English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Radio-Frequency Quadrupoles (RFQs) are multi-purpose linear particleaccelerators that simultaneously bunch and accelerate charged particle beams.They are ubiquitous in accelerator physics, especially as injectors tohigher-energy machines, owing to their impressive efficiency. The design andoptimization of these devices can be lengthy due to the need to repeatedlyperform high-fidelity simulations. Several recent papers have demonstrated thatmachine learning can be used to build surrogate models (fast-executingreplacements of computationally costly beam simulations) for order-of-magnitudecomputing time speedups. However, while these pilot studies are encouraging,there is room to improve their predictive accuracy. Particularly, beam summarystatistics such as emittances (an important figure of merit in particleaccelerator physics) have historically been challenging to predict. For thefirst time, we present a surrogate model trained on 200,000 samples that yields<6% mean average percent error for the predictions of all relevant beam outputparameters, solving the problem of poor emittance predictions by identifyingand including hidden variables which were not accounted for previously. Thesesurrogate models were made possible by using the Julia language and GPUcomputing; we briefly discuss both. We demonstrate the utility of surrogatemodeling by performing a multi-objective optimization using our best model as acallback in the objective function to select an optimal RFQ design. We considertrade-offs in RFQ performance for various choices of Pareto-optimal designvariables: common issues for any multi-objective optimization scheme. Lastly,we make recommendations for input data preparation, selection, and neuralnetwork architectures that pave the way for future development ofproduction-capable surrogate models for RFQs and other particle accelerators.

Neural Networks as Effective Surrogate Models of Radio-Frequency Quadrupole Particle Accelerator Simulations