Multiple linear regression approach to predict tensile properties of Sn-Ag-Cu (SAC) alloys

Abstract

In this study, a number of microstructural characteristics of SAC207 and SAC307 alloys were quantitatively extracted and utilized in a multiple linear regression (MLR) model. The MLR results well represent the ultimate tensile strength (σu) of both alloys, but not the elongation (δ) of the SAC207 alloy. The addition of the tertiary dendritic spacing considerably improved the MLR results for the SAC307 alloy. These findings, which are more consistent than those obtained by single linear regressions (SLR), might give technical supports in either regulating or improving tensile properties of solder alloys.