luminary-shift/Battery-sample

SHIFT-Battery: High-Fidelity Computational Fluid Dynamics Dataset for Battery Pack Thermal Design and Analysis

We're excited to introduce the SHIFT-Battery dataset — a high-fidelity conjugate heat transfer simulation dataset developed as part of the Luminary SHIFT Models initiative. This dataset enables the training and benchmarking of real-time physics AI models for battery cooling design and thermal performance analysis.

Website: shift.luminarycloud.com

Contact: shift@luminarycloud.com

Summary

Physics AI models can transform early stage battery cooling design by giving users real-time feedback on the physics-based thermal and hydraulic performance implications of design decisions, with minimal setup cost and expertise required. This is particularly important for multiphysics design problems, where multiple teams across a company are stakeholders in the final design. However, the lack of high-quality training data has been a barrier to their development and adoption. Luminary SHIFT Models provide access to both high-quality datasets and pretrained models for a variety of applications and industries.

SHIFT-Battery is a massive step forward in this direction: purpose-built for high-fidelity battery cold-plate cooling performance inference, without requiring CFD expertise or meshing. This dataset is based on thousands of parametrically varied serpentine cold-plate cooling channel geometries operating at and outside design power conditions, allowing systematic exploration of the design space for battery module cold-plates and their integration with the modules via a Thermal Interface Material (TIM).

This dataset supports training surface-based or volume-based battery thermal surrogate models, real-time inference systems, and exploring geometry-performance correlations for battery cooling design optimization.

Applications

• Rapid battery cold-plate design prototyping and thermal performance optimization
• Research in conjugate heat transfer inference, point cloud learning, or physics-aware generative models for thermal management
• Training and fine-tuning Physics AI models for battery cooling design

Attribution

Please attribute Luminary Cloud for the SHIFT-Battery model and dataset.

An article is being prepared so users can cite this dataset - we will update this accordingly when available. Until then you can use this citation:

@misc{shift_battery_2026,
  author = "{Luminary Cloud}",
  title = {SHIFT-Battery: High-Fidelity Conjugate Heat Transfer Dataset for Battery Cold-Plate Cooling Channel Design and Analysis},
  year = {2026},
  url = {https://huggingface.co/datasets/luminary-shift/Battery/}
}

Contents

This repository contains the SHIFT-Battery dataset with 50 unique cold-plate cooling channel geometry and operating condition configurations. We will continue to push newly computed samples to this repository periodically, on our path towards comprehensive battery cooling design space coverage. The data generation and organization within the repository is described below.

This repository provides access to the first 50 simulations from the larger SHIFT-Battery dataset, hence the qualifier "Sample". To access the full dataset (~1,300 cases), please see luminary-shift/Battery or visit the SHIFT website.

Geometry Creation

The geometry creation process transforms design parameters into a complete battery cooling assembly consisting of a serpentine cold-plate cooling channel and a battery module plate that contacts the cells through a Thermal Interface Material. Channel centerlines and cross-sections are computed from the input parameters (number of serpentine returns, number of parallel paths per pass, parallel path gap, starting radius, filling ratio, and profile exponent), generating curve data files for the channel path, parallel branches, and channel cross-section profile. Three-dimensional channel geometries are created by sweeping the cross-section profile along the serpentine path, generating the fluid channel volume, and extracting the wetted surface. The module plate creation process loads a parametric template, applies plate geometric parameters and TIM thickness, and generates the solid plate geometry. The plate geometry is then imported into a cloud CAD system where fillets are applied to specified edges before being exported. The final assembly stage imports the cold-plate as base geometry, patterns the cooling channel through the plate, subtracts the channel to create the coolant flow passage, and imports the battery module.

Figure: Geometry parametrization. Each cold-plate is defined by six design parameters: number of serpentine returns, number of parallel paths, parallel-path gap, starting radius, filling ratio, and profile exponent.

Figure: Sample cold-plate cooling channels generated by sweeping the parametric design space.

CFD Solver

The simulations are run on a fixed mesh with boundary layers applied to the cold-plate cooling channel walls. The simulation setup configures a conjugate heat transfer simulation with a water-glycol mixture as the working coolant. The fluid channel domain transports coolant while the solid plate domain conducts heat from the battery cells through the TIM, connected through a conjugate interface. Boundary conditions include a mass flow rate inlet with prescribed inlet temperature, a pressure outlet, conjugate fluid-solid interfaces between the coolant and the plate, and a volumetric heat source within the module representing the cell power dissipation.

Files

At the root level of the repository you will find directories containing the simulation data organized using unique sample IDs. Each directory corresponds to a unique cold-plate geometry configuration or operating condition. The directory structure looks like:

shift_battery_0025e284
shift_battery_00270a15
shift_battery_0070e63e
...

The full SHIFT-Battery dataset contains over 1,300 unique geometry-and-operating-condition configurations (i.e. a single geometry simulated at two different operating conditions counts as two configurations). In each directory you will find the following files:

• fluid_channel.stl: STL file of the serpentine cooling channel geometry (used as the geometry-token input for SHIFT-Battery surrogate models)
• merged_surfaces.stl: STL file with the combined battery cooling assembly geometry (cooling channel and module cold-plate)
• merged_surfaces.vtp: surface field solution file with temperature, heat flux, and wall pressure on the combined cooling assembly from the CFD simulation
• volume_fluid.vtu: volume field solution file with pressure, velocity, and temperature fields in the coolant
• params.json: JSON file containing the parametric design values including N_returns, N_parallelpath, parallel_path_gap, R_start, filling_ratio, n_profile, power_per_module, inlet_mass_flow_rate, and TIM_thickness
• metadata.json: JSON file containing simulation metadata including simulation_id, simulation_name, mesh_id, and project_id
• qoi.json: JSON file containing computed quantities of interest including pressure_drop_Pa (coolant pressure drop across the channel) and max_temp_K (peak module temperature)

Dataset Statistics

The distributions of the two main quantities of interest across the dataset are shown below.

Figure: Distribution of coolant pressure drop across the channel (pressure_drop_Pa) across the dataset.

Figure: Distribution of peak module temperature (max_temp_K) across the dataset.

Figure: Joint distribution of pressure drop and peak temperature, illustrating the design trade-off the surrogate must learn.

Downloading

You can use HuggingFace to gain access to the entire repository, but will require the associated storage available locally. Note you will need to have git lfs installed first, then run

git clone git@hf.co:datasets/luminary-shift/Battery

to ensure future git pull commands won't checkout full files, you'll want to ensure the skip is active in this repo

cd <path/to/repo>
git lfs install --skip-smudge --local

You can then pull down files you want to interact with in multiple ways:

• pull a specific file

git lfs pull --include="path/to/your/file"

• pull a directory

git lfs pull --include="path/to/file1,path/to/dir/*"

• pull, but exclude certain paths

git lfs pull --exclude="**/*.mp4"

• and remove those files and reset them to pointers when done using them:

rm path/to/your/file
git checkout -- path/to/your/file

Example Surrogate Predictions

Figure: Physics AI predictions.

Credits

The SHIFT-Battery dataset was developed by Luminary Cloud as part of the SHIFT Models initiative, providing comprehensive battery cooling design data for physics AI model training and battery thermal management optimization. Please attribute Luminary Cloud for the SHIFT-Battery model and dataset.

License

This dataset is distributed under the CC-BY-NC-4.0 license, which is also included in the dataset itself. By downloading the dataset you acknowledge the terms of this license.