CoE researchers to develop safer, more reliable machines with new grant

Mechanical failures in agricultural and industrial machinery could become easier to predict with the help of a new grant awarded to a team of Iowa State University researchers.

Hu

Chao Hu, an assistant professor of mechanical engineering, and his research team have received a $239,509 award from the NSF Partnerships for Innovation (PFI) program for a two-year project entitled “PFI-TT: Physics-based Deep Transfer Learning for Predictive Maintenance of Industrial and Agricultural Machinery.” The aim of this project is to integrate physics-based modeling and data-driven transfer learning for enabling practical and scalable predictive maintenance that makes rotating machines safer and more reliable. This will be the third NSF project that Hu leads as the principal investigator (PI).

If a rotating machine (e.g., motor, pump, or fan) fails unexpectedly, this could lead to unplanned downtime, reduce customer satisfaction, and cause human injuries and fatalities. These consequences not only impact the end user of the rotating machine, but also the machine manufacturer by tarnishing their reputation and potentially putting them at a competitive disadvantage. Thus, it is imperative for companies across industries to have a scalable and practical predictive maintenance solution that will provide advanced warning of failure so that maintenance actions can be scheduled during planned downtime, according to Hu.

Grace’s IIoT product family for predictive maintenance. From left to right: hot spot monitor, vibration & temperature node, panel-mount CloudGate/node, and web-based Maintenance Hub (computer, tablet and phone)

“What we want to create is a low-cost, easy-to-implement, massively scalable industrial internet of things (IIoT) platform,” said Hu. “This platform allows remotely monitoring hundreds or thousands of machines that operate in the field or on a production floor. With the ability to remotely monitor large quantities of machines, maintenance and reliability engineers will be able to track the health of very large assets and better plan maintenance schedules.”

LaFlamme

Other Iowa State researchers serving as the Co-PIs on this project include Simon Laflamme, Waldo W. Wegner Professor in Civil Engineering; Carey Novak, Industrial Specialist for Iowa State’s Center for Industrial Research and Service; and Matthew Darr, professor of agricultural and biosystems engineering. The project will also involve a consultant and technology commercialization expert – Andy Zimmerman, the CTO of Grace Engineered Products, who has expertise and experience in wireless sensing and communication, and IIoT product development and commercialization.

Novak

“The core of our IIoT platform is a new deep learning solution that provides more accurate and robust failure prediction than current solutions and can be easily deployed across multiple types of machines,” Hu said. “The solution will be implemented on Grace Engineered Products’ existing, industrially hardened IIoT platform, where a network of smart devices monitor machine vibrations and predict machine failures on the edge, and then send prediction results to a web-based Maintenance Hub that provides real-time analytics, dashboards and alert capabilities.”

Darr

The broader impacts of this project will include (1) enhanced economic competitiveness of the U.S. industrial and

Zimmerman

agricultural sectors from having more reliable, safer and lower-cost rotating machinery and (2) training of undergraduate and graduate students with innovation and technology translation skills in machine learning, IIoT, and predictive maintenance. The students are expected to receive real-world experience working on a multi-team technology translation project from idea stage through invention disclosure to licensing.

Funding for this research will begin on Aug. 1, 2019, and continue through July 31, 2021.