466884 A Tree Ontology for Multi-Scale Muti-Disciplinary Process Model

Monday, November 14, 2016
Grand Ballroom B (Hilton San Francisco Union Square)
Heinz A. Preisig and Arne Tobias Elve, Chemical Engineering, Norwegian University of Science and Technology (NTNU), Trondheim, Norway

We present an new approach to multi-disciplinary model representation and generation, which is based on an ontology that forms a tree with the branches shaped by the involved disciplines and the leaves being the detailed multi-facet models.

Each leave in the tree represents a network of capacities linked by transfers. What is ''living'' in these networks depends on the discipline. In physical systems this will be mass, energy, species detailed into phases that may be detailed even further. In control this will be signals and information, in material models it can be particles, molecules, atoms. The structure is completed by linking the leaves together where the discipline-specific networks are linked through connection networks. All networks build on a very small set of meta terms, the set of which is the result of intense research over the last couple of years not at least in connection with a European project that has the objective to model polyurethane foams from quantum to mechanical properties (http://modena.units.it). 

The process of building the ontology start with generating the tree structure. One of the keywords provides the branching the rest provides the key construction elements for the equation/variable system that eventually will represent the behaviour of the modelled process in a mathematical form.

The ontology is used in conjunction with a graphical user interface that implements a hierarchical network representation. The hierarchy being used here is different from the hierarchy of the ontology. In the graph model designer, the hierarchy serves the purpose of managing the model complexity, in terms of a large number of nodes and arcs in the respective discipline-dependent network.

The system generates automatically code for target applications that have the necessary solution methods implemented.

The presentation will focus on the philosphy of the approach and the range of applications with the example of materials modelling and controlled plants.

Extended Abstract: File Not Uploaded