A new analytical framework helps optimise manufacturing environments. The framework analyses, connects and coordinates the different steps that take place within manufacturing environments. The framework ensures that optimisations of individual steps advance the production process as a whole.
The analytical framework was developed by a team of researchers led by the US University of Buffalo. The framework is called STREAM and combines artificial intelligence with simulations, among other things. The analytical framework creates a public online library. Researchers and manufacturing industry professionals can share information and experiences related to data, models, simulators, controllers, analytics and empirical studies here.
Large number of processes
A manufacturing process consists of a large number of steps, which are interconnected. "A commercial production is the end result of a long chain of interconnected steps, which can span regions, sectors and different production processes," explains Hongyue Sun, assistant professor of industrial and systems engineering at the University of Buffalo.
Each individual step in the production process can be optimised. While this optimisation advances the individual step, the effect on the production process as a whole is not always positive. The STREAM framework links the different steps and ensures that optimisations benefit the production process as a whole.
"We are creating an analytical framework that connects and coordinates all these processes. The end result is a cyber-physical system that uses artificial intelligence and other tools to optimise and ultimately improve manufacturing systems," Sun said.
Chip production
As an example, the researcher cites chip production, a multi-step process. "This includes dozens of steps, including crystal growth, ingot cutting, wafer creation and polishing, lithography, etching and chemical-mechanical planarisation," says the researcher. "These phases have strong dynamics and dependencies. The processes in downstream phases are influenced by the processes in upstream phases, both in terms of quality and productivity."
As an example, Sun cites multiple wafer machines working together to produce hundreds of wafers. Real-time process and production information from the machines involved affect each other and determine the performance of the production system as a whole. The STREAM framework can coordinate these steps with each other, optimising chip production as a whole.
Grant
The project receives a $2.3 million grant from the US National Science Foundation (NSF), which supports various industries. The grant is part of the NSF's Future Manufacturing Research Grant CyberManufacturing project. Future Manufacturing supports basic research and education of future workforce. In this way, the NSF aims to overcome scientific, technological, educational, economic and social barriers and catalyse new manufacturing opportunities that do not yet exist.
Author: Wouter Hoeffnagel
Photo: Ulrike Leone via Pixabay
