Joint evaluation of fracture results from distinct test conditions, implying loading, specimen size and geometry

The reliability of the statistical assessment of test results is strong influenced by the reduced data size emanating from the low number of specimens included in each of the groups, characterized by test type, specimen geometry and size implied in the material sample being tested. Due to the traditional economical and time limitations, the available experimental results owned by a research group lack of parameter diversity so that they must be complemented by foreign results reported elsewhere in order to verify a model, a methodology or reinforced probabilistic conclusions. Generally, difficulties arise in assessing the parameters involved in the phenomenon due to the diversity of samples and test conditions supported by few data results. In this paper, a methodology based on a generalized local model, denoted GLM, is proposed to overcome this limitation allowing a joint evaluation diversified tests as those resulting from different loading conditions and specimen shape and size. It implies the primary failure cumulative distribution function, PFCDF, as the failure characteristic of the material to be derived from experimental results obtained from distinct test programs. In this work, the local methodology is applied for probabilistic assessment of cleavage fracture toughness data of steel A533B lent from an external experimental program carried out by Rathbun et al. (2006), aiming at analyzing the constraint influence on the cleavage fracture once the scale effect is recognized