[2D5] A novel printed sparse array for ultrasonic wide area corrosion monitoring utilising condition monitoring machine learning
M McInnes, C Dick, C Thring, D Irving and D Hughes
Novosound Ltd, UK
Corrosion costs industry millions in repairs and inspections every year. While there are many sensor systems designed for corrosion monitoring, no one system can provide the sensitivity and continuous full system coverage required to tackle these costs. Permanently installed ultrasound transducers are a valuable tool in the progression towards Industry 4.0; however, most only provide single-point location data. Guided wave ultrasonics can provide full asset coverage; however, there is a trade-off between range and sensitivity. This paper demonstrates a novel sparse array of sensors capable of detailed corrosion measurement and classification on a sub-mm scale. The density of sensors is made feasible by the easy-to-install, scalable, printed detection sensors and so can provide full asset coverage. The complexity of the guided wave interpretation is simplified by the application of machine learning techniques, demonstrating key progress to Industry 4.0 and the capability for an automated early warning system against hidden corrosion.
