Effect of short-term data on predicted creep rupture life: Pivoting effect and optimized censoring

Stefan Holmström, Pertti Auerkari

    Research output: Contribution to journalArticleScientificpeer-review

    7 Citations (Scopus)


    Fitting data to classical creep rupture models can result in unrealistically high extrapolated long-term strength. As a consequence, the standard strength values for new steel grades have frequently needed downward correction after obtaining more long-term test data. The reasons for non-conservative extrapolation include the influence of short-term data, which are easiest to produce but tend to pivot upwards the extrapolated values of creep rupture strength. Improvement in extrapolation could be expected by reducing this effect through model rigidity correction and censoring of very short-term data, but it may not be immediately clear how to justify the correction of particular models or censoring.

    Analogously to the instability parameter in the minimum commitment model for creep rupture, a rigidity parameter correction (RPC) is introduced to assess the pivoting effect of creep rupture models for the purpose of reducing potential to non-conservativeness in extrapolation. The RPC approach can be used with any creep rupture model for comparing the model rigidity and the potential benefit from censoring short-term data. The correction itself will never introduce non-conservatism, regardless of the model. The RPC approach is demonstrated by analyzing an ECCC data set for cross-welded 9%Cr steel (E911).
    Original languageEnglish
    Pages (from-to)103-109
    Number of pages7
    JournalMaterials at High Temperatures
    Issue number3
    Publication statusPublished - 2008
    MoE publication typeA1 Journal article-refereed


    • Creep rupture life
    • Optimized censoring
    • Pivoting effect
    • Short-term data


    Dive into the research topics of 'Effect of short-term data on predicted creep rupture life: Pivoting effect and optimized censoring'. Together they form a unique fingerprint.

    Cite this