Automation and control of laser wakefield accelerators using Bayesian optimisation

TL;DR

This paper demonstrates the use of Bayesian optimisation and machine learning to automate and improve the control of laser wakefield accelerators, achieving significant enhancements in electron beam charge.

Contribution

It introduces a machine learning-based automation method for optimizing laser wakefield accelerators by tuning multiple parameters simultaneously, revealing improvements overlooked by traditional methods.

Findings

80% increase in electron beam charge through subtle laser pulse tuning

Optimization of laser evolution enabled by the model

Simultaneous variation of 6 parameters improves accelerator performance

Abstract

Laser wakefield accelerators promise to revolutionise many areas of accelerator science. However, one of the greatest challenges to their widespread adoption is the difficulty in control and optimisation of the accelerator outputs due to coupling between input parameters and the dynamic evolution of the accelerating structure. Here, we use machine learning techniques to automate a 100 MeV-scale accelerator, which optimised its outputs by simultaneously varying up to 6 parameters including the spectral and spatial phase of the laser and the plasma density and length. Most notably, the model built by the algorithm enabled optimisation of the laser evolution that might otherwise have been missed in single-variable scans. Subtle tuning of the laser pulse shape caused an 80% increase in electron beam charge, despite the pulse length changing by just 1%.