Data-Driven Excellence: Fast and Accurate Digital Twin Models for Zero-Emission Long-Haul Vehicle Applications

Logistics and freight transport companies aim to enable and support managerial tools through their digital twin platform. Consequently, requirements for vehicle models in logistic digital twins differ, as their purpose are to support operational and logistical decision-making rather than vehicle system sizing, design, and control strategies. Integrating these models into broader, iterative workflows necessitates extremely short simulation times-much less than a second-while retaining the accuracy of physics-based vehicle simulation models. In this paper, a data-driven approach is proposed for developing a vehicle digital twin model to estimate three crucial parameters: (a) battery SoC, (b) energy consumption, and (c) charging events, solely based on datasets provided through a dedicated vehicle CAN bus. The proposed data-driven model is more than 50 times faster than the corresponding physical one developed in the MATLAB environment. Moreover, this significant reduction in the CPU time is allied to a high accuracy.