Bowling Green State University researchers are embarking on a project to leverage artificial intelligence to enhance process controls and energy efficiency in glass melting.
The three-year project is part of the Northwest Ohio Glass Innovation Hub, established in 2024 to bring together industry and academia to strengthen the state’s economy through research and innovation.
Mohammed Abouheaf, Ph.D., an associate professor in the BGSU College of Engineering and Innovation with research expertise in machine learning and autonomous and intelligent systems, will lead the research.
“In working with our industrial partners, we have identified that energy efficiency is a focus in glass manufacturing,” Abouheaf said.
“By integrating AI and machine learning into the glass-melting processes, our goal is to improve performance, which, in turn, will improve energy efficiency.”
The research is being supported by a $652,000 grant from the Northwest Ohio Innovation Consortium, the not-for-profit entity that established the Glass Innovation Hub.
The Glass Innovation Hub, funded through a $31.3 million grant from the Innovation Hubs programme administered by the Ohio Department of Development.
It is expected to increase state tax revenue by $25 million and produce more than 200 new graduates working in science, technology, engineering or maths fields to meet workforce demands.
Industry partners leading the initiative alongside BGSU and the University of Toledo include global organisations headquartered in the region.
The goal of the project is to develop and then integrate AI and machine learning-based methods into the glass manufacturing process to improve performance and energy efficiency.
BGSU is collaborating with Actual Reality Technologies, a Toledo, Ohio-based company which specialises in augmented intelligence and data modelling.
Tom Bush, CEO of Actual Reality Technologies, said: “By combining BGSU’s research with our applied AI systems, we’re transforming how glass is made – creating intelligent furnaces that learn, adapt and optimize energy use in real time.”
Among the challenges in glass-melting operations is the lack of sufficient sensor coverage to monitor temperature fluctuations and other critical variables in real-time.
Abouheaf said the extreme temperatures and corrosive environment inside glass furnaces make it unfeasible to install enough sensors for adequate measurements.
To address this data gap, Abouheaf and his team of researchers plan to develop algorithms to infer the values of unmeasured variables from the available sensor data.
The project’s most ambitious objective is the simultaneous optimisation of multiple, often competing, goals.
Traditional control systems usually optimise for a single objective, such as maintaining a specific temperature.
Through this multifaceted project, Abouheaf and the team plan to develop a data-driven multi-objective optimisation tool to balance energy efficiency, nitrous oxide emissions, control input constraints and boundary-condition robustness.
While the project’s primary focus is glass manufacturing, Abouheaf said the project has the potential for broader impacts.
“Successfully integrating AI into the manufacturing process to improve performance and energy efficiency would benefit multiple industries beyond glass manufacturing,” Abouheaf said.
"This work is incredibly important as companies continue to adapt to emerging technologies and production demands."
