Automating Rule‐Checking to Identify SysML Modeling Errors: A Preliminary Study in a Classroom Environment

Abstract Among other benefits, the adoption of Model‐Based Systems Engineering (MBSE) or Digital Engineering (DE) is expected to decrease the number of escaped errors in a project by enabling early error detection within the modeling environment. This contributes to reducing the costs associated with corrective activities, which are generally greater in later phases of development. This paper provides an empirical insight into error detection through a study of models developed by students in a graduate MBSE course, where they leveraged the use of automated rule checking within the modeling tool. The dataset covers 10 editions of the course, spanning 2016‐2023, and contains 601 models. The study shows that the term project models resulted in nearly zero latent errors when non‐stylistic rules are concerned, with most of the latent errors categorized as stylistic rather than fundamental violations.