DATA MODELING FOR MACHINE LEARNING AND DATA-CENTRIC ANALYTICS FOR SMART AEROSPACE ADDITIVE MANUFACTURING
Project Lead: Honeywell
Partners: MORF3D, University of California Los Angeles, University of Southern California, Missouri University Science &Technology, Identify 3D, Sentient Science Corp, Stratonics
Problem Statement: Challenges in contextualization of the enormous amounts of data in a meaningful timeframe in data modeling including gathering of right data; real-time data collection scalability; application of right machine learning tools; interoperability; interactions between edge and cloud; connectivity of smart factory to office; office/factory to cloud; cloud services which connect back with the office; factory and the worker/process engineer.
Project Goal: The goal of this project is to develop technologies on data modeling, machine learning and data-centric analytics for smart Aerospace additive manufacturing and to implement these innovations using data from working Aerospace manufacturing facilities.
Accelerate the implementation of critical data modeling for machine learning and data-centric analytics into Smart Manufacturing (SM) technologies for aerospace. The project will implement these innovations using data from working Aerospace manufacturing facilities.
Identify and collect available data sources within smart manufacturing (M7)
Accurately capture and predict defects through data modeling originating from the process of the manufacturing plant (M17)
Develop, refine, and integrate data analytics and machine learning into SM DMLS system (M18)
Set up the edge-cloud interoperability and establish a workflow to demonstrate the ML toolkit (M18)
Demonstration OF D2ML technologies using two working Aerospace manufacturing sites (M18)
Financial tracking, reporting, and project cost performance (M18)
Improve AM process development monitoring and build efficiencies reducing the need for multiple development iterations to establish acceptable build parameters
Energy usage will be reduced as development cycles are eliminated
Additive Manufacturing (AM) as an affordable alternative to more traditional, energy-intensive part manufacturing processes, reducing overall US manufacturing energy consumption
This program will enable the large data sets currently being created by AM equipment to be utilized to develop actionable insights and make affordable AM part build processes
Benefit to CESMII:
D2ML addresses secure, scalable, and interoperable on premise, edge, and off-premise cloud-technology integration
Application across aerospace industry
Help new AM startups accelerate their AM product expansion through SM technologies
Project Cost Share and Duration:
Project Duration: 18 months, CESMII Cost Share: 64%, Member Cost Share: 36%