Deep learning architecture for medium laser acceleration

Abstract

The invention discloses a deep learning architecture for medium laser acceleration, which relates to the technical field of nuclear technology systems, and comprises the following steps: step 1, designing and determining configuration parameters including structure size, resolution, boundary conditions, objective functions and the like, initializing variables including a light source, a dielectric material and the like, and obtaining physical property parameters through previous experiments. Metamaterial structure-based medium laser acceleration utilizes an optical metamaterial to regulate and control electric field distribution of light to achieve the purpose of particle acceleration, the form is complex, how to find a structure required by an accelerator is the research key point of the project, and the problem can be solved along with development of artificial intelligence in the field of metamaterials in recent years. A wider material and design space is searched through AI, a metamaterial structure is designed by using deep learning, an optimal scheme is searched by using deep learning, large device design, any light source layout and any ion type are supported, and a limited optimal scheme is searched based on process constraints.

Description

technical field [0001] The invention relates to the field of nuclear technology, in particular to a deep learning framework for medium laser acceleration. Background technique [0002] The basic principle of dielectric laser acceleration is to use the near field generated by the laser in the periodic dielectric structure to accelerate charged particles. Existing solutions include grating structures, photonic crystals, etc. At present, the main method in the field of dielectric laser accelerators is through repeated Experiments to find a suitable structure are time-consuming and inefficient, and artificially designed structures are too simple for particle acceleration to achieve ideal modulation effects. Metamaterials can perform arbitrary operations on light waves / microwaves to achieve specific functions. Through design With a suitable metamaterial structure, we can achieve better modulation effects on DLA and meet the needs of accelerated particles. We propose a deep learni...

AI Technical Summary

Problems solved by technology

[0002] The basic principle of dielectric laser acceleration is to use the near field generated by the laser in the periodic dielectric structure to accelerate charged particles. Existing solutions include grating structures, photonic crystals, etc. At present, the main method in the field of dielectric laser accelerators is through repeated Experiments to find a suitable structure are time-consuming and inefficient, and artificially designed structures are too simple for particle acceleration to achieve ideal modulation effects. Metamaterials can perform arbitrary operations on light waves / microwaves to achieve specific functions. Through design With a suitable metamaterial structure, we can achieve a better modulation effect on DLA and meet the needs of accelerated particles. We propose a deep learning architecture to design optimal smart metamaterials for dielectric laser acceleration.

Images