An incremental minimization principle suitable for the analysis of low cycle fatigue in metals: A coupled ductile–brittle damage model

The present paper is concerned with a novel variational constitutive update suitable for the analysis of low cycle fatigue in metals. The underlying constitutive model originally advocated in [1] accounts for plastic deformation as well as for damage accumulation. The latter is captured by a combination of two constitutive models. While the first of those is associated with ductile damage, the second material law is related to a quasi-brittle response. The complex overall model falls into the range of so-called generalized standard materials and thus, it is thermodynamically consistent. However, since the evolution equations are non-associative, it does not show an obvious variational structure. By enforcing the flow rule as well as the evolution equations through a suitable parameterization, a minimization principle can be derived nevertheless. Discretized in time, this principle is employed for developing an effective numerical implementation. Since the mechanical subproblems corresponding to ductile damage and that of quasi-brittle damage are uncoupled, an efficient staggered scheme can be elaborated. Within both steps, Newton’s method is applied. While the evolution of the quasi-brittle damage requires only the computation of a one-dimensional optimization problem, the ductile damage model is defined by a numerically more expensive tensor-valued variable. For further increasing the numerical performance of the respective minimization principle, a closed-form solution for the inverse of the Hessian matrix is derived. By numerically analyzing the prediction of mesocrack initiation in low-cycle fatigue simulations, the performance of the resulting algorithm is demonstrated.

► Low cycle fatigue model including ductile and quasi-brittle damage.
► The model is based on energy minimization.
► The model is thermodynamically consistent.
► An efficient staggered implementation is proposed.
► Closed-form, physically-based inverse of a 4th-order Hessian Matrix is proposed.