Research related to Linked Data

Graph databases, in contrast to traditional relational databases, can store a variety of data relationships without complying to a predefined data schema. They are therefore suited for capturing the large variety of data relationships which can exist in engineering, some of which cannot be anticipated in advance, in order to achieve complete data traceability across various product lifecycle aspects such as requirements, design, simulation, and bill-of-materials.

However, existing graph databases do not efficiently manage different versions of a graph. Their adoption in an engineering context is therefore limited. Koneksys is investigating different delta encoding mechanisms for efficiently saving different versions of a graph. Furthermore,  Koneksys is also investigating different approaches to support full-text search on large quantities of Linked Data (Big Linked Data), as well as to support horizontal scalability of graph databases supporting the W3C SPARQL query language standard.

Research related to Data Exchange in Simulation

Many simulation models, such as for structural, thermal, and aerodynamic analysis, are based on partial differential equations. These models are typically solved by a numerical approach called finite element analysis (FEA) which consists of describing a continuous system through many tiny discrete systems, whereby by the solution to  a system of complex partial differential equations is ultimately approximated by a solution to a system of simpler algebraic equations. Unfortunately, there is no standard description of FEA models. Existing FEA model descriptions are incomplete, proprietary, informal or a combination of these. As a result, interoperability between FEA software applications is compromised, and communication between engineers is inefficient. A standard for the description of FEA models is difficult to develop, as the geometry, mathematics and physics of finite elements can vary greatly.

Koneksys is performing research for NIST to propose a new finite element analysis specification based on the topological characteristics of finite elements that is formal, precise, and understandable to engineers. Such a description is not only suitable to support collaboration amongst simulation engineers, but also between simulation engineers and engineers from other disciplines such as systems engineers.