Combining Domain Expertise and Artificial Intelligence For Aerospace Manufacturing
By Claire Tremont, Manager of Communications and Digital Strategy
University of Connecticut College of Engineering faculty are joining forces with private and public partners to innovate aerospace manufacturing, improving precision-machining of aero-structure components past its current capabilities.
UConn, Raytheon Technologies Research Center, Siemens Industry and the Connecticut Center for Advanced Technology recently secured $2 million in funding from the United States Department of Energy for its project titled “Control of Aerospace Manufacturing Variability Using Physics-Informed Artificial Intelligence.”
The group of experts propose a digital twin framework that uses physics-informed artificial intelligence models to control variability in the precision manufacturing of large-scale, aero-structure components. The project will achieve high-throughput aerospace manufacturing through adaptive machine control using real-time process signals that include conventional sensors of the power, current and voltage of the machining spindle, fused with digital signals of audio and vibration by employing the hybrid manufacturing digital twin.
This project addresses a high-priority challenge in the aerospace industry – the need to develop a standardized process for optimizing speed, feed rate, efficiency, and productivity during the manufacturing of large-scale precision aero-structure components.
The DOE selected 20 projects to focus on cost-effective manufacturing processes and development of novel materials, especially those with high strength, enhanced conductivity, or high performance under extreme conditions.
The request began with a daunting task for Collins Aerospace or the Raytheon Technologies Corporation – to nearly double landing gear output without increasing the number of machines on the floor.
“That is the devil in manufacturing,” said George Bollas, Director of the Pratt and Whitney Institute for Advanced Systems Engineering. “When manufacturers increase yield, they also increase costs, inaccuracy, scrap or machining deficiencies, machine equipment failures, supply lead time and more. How to improve manufacturing productivity without the concomitant increase in manufacturing failures is a real challenge. We’re doing that by joining together four complementary partners, starting with a physics-based model to reduce those stressors, then using machine learning and artificial intelligence to use available machining data and improve model accuracy and system awareness, and then deploy data infrastructure and advanced controls to optimize the machining process.”
Bollas is also the Pratt & Whitney endowed chair professor in chemical and biomolecular engineering.
With this new technology, the already strong aerospace industry presence in the state of Connecticut will only improve.
“Connecticut has been at the forefront of advanced manufacturing and our aerospace manufacturers need the support of researchers to push the boundaries of production using emerging technologies,” Lavoie said. “I am pleased with the collaboration on this project and grateful to the Department of Energy for awarding the grant. With this project we’ll be able to connect machining and machine learning and new and old technologies that make our aerospace sector competitive, all to drive economic growth in Connecticut.”
The four teams are placing diversity, equity, inclusion and accessibility at the forefront alongside this research. They are integrating the DEIA mission by including people from groups underrepresented in STEM as senior personnel, undergraduate and graduate researchers and post-doctoral researchers. They also aim to enhance or collaborate with existing diversity programs at UConn, CCAT, RTX, and Siemens.
While the proposed project is focused on aerospace structures, digital twin technology has wide applicability to other manufacturing sectors including automotive/transportation, defense, energy, biomedical, agriculture and the retail sector.