Research Software
Open-source codes and apps from Dr. Han Hu's NED3 Lab at the University of Arkansas, for running boiling and flow experiments, processing sensor and image data, tracking bubbles with machine learning, building fast CFD surrogates, and modeling data-center energy use.
Search and filter software
Type a tool name or word, or pick a category, topic, method, what it comes with, or the lead developer.
Thrust I
Community Hub and Resources
Thermal-AI Commons
Lead developer ·
Collaborative infrastructure for interoperable, evidence-aware thermal-AI research through shared interfaces and evaluation practices. It does not represent universally validated models or public benchmarks.
Data-Center Water and Energy Atlas
Lead developer ·
Interactive research atlas for examining data-center water and energy exposure across locations and scenarios. A decision-support resource, not measured operational data.
Data-Center Cooling Research Tools
Lead developer ·
Curated discovery hub cataloging data-center cooling research tools across thermal-fluid mechanisms, chip/rack cooling, facilities, AI/control, metrics and sustainability. Entries are curation aids, not validation endorsements.
Thrust II
Experimental Protocols and Data Processing
BoilingLab
Lead developer ·
Open protocols and synchronized multimodal workflow materials for pool-boiling research, supporting reproducible acquisition and analysis.
FlowLab
Lead developer ·
Research toolkit for single- and two-phase liquid-cooling experiments: flow-loop protocols, heater-power, acoustic-emission and accelerometer measurements, example files and analysis notebooks.
AELab
Lead developer ·
Collaborative acoustic-emission research workspace for pool boiling, flow boiling, partial-discharge and sensing-system workflows, with reproducible tutorials and analysis guidance.
IRISApp
Lead developer ·
Desktop analysis and data-collection app for the InfraRed Imaging Station: organizes tube-fin experiments, imports FLIR thermal fields and Arduino logs, compares experiments, and fits thermal profiles and effective in-plane conductivity.
Thrust III
Machine Learning and Computer Vision
BubbleID
Lead developer ·
Computer-vision package for static and dynamic bubble analysis: bubble size, count, departure frequency, liquid-vapor interface profiles and velocity profiles.
How to cite
Dunlap et al., "BubbleID: A Deep Learning Framework for Bubble Interface Dynamics Analysis," Journal of Applied Physics 136, 014902 (2024).
SeqReg
Lead developer ·
Machine-learning libraries for multi-dimensional sequential data, from 0+1D time series to 2+1D videos.
How to cite
Dunlap et al., "Hit2flux: A Machine Learning Framework for Boiling Heat Flux Prediction Using Hit-Based Acoustic Emission Sensing," AI Therm. Fluids 1, 100002 (2025).
CFDTwin
Lead developer ·
Wizard-based desktop app for building neural-network surrogate models from ANSYS Fluent simulations: load .cas files, sample the parameter space, run batch simulations, train POD+NN models and validate them.
How to cite
Curl & Hu, "Physically interpretable surrogate modeling of thermal fields in electronics cooling using combined proper orthogonal decomposition and neural networks," AI Thermal Fluids (2026).
BubbleID-Flow
Lead developer ·
Flow-boiling image-segmentation workspace: flow-channel ROI preprocessing, high-speed image-sequence support, COCO annotation utilities, Mask R-CNN fine-tuning configurations and bubble-mask validation scripts.
Thrust IV
Energy Modeling
OpenDC-LCA
Lead developer ·
Physics-informed life-cycle screening workflow for data-center cooling. Outputs are screening evaluations, not ISO-conformant assessments or measured facility impacts.
These tools are shared by the NED3 Lab for research use, as is and without warranty. If they help your work, please cite the related papers. More details are on the NED3 Lab software page.