Predicting Fatigue Crack Behavior in Titanium with the use of Statistical Tools

The study of fatigue crack growth (FCG) in structural materials is aimed at residual life estimations in order to apply the fail-safe design criterion. Because of their excellent strength-to-weight ratio, titanium and its alloys are the most suitable metallic materials for use in the aircraft industry. Recently a new research topic is being developed, in which FCG behavior is predicted by means of statistical tools. In the present work, two of the most promising methods are tested in the description of FCG in unalloyed titanium sheet samples: 1) Artificial Neural Networks (ANNs) and 2) Stochastic Differential Equations (SDEs). These techniques are employed together with a set of 30 experimental curves obtained from constant amplitude FCG tests of center-cracked titanium specimens, conducted in laboratory air under room temperature. Additional numerical data obtained from a predictive damage accumulation model code, were also employed in the work with ANNs. Some possibilities of these tools in determining the residual life and the crack growth history, without disregarding the statistical nature of these phenomena, are explored in the development of the work.