Form to Flow: a cloud-based workflow automation system for introductory CFD courses

Introductory CFD courses at the undergraduate level are taught with an emphasis on the methodology of computational engineering and application to real-world problems, and with limited discussion of numerical methods and algorithms. The interaction of students with any CFD solvers used in such courses is therefore limited to running simulations, and does not include software or model development. However, two problems arise in this educational setting, both stemming from a lack of familiarity of beginner students with command line-based interfaces. Firstly, to run sufficiently high quality calculations of complex geometries and physics requires computing power beyond what is available to the students' personal computers, necessitating the use of high performance computing (HPC) systems which many undergraduates are not conversant with. Secondly, available industry and open-source tools that provide a graphical user interface (GUI) for students are of such complexity as to result in a large portion of class time spent in learning to interact with the tool.

We have developed a novel workflow automation system called Form to Flow (F2F) that solves both the above problems, by using a combination of Google tools and a cloud storage synchronization tool to connect students seamlessly to a HPC system that remotely runs CFD simulations, with interactions through a simple form. F2F is modular and solver-agnostic. We tested this system in the Computational Engineering course (ME123) at Stanford University in Spring 2024, and present results.