The BLINDFAST project aims to deliver an automatic inspection method for the installation of blind fasteners. Generally, in this type of bolts the back side of the assembly is not accessible, so a system that assesses the formed head at the back side with no access to it is...
The BLINDFAST project aims to deliver an automatic inspection method for the installation of blind fasteners. Generally, in this type of bolts the back side of the assembly is not accessible, so a system that assesses the formed head at the back side with no access to it is required.
The solution proposed consists of an in-line monitoring system that based on sensor signals acquired during the installation and conveniently analyzed provides an evaluation outcome on how the fastener was installed. The objectives defined to deliver such a solution are: (i) learn to generate defect installations, (ii) acquire in-line signals, (iii) extract sensitive signal features, (iv) implement and assess a classifier based on previous signal features.
Up to date, most relevant results achieved have been:
* A test bench with multi-signal acquisition capabilities is operative for fasteners installations
* Methods to create high, buckled and flared (formed) heads have been defined
* An algorithm based on a short dataset classifies OK/NOK installations due to variations in the grip-thickness pair
The new method will help production to decrease/eliminate time and cost intensive inspections and fasteners overinstallation in structures. The decrease of the number of installed fasteners will also contribute to weight savings and therefore, to a greener aviation.