3D打印AlSi10Mg铝合金内部缺陷聚类分析

Cluster analysis of internal defects of AlSi10Mg aluminum alloy by 3D printing

使用3D打印机制备了15mm^3的AlSi10Mg铝合金正方体。然后使用高精度纳微米CT扫描该立方体内部的打印缺陷,提取出大于0.0001mm^3缺陷的构形和尺寸等属性。先后使用因子分析和K-means聚类处理提取出的数据样本。分析结果:CT扫描得到348个缺陷样本,提取样本的16种属性,用因子分析计算因子和综合得分,找出最严重的前20个缺陷。利用K-means聚类把348个缺陷样本分为4类,构建矩阵散点图对聚类结果进行验证分析,聚类结果可靠。

A 15 mm^3 AlSi1OMg aluminum alloy cube was fabricated by using a 3D printer.Then the printing defects in the cube were scanned by high precision nano-micron CT,and the properties such as the shape and size of defects larger than 0.0001 mm^3 were extracted.Factor analysis and K-means clustering were used to extract data samples.Analysis results show that 348 defect samples are obtained by CT scanning,16 attributes are extracted from the samples,factor analysis is used to calculate factors and comprehensive sco res,and the most serious 20 defects are obtained.348 defect samples were classified into four categories by K-means clustering.Matrix scatter plots were constructed to verify the clustering results.The clustering results were reliable.