The high-temperature, low-cycle, thermal and mechanical fatigue behavior has been investigated for two high-strength, oxidation-resistant, superalloys used in aerospace propulsion systems. Experimental results were generated to support development of an advanced thermal fatigue life prediction method. Strain-controlled thermomechanical and loadcontrolled, strain-limited, bithermal fatigue tests were used to determine the fatigue crack initiation and cyclic stress-strain response characteristics of two superalloys, cast nickel-base B1900 + Hf and wrought cobalt-base Haynes 188.
Both in-phase and out-of-phase thermomechanical and bithermal fatigue experiments were conducted using computer-controlled, servohydraulic uniaxial fatigue machines equipped with diametral extensometry. The thermomechanical tests were conducted with triangular-wave diametral strain versus time and temperature versus time histories, with cycle times of 3 or 4 min. Bithermal experiments utilized a trapezoidal-wave temperature versus time profile wherein mechanical straining was imposed only during the isothermal portions at the maximum and minimum temperatures. Temperature was changed while the specimen was held at zero load, thus excluding mechanical strain application during thermal expansion straining. This testing procedure permits significantly more accurate partitioning of the mechanical and thermal expansion components of strain compared to conventional thermomechanical straining experiments. This is especially important when it is desired to further separate the mechanical strain into its elastic and inelastic components, the latter of which may be further partitioned into time-dependent and time-independent (creep and plasticity) components.
Bithermal temperatures of 483 and 871°C were employed for the cast B1900 + Hf nickel-base alloy and 316 and 760°C for the wrought Haynes 188 cobalt-base alloy. Thermomechanical fatigue tests were also conducted using maximum and minimum temperatures corresponding to those for the bithermal experiments. Lives cover the range from about 10 to 3000 cycles to failure. Comparisons are made with isothermal fatigue results obtained previously.