BubbleSH: A Dataset of Rising Bubbles with Deformable Interfaces

2026-07-08 · Source: Machine Learning · Field: Science & Research — Engineering & Applied Sciences, Artificial Intelligence & Machine Learning · Depth: Expert, quick

Summary

BubbleSH is a new dataset designed for modeling complex bubbly flows, featuring transient, three-dimensional bubble-swarm dynamics. Generated from high-fidelity direct numerical simulations of rising bubbles in a periodic domain, it provides time-resolved data on bubble trajectories, velocities, and shape evolution. A key innovation is the compact representation of bubble morphology using spherical harmonics, making the dataset lightweight yet physically expressive. BubbleSH aims to facilitate data-driven modeling for bubbly flow simulators, particularly those where bubble deformation and inter-bubble interactions are central. The dataset is characterized by bubble kinematics, morphology, and interaction patterns, and includes specific evaluation metrics for trajectory and shape prediction. Its sensitivity to local perturbations makes it ideal for generative models, and it serves as a benchmark for developing data-driven models of deformable, chaotic multiphase systems.

Key takeaway

For research scientists developing data-driven models for multiphase systems, BubbleSH provides a critical benchmark. You can use this compact, high-fidelity dataset to train and evaluate generative models, especially where bubble deformation and interactions are key. Utilize the provided metrics for trajectory and shape prediction to rigorously assess your model's performance. This dataset enables you to advance simulations of chaotic bubbly flows.

Key insights

BubbleSH offers a high-fidelity, compact dataset for data-driven modeling of complex, deformable bubbly flows using spherical harmonics.

Principles

• Spherical harmonics compactly represent bubble morphology.
• Bubble-swarm dynamics are sensitive to local perturbations.
• High-fidelity simulations enable robust data-driven models.

In practice

• Develop generative models for chaotic multiphase systems.
• Benchmark data-driven models for bubble trajectory prediction.
• Simulate bubbly flows with deformable interfaces.

Topics

• Bubbly Flows
• Multiphase Systems
• Direct Numerical Simulations
• Spherical Harmonics
• Generative Models
• Dataset Benchmark

Best for: AI Scientist, Research Scientist

Related on AIssential

• [D] Solving the "Liquid-Solid Interface" Problem: 116 High-Fidelity Datasets of Coastal Physics (Waves, Saturated Sand, Light Transport) — A new collection of 116 high-fidelity datasets has been created from the Arabian Sea to address the challenges modern generative models like Sora, Runway, and Kling face with complex shoreline physics.
• PyDMD: A Python Package for Dynamic Mode Decomposition (DMD) — PyDMD enables data-driven modeling of complex systems by decomposing time-varying data into dominant spatio-temporal modes.
• Learning Interface Breakup: A Geometry-Conditioned Latent Surrogate for Spray Formation — AMR refinement structure can serve as a compact, learnable representation for geometry-conditioned surrogate modeling of transient two-phase flows.
• Multipath Adaptive Gated Bottleneck Latent ODE with Raman Data Fusion for Cell Culture Process Forecasting — This framework combines ODEs, just-in-time fine-tuning, and data fusion for robust, multi-path bioprocess forecasting.
• Function-Space Decoupled Diffusion for Forward and Inverse Modeling in Carbon Capture and Storage — Fun-DDPS combines diffusion models and neural operators for robust, efficient, and physics-consistent CCS modeling.
• Learning universal approximations for partial differential equations with Physics-Informed Broad Learning System — PIBLS offers a backpropagation-free, least-squares optimization approach for PDEs, achieving faster, more accurate solutions than PINNs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning.