Micromechanics based fatigue life prediction of a polycrystalline metal applying crystal plasticity

• A statistically RVE consisting of a number of crystal grains is applied.
• The RVE is subjected to repeated symmetric tensile-compressive loading.
• The inhomogeneous strain fields are statistically analyzed.
• Three criteria are proposed for predicting LCF failure of a metal.
• A family of curves for fatigue lives can be predicted by varying the critical values.

The fatigue-life of a polycrystalline superalloy under symmetrical cyclic strain controlled loading at a temperature of 650 °C is investigated by numerical simulations on the micro-level, focusing on the inhomogeneous evolution of plastic deformation in a polycrystalline aggregate. A methodology (Zhang et al., 2011, 2013) to predict the low-cycle fatigue life by micro-level simulations along with statistical analysis is applied following the steps: (1) A statistically representative volume element (RVE) consisting of a number of crystal grains is constructed by Voronoi tessellation. Stresses and plastic strains are calculated by a crystal plasticity model including nonlinear kinematic hardening. (2) The RVE is subjected to repeated symmetric tensile-compressive loading. (3) The inhomogeneous stress and strain fields are statistically analyzed during the load cycles. (4) Failure by LCF is strain controlled and occurs if either of the quantities, standard deviation of longitudinal strain in tensile direction, maximum or statistical average of first principal strains in the RVE at the tension peak of cyclic loading reaches a respective critical value. (5) Using the present methodology, a family of failure curves for fatigue lives under different strain amplitudes can be predicted by varying the critical values. Finally, appropriate critical values can be identified by a respective cyclic experiment with only one strain amplitude.