Probabilistic Risk Assessment for Life Extension of Turbine Engine Rotors

Abstract

Maintaining the component service life beyond its historical limits requires the ability to accurately quantify component reliability and address the uncertainties in material responses. A probabilistic method for predicting the total fatigue life was developed and applied to determine the probability of failure of a Ti-6Al-4V turbine disk component. The total fatigue life incorporates a dual mechanism approach including the crack initiation life and propagation life while simultaneously determining the associated initial flaw sizes. A microstructure-based model was employed to address the uncertainties in material response and relate the crack initiation life with crack size. The propagation life was characterized using both small and large crack growth models to ensure accurate fatigue life prediction. Fatigue life predictions were found to correlate with experimental data at high stress levels. The risk assessment can be used to estimate the expected initial crack sizes from variability in material properties, which can further be used to establish an enhanced inspection planning.