Al-Si eutectic alloys with various amounts of phosphorus (less than 1ppm and 15ppm) were prepared by using high-purity aluminum and silicon. The generation mechanism of very-fine globular α-Al formed in commercial alloy containing P was investigated by using gravity casting and die casting.

  A large number of very-fine globular α-Al formed in the casting surface when molten metal on eutectic temperature was injected into the mold at gate speed of over 200m/s by using a die casting process. However, when the melt temperature is higher or lower than the eutectic temperature, the number of very-fine globular α-Al decreased. On the other hand, very-fine globular α-Al did not form when molten metal of eutectic temperature was poured into the mold by using a gravity casting. In addition, even if the temperature of molten metal added with P immediately after it has been filled into the mold is low temperature estimated to be eutectic temperature, very-fine globular α-Al did not form in the part where AIP flocculation did not progress completely. These results suggest that injecting the molten metal into the mold at eutectic temperature has two effects for generating very-fine globular α-Al. 1) In alloys containing P, the nucleus of α-Al is generated easily because eutectic Si is hardly generated by the flocculation of AlP, which then becomes the nucleus of the eutectic Si. 2) Many nuclei are easily generated and released from the mold by injecting the eutectic molten metal into the mold.

  Recently, full mold casting is used to manufacture many kinds of large-sized castings. However, this casting method involves peculiar problems, including the occurrence of foam residue defects. In this study, we proposed a new casting design system called foam residue traps to reduce foam residue defects. These traps are directly attached to the product part in the same way as overflows or risers to collect molten metal that includes a large amount of foam residue. In order to maximize the effect of foam residue traps, we also optimized the shape of trap by using computational fluid dynamics (CFD) simulation and a multi-objective optimization method based on a genetic algorithm. Eventually, the effectiveness of the optimized shape was demonstrated through experiments with actual full mold casting.