Spatial domain mapping from in-process sensor signals for visual inspection of multi-material stack drilling

Airliner assembly processes involve components being pre-assembled into a ‘stack’, which is then drilled through. Manufacturers have strict hole quality requirements and need confidence in hole quality, since defects such as burrs and delamination can affect structural integrity. Human experts can be empowered to perform hole quality inspection through the provision of useful information. Visual representations of signal features and their association with the spatial and temporal features in the hole quality is a powerful mechanism by which to facilitate quality inspection. This paper proposes a novel sensor signal integration framework to map sensor signals from the time domain to the relative spatial domain as indicated by the drill bit position. Kalman filter based rotational position estimation from fibre-optic signal and relative drilling depth estimation from laser signal provided the relevant spatio-temporal information for the mapping. The resulting spatial domain mapping enables visualisation of signals for the detection of any defect related anomalous patterns for a human expert to inspect hole quality. Its potential is demonstrated on a real-world drilling trial of different quality holes.