Case Study: Worcester Polytechnic Institute (WPI) advances cold spray performance prediction with Citrine Informatics

Worcester Polytechnic Institute builds cold spray prediction models with Citrine Informatics using 50x more documents

Worcester Polytechnic Institute (WPI), as part of a DARPA-funded research group, faced the challenge of predicting how metal alloys would perform in cold spray additive manufacturing, a process increasingly used for defense applications. With little existing data on how new materials behave in this process, they partnered with vendor Citrine Informatics to build a predictive model using their AI and machine learning capabilities.

Citrine Informatics rapidly built a complex database of cold spray properties by utilizing large language models (LLMs) to gather and integrate data. The solution resulted in a database containing approximately 50 times more documents, 23 times more experiments, and 10 times more parameters than previous efforts. This allowed Citrine Informatics to deliver robust predictions of cold spray performance, including uncertainty assessments, for a broad range of alloy classes.

Worcester Polytechnic Institute (WPI)