Combining Real Time Classifiers for Fast and Reliable Electronic Nose response analysis for Aerospace NDTs

摘要

Fast response and reliability are a prerogative in non-destructive tests specifically in aerospace industry for safety and efficiency reasons. Currently, composite panels bonding, in green aircraft concept, is lacking a validated NDT technique for the bond quality. E-noses equipped with PARC algorithms appear a promising choice to acquire speedily a complete pattern response maximizing reliability. In this paper, combining real time classifiers, we show how to obtain a rapid firsthand response with the possibility of increasing accuracy awaiting for the end of the e-nose measurement cycle. A reject option is casted on the base of classifier self-perceived reliability to nullify false negatives while keeping the false positive rate at minimum.