Managing Mesh-Based Data in a Semantic Database

Several numerical methods, such as finite difference method, finite element method, and control volume method, use spatial discretisation for estimating spatial differentials. Common to these methods is that they estimate continuous volume with discrete finite volumes or nodes. Use of semantic data model is an attempt to unify the representation of modelling data in computational systems.The objective of this work is to create a generic and comprehensive description of spatially discretised geometric model usually called a mesh. The semantic representation should be applicable to different numerical computational methods and modelling domains. The fundamental objective is to enable the use of different tools like finite element analysis and computational fluid dynamics tools to access mesh data from a common semantic database. The emphasis on this representation development is on minimising the amount of data that is needed to store explicitly for a computational mesh and all data that is needed to run a computation. This study shows that spatially discretised data can be represented in semantic form and that there are several approaches for representing the data. The fully semantic approach provides the most flexible data model from semantic point of view by allowing semantic reasoning in single element level, due to semantic representation of each of the nodes and elements of the mesh. On the other hand, the fully semantic approach is also the most resource intensive and is probably too inefficient for industrial applications. An approach between fully semantic data representation and plain tabular data was selected as the best compromise for the data model.