Numerical simulation of macrosegregation: a comparison between finite volume method and finite element method prediction

  • PDF / 441,469 Bytes
  • 14 Pages / 612 x 792 pts (letter) Page_size
  • 14 Downloads / 286 Views

DOWNLOAD

REPORT


I.

INTRODUCTION

MACROSEGREGATION, i.e., variations of average mass fractions at the macroscopic scale of a casting, plays an important role in many solidification processes. It can be induced by fluid flow in the mushy zone,[2–7] which redistributes segregated solute elements within the remaining liquid volume; by deformation of the solid skeleton in the semisolid state;[8,9] or by transport of equiaxed grains.[10,11,12] The movement of the fluid flow can be associated with thermal and/or solutal buoyancy,[3,4,5] shrinkage,[13,14] or surface exudations.[15] These two last phenomena usually lead to ‘‘inverse segregation,’’ i.e., to a solid at the surface of the casting, which is solute rich if the partition coefficient is smaller than unity. Natural convection gives rise to complex segregation maps, which are schematically summarized in Figure 1 for a rectangular cavity and two alloys: Sn-5 wt pct Pb and Pb-48 wt pct Sn. In this study, the situations where the cavity is cooled from the left lateral side have been considered. At the beginning of solidification, horizontal gradient of temperature and liquid solute mass fraction develop close to the chill. These two gradients lead to a horizontal gradient of the liquid density. As the gravity is perpendicular to this gradient, fluid flow driven by natural convection (both thermal and solutal) appears early during solidification. Since the partition coefficient of the solute of both alloys is smaller N. AHMAD, Postdoctoral Fellow, and J. RAPPAZ, Professor, are with Chaire d’analyse et simulations nume´riques, Ecole Polytechnique Fe´de´rale de Lausanne, CH-1015 Lausanne, Switzerland. J.-L. DESBIOLLES, Staff Member, T. JALANTI, Ph.D. Student, and M. RAPPAZ, Professor, are with Laboratoire de Me´tallurgie Physique, Ecole Polytechnique Fe´de´rale de Lausanne, CH-1015 Lausanne, Switzerland. H. COMBEAU, Assistant Professor, G. LESOULT, Professor, and C. STOMP, Postdoctoral Fellow, are with Laboratoire de Science et de Ge´nie des Mate´riaux Me´talliques, INPL, Ecole des Mines de Nancy, F-54042 Nancy, Cedex, France. Manuscript submitted July 11, 1997. METALLURGICAL AND MATERIALS TRANSACTIONS A

than unity, the interdendritic liquid becomes solute rich in both cases as solidification proceeds, i.e., as the temperature goes down. For the tin-rich alloy (Figure 1(a)), this liquid becomes heavier as an effect of the combined thermal and solutal fields and the resulting convection is counterclockwise (ccw). For the lead-rich alloy (Figure 1(b)), the effects of solute and temperature on the liquid density are opposite, but solute-induced convection dominates in this case and a clockwise (cw) vortex will form. When the cavity is cooled down from the bottom surface, stagnant layers of liquid form in the case of the tin-lead alloy (Figure 2(a)), whereas the situation of lead-tin is potentially unstable as the tinrich interdendritic liquid has a tendency to go up (Figure 2(b)). This can lead to Rayleigh–Be´nard cells in a first step or even to freckles (i.e., remelting of small channels

Data Loading...