Teaching Numerical Problem-Solving Methods to Undergraduate Engineering Students Using Specially Designed Finite Element Codes

Abstract

Computer-Aided Engineering is an essential solution in many engineering problems in today’s industry. Modules looking into this area in the engineering courses aim to further enhance both the theoretical and practical appreciation of the numerical problem-solving methods. In such modules, students are expected to appreciate how mathematics, numerical analysis and computational technology are combined to model and simulate the behaviour of physical systems. However, when it comes to teaching, there is a significant difference between how the background theory looks and what students experience as they work with a commercial software package as they cannot see what is going on in the background of software packages and how the outputs have been achieved. In this work, the authors have proposed a method to help students comprehend how the theory is related to software packages. This is done by providing students with specially designed Finite Element codes asking them to investigate, and tailor the codes for some basic but real-life applications. The method starts with 2D problems for elementary Finite Elements, and through a few steps helps students extend the codes to 3D cases to enable them to solve real-life applications by the FE codes they have tailored themselves. This approach enables engineering students make meaningful links between the background math and the target numerical problem-solving methods. According to student surveys taken over three academic years, 85% of students believe “User FE-codes helped understand how theory translates to problem-solving tools and FE software”. There was also significant enhancement in student performance on the associated assessments.