Tension and torsion low-cycle fatigue tests were carried out on thin tubular specimens of 316 stainless steel, tested at both room temperature and 600°C. Two types of experiments were performed: (1) conventional continuous isothermal tests, and (2) isothermal sequential tests with different sequences of tension/compression and torsional loading. Most of these tests were carried out under air environment at the equivalent total strain amplitude of Δεt/2 = ±0.80% in tension or Δγt/2 = ±0.80 √3% in torsion. Optical and scanning electron microscopy were used to analyze the failure modes, the orientation of microcracks and macrocracks which lead to failure, and crack density.
Continuous tests show two main results: (1) it is confirmed that, for the same equivalent von Misès strain, the fatigue lives in torsion are larger than those observed in tension; (2) detailed examinations of the microcracks' orientation showed that Stage I initiation followed by Stage II growth occurs only at room temperature or at elevated temperature under vacuum. At 600°C, however, Stage I crack initiation is bypassed due to the formation of external hematite (Fe2O3) and internal (Fe-Cr rich) oxides.
These results are used to explain qualitatively the significant deviations observed in the sequential tests from the Miner linear cumulative damage rule. A quantitative model based on a Monte Carlo simulation is developed to account for the results of sequential tests in which crack initiation and crack growth were coplanar. This model includes three fatigue damage laws: (1) a nucleation rate equation; (2) a push/pull or torsion crack growth law; and, (3) a criterion for crack coalescence. It is shown that this model is able to reproduce with good accuracy the results of sequential push/pull → torsion tests carried out at 600°C.