Unable to Take Lab Course to Graduate, Student Turns to Simulation

Reaction rate as a function of length and radius of tube flow reactors.

For many students, successful completion of coursework for an undergraduate degree in chemical engineering includes lab courses in the senior year. But the labs were shut down during March and April as COVID-19 spread across the world. It was a desperate situation for those wanting to graduate in June.

No labs? No problem, said Vasilios Halkias, who was enrolled in the Otto H. York Department of Chemical and Materials Engineering at the New Jersey Institute of Technology (NJIT).

Halkias knew from his first chemistry course that he was going to get into chemicals, just as he knew from his first physics course that he was going to be an engineer. A pandemic would not stand in his way. He, like all the other students, was taking his lecture-based classes online, but the labs required another solution.

Vasilios Halkias, recent graduate of NJIT with a bachelor’s degree in chemical engineering, and winner of the best paper in engineering.com/NAFEMS call for papers.

This is 2020, thought Halkias. We have software, the Internet….  The reactions we were going to test in the labs—we will simulate them. The software will provide more—and less. More, in that fluid flow and heat transfer will make more sense with colorful 3D graphics (than hastily drawn rough sketches drawn on the whiteboard—no offense to the professors reading this, says Halkias). Less, as in less risk of an equipment failure in the lab and less risk of experiencing a COVID-19 infection.

One of the lab experiments involved tubular flow reactors—a fancy name for what is essentially a tube into which fluids are piped into on one end, are mixed along its length (and possibly causing a reaction), and come out on the other end. Ideally the heat is contained and the mix is 100 percent—but that depends on various factors: the length, radius, types of liquids to be mixed, etc.

Just as the reaction can be controlled with valves on varying sizes of reactors in the lab to determine percent mixing, so it can be controlled with simulation software. COMSOL, the multiphysics solver, allowed for a variation of parameters far greater than would be possible in the lab.

We’ll accept the simulation as a substitute for the lab work, said NJIT, and Halkias was able to graduate. But that was not enough. Halkias knew chemical engineering students around the world were facing the same predicament. And not all of them would necessarily have the COMSOL application.

But through the ability of COMSOL to spin off applets that could run as executables, without a license for the COMSOL program, Halkias could help other fellow chemical engineering students graduate, too.

Winner of the engineering.com/NAFEMS call for papers and a $1,000 prize, Halkias turned in the winning entry for his work on chemical reactors. (Pictures courtesy of Halkias.)

The app derived from COMSOL would, in the words of another New Jersey scholar (Albert Einstein, who taught at Princeton), be as simple as possible, but no simpler. With it, just the necessary parameters are asked for and the solver runs and provides results. It’s already set up with fluid and material properties. You push the button to see the quality of the mix and the resulting temperatures.

The app is self-contained, a free-standing executable file that runs on a PC or workstation and can be freely shared. COMSOL charges more for the ability to create specialized apps from its general-purpose solver, but once created, the apps can be passed around and used at no cost.

To build the app, Halkias needed to do a little programming—and take care of a few bugs. For example, the app would not resize the view properly.

Halkias will have to wait until he gets back in the lab to verify the results with lab results.

An app was created from the COMSOL multiphysics program that can be used on any computer without the COMSOL software being loaded. (Picture courtesy of Halkias.)

The app makes some simplifications in the interest of simplicity and time. Flow is laminar all the time. Turbulent flow, which makes for more rapid mixing, is not modeled, so the app could be considered a conservative solution. If complete mixing occurs in a length of reactor with laminar flow, it will be sure to have occurred if there was turbulent flow.