Abstract
With the integration of CAD and FEA software packages, design engineers who are not skilled in finite element analysis are performing finite element modeling and analysis. Furthermore, in the analysis of a system, engineers often make numerous modeling simplifications and analysis assumptions depending on the trade-off between cost, accuracy, precision or other engineering analysis objectives. Thus, reusability or interoperability of engineering analysis models is difficult and often impractical due to the wealth of knowledge involved in the creation of such models and the lack of formal methods to codify and explicitly represent this critical modeling knowledge. Most institutions and organizations have started documenting these simplifications and assumptions, making them understandable for the other engineers within the organization. However, this does not allow a seamless exchange of data or interoperability with other analysis models of similar or dissimilar nature. This plays a very important role in today’s market, which is moving away from the traditional make-to-stock production model to a build-to-demand model. We address these issues in this paper by adopting and extending the computer science concept of meta-object, and applying it in novel ways to the domain of FEA and the representation of finite element modeling knowledge. We present a taxonomy for engineering models that aids in the definition of the various object analysis classes. A simple beam analysis example, followed by a more realistic injection-molded part example. The latter example involves injection-mold filling simulation, thermal cooling, and part ejection analyses which are subclasses for a generic manufacturing analysis meta-object class. Prototype implementations of automated support for this meta-object approach to finite element modeling is in progress.
Developing an Extended IFC Data Schema and Mesh Generation Framework for Finite Element Modeling
