A feature-mapping optimisation framework for CAD based design

Abstract

Much engineering attention is given to ensuring engineering structures are designed optimally. This includes minimising the use of raw materials to maximise both the structural and fuel efficiency of, for example, commercial aircraft. Similar efforts can be observed in the design of other vehicles for public transport.

Computer-aided engineering provides technologies that can support the corresponding engineering design and optimisation processes. However, the literature indicates that there is limited automated integration of structural topology optimisation techniques in commercial computer-aided design (CAD) workflows in the form of feature-based geometry. Consequently, it is challenging to exploit the full capabilities of these technologies in the development of optimal designs. An automated transition of optimisation results to parametric feature-based models is required to ensure they are seamlessly integrated in a commercial design cycle.

This thesis conducts a thorough review of available structural optimisation technologies, with particular attention given to the challenges of integrating these technologies in design workflows. Subsequently, a framework is proposed to develop the capabilities of feature-mapping optimisation techniques. This framework provides a synergised solution, facilitating topology and shape optimisation of structural designs, and providing a mechanism for translation to commercial CAD packages. This is supported by the development of mapping functions that can represent structural features with a parameterisation corresponding to extruded and swept feature-based templates available in commercial CAD packages. The mapping functions accurately represent the features in the optimisation environment. They are also analytically differentiable, providing sensitivity information corresponding to the feature’s geometric parameters. Extensions are added to the framework that facilitates design modification by refining existing or instantiating new, structural features.

A feature clean-up strategy is included in the framework to ensure the resulting design corresponds with a hierarchical feature-based CAD model, as commonly seen in the feature tree of CAD based designs. This is a critical step to ensure designs can be manipulated in downstream engineering activities. Shape optimisation is carried out in the final stage of the framework to refine the design while maintaining the feature connections established in the feature-based CAD model. This is facilitated within the same framework, mitigating the need to introduce additional software. The overall framework is realised using scripts and by exploiting the capabilities of commercially available finite element analysis software for calculating structural responses and design sensitivities. This ensures a range of structural responses is available to the designer.

Several example problems are progressed throughout the thesis, demonstrating the application of each stage of the framework in successfully constructing optimised designs in a CAD environment. The developed successive feature refinement and feature insertion techniques have been demonstrated to effectively improve structural performance through several examples. The fully parameterised, feature-based models are a testament to the potential of the framework in integrating structural optimisation in commercial engineering design workflows.

Date of Award	Jul 2023
Original language	English
Awarding Institution	Queen's University Belfast
Sponsors	Northern Ireland Department for the Economy
Supervisor	Trevor T Robinson (Supervisor), Adrian Murphy (Supervisor) & Cecil Armstrong (Supervisor)