A team of UT Austin engineers is completing a $11.3 million project to modernize the U.S. manufacturing industry through development of the nation’s first smart manufacturing platform for industrial networked information applications. The project, overseen by the Smart Manufacturing Leadership Coalition (SMLC), employs the latest advances in digital process controls and big data analysis to save energy through more efficient manufacturing operations.
The research team is putting preparing its final report, which will be delivered to the U.S. Department of Energy (DOE), principal funders of the research, early next year.
Project leaders say adoption of the research team’s recommendations – as demonstrated in two industrial test beds using actual data – in turn could help mitigate the environmental and public health effects associated with emissions released during electricity generation.
Based largely on the success of this project, UT Austin and a consortium of industry and other academic institutions received an additional award of $70 million from the DOE to establish five regional centers under the auspices of Clean Energy Smart Manufacturing Innovation Institute. Partners in the Gulf Coast center, which focuses on manufacturing improvements for the chemical, oil and gas sectors, include Texas A&M University and Tulane University, in addition to UT Austin.
Chemical Engineering Prof. Tom Edgar, director of UT’s Energy Institute, has served as the principal investigator for the project. Edgar also has led the university’s participation in the SMLC, which for 10 years has focused on accelerating the development and adoption of advanced sensors, data analytics, and controls in manufacturing. The coalition is an industry-academic collaborative organization comprised of nearly 200 partners from academia, industry, and non-profit organizations in more than 30 states.
Researchers have analyzed massive amounts of operating data from commercial plants – including unconventional data such as temperatures measured via infrared (IR) video feeds – and employed the results to build predictive models used for making optimal design and operating decisions in order to save energy.
“Manufacturers have tried to be efficient in their use of energy for years, but those efforts have tended to be passive and fragmented,” Edgar says. “Our work demonstrates that actively integrating advanced instrumentation and data sharing programs can significantly improve plant energy efficiency.”
The team found that combining high-performance computing with a cloud-based platform for data exchange reduces the costs of current manufacturing systems and technologies by half, Edgar notes.
“These innovations have strong potential to drive a resurgence in U.S. manufacturing and attract more investment and jobs from around the world,” he adds.
Several other UT Austin engineering faculty members, research staff and graduate students participated in the research, including Michael Baldea, Joe Beaman, D.K. Ezekoye and Vince Torres.
Partners in the collaborative research effort included the SMLC, Emerson Process Management, Praxair, General Dynamics, Schneider Electric,, American Institute of Chemical Engineers, National Center for Manufacturing Sciences, National Institute of Standards and Technology, University of California, Los Angeles, and Nimbis Services.
Central to the research team’s goals was to demonstrate how the implementation of sensor-driven data analytics and comprehensive performance metrics allow for a far-flung array of U.S. plant operators – from petrochemical and pharmaceutical companies to cosmetics and pet food manufacturers – to more efficiently use energy, water and materials.
Vince Torres, associate director of UT’s Center for Energy and Environmental Resources in the Cockrell School of Engineering, has been integrally involved in managing the project since its inception.
Throughout the project, researchers focused on developing and executing a work plan designed to take manufacturers to a new level of efficiency, he says.
“We needed to do something radically different, something that would really change the game,” Torres notes from his office at the Pickle Research Campus in north Austin.
“That’s where smart manufacturing was envisioned; we think we have answer to that challenge.”
The research team’s approach hinged on developing a platform that rendered practical applications to ensure it would meet industry needs, Torres says.
One of the test cases entailed optimizing heat-treating furnaces at a General Dynamics Army munitions plant; a second test, at a Praxair hydrogen production plant, centered on boosting the performance of steam-methane reforming furnaces.
For the Praxair test, engineers separated operational phases within the steam methane reforming units, with particular focus on how to better measure and control the temperature of tubes and methane gas flow in furnaces used to produce hydrogen.
The resulting work flows, created from sophisticated computer modeling, proved to be a ‘holy grail’ for researchers seeking ways to optimize the furnaces, Torres says.
“If they can accomplish the same thing by operating at a lower temperature,” he notes, “their tubes last longer, which means less down time – because they’re not changing them out as often – and more product can be manufactured.”
Customized versions of the workflow models can be used by a variety of manufacturers in their operations, Torres adds.
In the test case for General Dynamics, a contractor that manufactures artillery shells for the U.S. Department of Defense, researchers deployed real-time data analytics and modeling to optimize the forging of materials used to manufacture shell casings and commercial metal parts.
“Our job in this case was not to optimize the whole line, although developing the approach to do so was part of the project,” Torres notes, “but to focus on the most energy-intensive operations of the plant.”
Depending on the composition and quality of materials used in the process, the team’s models will be used to help optimize the temperature of the five zones of the furnace after the materials are forged but before they were hardened. The goal was to create a uniformity of temperature throughout the process.
Researchers also examined how to decrease the amount of ‘scrap’ – materials that were forged improperly, or otherwise did not meet quality standards, and would have to be thrown away.
“Scrap is really energy wasted,” Torres explains. “If you can reduce the scrap, you can reduce the amount of energy you’re using.”
Torres likens deployment of smart manufacturing systems that save energy and increase productivity to an amateur chef preparing a family dinner meal of twice-baked stuffed potatoes. The cook realizes he is wasting energy by cooking the potatoes in a conventional oven, so he decides to jump-start the process by microwaving them for 10 or 12 minutes. To reduce energy usage further, he cooks the potatoes in a toaster oven at 450˚ before lowering the temperature to 300˚ for the final 15 minutes before adding cheese and other ingredients.
The analogy is instructive, Torres says, with one big difference: unlike the engineers involved in UT’s smart manufacturing research, the cook has only primitive sensors at his disposal to how long and at what temperatures he should cook his potatoes to achieve the desired outcome – in this case, the creamy texture of a perfectly baked stuffed potato.
“That’s exactly what we would hope to accomplish in a follow-on phase,” Torres says, “optimize the whole process.”
“The additional instrumentation – the knowledge you get from those measurements, and how you use it – all contribute to a higher quality product, which is something that smart manufacturing helps to provide.”
As with the Praxair test case, the models produced by researchers, combined with a cloud-based platform for data exchange, provides manufacturers a framework they can adopt and tailor to suit their specific needs.
“The successful demonstration of smart manufacturing in these two test beds was a springboard for the Department of Energy to establish the new national institute on Smart Manufacturing headquartered in Los Angeles,” Edgar notes.