Evaluation of cumulative damage models for fatigue crack growth in an aircraft engine alloy

Abstract

Two models for evaluating crack growth in aircraft engine alloys under typical mission spectra were evaluated. Each model had the capability to determine the effects of frequency, stress ratio, temperature, and hold time on the crack growth rate. Data on an advanced alloy (AF115) were used to evaluate the hyperbolic sine (SINH) model and modified sigmoidal equation (MSE) model. Both models were found to have adequate capability and flexibility in modeling crack growth behavior over a wide range of conditions. The SINH model has been much more fully developed than the MSE model and is easier to apply to new materials. HE ability to predict fatigue crack growth rates in struc- tural alloys has become an integral part of the design and life management procedures for U. S. Air Force gas turbine engines. Increased performance requirements resulting in higher operating temperature and stresses over the last two decades have resulted in drastic changes in the design criteria applied to critical engine components. In the 1960s, most critical structural components (e.g., turbine disks, spacers, etc.) were limited by creep and stress rupture properties. Less than 1% of all rotating components were life limited by low- cycle fatigue. However, today's designs find well over 15% of these components limited in life by low-cycle fatigue. The present life management system for critical structural components whose life is limited by low-cycle fatigue is to retire these components from service at the end of their design life. The design life is determined by the time required to in- itiate a crack, therefore completely ignoring subcritical crack growth. Moreover, the design life is based on statistically safe material properties, using criteria such as 1 in 1000 com- ponents initiating a detectable crack. Statistically, 999 out of 1000 components will not have initiated a detectable crack when they are retired from service. Although this is clearly a conservative and safe design procedure, the replacement costs for these components are becoming prohibitive. Furthermore, analysis has shown that the average useful life of the com- ponents is often greater than 10 times their design life. Finally, the accurate determination of the tail end of the statistical distribution curve of component lives is very difficult, especially in cases where low-cycle fatigue life is heavily in- fluenced by the presence of defects in the material. For these reasons, the Air Force is implementing a retirement-for-cause (RFC) life management philosophy on some of its existing engines.