Predicting creep strain response from rupture data and a robust creep curve model

Creep strain evolution plays a critical role in design and in life assessment of components subjected to service at high temperatures. For instance in turbines, boilers and steam pipes the recommended limits are 1% or 2% of strain. Unfortunately, straightforward engineering methodologies for predicting long term creep strain are not readily available. Robust methods for creep rupture extrapolation have been developed for example in the recommended procedures of the European Creep Collaborative Committee (ECCC) and PD6605 of BSI. The recently developed logistic creep strain prediction (LCSP) model transfers robustness into creep strain modelling. The LCSP model is directly linked to the rupture model and applies a small set of creep curve shape parameters to adjust primary, secondary and tertiary creep properties. The stress and temperature dependence of these shape parameters is optimised by fitting the available strain data. In this work the LCSP model was acquired for P22 steel from a small data set (max 3000 h) together with standard data for time to 1% strain. The model was then used to predict time to strain (0.5 to 5%) for two other P22 data sets using only true time to rupture. The model was also inversely applied to predict time to rupture from values of time to given strain for one of the data sets. The approach appears to be very competitive in spite of its simplicity, and is thought to hold considerable promise for e.g. creep modelling of new materials and welds, and when using FEA in creep analysis.