DynaSolver predicts fluid flow around moving robot parts accurately
DynaSolver: A Neural Operator for Autoregressive 4D Flow Prediction with Dynamic Boundaries
Machine Learning
Summary
Predicting how fluids like air or water flow around moving shapes is hard and takes a lot of computing power. The authors created a new tool called DynaSolver that can quickly and accurately forecast these flows even when the shape is changing over time, like robot parts moving. They also made a new test set called RoboCFD-4D to check how well such predictions work with different robot shapes and conditions. DynaSolver learns from past flow states and moving boundaries to improve its predictions over time, reducing errors as it forecasts further ahead.
What this means in practice
- •For robotics engineers: Simulate how fluids interact with moving robot parts to improve robot designs and control without slow physical testing.
- •For aerospace design teams: Produce fast and accurate airflow predictions around deforming aircraft components during maneuvers for better aerodynamic assessment.
Authors
Rui Qian, Shibo Dai, Mengjun Lin, Yuk-Kit Hou, Saining Zhang, Bohan Li, Junhao Chen, Hao Zhao, Chaojian Li
Abstract
Neural operators provide efficient surrogates for computational fluid dynamics (CFD), but most are designed for static geometries and are not equipped to model unsteady flows induced by continuously moving boundaries. We study dynamic-boundary 4D CFD prediction for articulated moving geometries and introduce RoboCFD-4D, a temporally resolved robotic CFD benchmark spanning diverse morphologies and flow conditions. To address the computational and autoregressive challenges of this setting, we propose DynaSolver, a neural operator that explicitly conditions on boundary evolution and causal flow history. DynaSolver employs an efficient physical-token formulation that amortizes point-to-token projection across the operator stack, substantially reducing repeated point-space spatio-temporal computation while retaining accurate point-level reconstruction. We further introduce Replay-Augmented Parallel Training, which exposes the model to autoregressively generated histories while preserving parallel dense supervision, narrowing the gap between ground-truth-conditioned training and autoregressive rollout. Experiments on RoboCFD-4D show that DynaSolver achieves the best full-rollout accuracy among a broad range of learned CFD surrogates, with its advantage increasing as rollout proceeds, indicating substantially reduced autoregressive error accumulation. Project page: https://made-gpt.github.io/DynaSolver-ProjectPage.