Prognosis of Long-Term Load-Bearing Capability in Aerospace Structures: Quantification of Microstructurally Short Crack Growth
Final rept. 1 Aug 2010-31 Jul 2013
CARNEGIE-MELLON UNIV PITTSBURGH PA DEPT OF MATERIALS SCIENCE AND ENGINEERING
Pagination or Media Count:
Carnegie Mellon and Cornell Universities collaborated to investigate the behavior of microstructurally small cracks under cyclic loading with the aim of improving our understanding of the substantial fraction of fatigue life occupied by the crack initiation phase. Heavy use was made of advanced tools such as High Energy Diffraction Microscopy HEDM for 3D mapping of microstructures in the materials of interest, namely Ni-based superalloys such as LSHR and Ren 88DT. We worked closely with partners at AFRL and GE Global Research to obtain suitable specimens for characterization and simulation. LSHR samples with multiple microcracks provided valuable information that established that fatigue cracks start on or close to coherent twin boundaries annealing twins in large grains where the twin-parallel slip systems are well aligned with respect to the loading direction. This conformation was then tested via finite element simulation to determine which criterion could be most reliable for predicting the development of microcracks. By employing a constitutive relation that accounts for the effect of plastic deformation gradient, which arise from accommodating the evolution of slip close to twin boundaries. The results show that simulation can effectively reproduce the concentration of slip parallel to twin boundaries that is characteristic of crack initiation. Some work was also performed to determine the appropriate size of a simulation volume when a particular location in a sample is known to be of interest. In the example of the HEDM image of LSHR, the size of the volume around the location of a microcrack was investigated with the result that for convergence of the elastic stress and strain fields, more than 250 grains around the microcrack needed to be included in the simulation volume.