STOCHASTIC MUTISCALE/MULTISTAGE MODELING OF ENGINE DISKS

Turbine disks are amongst the most critical components in aero- and naval-vessel engines. They operate in a high pressure and temperature environment requiring demanding properties. Nickel-based supperalloys which have high creep and oxidation resistance at high temperatures are widely used as the material of turbine disks. The elevated-temperature strength of this supperalloy and its resistance to creep deformation significantly depend on the volume fraction, size and antiphase boundary energy of the ?’ phase as well as on the grain size and texture. Future propulsion systems will require turbine disks with an increased material temperature capability and with optimized dual microstructures presenting high creep resistance and dwell crack growth resistance in the rim region and high strength and fatigue resistance in the bore and web regions. In the proposed STTR project, a state of the art, multi-fidelity, and efficient multiscale and multistage process modeling and simulation methodology will be developed together with a computer software package for advanced dual microstructure nickel-base supperalloy turbine disks. The proposed methodology is based on an integration of realistic microstructure evolution modeling, dislocation dynamics, crystal plasticity theory, finite element deformation and thermal processing simulation, and probabilistic, statistical and statistic learning methodologies. The proposed developments significantly advance the science of multiscale modeling by connecting the microstructure uncertainties to macroscale processing control, and further, to the resulting variability of material properties. Innovative techniques in data-driven representation of microstructure uncertainties will be employed together with adaptive sparse grid collocation based techniques for modeling uncertainty propagation in multiscale materials simulations. Moreover, a validated model that optimizes the processing technology to produce complex gas turbine engine components with controlled microstructures, defect populations and desirable mechanical properties will be developed, thus providing reliable guidance for industrial manufacture.