- Bridget Fitzpatrick
- Dr. Babatunde Ogunnaike
- Tom Burke
- John Rezabek
As an industrial practitioner, researcher and educator, Dr. Babatunde Ogunnaike has built career out of redefining models. The model-predictive control (MPC) polymer reactor strategies he helped pioneer at DuPont in the 1980s went on to become standard practice across industry. Today, the William L. Friend Chaired Professor of Chemical Engineering at the University of Delaware is working to model and control the incredible complexities of biological processes with an eye to streamlining the production of biopharmaceuticals, and enabling the practice of personalized medicine at scale. But the models he’s perhaps most proud of are those he developed to encourage the cross-pollination of academic and industry points of view—to the ultimate betterment of both.
This last model innovation dates back to when Ogunnaike was finishing his doctorate in chemical engineering at the University of Wisconsin. He was about to return to his native Nigeria to teach at his undergrad alma mater Lagos University, when he realized he was ill-equipped to give his students the industrial perspective that most of them would need. So he improvised, signing on for what he now calls an “industrial post-doc,” a year of real-world experience with Shell, split mostly between the company’s Houston engineering center and its Norco petrochemical complex in nearby Louisiana.
“It really rounded out my training, but was unheard of at the time,” Ogunnaike says. Most PhD candidates headed for academia opted for an academic post-doc, and those headed for industry just went. The experience stuck with Ogunnaike when he returned to the U.S. to take a position in DuPont’s central engineering organization, where he was soon heading the group doing pioneering work in process systems engineering, including optimization, modeling, control, data analysis and process analytical technique development.
But he was also fully engaged at the University of Delaware, teaching and writing textbooks on process control and statistical analysis. “I stayed on at DuPont because at that time we had arguably the best collection of process systems engineering people in the world—all under one roof. When I came to interview, I’d never seen a group like that in my life, and I thought ‘Academia is going to have to wait.’ Groups don’t come up like that but once in a while.”
While at DuPont, Ogunnaike approached his boss, the late Dave Smith, and proposed formalizing the industrial post-doc experience that had served him so well. “I sold the concept and we finagled the funds to finance a number of people who had already accepted an academic position to come work in the group for a year.” And while the program hasn’t withstood the trials of time, it "graduated" a number of highly respected leaders in the field, including Frank Doyle, Richard Braatz, Michael Doherty and Jay Lee. Other leading academics also joined the group for a time, including fellow Process Automation Hall of Fame members Thomas McAvoy and James Rollins.
“The fresh PhD grads left us with a portfolio of problems that now had a ‘so what’ attached to them,” Ogunnaike says. That understanding continues to shape the questions they ask and the type of meaningful research they pursue, he says.
It was during his time at DuPont that Ogunnaike’s dormant desire to better understand the behavior of biological systems was sparked by a group of neuroscientists, who were studying neurons in order to understand the science of how actual neural networks do what they do.
“Our ability to control biopharmaceuticals manufacturing is some 50 years behind what we can do it the mainstream chemicals industry,” Ogunnaike says. This is due in part to regulatory requirements but more so because of how tremendously complicated biological systems are. This complexity arises from how difficult it is to influence what goes on inside cells purely by manipulating the nutrients or conditions in the bioreactor. “We’re recreating the concept of controllability, but it’s taking all the tricks in the book—all we know about modeling, about multivariable control, about state estimation, plus new nuances specific to biological systems.” Ultimately, his work has implications for personalized medicine as well, which cries out for the leadership of process systems engineering professionals to manage the tremendous amounts of data that such efforts will entail.
“It’s fun to look at biological systems from the perspective of a control engineer,” Ogunnaike says. “And I’m having a ball doing it.”
