A Hands On Approach To Computational Methods In Engineering

Abstract

The Pratt School of Engineering at Duke University has been actively focusing on the development of a variety of mechanisms to provide undergraduate engineering students with an earlier, more practical, experience with engineering concepts. Through these mechanisms, students are now exposed to elements from each of our four departments, which gives the students a clearer understanding of the field of engineering that they want to pursue. As a part of this ongoing effort, all engineering undergraduate students are now required, during their first year, to take a course on computational methods in engineering, EGR 53L. Because of the newfound prominence of EGR 53L in the curriculum, as well as the wide spectrum of student interests and backgrounds, the course has undergone several significant changes in the past five years aimed at improving the student experience and exposure to engineering. This paper outlines several key facets of our approach to redesigning the course and also reports on outcomes and student assessments of the hands-on portion of the revised experience.

Undergraduate engineering students at Duke have long been required to take a full-credit course in either computational methods or computer programming. A full-credit lab course at Duke, denoted by the “L” after the course number, is the equivalent of four credit hours. As recently as five years ago, there were several different offerings that would fulfill this requirement. Over the past 20 years, several programming languages have been used in these courses – Fortran, Pascal, C, C++, and Java – and students could take the courses through the computer science department or the school of engineering. Furthermore, students could use AP credit to satisfy the requirement. This model led to a large disparity in the experiences of undergraduates – sometimes even those students pursuing the same field of engineering. Also, given the foundational requirements of chemistry, mathematics, and physics, undergraduate engineering students would often take no courses in engineering until their sophomore year. As a result, in many cases students would drop out of the school of engineering to pursue degrees in the social sciences, natural sciences, or humanities, without ever having experienced what engineering could be.

Beyond that, instructors in future classes could not count on any uniformity of programming skill, experience, or even language among students. A faculty member who wanted students to complete an assignment with a computational tool often had to find time in their course to instruct students on how to use that tool, reducing the amount of time spent on analyzing the primary material in the course. Furthermore, in addition to the traditional emphasis on programming and simulation in such courses, we felt that a true introduction to computation should devote attention to other, equally important, computational issues in engineering, for example, computerized data acquisition and actuation and control of elementary devices. All of these factors led to a comprehensive redesign of the first-year computational methods course offered in the engineering school. Each department then changed its curriculum requirements to mandate that all engineering students take the new course - neither computer science courses nor