On the Search for Optimal Damage Precursors

Abstract

A new approach to predict the service life of critical components via study of damage precursors is emerging and is the topic of this paper. To date most service life predictions are based on measurements of damage indicators and their growth towards criticality or failure, e.g. fatigue crack length, material loss due to corrosion or wear, etc. For many cases of concern, especially fatigue, the critical, direct damage indicator, Di, i.e. crack length, increases exponentially towards criticality, roughly dDi/dN α Di n where n can exceed 4. This makes lifetime estimates based on measurements of damage, e.g. around half-life, or even at 80 percent life, difficult and inaccurate, even for cases where the damage progress is well modeled and understood. To improve the accuracy and reliability of lifetime prediction, efforts are now underway to determine the state awareness of a critical component during service, based on property characterizations, in addition to the measurements of the direct damage indicators, such as crack length, acoustic emission, ultrasound signals, eddy current measurements, etc. These characterizations will include indirect damage indicators, i.e. precursors, and allied or affiliated damage indicators. Precursors, e.g. to fatigue crack formation, include electrical resistivity changes, dislocation distribution changes and crazing. For affiliated damage indicators residual stress relaxation or development, phase changes, electrical property (resistivity, dielectric constant, permeability), microstructural characterization, must be considered. The selection of the optimal combination of direct and indirect damage indicators will be application specific. For example for ultra-high strength steels and aluminum alloys, with high damage exponents n, where the critical crack length may also be very small and difficult to detect, measurements of reliable precursors and indirect damage indicators could significantly improve service life prediction. It is proposed to assess the efficacy of damage indicators on the basis of their Di/Df vs. Ni/Nf, i.e. damage ratio vs. life fraction curves, (referred to as damage indicator ratio curves, or DIR curves) searching for indicators with DIR curves that best meet the needs of the application.