Product variability analysis by geometrical hidden features

This work deals with the study of some components of uncertainty deriving from the manufacturing process. Starting from geometrical data extracted from a workpiece, "technological fingerprints" left from the process itself on the workpiece are identified and characterised. The set of measured points brings with itself the history of the product development, including the variability associated to the manufacturing process. Indeed, it contains a mixture of different elements, that should be analyzed and identified in the analysis of manufacturing, in order to gain the whole information necessary to improve the quality. This has been done from the macro-geometrical point of view, with the decomposition of complex geometries into simple ones with no prior knowledge about the features. In the micro-geometrical discrimination, I have introduced a fairly new technique to discover and discriminate the hidden components of variability - starting from the ISO normative on filtering - by means of independent component analysis (ICA).

Autorentext

Martina Gandini graduated in applied math in 2005 and obtained a Ph.D. in production systems and business economy in 2010. Her work deals with the improvement of mathematical aspects founding the basis of the Geometrical Product Specification and Verification framework.