An evaluation of back stress determination techniques in metals

The desire to develop unified creep-plasticity (UCP) models come from the necessity to design advanced nuclear reactor components for service under conditions which include combined creep and low cycle fatigue. These models should also be physically based since they would be used to extrapolate from laboratory data to predict long service lives. An approach to UCP modelling centers on the hypothesis that the inelastic strain rate is determined by a balance between the competing processes of work hardening and recovery. One class of UCP models is characterized by a power law relationship between strain rate and stress. A state variable common to these models characterizes the isotropic hardening and is allowed to evolve with history according to simultaneous work hardening and recovery. In order to treat behaviours unique to unloading or reverse loading conditions, several models also include a kinematic hardening variable which is also allowed to evolve according to a balance of work hardening and recovery. Such a treatment of inelastic deformation can mathematically treat a wide variety of behaviors. The measured response of 316SS and A800 indicates that the kinematic variable must, in steady state, be taken as a constant fraction (about 0.8) of the applied stress. This experimental result makes it impossible for the simple power law type expression to properly predict the commonly observed power law breakdown behavior in most metals and alloys. It is proposed that an expression for total inelastic strain rate involving the sum of two separate strain rate contributions is more appropriate. Acknowledging that separate expressions and separate mechanisms dominate low stress (engineering service) conditions and high stress (laboratory test) conditions requires that more emphasis be placed on long time, low stress laboratory testing. (orig.)