STATISTICAL MODELLING OF INTERGRANULAR BRITTLE FRACTURE IN A LOW ALLOY STEEL

Abstract

The intergranular brittle fracture (IBF) behaviour of a low alloy steel 16MND5 (A508 Cl. 3) was investigated. A temper embrittlement heat treatment was applied to the material to simulate the effect of local brittle zones (ghost lines) which can be found in the as‐received material condition. An increase in the Charpy V toughness transition temperature and a significant decrease in the fracture toughness measured on CT‐type specimens were observed in the embrittled material, as compared to the reference material which was submitted to the same austenitizing and tempering heat treatment, but which was not subjected to the temper embrittlement treatment. Tensile tests on notched specimens were carried out to measure the Weibull stress and scatter in the results. A statistical model, the Beremin model, originally proposed for brittle cleavage fracture was applied to IBF. It is shown that this model is not able to fully account for the results, in particular for the existence of two slopes in a Weibull plot. Systematic fractographic observations showed that the low slope regime in this representation was associated with the existence of MnS inclusions initiating brittle fracture, while the larger slope was related to microstructural defects. Initiation of IBF from MnS inclusions can occur when the material is still elastically deformed while the second population of microstructural defects is active in the plastic regime. A modified statistical model based on the Beremin model and taking into account these specific aspects is proposed in the framework of the weakest link theory. The parameters of this model are identified from test results on notched specimens. It is shown that this model is able to predict the temperature dependence of fracture toughness and the scatter in the experimental results.