Using parametric sensitivity analysis to detect design intent in CAD assemblies

Shaheer Zubairi

    Research output: Thesis (awarded by external institution)Doctoral Thesis

    Abstract

    The intellectual arrangement of parts in an assembly is a difficult task. Modern CAD environments contain tools which allow CAD part models to be brought together and problems such as clashes to be discovered. Clashes occur when components in an assembly unintentionally violate others. If clashes are not identified and designed out before manufacture, the physical parts will not assemble together without rework. This work introduces a novel approach for eliminating clashes by identifying which parameters defining the part features in a CAD assembly need to change, and by how much, to eliminate the clashes. Consideration is given to the fact that it is sometimes preferable to modify some components in an assembly rather than others, and that some components in an assembly cannot be modified as the designer does not have control over their shape. One of the interesting insights offered by this work is that certain aspects of design intent related to component interfaces in an assembly can be enforced by identifying the faces between the different components, and understanding the effect of parameters which define the model has on them. The work presented in this thesis determines which parameters should be related to one another and, more importantly, how parametric sensitivities are used to: (a) identify parametric relationships between different parameters in the CAD assembly, and (b) constrain the assembly using the identified relationships to define the design intent of the assembly. The developed processes advances the state-of-the-art CAD systems by explicitly determining the relationships between the parameters and eliminating clashes in CAD models. A well-captured design intent for assembly models will enable the designer to design out manufacturing and assembly difficulties at an early product development stage. The approaches have been tested on a number of example models in this thesis.
    Original languageEnglish
    QualificationDoctor of Philosophy
    Awarding Institution
    • Queen's University Belfast
    Supervisors/Advisors
    • Robinson, Trevor , Supervisor, External person
    • Armstrong, Cecil, Supervisor, External person
    Award date1 Dec 2014
    Publisher
    Publication statusPublished - 2014

    Fingerprint

    Dive into the research topics of 'Using parametric sensitivity analysis to detect design intent in CAD assemblies'. Together they form a unique fingerprint.

    Cite this