An uncertainty-aware strategy for plasma mechanism reduction with directed weighted graphs
- Sandia National Lab. (SNL-CA), Livermore, CA (United States)
- Donghua University, Shanghai (China)
- Princeton Plasma Physics Lab. (PPPL), Princeton, NJ (United States)
In this work, we present a framework for the analysis and reduction of plasma mechanisms by means of weighted directed graphs, in which reactions and species are both treated as nodes. The methodology consists of two distinct analyses. The first, which is qualitative, relies on graph spatializations via force-directed algorithms to discover the predominant global patterns in the chemical model. The second ranks the reactions based on their shortest paths' lengths from/to the species of interest and their relative contributions to the power balance. Further, this quantitative investigation enables a strategy for mechanism reduction that is fully automatized, as it does not require any expert knowledge, highly effective, as it generates reduced mechanisms that are highly accurate while relying on a small number of processes, and easily interpretable, as the algorithm justifies the importance of the retained reactions by outputting their related chemical pathways. Additionally, the work proposes a methodology extension that employs ensembles of graphs to improve the robustness of the reduced mechanism to reaction parameter uncertainties. The approach, here tested for steady-state predictions of a plasma system characterizing negative hydrogen ion sources, is general and can be used in a wide variety of applications outside the particular nuclear fusion context demonstrated in this work.
- Research Organization:
- Princeton Plasma Physics Laboratory (PPPL), Princeton, NJ (United States); Sandia National Lab. (SNL-CA), Livermore, CA (United States)
- Sponsoring Organization:
- USDOE Office of Science (SC), Fusion Energy Sciences (FES); USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR). Scientific Discovery through Advanced Computing (SciDAC); National Natural Science Foundation of China (NSFC)
- Grant/Contract Number:
- NA0003525; AC02-09CH11466; 12105041
- OSTI ID:
- 1974432
- Journal Information:
- Physics of Plasmas, Vol. 30, Issue 4; ISSN 1070-664X
- Publisher:
- American Institute of Physics (AIP)Copyright Statement
- Country of Publication:
- United States
- Language:
- English
Similar Records
A graph-based approach to developing adaptive representations of complex reaction mechanisms
On the application of betweenness centrality in chemical network analysis: Computational diagnostics and model reduction