Industry 5.0 - Machine Learning for CNC Machine Tool Health Monitoring
Computer numerical control (CNC) milling delivers high precision manufacturing. However, unexpected tool wear and breakage causes unplanned downtime, scrap, and safety risks. High-end tool condition monitoring suites work, but they are costly to buy, disruptive to fit, and hard for many small- and medium-sized enterprises to justify. This project asks a focused question: can a single low-cost inertial measurement unit, fixed securely to the machine, produce clean, labelled signals that separate core states like idle, spin-up, movement, and cutting etc. This creates a practical starting point for thresholding and later machine-learning classification. The poster sets out an end-to-end pipeline from raw motion to interpretable features, and tests where a simple sensor is enough and where it is not. I hope you enjoy the poster.
Please sign in
If you are a registered user on Laidlaw Scholars Network, please sign in