2014

Cetin, Ali; Roiko, Andrew; Lind, Minna

Advancements in applied statistics with great relevance to defect sampling and analysis are presented. Three main issues are considered; (i) proper handling of multiple defect types, (ii) relating sample data originating from polished inspection surfaces (2D) to finite material volumes (3D), and (iii) application of advanced extreme value theory in statistical analysis of block maximum data. Original and rigorous, but practical mathematical solutions are presented. Finally, these methods are applied to make prediction regarding defect sizes in a steel alloy containing multiple defect types