An Investigation into Hybrid Topology Optimisation of Vehicle Structures

Abstract

Finite Element Analysis (FEA) based topology optimisation (TO) is now widely used across a range of industries from civil to automotive engineering. It efficiently identifies the optimal load-path locations (topology) based on a user defined design volume. Current TO algorithms are well developed for linear static applications. A vehicle structure is however significantly different as it exhibits both linear and non-linear deformation during an impact scenario. Throughout this thesis linear and non-linear behaviour refers to both material and geometry. No current state of the art TO software simultaneously considers coupled linear/non-linear behaviour within a single system. The contribution of this thesis is a novel hybrid methodology which combines existing TO algorithms within a single optimisation process to cater for this combined behaviour. After an initial investigation into the capabilities of existing, commercially available TO algorithms, the Variable Density Method (VDM) and Solid Isotropic Material with Penalisation (SIMP) methods were selected for linear behaviour, while a Bi-directional Evolutionary Structural Optimisation (BESO) based method BEETS was selected for non-linear behaviour. A number of hybrid optimisation parameters were created, defined and evaluated via two case studies. The first case study was a 2D plate study which demonstrates the proof-of-concept, and the second case study a Research Council for Automobile Repairs (RCAR) bumper test based study used to critically assess the performance of the methodology applied to a real-world vehicle structure scenario. The results demonstrated that a specific aspect of the developed methodology, the interface parameters, were the most influential upon the resulting topology. It was determined that increasing data exchange between the linear and non-linear solvers improved the structural performance of the optimised topology. Overall, the main conclusion from this PHD is that a hybrid topology optimisation methodology, combining linear and non-linear optimisation within the same process, can be advantageous in certain circumstances, highlighting the importance of work in this field in the pursuit of obtaining lightweight and safe vehicle structures.
 

Date of Award	2019
Original language	English
Awarding Institution	Coventry University
Sponsors	HORIBA MIRA Ltd.
Supervisor	Jesper Christensen (Supervisor), Christophe Bastien (Supervisor) & Waldemar Golinksi (Supervisor)