Unreliability in the Prediction of Creep Rupture Life and Its Avoidance

Abstract

Abstract. Methods of modeling and extrapolating creep rupture data are critically examined, contrasting conventional models evaluated by statistical methods with models produced by the parametric, numerical isothermal datum (P-NID) method developed by the writer. The conventional method used as the primary basis for this comparison is the procedure recommended by the European Creep Collaborative Committee (ECCC). The differences are discussed in terms of the reliability of predictions for the long-term properties of engineering materials, upon which industry standards and plant safety depend. The article illustrates how conservative adherence to established modeling and evaluation methods, which is common among those active in this field, has perpetuated the unreliability of predictions; and demonstrates that such problems are avoided by applying the strict criteria of the P-NID procedure.