【正文】 force is supplied by the frictional force between the feedstock and wheel groove, plex metal flow is observed in the groove of the wheel. However, internal defects in products, resulting from entrainment of impurities such as the oxide films on the feedstock, remain problematic. In this report, we were able to observe dead metal invading products both experimentally and by numerical simulation. We also attempted to remove the impurities deposited in front of the dead metal and along the shoe. Optimizing the clearance between the wheel and the abutment proved effective in removing these impurities, and use of a dividedflow die can decrease the admixture of impurities in extruded products.1. IntroductionConform extrusion is a continuous process in which the extrusion force is supplied by the frictional force between the feedstock and the groove of the wheel [1],[2]. Figure 1 is a schematic illustration of conform extrusion: 1(a) illustrates a type of conform extrusion in which the products are extruded from the die perpendicular to the wheel and Fig. 1(b) is another type, in which the products area extruded tangentially from the wheel. Figure 1(c) shows a crosssection of plane A in Figs. 1(a) and (b). The feedstock arrives coated with an oxide film, which risks causing problems if it is entrained into the extruded product. It is thought that defects in the product are caused by impurities being deposited in front of the dead metal [3],[4], making it important to observe the metal flow in the groove of the wheel. It is thought that this product defect is caused by impurities that are deposited in front of the dead metal and along the shoe. In this report, we were able to observe, both experimentally and by numerical simulation, the dead metal invading the products. Observation of the metal flow in the groove of the wheel suggested that the clearance between the wheel and the abutment affects the geometry of dead metal in front of the die.2. Experimental procedureFigures 2(a) and (b) show the experimental equipment for perpendiculartype conform extrusion. Figure 2(c) shows tangenttype extrusion. These experimental Conform machines lack a scraper, normally installed in industrial machines to strip the dregs from the bottom of the groove. The diameter of the wheel is 300 mm. The radius of the groove crosssection is 11 mm and its height is 25 mm. The rotational frequency of wheel is rpm. The extrusion ratio R is 5 or 10 (., the diameter of the cylindrical product D is or 8 mm). The extrusion ratio is defined as the crosssectional area of the groove divided by that of the product.The feedstock is made from Colour Clay (modelling clay), 22 mm in diameter. The feedstock for observing the velocity field is made of alternating black and white disks, as shown in Figure 3(a). The front part of this feedstock is made of grey Colour Clay to represent the dead zone. Feedstock for experiments to observe the stream line of metal is shown in Figure 3(b).3. Numerical simulationThe mercial finite element code Flow3D is used in the study. Figure 4 shows a sample of circular cylindrical coordinates and a mesh with grid location points for tangenttype conform extrusion. The conditions for calculation and experiment are listed in Table 1. The friction factor between the material and wheel is set very high since the wheel groove is covered by material film due to the lack of a scraper.4. Results of deformation and velocity field. Accumulating point and velocity fieldFigure 5 shows the material flow in the longitudinal section during roundbar extrusion (extrusion ratio R = 5, 10 in for the tangent type). There is a dead zone along the shoe in front of the die. In the experiment, the dead zone is grey, but it is blue in the numerical simulation. The angle of the accumulating point is measured in each case. The angle in the numerical simulation is not identical to that in the experimental result, but the same pattern is observed: the angle of the accumulating point increases at larger extrusion ratios. In the numerical simulation, the feedstock is thickened immediately before the accumulating point, but this was not observed experimentally. In the experiment and numerical simulation, the lower half of the feedstock is deformed before the accumulating point. However, the lower half, past the accumulating point, is deformed slightly. On the other side, there is a strong shearing zone along the dead zone in the upper half.. Geometry of dead zone and stream line at the surfaceFigure 6 shows the geometries of the dead zone in tangent type and R= geometry of the dead zone, which is grey in the experiment, is similar to that in the calculation. The area of the dead zone increases as the angle of crosssection θ decreases. But in θ = 0176。, the area of the dead zone bees smaller than that in θ = 10176。 because the material is extruded through the die hole which is present at the centre of the crosssection. In the calculation, the point whose velocity is fastest in every crosssection is the bottom of the groove, since the material there sticks to the wheel. However, the velocity of the bottom area in θ = 30176。, which is near the accumulating point, is not equal to that for θ = 20176。. The material slips off the wheel at θ = 20176。 and sticks to the wheel again at θ = 10176。.Figure 7 shows the material flow along the dead zone in the perpendicular type. The lower half is similar to the original geometry of the feedstock. There is a slight deformation in the lower half by pression of the dead zone, which increases as θ decreases. Near the accumulating point, mainly the upper half is deformed and fills up the section between the wheel and the shoe. Due to pression of the dead zone, the material overflows the groove from the clearance at the corner. The upper half of the feedstock flows in the direction of each side along the dead zone.Figure 8 shows how to measure the position