Elasto-Plastic Numerical and Non-Numerical Determination of Nominal to Local Deformation Transfer Functions for Strain Controlled Fatigue Testing in Air and High Temperature

Abstract Strain life fatigue curves are usually determined by directly controlling the strain amplitude of the standard (round or hourglass) specimen. This constitutes a well-proven straight forward experimental procedure for tests in air at ambient temperature. Depending on the individually designed test rig the local strain control parameter is not directly measurable any more in case of more complex (in medium at elevated temperatures) Environmentally Assisted Fatigue (EAF) testing. For this kind of testing the nominal deformation is controlled by way of a linear variable differential transformer (LVDT) at the shoulder and not within the gauge length (gauge position). As a consequence, an appropriate transfer function has to be derived accounting for the elasto-plastic correlation of the global (“shoulder”) and local gauge length (“gauge”) deformation. This task is fulfilled by elasto-plastic finite element analyses (FEA) of the specimen with a suitable material model and the specification of the shoulder deformation as the load model. The transfer functions are determined for 1.4404 (316L) austenitic stainless steel base and weld material and related material parameters. The application of the Twice Yield method with the implicit assumption of cyclically stabilized material behavior constitutes a very economic way of elasto-plastic analysis. Sufficient accuracy compared to the application of more sophisticated kinematic hardening constitutive material models has to be shown. Tests in air at ambient temperature with both LVDT (“shoulder”) and local gauge length (“gauge”) deformation measurement deliver an appropriate reference and validation. Here, the calibration methodology is specifically applied to EAF testing but can easily be transferred to notched specimen and component testing. After applying the required displacement boundary condition in a monotonic static analysis, the strain relationships are the direct result of the FEA calculation considering the twice yield method. As an alternative, a non-numerical approach is to replicate a recorded global strain signal from fatigue testing in air and using this as control function for strain control in EAF testing. Thereby, the hardening, softening, non-Masing and secondary hardening behavior of the austenitic stainless steels can then be replicated with an analytical function instead of FEA path prediction.