Automation and control of laser wakefield accelerators using Bayesian optimisation

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%.

arXiv comment10 pages, 5 figures

References in corpus (6)

• SciPy 1.0--Fundamental Algorithms for Scientific Computing in Python
• Developments in laser-driven plasma accelerators
• A spectral, quasi-cylindrical and dispersion-free Particle-In-Cell algorithm
• Bayesian optimization of a free-electron laser
• Coherent control of plasma dynamics
• Observation of Laser Power Amplification in a Self-Injecting Laser Wakefield Accelerator

Cited by in corpus (28)

• Machine Learning in Nuclear Physics
• The data-driven future of high energy density physics
• 2022 Review of Data-Driven Plasma Science
• Spiderweb nanomechanical resonators via Bayesian optimization: inspired by nature and guided by machine learning
• Proton beam quality enhancement by spectral phase control of a PW-class laser system
• Physics-informed Gaussian Process for Online Optimization of Particle Accelerators
• Bayesian optimization of laser-plasma accelerators assisted by reduced physical models
• A laser-plasma platform for photon-photon physics
• Exploring Sensitivity of ICF Outputs to Design Parameters in Experiments Using Machine Learning
• Energy Compression and Stabilization of Laser-Plasma Accelerators
• Unsupervised Discovery of Inertial-Fusion Plasma Physics using Differentiable Kinetic Simulations and a Maximum Entropy Loss Function
• Colliding Pulse Injection of Polarized Electron Bunches in a Laser-Plasma Accelerator
• Beyond optimization -- supervised learning applications in relativistic laser-plasma experiments
• A multifidelity Bayesian optimization method for inertial confinement fusion design
• Online Bayesian Optimization for a Recoil Mass Separator
• Bayesian optimization of electron energy from laser wakefield accelerators
• Restoring betatron phase coherence in a beam-loaded laser-wakefield accelerator
• Machine learning phases of an Abelian gauge theory
• Multipoint-BAX: A New Approach for Efficiently Tuning Particle Accelerator Emittance via Virtual Objectives
• Optimization of transformer ratio and beam loading in a plasma wakefield accelerator with a structure-exploiting algorithm
• Visualization of high-intensity laser-matter interactions in virtual reality and web browser
• Characterization and automated optimization of laser-driven proton beams from converging liquid sheet jet targets
• Performance envelope of laser wakefield accelerators
• AI-Assisted Transport of Radioactive Ion Beams
• Adaptive Machine Learning for Time-Varying Systems: Low Dimensional Latent Space Tuning
• Optimized matching conditions for self-guided laser wakefield accelerators
• Neural-network-based longitudinal electric field prediction in nonlinear plasma wakefield accelerators
• Combined tools for Particle-In-Cell simulations performed with transversely asymmetric chirped lasers