Introduction

The overarching motivation of the work reported in this paper is to improve fatigue life prediction for aerospace components. Traditionally, fatigue assessment is heavily dependent on testing and statistical analysis. With the introduction of new materials and methods of manufacture, such as composites and Additive Manufacturing, the material property and component geometry variation places significantly greater demands on the numbers of tests required to obtain satisfactory statistical confidence in the test results. In the engineering sector, there is an increasing trend to reduce testing, and thus testing costs, by relying more on validated computational modelling. In the context of fatigue assessment, this presents a problem, because fatigue understanding is largely based on experimental results and empirical models, such as the NASGRO crack growth equation. Empirical models can be used in a computational context, but still require compelling data evidence to validate their applicability.
A more desirable objective is to create a modelling capability that is based on well-established physical principles, and to link the fatigue assessment to measurable attributes of the material properties and component geometry. That is the objective for this present paper, which is entirely based on computational modelling. Two types of component geometry feature are modelled: surface roughness and porosity. The material is modelled as a simple piece-wise linear elastoplastic material, and the computational specimen geometries are subjected to cyclic loading, at load levels that would be below the yield threshold for specimen without roughness or porosity. We assess the results of these computations, and compare with the empirical and pragmatic methods of fatigue assessment. Although it is difficult to arrive at a firm conclusion at this stage, we believe that the computational models presented here enhance our understanding and could form part of a future fatigue assessment methodology. The reader might be critical of the simplifications and approximations we have made, but, as Einstein is frequently paraphrased as saying, “Everything should be kept as simple as possible, but not simpler” – a paraphrasing of Occam’s Razor. The investigation of further model enhancement and complexity will be the subject of future work.
The structure of this paper is as follows. In the present section, we provide the context and motivation for this work. In Section 2, we review existing understanding and the prior work on which the present work is predicated. In Section 3 we discuss the methodology which underpins this present work, using probing questions, based on Bacon’s [1] inductive principles. Section 4 explains the construction of the computational models used. Section 5 presents qualitative results to illustrate the interactive effects of surface roughness and subsurface porosity to create a pattern of stress and equivalent plastic strain (PEEQ) in the subsurface layer. In Section 6, we analyse this data and present results in answer to the probing questions set out in Section 3. Section 7 provides a detailed discussion of the methodology, the modelling methods used, the results obtained and weaknesses in our approach that could be addressed in future work, and proposes a potential route map for an analytical approach to fatigue life assessment for Additive Manufactured aerospace components. The conclusions to this paper are set out in Section 8.