Grain structures in aluminum alloy GTA welds
- PDF / 2,432,539 Bytes
- 9 Pages / 594 x 774 pts Page_size
- 65 Downloads / 227 Views
THE properties of a weld are related to its grain structure. For example, welds in which the grains grow straight to meet at a distinct centerline are more likely to exhibit hot cracking than welds in which grains gradually curve to follow the heat flow. l Grain refinement also reduces the sensitivity to hot cracking. 2,3 Such factors have led to the investigation of grain structures in GTA welds in several materials. In aluminum alloys, unlike steel welds: the solidification structure is readily revealed, and a variety of structures have been observed by Japanese workers, s,6 Their explanations for various structural transitions are based mainly on the changes in growth rate and temperature gradient, with different welding conditions, affecting the amount of constitutional undercooling at the advancing solid-liquid interface. For example, the transition to an equiaxed structure was ascribed to a sufficiently long constitutionally undercooled zone: ,6 Both theory and experiment have shown that solidification morphology in any given alloy depends on the ratio of G / R where R is the solidification velocity and G the temperature gradient normal to the macroscopic interface. For different alloys the critical ratio of G / R (or G / R l/z) depends mainly on the freezing range) Hence if constitutional undercooling is determining the transition to equiaxed grains, the critical G / R ratio should increase as the freezing range of the alloy increasesl Using the reported 6 length of the constitutionally supercooled zone to calculate the temperature gradient along the weld centerline Go, and the reported structural transition to determine the critical solidification velocity (taken as the welding speed V at the chosen6 heat input of 1400 J/s) the ratio of G / V for the transition to equiaxed grains may be calculated. Figure 1 shows the calculated G / V vs the freezing range. Contrary to solidification theory, no correlation between the two is apparent. This is inconsistent with the proposal that the transition to equiaxed grains is T. GANAHA, formerly Graduate Student, University of Waterloo, is now Welding Engineer, Guelph Engineering Co. Ltd., Guelph, Ontario, Canada. B. P. PEARCE and H. W. KERR are Research Assistant and Professor, respectively, Department of Mechanical Engineering, University of Waterloo, Waterloo, Ontario, Canada. Manuscript submitted December 27, 1979.
dependent on the degree of constitutional undercooling. The above analysis, as well as the original analyses5,6 may be criticized on the basis that the temperature gradients and solidification velocities are those along the weld centerline, whereas the equiaxed transition may take place closer to the fusion line. More accurate calculations of G and R are available7 which could be used to refine the calculations, but other grain refinement mechanisms should also be considered. In ingots, other observed mechanisms include dendrite arms breaking off via fluid flow or remelting, nucleation at the surface, and bulk heterogeneous nucleation?,4 A more detailed
Data Loading...
