The Research Group
- DARPA-funded program
- WPI (PI: Danielle Cote), Citrine Informatics, and others
Rubble to Rockets (R2)
DARPA-PS-24-08
HR0011-25-9-0028
PM: Hunter Martin
The Problem
- Metal alloy performance in additive manufacturing processes can differ greatly from traditional methods
- Cold spray additive manufacturing is increasingly used in defense applications, but not much data for many materials in how new alloys behave
- Build a model to predict cold spray performance from traditional performance
The Process
- Rapidly built complex database of cold spray properties using large langue models (LLMs)
- Integrated databases of traditional alloy properties and simulation tools
- Assessed uncertainty of cold spray performance from data and simulations
The Outcome
- Database includes ~50x more documents, ~23x more experiments, and ~10x more parameters than previous cold spray databases
- Robust prediction of cold spray performance, including uncertainties
