I’m currently in my fourth and final year studying towards an EngD in machining science as a member of the Industrial Doctorate Center (IDC) in machining science. I spend my time between the Dynamics Research Group (DRG) and the Advanced Manufacturing Research Centre (AMRC) at the University of Sheffield, along with Safran Landing Systems based in Gloucester.
- In-process monitoring
- Machining dynamics
- Signal processing
- Artificial intelligence
- Prognostic modelling
- Probabilistic risk assesment
- Real-time data analysis
My work focuses on the finish machining of Ti-5553 for safety-critical landing gear components within Safran Landing Systems. The aim of this project is to correlate in-process monitoring signals with process outputs such as tool wear and material surface finish, ideally allowing intelligent decisions and predictions to be made about the process without the intervention of an operator. This in turn will increase process efficiency and lead to immediate, tangible results and savings for Safran.
The below figure shows how a number of time domain features of an acoustic emission (AE) signal can be used to identify tool wear by clustering data into wear states (four in this example). An SVM can then be used to cluster incomming AE data into one of the designated states.
The below figure shows an example SVM classification with nine states achieving a classification accuracy of 93.9%. Dots marking correct classifications while crosses represent missclasifications.
Once clustering has been proven possible, state predictions can be performed using trained dynamical models.
An Alicona optical microscope is used to correlate AE data with the wear found at each cutting edge of a multi-flute tool. Below is an image showing a 3D scan of a tool flute with corresponding radius and angle measurement.
- N. Ray, K. Worden, S. Turner, J-P. Villain-Chastre & E.J. Cross, "Relationship between 3D tool geometry, in-process acoustic emissions and workpiece surface integrity in finish end milling”, Condition Monitoring 2015 conference (CM2015) BINDT
- N. Ray, K. Worden, S. Turner, J-P. Villain-Chastre & E.J. Cross, "Tool wear state clustering in milling based on recorded acoustic emission”, 6th European Conference on Structural Control (EACS 2016)
- N. Ray, K. Worden, S. Turner, J-P. Villain-Chastre & E.J. Cross, "Tool wear prediction and damage detection in milling using hidden Markov models”, International conference on noise and vibration engineering (ISMA 2016)
- R. Fuentes, E.J. Cross, N. Ray, N. Dervilis, T. Guo & K. Worden, “In-Process Monitoring of Automated Carbon Fibre Tape Layup using Ultrasonic Guided Waves”, IMAC XXXV conference on structural dynamics