Search

Solidification cracking has been a serious problem in austenitic stainless steels. Since solidification cracking often occurs at the solidification grain boundary, the grain boundary character distribution (GBCD) in the weld metal may affect the solidification cracking susceptibility. On the other hand, the columnar grains of the weld metal are formed by the epitaxial growth from the grains in the base metal. Therefore, it is expected that microstructure and cracking susceptibility of the weld metal can be reduced by controlling the microstructure of the base metal. In this study, the effect of microstructure of the base metal on the microstructure and solidification cracking susceptibility of the weld metal was examined. The GBCDs of the base metal of the type 310S stainless steel were changed by the thermomechanical process. The solidification cracking susceptibility was evaluated by the Trans-Varestraint test. The thermomechanically processed specimen had larger ratio of coincidence site lattice (CSL) boundaries compared with those of the as-received specimen. Moreover, the specimen was found less susceptible to solidification cracking than was the as-received material. The GBCD analysis indicated that the complex shape and distribution of the grain boundaries in the thermomechanically processed specimen affect the initiation and propagation of solidification cracks.

We have investigated the martensite transformation phenomena of a Cr-Ni type weld metal, especially the effects of stress and plastic strain that exist before the martensite transformation start. Tensile specimens machined from the weld metal were firstly heated up to 800 °C to austenitize the specimens and then cooled down. During the cooling, we applied stress and plastic strain before the martensite transformation starts. The results show that the stress increases the M s temperature, while the plastic strain decreases the M s temperature. In addition, the transformation-induced plasticity was also investigated. Finite element method (FEM) analysis was conducted considering these effects, and the results show that these effects on the residual stress and the weld distortion are considerably large. The FEM analysis shows that the compressive residual stress in the weld metal is likely to be decreased by the transformation-induced plasticity. We also conducted hydrogen diffusion analysis for the case that the retained austenite exists in the weld metal by varying the M s temperature in order to consider the appropriate time period to keep a cold cracking test specimen before the crack observation.

The hydrogen-releasing properties from weld metals that mainly consist of martensite and contain some amount of retained austenite were experimentally investigated in order to determine the physical properties such as the activation energies for the simulation of the hydrogen distribution in welded joints and to consider the time period to age a cold cracking test specimen before crack observation. The numerical calculations for single-pass welding show that in the case of the 2.5 % and 4.2 % retained austenite, the hydrogen contents in the martensite become sufficiently low in 2 days. In the case of 16 % retained austenite, about 40 % of the initial hydrogen content still exists in the retained austenite 7 days after the completion of the welding. But the hydrogen content in the martensite becomes less than 1 % of the initial content in 7 days. Hence, the 7-day-ageing time before crack observation is considered to be adequate. The numerical calculations for multi-pass welded joints show that the 7-day ageing is also considered to be appropriate.

Hydrogen evolutions for martensitic-austenitic weld metals were measured at 45 °C. The specimen without the austenite released hydrogen in about two days, and this result agrees with the 72 h holding time required by JIS Z3118. In the case of the specimens that contain the retained austenite, hydrogen release continued for about 100 days or more. We derived the analytic solution of the McNabb & Foster diffusion equation by regarding the austenitic microstructures as trapping sites to compare the experimental and calculation results, and showed that both of them are in good agreement. We also showed that the initial ratio of free and trapped hydrogen atoms is not necessarily equal to the thermal equilibrium condition at the measurement temperature, which can be analysed using the present solution.