Abstract
Fatigue tests with two levels of cyclic stresses were conducted on mild steels (aluminum killed steels). The higher stress levels were 1.20 and 1.42 times of the endurance limit and the lower ones were 0.80, 0.57 and 0.38 times. Two cycles of higher stresses and various cycles of lower ones in a loading period were repeatingly applied. Fatigue lives in terms of number of loading cycles, Zf, lineally decreased by log nL, where nL is the total number of lower stress cycles in a loading period. When the frequency of higher stress cycles was low, 1Hz in the present study, the lives to failure also decreased compared with the ordinary frequency. It was concluded that the modified Miner's method was unable to be applied to these experimental results, because the values estimated by that method varied widely with nL. Emphasis was placed on the need for a modified estimation method in which the rate of fatigue damage by lower stess cycles is not constant but decreases as increase of the lower cycles in a loading period.
