An Ultrafast Crack Growth Lifing Model to Support Digital Twin, Virtual Testing, and Probabilistic Damage Tolerance Applications
Juan Ocampo,Harry Millwater,Nathan Crosby,Beth Gamble,Christopher Hurst,Michael Reyer,Sohrob Mottaghi,Marv Nuss
DOI: https://doi.org/10.1007/978-3-030-21503-3_12
2019-07-03
Abstract:New aeronautical technologies like the Airframe Digital Twin, Virtual Fatigue Testing, and Probabilistic Damage Tolerance Analysis require a very large number of crack growth evaluations with a comprehensive number of random variables in order to accurately predict the fatigue life, the structural risk, or the remaining useful life of a structure. Current state-of-the-art crack growth methodologies and probabilistic methods do not make these new technologies possible due to limitations on computational speed, number of random variables, and statistical tools. In this work, a new computational strategy is developed and demonstrated such that several random variables directly affecting the crack growth analysis can be considered. This approach provides the opportunity for a more comprehensive and accurate digital twin evaluation, virtual testing prediction, and risk assessment, hence, improving aircraft design, safety, and reliability.Under Federal Aviation Administration (FAA) Funding, This methodology focused on the development of an ultrafast numerical crack growth algorithm that consists of: (a) a constant amplitude equivalent stress derived from a variable amplitude loading spectrum, and (b) an adaptive step-size Runge-Kutta ordinary differential equation (ODE) solver.Several examples with the Airframe Digital Twin, Virtual Fatigue Testing, and Probabilistic Damage Tolerance applications will be demonstrated using through, corner, and surface cracks at a hole under representative loading spectra. The crack size versus cycles results from this new approach will be compared against results obtained from commercial lifing software codes. All results to date indicate the comparison is within a few percent. The probabilistic crack growth analysis has been parallelized using OpenMP in order to fully utilize multi-core computers. This approach provides a more comprehensive and accurate risk assessment, hence, improving aircraft safety and reliability.