Abstract:The precise simulation of turbulent flows is of immense importance in a variety of scientific and engineering fields, including climate science, freshwater science, and the development of energy-efficient manufacturing processes. Within the realm of turbulent flow simulation, direct numerical simulation (DNS) is widely considered to be the most reliable approach, but it is prohibitively expensive for long-term simulation at fine spatial scales. Given the pressing need for efficient simulation, there is an increasing interest in building machine learning models for turbulence, either by reconstructing DNS from alternative low-fidelity simulations or by predicting DNS based on the patterns learned from historical data. However, standard machine learning techniques remain limited in capturing complex spatio-temporal characteristics of turbulent flows, resulting in limited performance and generalizability. This paper presents a novel physics-enhanced neural operator (PENO) that incorporates physical knowledge of partial differential equations (PDEs) to accurately model flow dynamics. The model is further refined by a self-augmentation mechanism to reduce the accumulated error in long-term simulations. The proposed method is evaluated through its performance on two distinct sets of 3D turbulent flow data, showcasing the model's capability to reconstruct high-resolution DNS data, maintain the inherent physical properties of flow transport, and generate flow simulations across various resolutions. Additionally, experimental results on multiple 2D vorticity flow series, generated by different PDEs, highlight the transferability and generalizability of the proposed method. This confirms its applicability to a wide range of real-world scenarios in which extensive simulations are needed under diverse settings.

Learning nonlinear operators in latent spaces for real-time predictions of complex dynamics in physical systems

Learning in latent spaces improves the predictive accuracy of deep neural operators

Diffeomorphic Latent Neural Operators for Data-Efficient Learning of Solutions to Partial Differential Equations

Learning Semilinear Neural Operators : A Unified Recursive Framework For Prediction And Data Assimilation

Latent Neural Operator for Solving Forward and Inverse PDE Problems

Neural Operator: Learning Maps Between Function Spaces

Neural Operator Learning for Long-Time Integration in Dynamical Systems with Recurrent Neural Networks

Operator Learning: Algorithms and Analysis

Smooth and Sparse Latent Dynamics in Operator Learning with Jerk Regularization

Learning the intrinsic dynamics of spatio-temporal processes through Latent Dynamics Networks

Wavelet neural operator: a neural operator for parametric partial differential equations

Neural Operator: Is data all you need to model the world? An insight into the impact of Physics Informed Machine Learning

Multiscale Neural Operator: Learning Fast and Grid-independent PDE Solvers

Learning integral operators via neural integral equations

Neural Operators for Accelerating Scientific Simulations and Design

Wavelet Neural Operator for solving parametric partial differential equations in computational mechanics problems

Optimal deep learning of holomorphic operators between Banach spaces

Physics-Informed Deep Neural Operator Networks

Learning Transient Partial Differential Equations with Local Neural Operators

Exploring Physical Latent Spaces for High-Resolution Flow Restoration

Physics-enhanced Neural Operator for Simulating Turbulent Transport