evaluated with both specimen and component tests. An automotive cradle mount is used as an illustrative component. Comparison of predicted and experimental fatigue lives in both specimen and cradle mount variable amplitude load tests indicate satisfactory predictions in both cases.     

A unified statistical model for S‐N fatigue curves: probabilistic definition

In recent years, experimental tests exploring the gigacycle fatigue properties of materials suggest the introduction of modifications in well‐known statistical fatigue life models. Usual fatigue life models, characterized by a single failure mechanism and by the presence of the fatigue limit, have been integrated by models that can take into account the occurrence of two failure mechanisms and do not consider the presence of the fatigue limit. The general case, in which more than two failure mechanisms coexist with the fatigue limit, has not been proposed yet. The paper presents a unified statistical model which can take into account any number of failure mechanisms and the possible presence of the fatigue limit. The case of S‐N curves with different fatigue life distributions coexisting for the entire stress range covered by fatigue tests is also considered. The adaptability of the statistical model to the S‐N curves proposed in the open literature is demonstrated by qualitative numerical examples.     

Comparison of different methods for presenting variable amplitude loading fatigue results

This paper discusses eight methods for presenting fatigue test results for variable amplitude loading and their comparison with constant amplitude loading. While the maximum amplitude method compares constant and variable amplitude loading results by the Woehler and Gassner curves, all other seven methods try to transform the variable amplitude results into the Woehler curve by applying different equations. The advantage of the maximum amplitude method is the direct comparison of the maximum amplitude of the spectrum with the yield strength and with the high‐cycle fatigue strength, which is an important step in structural design. Among the other methods, the best results were obtained by following: most damaging, half damage and mean damage amplitudes. However, the presentation of constant and variable amplitude results by these methods in one scatter band is possible only when the real damage sum is close to D = 1.0.