Adaptive Level Grouping for Complexity-Reduced, Collisional-Radiative Models

Accurate time-dependent modeling of atomic kinetics may require many atomic levels and transitions that can exacerbate computational times. Complexity-reduction algorithms were therefore applied to ameliorate the compute times associated with solving the atomic levels' coupled rate equations. However, recent analysis of these reduction algorithms show that adaptive level-grouping strategies are needed to capture the atomic kinetics evolution with limited user intervention in the construction of atomic groups.¹ In this work, a complexity-reduced, collisional-radiative model was improved by incorporating clustering algorithms to automate the level-grouping process. The grouped levels were then utilized in the CR model through Boltzmann group descriptions developed by Le et al.² Preliminary results will show the ability of clustering algorithms to construct atomic groups that lend well to the use of the Boltzmann grouping technique.

¹ Abrantes et al. JQSRT 216, 47-55 (2018)
² Le et al. Phys. Plasmas 20, 1-19 (2013)

Distribution A: Approved for public release; unlimited distribution, PA (Public Affairs) Clearance Number 18366