Design optimisation of electric vehicle structures for crashing safety

Abstract

A key contribution of this thesis is the development of a high-fidelity multi-physics model that predicts the behaviour of cylindrical lithium-ion batteries (18650 and 26650 formats) under mechanical impact. The model integrates detailed experimental characterisation with finite element analysis to capture the complex interplay of thermal, electrochemical, and mechanical responses during dynamic loading. By accurately simulating failure modes and damage progression and validating the results against experimental drop-weight impact tests, the framework provides a robust predictive tool for assessing battery integrity and mitigating the risk of thermal runaway in electric vehicle applications.

Building upon this, the thesis introduces a three-dimensional thermo-elastoplastic topology optimisation method using a reaction–diffusion-based level set approach. This method is capable of optimising structural layouts under combined mechanical and thermal loading, accounting for nonlinear material behaviour. The proposed method is demonstrated on benchmark problems and applied to battery compartments, ensuring enhanced structural integrity and energy absorption.

Further, the research develops a data-driven optimisation framework combining genetic algorithms and neural networks for the structural design of EV battery modules. This approach effectively addresses varying boundary conditions due to different battery placements, enabling compliance minimisation and impact resistance through surrogate modelling.

Collectively, the methods developed in this thesis provide a systematic strategy for designing safer, lighter, and more efficient EV battery enclosures. The findings contribute to the broader field of structural optimisation and battery safety, offering tools and insights essential for advancing next-generation electric mobility solutions.