Advances in Scientific Machine Learning for Coupled Fluid Flow and Transport

2026-06-17 · Source: Takara TLDR - Daily AI Papers · Field: Science & Research — Artificial Intelligence & Machine Learning, Mathematics & Computational Sciences, Engineering & Applied Sciences · Depth: Expert, medium

Summary

The chapter "Advances in Scientific Machine Learning for Coupled Fluid Flow and Transport" reviews recent progress in Scientific Machine Learning (SciML) for modeling complex fluid flow and transport phenomena, specifically those governed by incompressible Navier-Stokes and scalar transport equations. These systems, exemplified by turbidity currents and thermal convection, are characterized by strong nonlinear coupling and multiscale behavior, making high-fidelity simulations computationally expensive. The review covers efficient surrogate modeling techniques, including linear reduced-order methods like Singular Value Decomposition and Dynamic Mode Decomposition, alongside nonlinear neural network approaches such as Physics-Informed Neural Networks (PINNs) and $\beta$-Variational Autoencoders ($\beta$-VAEs). It integrates these models with High Performance Computing strategies, including Adaptive Mesh Refinement/Coarsening (AMR/C) and scientific floating-point data compression. The authors also introduce new contributions: PINN-based surrogate modeling for turbidity currents and $\beta$-VAE-based extraction of disentangled nonlinear modes from thermal flows, demonstrating SciML's capacity for fast, accurate, and cost-effective approximations.

Key takeaway

For research scientists or ML engineers developing models for coupled fluid flow and transport, you should explore Scientific Machine Learning (SciML) techniques to significantly reduce computational expense. Consider integrating Physics-Informed Neural Networks (PINNs) or $\beta$-Variational Autoencoders ($\beta$-VAEs) with High Performance Computing strategies like Adaptive Mesh Refinement/Coarsening (AMR/C). This approach enables faster, accurate approximations for systems like turbidity currents, allowing you to accelerate research and development in complex fluid dynamics.

Key insights

SciML, combining reduced-order models and neural networks with HPC, efficiently approximates complex coupled fluid flow and transport, significantly cutting computational costs.

Principles

• Strong nonlinear coupling increases simulation cost.
• SciML reduces computational cost for complex systems.
• Real-time prediction in SciML is problem-dependent.

Method

Combine linear reduced-order models (e.g., DMD) and nonlinear neural networks (e.g., PINNs, $\beta$-VAEs) with HPC strategies like AMR/C and data compression to build efficient surrogate models for coupled fluid flow.

In practice

• Model turbidity currents using PINNs.
• Extract nonlinear modes from thermal flows via $\beta$-VAEs.
• Apply AMR/C for high-performance fluid simulations.

Topics

• Scientific Machine Learning
• Fluid Dynamics Modeling
• Physics-Informed Neural Networks
• Variational Autoencoders
• High Performance Computing
• Reduced-Order Models

Code references

• divelab/AIRS
• marianlupascu/OT-RF

Best for: AI Scientist, Research Scientist, Machine Learning Engineer

Related on AIssential

• Advances in Scientific Machine Learning for Coupled Fluid Flow and Transport — SciML offers efficient surrogate models for complex, computationally expensive coupled fluid flow and transport systems.
• Differential Equations V — Physics Informed Neural Networks (PINNs) — Physics Informed Neural Networks embed physical laws into neural networks for fast, accurate, and physically consistent simulations.
• Engineering Hybrid Physics-Informed Neural Networks for Next-Generation Electricity Systems: A State-of-the-Art Review — Physics-informed machine learning integrates domain physics into models, enhancing accuracy and efficiency for data-scarce electricity systems.
• FLAME 2023: Diving into the Future of Fluid Dynamics and Machine Learning — AI techniques are transforming simulation by integrating physics-based modeling with data-driven machine learning.
• Physics-informed convolutional neural networks for fluid flow through porous media — Physics-informed CNNs accurately predict fluid flow in porous media, significantly accelerating traditional LBM simulations.
• Droplet-LNO: Physics-Informed Laplace Neural Operators for Accurate Prediction of Droplet Spreading Dynamics on Complex Surfaces — PI-LNO uses Laplace transforms and physics-informed neural networks to model droplet spreading dynamics efficiently.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Takara TLDR - Daily AI Papers.