Modelling thermomechanical fatigue life

Abstract

In the present study five different widely used life models were modified to allow for life prediction under thermomechanical fatigue (TMF) loading conditions. In order to assess the predictive capabilities of the models, life prediction results were compared with TMF data obtained on a stainless steel and a SiC particulate reinforced aluminium alloy. The analysis of the data obtained demonstrated that TMF tests are not sufficient per se to evaluate a life model. It is emphasised that TMF tests chosen for validation of life models should be conducted under loading conditions such that coupling effects between the various damage mechanisms become apparent. The data analysed indicate that appropriate TMF test conditions can be estimated for the case of creep-affected damage evolution. By contrast, selection of test conditions that would clearly reveal other potential couplings, say between oxidation and plasticity, is less straightforward. Unless the tests used to validate a model are designed to account for such effects, non-conservative life prediction may result for actual components despite a seemingly excellent correlation of model predictions with experimentally obtained life data.