Machine Learning for Predictive Capability for Femtosecond Laser Processing

Lead Research Organisation: University of Southampton
Department Name: Optoelectronics Research Centre

Abstract

Laser materials processing is a £10 billion global industry that continues to experience rapid growth of 10% per year. Femtosecond pulse lasers are expected to become a key industrial manufacturing strategy for improving the resolution of materials processing, as the extremely short time scales for the laser pulses mean that the material removal method is predominantly ionisation, rather than melting.

However, this highly nonlinear interaction results in significant challenges for the prediction of the optimal laser and material parameters given any particular application. Due to the computational requirements, simulations that are based on first principles have yet to be accurately scaled-up to experimentally useful dimensions.

However, recent advances in neural networks, in particular conditional adversarial networks, have shown the profound ability for a computer to "learn" the properties of extremely complex processes.
The objective of this project is the creation of an all-inclusive femtosecond light-matter model for laser ablation, in particular to accelerate process optimisation. This project will therefore require the latest in machine learning algorithms, computational hardware, experimental automation, and data processing, in order to collect the required terabytes of training data required for such a model.

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/S515590/1 01/10/2018 30/09/2022
2115855 Studentship EP/S515590/1 27/09/2018 26/03/2022 Michael McDonnell