ANS is committed to advancing, fostering, and promoting the development and application of nuclear sciences and technologies to benefit society.
Explore the many uses for nuclear science and its impact on energy, the environment, healthcare, food, and more.
Explore membership for yourself or for your organization.
Conference Spotlight
2026 ANS Annual Conference
May 31–June 3, 2026
Denver, CO|Sheraton Denver
Latest Magazine Issues
Mar 2026
Jan 2026
Latest Journal Issues
Nuclear Science and Engineering
April 2026
Nuclear Technology
February 2026
Fusion Science and Technology
Latest News
Two new partnerships forged in AI and nuclear sectors
The nuclear space is full of companies eager to power new AI development. At the same time, many AI companies want to provide services to the nuclear industry. It should come as no surprise, then, that two new partnerships have recently been announced that further bridge the AI and nuclear sectors.
AtkinsRéalis has announced a partnership with Nvidia that aims to leverage Nvidia’s technologies to deploy “nuclear-powered, large-scale AI factories.” Centrus Energy has announced a partnership with Palantir Technologies to use Palantir’s software in support of Centrus’s plans to expand enrichment capacity.
R. Le Tellier, D. Fournier, J. M. Ruggieri
Nuclear Science and Engineering | Volume 163 | Number 1 | September 2009 | Pages 34-55
Technical Paper | doi.org/10.13182/NSE163-34
Articles are hosted by Taylor and Francis Online.
This paper describes a new approach for treating the energy variable of the neutron transport equation in the resolved resonance energy range. The aim is to avoid recourse to a case-specific spatially dependent self-shielding calculation when considering a broad group structure. This method consists of a discontinuous Galerkin discretization of the energy using wavelet-based elements. A t-orthogonalization of the element basis is presented in order to make the approach tractable for spatially dependent problems.First numerical tests of this method are carried out in a limited framework under the Livolant-Jeanpierre hypotheses in an infinite homogeneous medium. They are mainly focused on the way to construct the wavelet-based element basis. Indeed, the prior selection of these wavelet functions by a thresholding strategy applied to the discrete wavelet transform of a given quantity is a key issue for the convergence rate of the method. The Canuto thresholding approach applied to an approximate flux is found to yield a nearly optimal convergence in many cases. In these tests, the capability of such a finite element discretization to represent the flux depression in a resonant region is demonstrated; a relative accuracy of 10-3 on the flux (in L2-norm) is reached with less than 100 wavelet coefficients per group.