Abstract
In this study, creep-fatigue tests were carried out under various total strain ranges and tensile hold times using AISI 316 stainless steel. Using the 13 data points obtained from the tests, creep-fatigue life was predicted by two methods, the modified Coffin-Manson method and the artificial neural network method. And then the 3 verification data were carefully chosen to evaluate the predictability of each method. The predicted lives were compared with the experimental results also and the following conclusions were obtained within the scope of this study. The creep-fatigue life prediction by the modified Coffin-Manson method had the average error of 16.2%. The average error in the artificial neural network method was 4.6%. As the result, the artificial neural network method with the adaptive learning rate was found to be far more accurate and effective than the modified Coffin-Manson method. The validity of artificial neural network method checked with 3 verification data points was proved to be very satisfactory also.