There is no standard way to store 3D-printable concrete data. Research papers use inconsistent naming, units, and formats. Combining datasets from different labs requires painful manual reformatting. Machine learning on concrete data starts with months of data cleaning.
Open3DCP defines a canonical flat schema where every material, process parameter, and test result has exactly one column name, one unit, and one meaning. Mass-percent basis. Standards-aligned naming. A dataset in this format goes straight into ML training with a single SELECT.
Unlike conventional concrete databases, Open3DCP has first-class columns for print process parameters (nozzle geometry, layer timing, print speed), fresh-state rheology (yield stress, thixotropy, open time), and interlayer bond properties. Built for additive construction from the ground up.
Open3DCP defines the format, not the data. We publish column names, types, units, and engineering context so that datasets from any lab or research group can be structured in a common format. Adopt the columns relevant to your work and leave the rest null.