Curated Process–Property Dataset for AI-Enabled Minimal Viable Testing in Metal Additive Manufacturing

Description

This dataset was developed for the project “AI-Enabled Minimal Viable Testing for Additive Manufacturing”. It contains curated data on metal additive manufacturing collected and harmonised from publicly available research datasets and open-access literature. The dataset links material, manufacturing process, processing conditions, build orientation, porosity, surface condition and post-processing information with mechanical performance outcomes, including ultimate tensile strength, yield strength, elongation, elastic modulus and fatigue life where available. Data from heterogeneous sources were cleaned, standardised and transformed into a consistent schema suitable for statistical analysis and machine-learning modelling. Literature-derived records were supported by an LLM-assisted extraction and auditing workflow, while source provenance is retained to allow records to be traced back to their original publications or datasets. The dataset is intended to support research into AI-enabled minimal viable testing for metal additive manufacturing, including predictive modelling, feature-importance analysis, sensitivity analysis, uncertainty assessment and the identification of reduced but representative mechanical testing strategies. Missing values indicate that the corresponding information was not reported or could not be reliably extracted from the original source.

Institutions

• University of Warwick

  England, Coventry

Categories

Licence