Joseph Fields

Physics-Informed Machine Learning of Metal Cutting Performance

Supervised by: Prof. Neil Sims, Dr Javier Dominguez-Caballero, and Dr Dan Östling

I have decided to pursue this EngD as I enjoy solving practical problems and learning new things. I'm interested in machine learning and how it can be directly applied to industrial applications.

The aim of this project is to be able to predict when tools are going to fail without continuous inspection. Physics-informed machine learning will hopefully reduce the amount of tool wear data required, as there are a wide range of shapes and materials used. This project is also interested in trying to predict material surface finish as part of this.

Sandvik have a new intelligent tool holder, which provides a wide range of information when machining. I'm excited to see what insights can be gained from it.

This research will be used to optimise tool lifespan, prevent tool breakages from damaging workpieces and provide intelligent predictions of material surface finish.