Parametrised post-processing of topology optimisation results

    Student thesis: Doctoral ThesisDoctor of Philosophy


    Topology Optimisation (TO) is a process that is becoming increasingly reliable and necessary in the pursuit of highly efficient components comprising of low mass with a high structural performance. The ability to maximise a component’s structural characteristics has yielded many variations of computational topological solvers over the years. Over time many different methodologies have been used to generate suitable manufacturable solutions. Despite this, a gap between the generation of TO solutions and the creation of ready-to-manufacture solutions still exists today. The ability to refine and represent a TO solution as a fully manufacturable design is a procedure known as Post-Processing (PP). PP of TO results (e.g. from variable density to manufacturable structures) does however remain a heavily heuristic process where the end-results (and consequently the “efficiency” of the optimised product) can vary significantly as a function of the individual designer/engineer. This “variation” makes the use of TO prohibitive in certain instances. In order to address this issue, this thesis presents a systematic and repeatable approach to parameterised PP of TO results. This method, developed into a software tool, considers the refinement of a TO solution specific to a series of user-defined geometrical features. A stencil method is utilised, which scans the TO solution to extract geometrical features against users’ design requirements. In addition to presenting the methodology, this thesis also investigates different parameter variations; such as geometry update sequence, search radii, stencil shape and type and their influence on the generated post-processed result. Definition of algorithm parameters is provided, together with suggested user-defined settings to enable the derivation of consistent refinements of TO results. The code is applied to more “industry-standard” models to identify its practicality as a usable methodology to refine optimisation results files, with focus taken to improving the designs of 2D TO solutions and further work focussing on refining more complex components.
    Date of AwardJul 2020
    Original languageEnglish
    Awarding Institution
    • Coventry University
    SupervisorJesper Christensen (Supervisor), Christophe Bastien (Supervisor), Stratis Kanarachos (Supervisor) & Alexis Wilson (Supervisor)

    Cite this