Validation of a statistical criterion for intergranular brittle fracture of a low alloy steel through uniaxial and biaxial (tension-torsion) tests

Abstract

Criteria initially developed by Beremin (1-5) for brittle cleavage fracture are applied to intergranular brittle fracture of a MnNiMo steel submitted to a temper embrittlement heat treatment. These models which are based on the maximum principal stress as a damage loading parameter and on the Weibull statistics are firstly reviewed and discussed. In particular the effect of plastic strain and of temperature on the intergranular fracture stress is emphasized. Then the results of fracture tests on smooth tensile specimens and on notched bars are used to identify the parameters of the models. Finally these models are applied to multiaxial tension-torsion tests carried out on thin tubular specimens. In these tests the observation of the orientation of the fracture surfaces clearly shows the importance of the maximum principal stress. These tests underline also the importance of one of the basic assumptions of the models which is the necessity of plastic deformation to initiate brittle fracture. In particular the existence in the r-tr plane of two domains for fracture, one controlled by the maximum principal stress, the other by plastic yielding is discussed. The results of further tests in which the material was submitted to a predeformation at room temperature before being tested at -196°C are also presented and used to discuss the effect of plastic deformation on intergranular fracture. In all cases it is shown that the Beremin criteria account very well for the scatter in the results and for the size effect observed in the experiments.