A community-driven collection of open benchmarks, analysis tools, and validation infrastructure for computational simulations in engineering and geosciences.

The Simulation-Benchmarks organization develops and maintains standardized benchmark cases, evaluation scripts, and shared platforms that enable rigorous comparison of simulation codes across research groups. Projects span subsurface flow, structural mechanics, and fluid dynamics, with a focus on reproducibility and open science.

Shared infrastructure for model validation within the NFDI4ING (National Research Data Infrastructure for Engineering Sciences) consortium. Provides a registry of benchmark cases, example implementations, and documentation for the engineering simulation community.

• Standardized benchmark definition and contribution workflow
• Published documentation via ReadTheDocs / MkDocs
• CI-integrated result verification

A convergence study for linear elasticity using a plate-with-hole geometry that admits an analytical solution. Serves as a reference benchmark for structural mechanics codes.

The Taylor-Couette problem — Navier-Stokes flow between two concentric rotating cylinders — as a reference benchmark for CFD solvers.

Analysis scripts and reference data for the 11th SPE Comparative Solution Project (SPE11) — a community benchmark organized by the Society of Petroleum Engineers for comparing subsurface CO₂ storage simulators.

• Distance metrics for quantitative result comparison
• Geometry files and grid variants (SPE11A, SPE11B, SPE11C)
• Simulator input decks and submission templates
• Thermodynamic utilities for phase properties and CO₂ solubility
• Jupyter notebooks demonstrating evaluation workflows

Official benchmark specification: https://spe.org/csp/spe11/

Machine learning tools for geological carbon storage (GCS) simulation, built on top of the SPE11 benchmark data and the DarSIA image analysis library.

• Configurable training pipelines (configs/)
• Python source package (src/ml4gcs/)
• Jupyter notebooks for interactive exploration
• Supports local development and NFDI JupyterHub environments

Each repository contains its own setup instructions. Most projects use either Conda/Mamba or plain pip for dependency management, and Jupyter notebooks as the primary interface for running analyses.

# Example: clone and set up the SPE11 analysis scripts
git clone https://github.com/Simulation-Benchmarks/11thSPE-CSP.git
cd 11thSPE-CSP
# follow README in that repository for environment setup

Contributions are welcome across all repositories — additional grid files, simulator input decks, new benchmark cases, and bug reports are particularly appreciated. Open an issue or pull request in the relevant repository to get started.

Individual repositories carry their own licenses (e.g., MIT for NFDI4IngModelValidationPlatform). See each repository's LICENSE file for details.