A Generic Parametric Modeling Engine Targeted Towards Multidisciplinary Design: Goals and Concepts

Abstract

This paper presents the design concept of a generic parametric modeling engine that is completely decoupled from geometry generation. Driven by requirements extracted from preliminary multidisciplinary airplane design, the presented software architecture provides a platform that enables an interplay of different modeling and simulation tools on the one hand, and their efficient execution in a parametric tree on the other hand. An integrated plugin system allows users to define custom plugins exposing arbitrary types and functions. All geometric functionality is provided via plugins, decoupling it entirely from the parametric engine. First, we specify the goals that the software framework needs to fulfill, elaborating on the requirements encountered in early aircraft design. Then, we describe the software architecture and its modules, realized as a C++ library. As such, the software is a back-end that can be used by third party developers to create user-friendly and interoperable tools. The core of the framework is a parametric engine called grunk with its integrated plugin system and serialization functionality. The key feature of grunk is the possibility for users to define custom types in plugins and their use in the parametric tree. Geometric modeling functionalities are provided through the plugins grocc and geo: the first integrating OpenCascade Technology's functionalities and the latter extending it. A major feature on the geometry side is the provision of derivatives through algorithmic differentiation, making the framework particularly suitable for gradient-based optimization applications. Finally, we demonstrate the use of the software via examples and show the results.