Using a Smart Recognition Framework for the Automated Transfer of Structural Whole Engine Models

Abstract

The development of adequate simulation models from geometric CAD assemblies is one of the most important tasks in early design phases. With this step requiring a lot of manual effort, the desire for a process efficiency improvement via an automated solution rises. In order to derive information about the assembly to build Finite-Element (FE) models, various different steps have to be taken which require visual assessment and engineering evaluation, knowledge and judgement. The approach described in this research mimics the engineer's logic and way of thinking to automate these steps. Thereof, the recognition of entities plays a fundamental role for further processing. To achieve the desired recognition, methods have been developed to retrieve criteria like form, function, context and positioning from the available geometry data. The developed recognition framework supports and provides a component categorization so that specifically optimized process chains for each category can be implemented, depicting a more robust and reasonable overall process.

Full document