The finite element method is one of the most widely used computational methods in engineering, with interest for both theoretical studies and applications across a wide range of fields. But until recently, there was no English-language online training in the finite element method. Krishna Garikipati, U-M Professor of Mechanical Engineering, and Mathematics, and Associate Director for Research at the Michigan Institute for Computational Discovery and Engineering (MICDE), is working to change that.

First off, about a year ago, he posted 50 hours of lectures on the subject on Open.Michigan, which have garnered more than 100,000 views.

Then this fall, the material was organized into a Massive Open Online Course (MOOC) on Coursera, titled “The Finite Element Method for Problems in Physics.” The class started in October, and has more than 12,000 students enrolled from more than 150 countries. Twelve percent of them already hold PhDs, and 33 percent have Masters degrees.

While Garikipati said the MOOC seemed like a potentially popular offering — because of the appeal to so many fields, such as structural mechanics, electromagnetics, nanotechnology, biomedical engineering and materials science, as well as the attraction for academia and industry — he didn’t anticipate the level of enrollment that materialized.

“It seemed an obvious thing to try,” he said. “12,000 is much more than I expected.”

The MOOC is aimed at a graduate-level audience in engineering, physics and mathematics, as well as practicing engineers interested in learning the basis of a method they may otherwise use as a black box. The class exposes students to the mathematical basis of the finite element method, its formulation, and its implementation using the open-source finite element library deal.ii.

The Coursera class  is 15 weeks long, with students watching about 3 hours of video lectures per week. Course work primarily involves short quizzes and longer computer coding assignments. Gregory Teichert, a PhD student in Garikipati’s Computational Physics research group has recorded 5 hours of tutorials to help students work on the programming assignments, which use templates derived from code developed in the Computational Physics group.

“The aim of the course is to bring students to the point where they are able to write computer code to actually solve problems with the finite element method,” Garikipati said.

The course will be available on demand after the first session is over at the end of January.

Garikipati added that the material aligns well with the education in scientific computing being offered by MICDE, the Ph.D in Scientific Computing, and the CDE Certificate.