Multiscale data-driven failure prediction of hydrogen composite vessels under static and dynamic impact loading

As a key part of the fuel cell vehicles (FCVs), the development of safe, lightweight hydrogen storage vessels is critical for the range of the vehicles. Composite storage vessels has the potential to replace traditional metal vessels, enabling high hydrogen storage density per unit mass. However, due to the poor impact resistance and complex damage mechanisms of carbon fibre reinforced polymer (CFRP) composites, safety is still concerned once it is subjected to transverse impact. Therefore, it is imperative to design and optimise the hydrogen storage vessels based on safety requirements. How to replace the expensive and harsh dynamic test method, accurately predict the complex thermomechanical response of the composite storage vessel under the dynamic impact loading is critical for the development and deployment of new hydrogen composite vessels. Therefore, this EU Marie Curie fellowship aims to:
(1) For the first time, create systematic and fundamental understanding of the failure strength and mechanisms of composites under much broader multiaxial loading conditions using a patented test rig;
(2) Establish high fidelity RVE-based FE modelling of composite with ML identified uncertain material parameters to generate a full spectrum of failure predictions under multiaxial loading conditions;
(3) Synergise the experimental data and numerical predictions to train data-driven ML tools for predicting the material failure envelope and informing the development of a modified failure criterion;
(4) Implement the modified failure criterion and develop a multiscale virtual design and test tool for hydrogen composite vessels under impact loading as well as evaluate their CAI or TAI performance against experiments;
(5) Develop and maintain a dedicated platform for dissemination of the developed virtual tool and promote virtual design and test of hydrogen composite vessels backed up by physical tests