Benchmark to Evaluate Revised Censoring Procedures for ASTM E1921

Abstract

Abstract ASTM E1921, Standard Test Method for Determination of Reference Temperature, T0, for Ferritic Steels in the Transition Range, contains a censoring criterion, KJΔa, that limits the amount of ductile tearing prior to cleavage in specimen data used to calculate the reference temperature, T0. A frequent complaint from users is that the requirements for both determining and implementing KJΔa are unclear and difficult to apply. These provisions require that KJΔa be determined using either the maximum valid KJc data from another result within the data set (i.e., one with an acceptance amount ductile tearing prior to cleavage) or a K-value based on the material’s JIc value (i.e., KJIc). Both options are potentially problematic as the JIc value is often unknown and assigning other test data as the censoring value can cause situations where this value is higher than the measured KJc value for that specimen. A benchmark analysis evaluated two alternative options for ductile-tearing censoring requirements proposed to rectify these inadequacies. The first option uses an absolute ductile-tearing-limit criterion and is simpler to apply. The second option employs a ductile tearing limit that varies with the KJc result but is designed to minimize the conservative bias associated with excessive ductile tearing. Data sets for low upper-shelf materials with sufficient data to calculate valid T0 values, yet with a significant percentage of data with measurable ductile tearing prior to cleavage were selected for evaluation. The benchmark results illustrated that although there is no obviously better censoring option, the option that uses an absolute ductile-tearing-limit is simpler to implement, whereas the T0 results are generally consistent with those calculated for the option with the variable ductile-tearing-limit.