WELDING RESEARCH APRIL 2015 / WELDING JOURNAL 141-s based on the Scheil additive rule (Ref. 41): where ta is the incubation time required to reach a specified state on a TTT diagram for isothermal reactions, t is the time to that stage for the nonisothermal reactions, and dt is the time interval at temperature T. In this procedure, the total time to reach a specified state of transformation for nonisothermal reactions is obtained by adding the fractions of time to reach this stage isothermally until the sum reaches unity. The inclusion of the weld deposit compositions also leads to the calculation of the appropriate part of the phase diagram needed to obtain paraequilibrium compositions for kinetic analysis. These data are combined with austenite grain parameters and the computed cooling curves from the heat transfer and fluid flow model to calculate the volume fractions of allotriomorphic, Widmanstätten ferrite, acicular ferrite, and martensite (Ref. 23). The modeling results are used to better understand the experimentally observed microstructures of the weld metal in complete-joint-penetration hybrid laser-GMA welding. Results and Discussion Calculated Temperature and Velocity Fields The effects of welding speed and laser arc separation distance on weld profiles and temperature and velocity fields are studied separately for complete joint-penetration hybrid laser- GMA welding. Both these parameters have been shown to significantly affect the weld quality and welding productivity. Figure 1 shows the calculated temperature and fluid flow fields in three dimensions when the welding speed increases from 20 to 30 mm/s with a constant laser arc separation distance of 1 mm. The temperature is indicated by contour lines and the velocity field is represented by arrows. There is an intense temperature gradient on the top and bottom surfaces of the weld pool because the temperature of the liquid metal at the keyhole wall equals the boiling point, while the liquid metal at the weld pool boundary remains at the solidus temperature. The molten metal moves radially outward for both the top and bottom surfaces because of the Marangoni convection produced by the spatial gradient of surface tension resulting from the temperature gradient. The maximum outward flow velocities of the top surface are 155.6 and 194.2 mm/s for welding speeds of 20 and 30 mm/s, respectively. The momentum is then transferred by viscous dissipation to the inner layers of the weld pool adjacent to the surface. The comparison between these calculated and experimental weld pool geometries is shown in Fig. 2. The top surface width decreases from 9.6 to 7.0 mm, and the bottom surface width decreases from 3.6 to 2.0 mm when the welding speed increases from 20 to 30 mm/s. The boiling point contours indicate the cross-sectional geometry of the keyhole, while the solidus temperature contours indicate the fusion zone boundary. It is observed that the widths of the top and bottom surfaces, as well as the shape of the fusion zone predicted by the heat transfer and fluid flow model, agreed well with the corresponding experimental results. The weld pool dimensions decrease significantly with the increasing welding speed as the top surface widens due to Marangoni convection. Furthermore, the bottom width is also larger than the minimum weld profile width at the middle of the plate thickness, indicating the significant effect of convective heat transfer. Figure 3 shows the calculated temperature and fluid flow fields when the laser arc separation distance increases from 1 to 5 mm with a welding speed of 40 mm/s. When the laser arc separation distance is changed, the nominal total heat input of the welding process is constant, while the heat input decreases from 0.63 to 0.41 kJ/mm with a welding velocity increase from 20 to 30 mm/s, as shown in Fig. 2. The computed length of the weld pool increases from 18 to 23 mm with the increasing laser arc separation distance. The heat from the laser and arc as well as the droplet is more concentrated near the laser beam incident point at shorter laser arc separation distance. The heat distribution over the top surface of the weld pool significantly changes when the arc axis is further separated from the laser beam. The top part of the liquid weld pool is stretched and the distance along the welding direction from the maximum weld width to the laser beam incidental point is significantly increased. Figure 4 shows the comparison between these calculated and experimental weld pool geometries. The top surface width increases from 5.1 to 6.8 mm, and the bottom surface width increases from 1.2 to 1.6 mm when the laser arc separation distance increases from 1 to 5 mm. It can be observed from Figs. 1 and 3 that the effects of welding speed and laser arc separation distance on the weld profiles are different. When the welding speed increases from 20 to 30 mm/s, the weld pool shrinks significantly in all locations along the weld depth. However, only the top part of the weld pool significantly increases when the laser arc separation distance increases from 1 to 5 mm. The bottom part has a slight change and the middle part of the weld pool is almost unaffected by the laser arc separation distance. Changes in these processing conditions impact the resulting weld pool geometries in different ways. For example, the increase of the welding speed reduces the net energy absorbed by the workpiece. On the other hand, the increase of the laser arc separation distance mainly influences the energy distribution of the heat sources. Cooling Rates The cooling rates over the austenite decomposition range from 1073 to 773 K within the weld fusion zone are calculated using this same heat transfer and fluid flow model. The calculated cooling rates are then used in the modeling of the weld metal microstructures of low-alloy steels. Figure 5 shows the calculated cooling curves between 773 and 1073 K at different locations on the top surface of the weld with a welding speed of 20 mm/s. It can be seen that the cooling rates are almost independent of position. The cooling rates within the fusion zone and the cooling rate in the heat-affected zone (HAZ) are both at a level of approximately 20 K/s. These similarities in cooling rates allow a single cooling condition to be assumed for each horizontal plane across the weld depth. The comparison of the calculated ∫ t dt ta ( = T ) 1 (13) 0
Welding Journal | April 2015
To see the actual publication please follow the link above