Investigation of the Characterization of Material Inhomogeneity and Its Effect on the Ductile-to-Brittle Transition in Ferritic Steel

摘要

It has become routine to characterize the ductile-to-brittle transition in ferritic structural steels in terms of the Master Curve approach and the To reference temperature. A standard procedure has been developed in ASTM E1921 which utilizes as few as six small specimens to develop both the to reference temperature and the expected confidence bounds. This procedure implicitly assumes that the material is homogeneous and the six specimens tested are a faithful representation of the material to be utilized in the structural application. In many important applications the structure can be very large and the specimens can be cut from material removed from a single location during fabrication. The question as to whether the specimen results can be used to accurately or conservatively assess structural performance requires additional consideration. Recently an annex to E1921 has been proposed with procedures to investigate whether a material is homogeneous or inhomogeneous. The work presented in this paper describes a research program which applied a simple Monte Carlo procedure to address deficiencies in the proposed annex. Idealized input data sets representing both homogeneous and bimodal materials were analyzed using the proposed procedure to examine the likelihood that the proposed criteria for identifying the homogeneity of the material is acceptable. The assumed "test temperatures" and the number of data input to the analysis strongly influences the accuracy of the results of the annex procedure. Based on the results of the analyses, proposals are presented for modifying the proposed homogeneity annex.