Flow-based functional (FBF) approaches model a system by focusing on flows (of mass,energy, or information) and on the actions performed by components on the considered flows. In this paper, we show how these approaches are particularly suited to hierarchical representation, by proposing a general technique that automatically derives a hierarchical FBF representation in polynomial time, starting from a flat FBF model. We also briefly illustrate a diagnostic reasoning example to show the usefulness of the derived hierarchical representation. The proposed technique concentrates on a set of functional primitives which is common to all FBF approaches, and is thus applicable to any FBF approach.
