Fatigue life prediction of corrosion-damaged high-strength steel using an equivalent stress riser (ESR) model. Part II: Model development and results

The fatigue life of metallic aircraft structural components can be significantly reduced by environmentally induced corrosion. However, there have historically been no analytical methods to quantify the specific fatigue life reduction of individual unfailed corroded components with any reasonable degree of confidence. As part of a NAVAIR high-strength steel corrosion–fatigue assessment program, methods were studied to predict the impact that corrosion-induced surface roughness has on the fatigue life of high-strength steel aircraft components. The steels of interest produce general corrosion in patches as well as localized material loss similar to pitting. In addition, this type of corrosion has characteristic features over a wide range of scales. Consequently, traditional finite element analysis approaches are not well suited to this problem, since the mesh required to accurately reflect the fine details distributed over the entire corrosion patch make computation unrealistic. Therefore, approximate methods were developed that allow localized regions of interest of high stress to be identified. Subsequently, a simple notch metric formula is employed to approximate the stress riser in these regions of interest. Finally, an extension of Peterson’s fatigue notch sensitivity theory is applied to these small “notches” that has the result of suppressing the effect of smaller notches compared to larger notches in the prediction of life. Each region of interest is assigned a probability of crack initiation as a function of fatigue cycles, based on a probabilistic strain–life analysis using the predicted notch factor. The net life (to crack initiation) for the component is then the product of the survivabilities of all of the individual regions of interest on the component surface. Tests on corroded fatigue specimens have been conducted to both calibrate the parameters in the Peterson model as well as to test the life prediction capability of the approach. Predictions from the resulting model have demonstrated that an empirical approach to corrosion surface damage can be utilized to generate probabilistic life predictions that have substantial engineering value in assessing the residual fatigue life of corroded AF1410 steel components, and that the modeling technique can capture the significant corrosion features that cause fatigue cracking in most cases, especially for more severely corroded surfaces.