An assessment of a strain‐life approach for fatigue crack growth

Abstract

PurposeFatigue crack growth models based on elastic‐plastic stress‐strain histories at the crack tip region and strain‐life damage models have been proposed. The UniGrow model fits this particular class of fatigue crack propagation models. The residual stresses developed at the crack tip play a central role in these models, since they are applied to assess the actual crack driving force. This paper aims to assess the performance of the UniGrow model based on available experimental constant amplitude crack propagation data, derived for several metallic materials from representative Portuguese bridges. It also aims to discuss key issues in fatigue crack growth prediction, using the UniGrow model, in particular the residual stress computation and the suitability of fatigue damage rules.Design/methodology/approachThe UniGrow model is assessed using data derived by the authors for materials from Portuguese riveted metallic bridges. Strain‐life data, from fatigue tests on smooth specimens, are used to propose a convenient fatigue damage model. Predicted crack growth rates are compared with experimental crack propagation data obtained by authors using fatigue tests on compact tension specimens. Since the UniGrow model is a residual stress‐based propagation model, elastoplastic finite element analysis is proposed for comparison with the analytical approach implemented in the original UniGrow model.FindingsThe use of the Smith‐Watson‐Topper damage parameter overestimates the stress R‐ratio effects on crack propagation rates, mainly if the material shows crack propagation rates with small to moderate sensitivity to stress R‐ratio, which is the case of the materials under investigation in this paper. Alternatively, the application of the Coffin‐Manson damage law leads to consistent fatigue crack growth predictions for the investigated range of positive stress R‐ratios. The stress R‐ratios effects may be solely attributed to the residual stresses. Their estimation, using an analytical approach, may lead to inconsistent results, which is demonstrated by an alternative elastoplastic finite element analysis.Originality/valueContributions for more accurate predictions of fatigue crack propagation rates, for several stress ratios, using a strain‐based approach is proposed. This approach is valuable since it may be used to reduce the time consuming and costly fatigue crack propagation tests. Furthermore, the proposed approach shows potential for an unified crack initiation and propagation approach.