Machine learning-powered simulations are giving researchers a new window into the processes that create some of the heaviest elements in the universe.
Where does the gold in jewelry, the uranium in nuclear fuel, and many of the heaviest elements in the universe come from? Scientists believe these materials were created during some of the most violent events in the universe, but simulating these processes in detail remains a major computational challenge.
Now, GSI/FAIR researchers and their international collaborators have developed a machine learning-based model. This model provides deeper insight into how elements form during extreme events such as global warming. neutron star merger. For the first time, the research team incorporated deep learning neural networks into fluid dynamics simulations to model the energy released during r-process nucleosynthesis. Their discovery is Physical Review D.
Many chemical elements are produced in powerful astrophysical phenomena such as supernova explosions and neutron star mergers. These environments generate enormous amounts of energy and free neutrons, enabling rapid neutron capture processes, or r-processes, that produce many of the elements heavier than iron. During this process, the atomic nucleus rapidly absorbs neutrons, which then change into protons, building up increasingly heavier elements.
“Researchers around the world are trying to understand these complex reactions through theoretical simulations. However, modeling all the parameters requires incredible computational power, so models often need to be simplified,” said Dr. Oliver Just, lead author of this publication and a researcher in GSI/FAIR’s Nuclear Astrophysics and Structures department. “Our new model RHINE is artificial intelligenceprovides an efficient alternative. ”
RHINE uses deep learning for r process heating
RHINE (Implementation of r-process heating in fluid dynamics simulation using neural networks) is applied. machine learningspecifically a deep learning neural network, represents the energy released by a nuclear reaction during the r process within a fluid dynamics simulation. This release of energy, known as heating, can have a significant effect on the movement and velocity distribution of the material released during the explosion. It can also affect the electromagnetic signals produced by these phenomena, such as kilonovas observed after neutron star mergers.
“First, the ML model is trained using a large number of reference calculations generated on the complete set of nuclear reactions. The model is then employed to perform hydrodynamic simulations to approximate the heating rate during the reaction.” rwith minimal effort,” explained Dr. Zewei Xiong, a scientist in the Astronuclear Physics and Structures Department at GSI/FAIR, who played a central role in designing the machine learning model.
“A detailed comparison validated the ML scheme against reference data. The high degree of agreement suggests that the use of ML models can save a huge amount of computational time. We also inferred the following from our results. r-Process heating is an important effect and should be better considered in future modeling. ”
The researchers say RHINE will enable more detailed simulations in the future, which could help link experimental results from the upcoming FAIR facility with astronomical observations of star explosions and neutron star mergers.
References: “Implementation of r-process heating in fluid dynamics simulations using neural networks,” by Oliver Just, Zewei Xiong, and Gabriel Martínez-Pinedo, April 16, 2026. Physical Review D.
DOI: 10.1103/gl2l-7f3g
RHINE source code is publicly available and available for use. Among other things, this project was co-funded by the European Research Council (ERC).
Never miss a breakthrough: join the SciTechDaily newsletter.
Follow us on Google and Google News.
