Modelling of multiple, semantically-coupled flow-graphs

Various monitoring applications specifically in the process industry but not restricted to it need explorable models representing multiple flows and their properties. Most (plant) models share two characteristics: They mix structural and behavioral aspects and they are tailored for specific applications - either with respect to human interpretation or for easy automated treatment. The under-lying physics are however often rather shrouded. Thus the implicit assumptions and prerequisites in many cases are not obvious, making it difficult to add new software layers or to re-engineer. This paper discusses systems dealing with multiple, balanceable flows, which are semantically (i.e. usually physically) coupled, independently from their application domain. Specifically the various flow characteristics and their interaction are contrasted. By providing check-lists for these properties design decisions and their consequences are made explicit. The second part of this paper derives and discusses the graph representation in a general way. Starting off from several graphs (one per balanceable flow) finally one single hierarchically structured graph is derived. Additionally some aspects for implementations are discussed briefly.