English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In this study, a fast and stable machine-learned hybrid algorithm implementedin TensorFlow for the integration of stiff chemical kinetics is introduced.Numerical solutions to differential equations are at the core of computationalfluid dynamics calculations. As the size and complexity of the simulationsgrow, so does the need for computational power and time. Many efforts have beenmade to implement stiff chemistry solvers on GPUs but have not been highlysuccessful because of the logical divergence in traditional stiff solveralgorithms. Because of these constrains, a novel Explicit StabilizedVariable-load (STEV) solver has been developed. Overstepping due to therelatively large time steps is prevented by introducing limits to the maximumchanges of chemical species per time step. To prevent oscillations, a discreteFourier transform is introduced to dampen ringing. In contrast to conventionalexplicit approaches, a variable-load approach is used where each cell in thecomputational domain is advanced with its unique time step. This approachallows cells to be integrated simultaneously while maintaining warp convergencebut finish at different iterations and be removed from the workload. To improvethe computational performance of the introduced solver, specific thermodynamicquantities of interest were estimated using shallow neural networks in place ofpolynomial fits, leading to an additional 10% savings in clock time withminimal training and implementation requirements. However ML specific hardwarecould increase the time savings to as much as 28%. While the complexity ofthese particular machine learning models is not high by modern standards, theimpact on computational efficiency should not be ignored. The results show adramatic decrease in total chemistry solution time (over 200 times) whilemaintaining a similar degree of accuracy.

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