Learning to correct spectral methods for simulating turbulent flows
This work addresses the challenge of simulating turbulent flows more accurately for applications in science and engineering, though it is incremental as it builds on existing hybrid approaches.
The authors tackled the problem of improving numerical simulation of partial differential equations (PDEs) in fluid dynamics by combining spectral methods with machine learning, resulting in models that are 2-4 times more accurate than standard spectral solvers at the same resolution but with about twice the runtime.
Despite their ubiquity throughout science and engineering, only a handful of partial differential equations (PDEs) have analytical, or closed-form solutions. This motivates a vast amount of classical work on numerical simulation of PDEs and more recently, a whirlwind of research into data-driven techniques leveraging machine learning (ML). A recent line of work indicates that a hybrid of classical numerical techniques and machine learning can offer significant improvements over either approach alone. In this work, we show that the choice of the numerical scheme is crucial when incorporating physics-based priors. We build upon Fourier-based spectral methods, which are known to be more efficient than other numerical schemes for simulating PDEs with smooth and periodic solutions. Specifically, we develop ML-augmented spectral solvers for three common PDEs of fluid dynamics. Our models are more accurate (2-4x) than standard spectral solvers at the same resolution but have longer overall runtimes (~2x), due to the additional runtime cost of the neural network component. We also demonstrate a handful of key design principles for combining machine learning and numerical methods for solving PDEs.