Predicting creep rupture from early strain data

To extend creep life modelling from classical rupture modelling, a robust and effective parametric strain model has been developed. The model can reproduce with good accuracy all parts of the creep curve, economically utilising the available rupture models. The resulting combined model can also be used to predict rupture from the available strain data, and to further improve the rupture models. The methodology can utilise unfailed specimen data for life assessment at lower stress levels than what is possible from rupture data alone. Master curves for creep strain and rupture have been produced for oxygen-free phosphorus-doped (OFP) copper with a maximum testing time of 51,000h. Values of time to specific strain at given stress (40-165 MPa) and temperature (125-350 degrees C) were fitted to the models in the strain range of 0.1-38%. With typical inhomogeneous multi-batch creep data, the combined strain and rupture modelling involves the steps of investigation of the data quality, extraction of elastic and creep strain response, rupture modelling, data set balancing and creep strain modelling. Finally, the master curves for strain and rupture are tested and validated for overall fitting efficiency. With the Wilshire equation as the basis for the rupture model, the strain model applies classical parametric principles with an Arrhenius type of thermal activation and a power law type of stress dependence for the strain rate. The strain model also assumes that the processes of primary and secondary creep can be reasonably correlated. The rupture model represents a clear improvement over previous models in the range of the test data. The creep strain information from interrupted and running tests were assessed together with the rupture data investigating the possibility of rupture model improvement towards lower stress levels by inverse utilisation of the combined rupture based strain model. The developed creep strain model together with the improved rupture model is foreseen to give a sound basis for life predictions of the OFP copper overpack canister for the spent nuclear fuel. (C) 2008 Elsevier B.V. All rights reserved.