P91 steels are widely used in high temperature components for power generation. Creep data is often generated through accelerated short term creep tests, for practical reasons, via increasing stress or temperature though this may alter the creep behaviour. Through normalising the creep test stress by tensile strength the Wilshire models reduce the batch to batch scatter in the creep data and enable the prediction of long term creep data from relatively short term test results. In this work it is shown that the Wilshire models fitted to uniaxial creep rupture data can be used to predict failure in both as cast and service exposed multiaxial tests. This is provided that the equivalent stress is the rupture controlling stress, as is the case for the P91 tests examined, and the tensile strength is measured as part of the test programme.

This content is only available via PDF.
You do not currently have access to this content.