Recognition of thermal analysis cooling curves and refinement of molten Al-7Si alloys

李小春,李言祥,王强
DOI: https://doi.org/10.3969/j.issn.1005-0299.2003.02.022
2003-01-01
Abstract:A new way of intelligent evaluation of melt quality by pattern recognition of thermal analysis cooling curves has been proposed to evaluate the refinement efficiency of molten Al-7Si alloys. The cooling curve segment of the primary freezing stage (from the freezing temperature of primary aluminum to the freezing temperature of aluminum-silicon eutectic) is used for identification, and the relationship between the cooling curve-segment and the grain size (refining efficiency) is established using the relational database. The prediction of Al-7Si alloy refinement efficiency is realized by the pattern recognition of cooling curve based on the database. The experimental results show that when the comprehensive deviation parameter Ω which characterizes the similarity between two cooling curves approaches zero, the grain size difference converges to zero. Under a given solidification condition, the grain refinement efficiency can be precisely estimated by comparing the cooling curve of the melt with those cooling curves in the database to find out the most matching one.
What problem does this paper attempt to address?