Ultrasonic acoustic-emission leak detection for wellhead valves
MSc thesis (2024): DSP and machine learning on ultrasonic acoustic-emission signals to detect leaks in wellhead valves, end-to-end from research to a working proof of concept, R&D at Mecantronic.
Acousticssoftwareacousticssignal-processingmachine-learning
This project combined DSP with machine learning to detect leaks in wellhead valves from their ultrasonic acoustic-emission signature, developed as MSc thesis research and development work at Mecantronic in 2024.
The work spanned the full pipeline: acquiring and characterizing ultrasonic AE signals, designing features and models to separate leak signatures from noise, and validating the approach end-to-end into a working proof of concept.