Probabilistic Fatigue Crack Growth Analysis under Stationary Random Loading with Spike Loads
Shan Jiang,Wei Zhang,Zili Wang
DOI: https://doi.org/10.1109/access.2018.2810826
IF: 3.9
2018-01-01
IEEE Access
Abstract:The in-service loading condition of many engineering structures is generally composed of a stationary random loading caused by the mechanical vibration, and the spike loads due to occasional events, such as the sudden shock and accidental turbulence. In this paper, a physical-based method is proposed to evaluate the reliability of structure subjected to the stationary random fatigue loading superimposed by occasional spike loads. First, since the interaction effects of stationary random loading are approximately stable, the realistic random loading can be transferred to an equivalent constant amplitude loading. This equivalent transformation method can avoid the complicated cycle-by-cycle calculation. This approach is derived from the two-parameter fatigue crack growth model, in which the driving parameters of the fatigue crack growth are the stress intensity factor of peak load and the stress intensity factor range. Second, the spike loads lead to the high nonlinearity of interaction effect, which can be accounted for by the plasticity. Therefore, the generalized Willenborg model is employed to calculate the fatigue crack propagation under the spike loading effects. Then, the extensive experimental data of aluminum alloys are used to validate the proposed method, in which the indeterminacies of material parameter and spike loads are considered. It is observed that all the testing data are contained within the prediction 90% confidence interval bounds. In addition, a Monte Carlo simulation example of fatigue life reliability assessment under stationary random loading with spike loads is performed. Two scenarios of different spike load distributions are discussed. In the first scenario, the spike loads are applied at a fixed time interval, while in the other scenario, the spike loads occur with varying time period. The results indicate that the proposed approach can appropriately evaluate the fatigue reliability of the structure under stationary random loading with spike loads.