July 16, 2024Machine Learning Science and TechnologyOpen Access

Physics-Informed Neural Network for Turbulent Flow Reconstruction in Composite Porous-Fluid Systems

View Full Paper

Ask AI

Bookmark

Authors

SJSeohee Jang MJMohammad Jadidi SRSaleh Rezaeiravesh

Discussion

Member takes

Overview

Computational simulation demonstrates physics-informed neural networks reconstruct turbulent flow across porous-fluid systems, highlighting the value of interface training data for complex dynamics.

Key Points

Physics-informed neural networks accurately reconstruct complex turbulent flow within composite systems using Reynolds-averaged Navier-Stokes physics constraints.
Interface training data yields 18.04% and 19.94% errors for second-order statistics, improving prediction accuracy by 7% relative to unguided models.
Simulations demonstrate that incorporating turbulent second-order statistics and velocity gradient treatments improves flow prediction at porous boundaries.

Cite This Study

Jang et al. (2024) studied this question.

synapsesocial.com/papers/68e60139b6db643587595113 https://doi.org/10.1088/2632-2153/ad63f4

View Full Paper

Ask AI

Bookmark

July 16, 2024Machine Learning Science and TechnologyOpen Access

Physics-Informed Neural Network for Turbulent Flow Reconstruction in Composite Porous-Fluid Systems

View Full Paper

Ask AI

Bookmark

Authors

SJSeohee Jang MJMohammad Jadidi SRSaleh Rezaeiravesh

Discussion

Member takes

Overview

Computational simulation demonstrates physics-informed neural networks reconstruct turbulent flow across porous-fluid systems, highlighting the value of interface training data for complex dynamics.

Key Points

Physics-informed neural networks accurately reconstruct complex turbulent flow within composite systems using Reynolds-averaged Navier-Stokes physics constraints.
Interface training data yields 18.04% and 19.94% errors for second-order statistics, improving prediction accuracy by 7% relative to unguided models.
Simulations demonstrate that incorporating turbulent second-order statistics and velocity gradient treatments improves flow prediction at porous boundaries.

Cite This Study

Jang et al. (2024) studied this question.

synapsesocial.com/papers/68e60139b6db643587595113 https://doi.org/10.1088/2632-2153/ad63f4

View Full Paper

Ask AI

Bookmark