Commercial fusion energy plants and advanced compact radioactive sources may rely on high-intensity, high-repetition-rate lasers that can be fired multiple times per second, but it is difficult for humans to react to changes in these shot rates. This can be a limiting factor when doing so.
Applying advanced computing to this problem, an international team of scientists from Lawrence Livermore National Laboratory (LLNL), Fraunhofer Institute for Laser Technology (ILT), and Extreme Optical Infrastructure (ELI ERIC) collaborated to We conducted experiments to optimize the intensity and high-intensity light energy. – Repetition rate laser using machine learning.
“Our goal was to demonstrate robust diagnostics of laser-accelerated ions and electrons from solid-state targets at high intensities and repetition rates,” said Matthew Hill, principal investigator at LLNL. . “Supported by rapid feedback from the machine learning optimization algorithm to the laser front end, we were able to maximize the total ion yield of the system.”
The researchers trained a closed-loop machine learning code developed by LLNL's cognitive simulation team on the laser-target interaction data to optimize the laser pulse shape and allow it to be adjusted while the experiment was running. The data generated during the experiment was fed back to a machine learning-based optimizer, allowing it to fine-tune the pulse shape on the fly.
The laser was fired every 5 seconds and the laser intensity was always greater than 3×10.twenty one W/cm² at focus, I had to stop after 120 shots and replace the copper target foil. During this time, researchers also inspected diagnostic equipment for damage and assessed debris accumulation from vaporized targets. The team conducted experiments at ELI for three weeks, with experiment runs lasting about 12 hours per day, during which up to 500 laser shots were fired.
The experiment took place at the ELI Beamline facility in the Czech Republic, where researchers utilized the state-of-the-art High Repetition Rate Advanced Petawatt Laser System (L3-HAPLS) to generate protons within ELIMAIA laser plasma ions. accelerator. Focused on the goal of applying machine learning to high-speed laser experiments, the team simplified aspects of the experiment as much as possible, including using a robust and simple copper foil target.
“By leveraging HAPLS and pioneering machine learning techniques, we have embarked on an exciting journey to further understand the complex physics of laser-plasma interactions,” said Fraunhofer ILT Managing Director. said one Konstantin Hefner.
More than 4,000 shots were fired during the campaign, allowing statistical analysis to be performed on the results, demonstrating optimization of ion yield over the already impressive nominal baseline performance.
For experimental physicists, using machine learning was a new experience. “It becomes a spectator sport,” Hill said. “We looked at the data coming in and tried to guess what the optimizer would do. This is very different from experimenting with manual intervention.”
LLNL becomes L3-HAPLS user
The L3-HAPLS laser has excellent laser performance repeatability with very stable alignment, focal spot quality, and high repetition rates that facilitate the generation of secondary sources such as electrons, ions, and X-rays. It has the ability to generate powerful laser pulses.
“The success of such a complex experiment demonstrates the state-of-the-art quality and reliability of the L3-HAPLS laser system,” said Bedrich Rus, principal laser scientist at ELI Beamlines.
LLNL developed the HAPLS laser as part of a bilateral agreement with ELI Beamlines, and the first light was emitted from the system after it was delivered and installed in the Czech Republic in 2017. This was only his second user experiment at the facility, which was granted time through a competitive competition. Currently, it is published twice a year and receives hundreds of applications.
long preparation pays off
In addition to Hill, from Elizabeth Grace, Franziska Treffert, James McLaughlin, Isabella Pagano, Avik Sarkar, Raspberry Simpson, Blagoje Djordjevic, Matthew Selwood, Derek Mariscal, Jackson Williams, and Tammy Ma. The LLNL team spent about a year preparing for the experiment. Fraunhofer ILT and ELI beamline teams. In addition to diagnosing regional facilities, the Livermore team is responsible for laboratory-directed research and development programs, including the REPPS magnetic spectrometer, the PROBIES ion beam imaging spectrometer, the repeatedly evaluated scintillator imaging system, and the repeatedly evaluated X-ray spectrometer. We utilized some of the equipment developed below. .
/Open to the public. This material from the original organization/author may be of a contemporary nature and has been edited for clarity, style, and length. Mirage.News does not take any institutional position or position, and all views, positions, and conclusions expressed herein are those of the authors alone. Read the full text here.
