Core Strengths of Purdue Engineering in Data and Engineering Applications
Data science: key to bringing new and better materials to market faster
Materials have always been crucial to civilization; they have defined eras, from the Stone, Iron and Bronze ages to the age of silicon. Today, data science tools are helping engineers and scientists discover and deploy materials faster than ever.
Aided by these tools, the field of materials science and engineering is growing in terms of societal impact and interest. Two signs of this progress are:
- The listing by financial publication Barron’s of materials science as one of three “technologies that are poised to change the world” and “are just beginning to make the leap from lab to market”
- The Materials Genome Initiative (MGI) for Global Competitiveness, and the associated efforts by funding agencies to support this national priority, such as the National Science Foundation (NSF) program “Designing Materials to Revolutionize and Engineer our Future (DMREF)”
What problem has data science addressed? Until recently, it took far too long to move new advanced materials from the laboratory to the marketplace – think 20 years for widespread adoption of the lithium ion battery. The traditional Edisonian approach of trial and error, enhanced by intuition, domain knowledge, and physics modeling, is not effective in most materials discovery and optimization efforts, due to the astronomical number of possible combinations of compositions and processing one can attempt.
“Data science tools – particularly machine learning – are enabling us to combine experimental and theoretical results in statistically meaningful ways, accounting for uncertainties and discovering how materials behave in a computer even before the materials are made,” says Alejandro Strachan, professor of materials engineering in Purdue’s College of Engineering, which is in the vanguard of the trend.
Much as Amazon recommends books and Netflix suggests movies based on a consumer’s history, engineers can reduce traditional trial and error by using machine learning to recommend experiments or simulations to attempt. Purdue researchers are studying how to incorporate laws of physics into models to pare the amount of data needed to make meaningful decisions.
Purdue Engineering faculty, along with graduate and undergraduate students, are collaborating to propel materials engineering to the next level, benefiting industry and society. What’s more, the open and free nanoHUB platform (a 2020 R&D 100 winner) is enhancing the impact of data science efforts at Purdue and around the world. nanoHUB enables online interactive computing online using a standard web browser, so there’s no need to download or install any software. This frictionless access to powerful tools is accelerating innovation in education and research.