13th International Conference on Fracture June 16–21, 2013, Beijing, China -8- a) Longitudinal section b) Cross section Figure 6. Typical microstructure of the weld produced at 129 J/mm and 5.27 mm/s(WS 1, HI 2) 4. Discussion To analyze the microstructural observations, all the test conditions described in section 3 were reported in a graph of the effective welding power (efficiency current voltage) versus the welding speed (Fig. 7a). Such a graph was used by Dye et al. to study the effect of the welding parameters on the microstructure of autogeneous TIG welds in IN718 [15]. For the purpose of result presentation, we used a power efficiency factor of 75% as proposed by Dye et al. In Fig. 7a, the microstructural features of the welds are represented by different symbols. Empty circles indicate that the weld is not fully penetrated. Welds without a straight centerline grain boundary, such as the ones shown in Fig. 3 and 6, are presented by full circles. Welds showing a straight centerline grain boundary (CLGB) are presented by full rectangles. Finally, welds with a longitudinal grain (LG) are reported using empty rectangles. To complete the map, we performed additional tests at different welding speed and welding power. In the map of Fig. 7a, the “weldable area” is defined by combinations of welding speed and welding power leading to a fully penetrated weld, free of CLGB and of LG. The “weldable area” was circumscribe using three straight lines. Two lines of constant heat input represented the frontier for the formation of a centerline grain boundary (HI = 134 J/mm) and for incomplete penetration (HI = 112 J/mm). The boundary for the formation of a longitudinal grain was fixed to a constant weld speed of 3.39 mm/sec (vertical line). According to these boundaries, the weldability area identified is very narrow. Dye et al. predicted theoretically a weldabilty diagram for the autogeneous TIG welding of 2 mm thick IN718 sheets. In Fig. 7b, we adapted their predictions to our sheet thickness. To fit our experimental results at a welding speed of 4.23 mm/sec, the ratio Δx/ characterizing the CLGB criterion was adjusted to 0.75. Other material and process related constants were considered identical to Dye et al. A comparison of our experimental map (Fig.7a) with the theoretical map of Dye et al., reveals three significant observations. First, the incomplete penetration criterion of Dye et al. is adapted to our weld configuration. Secondly, the criterion proposed for the formation of a centerline grain boundary is not well adapted. According to Dye et al., as the welding speed decreases, a microstructure free of centerline grain boundary could be obtained for a wider range of 1000 m 1000 m
RkJQdWJsaXNoZXIy MjM0NDE=