PSI - Issue 13

Ghassen Ben Salem et al. / Procedia Structural Integrity 13 (2018) 619–624

623

Ben Salem Ghassen / 00 (2018) 000–000

5

Figure 6. (a) SEM examination of fracture surface (b) and Maximum principal stress profiles in the notch at failure for specimen 760AC

than cleavage fracture, mainly located around the specimen surface. To identify the initiating failure mechanism for this test, numerical simulation was compared to the SEM analysis. A F.E.A (Finite Element Analysis) of the tensile tests was developed in CAST3M [16] with an axisymmetric model taking into account the di ff erent layers presented in section1. The hard layer was modeled with a perfect plastic model and a yield stress of 2000 MPa [1]. The ferritic base metal, the ferritic HAZ, the decarburized HAZ and the austenitic buttering were characterized by tensile tests on smooth specimens. Maximum principal stress profile at fracture for specimen 760AC (Fig.6(b)) indicated a stress concentration on the interface between the hard layer and the buttering. The maximum value was reached at the interface near the specimen surface. On the ferritic side, principal stress values remained below 1350 MPa which is lower than the threshold stress of 1375 MPa for 18MND5 ferritic steel [17]. The stress concentration at the specimen surface in the interface is in agreement with the intergranular fracture observed in Fig.6(a) and sustains the scenario of an intergranular initiation at the MA interface. On the other hand, specimen ”760AM” configuration developed a stress concentration in the hard layer and higher stresses on the ferritic side whereas specimen ”760M” configuration seems to promote cleavage fracture in the ferritic HAZ. Thus, specimen 760AC configuration seems to be the most suitable for an intergranular initiation at the MA interface and was chosen for the threshold stress identification.

4.2. Threshold stress identification

Five additional NT tensile specimens with the same positioning as specimen 760AC were tested at the same temperature. Same SEM examinations combined with numerical simulations were carried out to confirm that in tergranular initiation at the MA interface was responsible for each specimen failure. The maximum principal stress ( σ I ) value at the interface at fracture was also numerically calculated for each specimen and tests were ranked by increasing σ I values. An experimental probability of failure was calculated for each test following :

i − 0 , 5 N

(1)

P r ( i ) =

where i is the rank of the specimen and N the number of tests. Then a 3 parameters Weibull law was used to define the threshold stress ( σ th ) below which an intergranular fracture at the MA interface (and brittle fracture of the specimen in general) cannot occur. According to Fig.7, σ th was estimated to be 1620 MPa, the other weibull law parameters being the weibull modulus m AE and the scale stress σ u AE . Fitted m AE and σ u AE values are relatively low due to the high dispersion in σ I values caused by the unevenness of the MA interface. The threshold stress value obtained is consistent with the maximum principal stress profiles at failure obtained by numerical simulation of the previous CT specimen tests. Specimens which showed an inter granular fracture at the MA interface at di ff erent temperatures developed in fact principal stresses higher than the threshold stress of 1620 MPa. This study focused on the brittle behavior of SS DMW in the brittle-to-ductile transition. Its main objectif is to characterize the fracture resistance of the 18MND5 / 309L interface where a high variety of microstructures is formed during the welding process. Moreover, carbon di ff usion from the low-alloy ferritic steel (18MND5) to the weld metal side (309L) during the PWHT creates a hardened layer of carburized martensite and austenite. This hard layer seems to be a weak zone in the DMW because of the local hardness gradient and metallurgic singularities. Fracture toughness tests were carried out at temperatures between − 120 ◦ C and − 50 ◦ C, with a precrack 5 5. Conclusion