There is an economic need to extend the useful life of jet engine rotors. Retirement-for-Cause (RFC) is a lifing method that allows for the continued operation of components passed traditional life limits. Under RFC, an extension of damage tolerance, components are deemed safe for a further service interval using non-destructive inspections (NDI) for crack like defe cts. As components are kept in-service beyond their designed service life it is essential that the probability of failure (POF), or risk, of continued service is known. Under current FAA rotor design certification practices the component POF is analyzed using a probabilistic framework focused on only the life limiting crack location. This method generates conservative approximations of the operational risk, also known as a relative risk. The proposed method for a feature-based discretization allows for a transition from the relative risk towards an absolute risk. The general guidelines, for the discretization, have been established through performance of probabilistic assessments of a representative turbine disk. The discretization is performed by initially separating the representative turbine disk into various features. These features are then discretized through the introduction of defect locations in response to the stress gradient topology. Once the discretization of the disk is completed, a fracture mechanics-based probabilistic assessment is performed utilizing DARWIN®. DARWIN® is a fracture mechanics based probabilistic assessment software package developed by Southwest Research Institute, SwRI®. The POF of the features are obtained through the statistical combination of the defect location POFs. The representative turbine disk POF is likewise obtained by the statistical combination of the feature POFs. The probabilistic assessment results for the two methods, the life limiting and discretization, are compared for the representative turbine disk.
Michael Thomas, Jace Carter, Lloyd Matson, Tarun Goswami. Feature-Based Discretizationofa Turbine Disk for Probabilistic Risk Assessment. Mechanics, Materials Science & Engineering Journal, 2018, 14, 10.2412/mmse.21.86.70.hal-01965611