(Elsevier, 2025-10-01) Penalva Oscoz, Mariluz; Gil del Val, Alain; Martín, Ander; Villanueva Roldán, Pedro; Uralde Jiménez, Virginia; Veiga Suárez, Fernando; Ingeniería; Ingeniaritza; Universidad Pública de Navarra / Nafarroako Unibertsitate Publikoa
Fastening is a crucial operation in the aircraft manufacturing cycle, and the demand for automated solutions has grown in recent years. Blind rivets are particularly suitable for automation due to their ease of use. However, fastening with blind rivets requires indirect evaluation of the formed head for in-line quality monitoring. This study presents two approaches to address this problem. Firstly, an analysis of the drilling-riveting chain assesses the impact of the previous operation on riveting outcomes. Secondly, time-dependent signals from the riveting process are coded into images and analysed using deep learning techniques. Despite some limitations, both methods for monitoring blind riveting have demonstrated high precision and accuracy values above 0.9, with 1 indicating perfect precision or accuracy, suggesting that they can reliably predict the quality of rivet installations.
