TechConnect Innovator Spotlight:
|
Machine Learning for Non-Destructive Evaluation in Manufacturing Environments, Citrine Informatics
TECHNOLOGY SUMMARY
Citrine’s software platform uses data aggregation and machine learning to help the world's largest companies turn their manufacturing into more reliable products.
Primary Application Area: Manufacturing, Instrumentation
Technology Development Status: Prototype
TECHNOLOGY DETAILS:
Citrine uses a number of machine learning algorithms combined to specifically analyze in-line metrology and process log files to predict materials quality. To achieve this, we had to develop a custom data schema to accommodate complicated materials, chemicals, and product data in a systematic way. We also have purpose built data infrastructure in the cloud that handles this data at scale and is able to process it in near-real-time. As new production data is collected, it can be evaluated rapidly against the models, and as new performance and lifetime data is collected, our systems ingest it and rebuild models on the fly so performance expectations always include the most up to date information possible. This is enabled by the highly tuned software libraries developed in house to specifically address materials, chemicals, and physical product performance.
FIGURES OF MERIT
Value Proposition: Advanced physical products are defined by the materials the comprise them. And understanding how those materials will perform over the course of their lives is very important. While statistical process control has become the norm in most manufacturing environments, and industrial internet of things (IIOT) is emerging as a useful set of tools in the management of machines, the quality assurance world has not fully benefitted from the revolution that data science has enabled. Citrine uses a purpose built data structure for materials and chemicals and optimized machine learning algorithms to connect long term performance to in-line metrology and historical accelerated life and other destructive testing to predict materials lifetimes. This enables manufacturers to have more confidence in the performance of every part they produce.
SHOWCASE SUMMARY
Organization Type: Mid-stage Startup (A or B)
Website: http://www.citrine.io
GOVT/EXTERNAL FUNDING SOURCES
Vetted Programs/Awards: StartX
SBIR/STTR Awards: $150,000, NSF, July 2014
$1,150,000, DOE, Jan 2016
External Funding to Date: $1.05 million seed stage venture capital