2019
DOI: 10.48550/arxiv.1905.09395
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

The Stabilized Explicit Variable-Load Solver with Machine Learning Acceleration for the Rapid Solution of Stiff Chemical Kinetics

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1

Citation Types

2
2
0

Year Published

2021
2021
2022
2022

Publication Types

Select...
1
1

Relationship

0
2

Authors

Journals

citations
Cited by 2 publications
(4 citation statements)
references
References 0 publications
2
2
0
Order By: Relevance
“…This made the DNN both faster to train and much more accurate, as well as easier to modify. This approach is in line with the work of other authors [41,61,[79][80][81].…”
Section: Deep Neural Network For 1d Sts Euler Shock Flow Relaxationsupporting
confidence: 89%
See 1 more Smart Citation
“…This made the DNN both faster to train and much more accurate, as well as easier to modify. This approach is in line with the work of other authors [41,61,[79][80][81].…”
Section: Deep Neural Network For 1d Sts Euler Shock Flow Relaxationsupporting
confidence: 89%
“…The machine learning solution time tends to diverge as soon as the tolerance is decreased. As already observed in [41,61], this behavior is connected to the nature of the solution methods for initial value problems (IVPs). Stiff chemistry solvers fall into this class, in which the usage of machine learning applied to the primary state variables is difficult but still suitable for secondary property prediction.…”
mentioning
confidence: 73%
“…This made the DNN both faster to train and much more accurate, as well as easier to modify. This approach is in line with the work of other authors [55,56,36,57,37].…”
Section: Deep Neural Network For 1d Sts Euler Shock Flow Relaxationsupporting
confidence: 89%
“…Even with relative prediction errors reaching as low as 10 − 5, the solver solution slowly diverges from a physically meaningful value. While it is true that evaluating the ML based predictions are exceptionally quick, the nature of the IVP does not allow for efficient correction of the ML predictions as observed also in [36,37].…”
Section: Matlab-python Interfacementioning
confidence: 99%