Reconstructing velocity and pressure from sparse noisy particle tracks using Physics-Informed Neural Networks

Abstract

Volume-resolving imaging techniques are rapidly advancing progress in experimental fluid mechanics. However, reconstructing the full and structured Eulerian velocity and pressure fields from sparse and noisy particle tracks obtained experimentally remains a significant challenge. We introduce a new method for this reconstruction, based on Physics-Informed Neural Networks (PINNs). The method uses a Neural Network regularized by the Navier-Stokes equations to interpolate the velocity data and simultaneously determine the pressure field. We compare this approach to the state-of-the-art Constrained Cost Minimization method [1]. Using data from direct numerical simulations and various types of synthetically generated particle tracks, we show that PINNs are able to accurately reconstruct both velocity and pressure even in regions with low particle density and small accelerations. PINNs are also robust against increasing the distance between particles and the noise in the measurements, when studied under synthetic and experimental conditions.

References in corpus (7)

• A public turbulence database cluster and applications to study Lagrangian evolution of velocity increments in turbulence
• Flow over an espresso cup: Inferring 3D velocity and pressure fields from tomographic background oriented schlieren videos via physics-informed neural networks
• Evolutional Deep Neural Network
• Synchronization of turbulence in channel flow
• Assimilation of wall-pressure measurements in high-speed flow over a cone
• Broken mirror symmetry of tracer's trajectories in turbulence
• GotFlow3D: Recurrent Graph Optimal Transport for Learning 3D Flow Motion in Particle Tracking

Cited by in corpus (9)

• Multi-scale data reconstruction of turbulent rotating flows with Gappy POD, Extended POD and Generative Adversarial Networks
• Ensemble flow reconstruction in the atmospheric boundary layer from spatially limited measurements through latent diffusion models
• Filtered Partial Differential Equations: a robust surrogate constraint in physics-informed deep learning framework
• Periodically activated physics-informed neural networks for assimilation tasks for three-dimensional Rayleigh-Bénard convection
• Generative Adversarial Networks to infer velocity components in rotating turbulent flows
• Interpretable structural model error discovery from sparse assimilation increments using spectral bias-reduced neural networks: A quasi-geostrophic turbulence test case
• Inferring parameters and reconstruction of two-dimensional turbulent flows with physics-informed neural networks
• Neural Network Complexity of Chaos and Turbulence
• Reconstruction of three-dimensional turbulent flows from sparse and noisy planar measurements: A weight-sharing neural network approach