Engineering Fracture Mechunics Vol. 23. No. 2, pp. 345-358. 1986 Printed in Great Britain. 0013-7944186 $3.00 + .oO Per&tamonPrw Ltd. zyxwvuts FRACTURE MECHANICS APPLIED TO NONISOTHERMAL FATIGUE CRACK GROWTH E. H. JORDAN? University of Connecticut, Storrs, CT 06268, U.S.A. G. J. MEYER% McGraw Edison Company, Worthington Compressors Division, Buffalo, NY 14240, U.S.A. Abstract-Twelve nonisothermal fatigue crack growth tests were performed on Hastelloy- X tubular specimens in which strain and temperature varied simultaneously. Conditions were selected to include nominally elastic and nominally piastic conditions and temperatures up to 982°C. A number of parameters, including the stress intensity factor. strain intensity factor, and J-integral, were examined for their ability to correlate the data. There was no decisive difference between the success of the three parameters. Each parameter correlated data from different strain ranges to within no worse than a factor of 2.1 on daidn. The effect of strain temperature cycle shape was investigated and found to be moderate, while a strain hoid of 1 min had very little effect. An attempt was made to predict nonisothermal test results from isothermal data. These predictions were better than those made by using peak test temperature isothermat data but still not within scatter. INTRODUCTION MANY engineering components operating at elevated temperatures spend a significant portion of their useful life propagating fatigue cracks to a critical size. For a significant fraction of engineering components in high-temperature service, the primary loads are generated by non- uniform heating during transient operating conditions. in such instances, fatigue cracks ex- perience simultaneous variations of temperature and strain. Simultaneous cyclic variations of strain and temperature are often referred to as thermomechanical mechanical fatigue (TMF) loading. Predicting fatigue crack growth for TMF loading conditions is an important engineering problem, especially in the gas turbine engine industry. This study was undertaken to examine the usefulness of various fracture mechanics ap- proaches in correlating TMF crack growth data. An attempt was also made to see if TMF growth rate data could be predicted from isothermal data. Such a prediction would be of en- gineering value in that isothermal testing is much less expensive and t?frere is much more iso- thermal data available at present. In some applications cracks grow from hot regions with nominally inelastic strains into cooler regions with nominally elastic strains. Accordingly, this program was directed at ac- cessing the usefulness of various fracture parameters for correlating TMF crack growth for both elastic and nominally plastic conditions. In order for a fracture mechanics parameter to be of use to the engineer, it must allow the prediction of fatigue crack growth rates in complex components utilizing data from simple specimens. The most important requirement for a parameter is that it can be used to predict crack growth rates independent of geometry. Accordingly, crack growth tests were run at several strain ranges such that crack growth rates at a single parameter value but different crack lengths could be compared. The success of a parameter is then related to the degree to which the crack growth rate is a single-valued function of the chosen parameter. The data was used to compare the usefulness of three fracture mechanics parameters. These parameters were the stress intensity factor range A&, the strain intensity factor range AK,, and the J-integral range AhJ. The test conditions included temperatures and strains for which the material experienced significant inelastic strain and also showed rate-dependent and load-history-dependent stress-strain response. Under such conditions none of the proposed tAssociate Professor, University of Connecticut and consultant to United Technologies corporation, Pratt and Whitney Aircraft Group. ~Formeriy Senior Analytical Engineer, United Technologies Corporation, Pratt and Whitney Aircraft Group, East Hertford, CT. Presently Product Development Engineer. 345