Nonintrusive Condition Monitoring of FCEV Using System-Level Digital Twin Model

High maintenance costs and short lifecycles are the two main issues in fuel cell (FC) electric vehicle (FCEV) development. With the development of digital technology, the concept of digital twins provides new opportunities for monitoring and predicting the behavior of FCEV. However, the constitution of the digital model of FCEV faces many challenges caused by the multiphysical domain. In this article, a digital model is built for condition monitoring of FCEV in both short term and long term. The built model reflects the interaction of multiphysical domains by representing the FCEV using the energetic macroscopic representation (EMR) method. Moreover, by considering different road conditions, the effect of different driving modes can be shown through the energy flow between the battery and the FC, and the model can give a visible reflection of the electric, magnetic, mechanical, and chemical inside the vehicle. To further investigate the effect of all the above factors on the long-term operation status, the echo state network (ESN) is used to predict the degradation phenomenon of the proton exchange membrane FCs (PEMFCs), which is used for prognostic purposes. At last, the digital twin test bench is shown with a real-time (RT) simulation platform of the Opal-RT.