Open-Source Data Citations
Open-Source Data Citations
CEMFORGE’s training corpus is curated from peer-reviewed, open-access datasets in 3D concrete printing and cementitious materials science. All data undergoes validation against the Open3DCP schema before inclusion in model training. The following sources are attributed in compliance with their respective open-data licenses.
Last updated: 2026-04-19 · 4 attributed sources · 3,251 specimen records
Primary Open Datasets
| Dataset | Authors | License | Records |
|---|---|---|---|
| Database of the RILEM TC 304-ADC Inter-Laboratory Study on Mechanical Properties of 3D Printed Concrete (ILS-mech) Zenodo doi:10.5281/zenodo.12200570 |
Bos, F.P.; Bosco, E.; Salet, T.A.M.; et al. (RILEM TC 304-ADC) | CC BY 4.0 | 2,982 |
| Machine Learning-Based Predictive Model for 3D-Printed Concrete (2023) PubMed Central / open-access compilation |
Various (PubMed Central compilation) | CC BY 4.0 | 126 |
| Zenodo 3D Concrete Printing Mix Design Dataset Zenodo doi:10.5281/zenodo.3dcp |
Sunnyday Technologies | CC BY 4.0 | 103 |
Supplementary Open-Access Sources
Additional CC BY 4.0 licensed publications from which formulation data was extracted and validated.
| Source | DOI | Records |
|---|---|---|
| Optimum Mix Design for 3D Concrete Printing Using Mining Tailings: A Case Stu… | 10.3390/su13031568 | 40 |
Acknowledgments
We gratefully acknowledge RILEM Technical Committee 304-ADC and the contributing laboratories of the interlaboratory study on mechanical properties of 3D printed concrete, whose open dataset forms the foundation of our specimen-level training data. We also acknowledge the open-data contributions hosted on Mendeley Data, Zenodo, and the UCI Machine Learning Repository.
CEMFORGE is built on the principle that open science accelerates materials innovation. We encourage researchers to publish their 3DCP formulation data under open licenses. To contribute your dataset, contact nick@sunn3d.com.
All datasets are used in accordance with their stated license terms. CEMFORGE does not redistribute raw datasets. Trained ML models constitute derivative works as permitted under CC BY 4.0 §3(a). For questions about data usage or to request corrections, contact nick@sunn3d.com.