BRITTLE FRACTURE ANALYSIS OF DISSIMILAR METAL WELD JOINTS

Demonstration of large components integrity is based on the demonstration that they could never undergo brittle fracture. In parallel, considering a conventional defect somewhere in the component, failure must prevent ductile crack propagation. Connections between a ferritic component and an austenitic one have to respect these rules. The considered case is a Ni base alloy weld joint between a ferritic pipe and an austenitic one. For brittle fracture exclusion, the aim of the present study is to show that in the same loading conditions, the weld joint is less sensitive to the brittle fracture than the surrounding ferritic part of the component. That is to say that the demonstration should be focused on the ferritic base metal which is the weakest material. For that purpose, Chapuliot developed a stress-based criterion model, using a threshold stress (oh) below which the cleavage cannot occur. It can be used to define the brittle crack occurrence probability, which means that it is possible to determine the highest loading conditions without any brittle fracture risk. For the experimental part of this demonstration, two different mock-ups with DMW are used. The first one is composed of a 16MND5 (A508 cl. 3) steel pipe welded to a 316L stainless steel pipe with Ni base alloy 82. As for the second one, the materials are the 18MND5 (A533) steel, the 316L stainless steel and the Ni base alloy 52. Conventional defects have been considered in the ferritic part, close to the weld joint in the heat affected zone, and far away from the weld joint in the ferritic part. Two hundred specimens have been taken from the mock-ups: special tensile specimens, compact tensile specimens, single edge notch tension specimens, notched tensile specimens, smooth tensile specimens. All the materials have been characterized at -125 degrees C/-130 degrees C and -170 degrees C/-175 degrees C, even the heat affected zones. Finite element calculations have been done to complete the experimental investigations. The first results show that, due to the mismatch of the materials, the brittle fracture risk is much lower in the HAZ. Thus, DMW HAZ could not be a weak part concerning brittle fracture. This paper presents the criterion, the experimental work and the analyses made to evaluate the conservatism of the homogeneous ferritic case compared to the DMW.