Abstract
The failure life of metals subjected to thermal stress cycling is often discussed referring to the lifetime in strain-cycling. In this case, the comparison of both strength is made, taking the mean temperature of thermal cycling test is the same as the test temperature of strain-cycling. In this mode of comparison of strength for these two types of tests, a question arises as to whether the mean temperature may be used as the common basis of test condition. There opens another question to research; in this comparison, locarization of plastic strain due to uneven temperature distribution along the length of specimen is not taken into account. In this connection, special attention is paid to these particular points in the present study.
Thermal cycling tests on AISI 318 type stainless steel are performed by maintainig mean temperature of 650°C for various temperature amplitude. Strain-cycling tests are also made under steady temperature of 650°C which is equivalent to mean temperature of thermal cycling test, where temperature distribution along the length of specimen is uneven, being similar to the case of thermal cycling tests by using the same heating setup. From these two types of tests, it is found that, in the range of comparatively large value of total strain amplitude, both data are very close in log-log plots of total strain amplitude and number of cycles to fracture, but they are otherwise in the range of small total strain amplitude. A similar relation holds even when the total strain is replaced by plastic component of strain. From this, it is possible to mention that the difference of lives for both series of tests of thermal cycling and strain-cycling which has been emphasized by other investigators consists partly in the difference of temperature distribution along the length of specimens in both tests.
