A Validated LBM Dataset and Pipeline for Surrogate Modeling of Turbulent 3D Obstructed Channel Flows
2026-06-15 • Machine Learning
Machine Learning
AI summaryⓘ
The authors created a reliable method to generate data for simulating 3D turbulent flows in channels, using a proven computational solver checked against real experiments. This method helps compare different neural networks that predict fluid behavior, focusing on tasks like forecasting and improving resolution. They plan to evaluate how well these networks capture important physical details and how efficient they are compared to traditional simulations. The authors invite feedback on their approach and future testing plans for these neural models.
neural operators3D turbulent flowlattice Boltzmann solvercumulant collision operatorsReynolds numberFourier Neural OperatorU-Netturbulent energy cascadegrid convergencesurrogate models
Authors
Lukas Schröder, Shubham Kavane, Harald Köstler
Abstract
Evaluating neural operators for 3D turbulent flow requires validated datasets with physical benchmarks. We present a reproducible pipeline generating training data for 3D channel flows around generated geometries at Re=1,000-10,000. Our lattice Boltzmann solver with cumulant collision operators is rigorously verified against experimental measurements (Strouhal number, drag coefficients, turbulent fluctuations) with comprehensive grid convergence studies at resolution 1024x512x512. Building upon an established framework, this validated pipeline enables standardized surrogate model comparison. We outline planned systematic evaluation of Fourier Neural Operator and U-Net variants on forecasting, super-resolution, and error correction tasks, using physics-informed metrics to assess turbulent energy cascade representation. Future work will compare computational efficiency between numerical solvers and neural surrogates, exploring practical application. We seek community feedback on our validation approach, planned benchmark methodology, and evaluation priorities for neural operators in turbulent flows.