Robust prediction of full creep curves from minimal data and time to rupture model

Abstract

A description of creep strain evolution is frequently needed in design or life assessment of components subjected to service at high temperatures. Unfortunately, long term creep strain data are not as easily available as rupture data or rupture models. One of the most demanding tasks in this context is to predict the creep behaviour reasonably accurately beyond the range of available data. Much effort has been invested into developing reliable methods to extrapolate creep rupture data, for example in the recommended procedures of the European Creep Collaborative Committee (ECCC) and PD6605 of BSI. However, for strain no such tools are currently available. Here a new and robust creep strain model is suggested to provide the whole creep curves based on the corresponding creep rupture model. This logistic creep strain prediction (LCSP) model defines the creep curve only with three additional parameters to those of the corresponding rupture model. In its basic form the LCSP model optimises a non-linear asymmetric logistic transition function fitted in logarithmic strain against a time temperature parameter (TTP), giving time to specified strain. Unique features of the model include its simple inverted expression for strain and strain rate. The fitting effectiveness of the new method is shown to match all the contesting creep strain models of the ECCC intercomparison activity on a single heat data set of the steel P22 (10CrMo9-10). The model is also shown to produce accurate predictions of the stress to 1% strain up to 100 000 h, when compared to the values given in DIN 17243.