Data-Based Energy Demand Prediction for Hybrid Electrical Vehicles

authored by: Daniel Fink, Oliver Maas, Daniel Herda, Zygimantas Ziaukas, Christoph Schweers, Ahmed Trabelsi, Hans Georg Jacob
Abstract: To achieve a resource-efficient automotive traffic, modern driver assistance systems minimize the vehicle’s energy demand through speed optimization algorithms. Based on predictive route data, the required energy for upcoming operation points has to be determined. This paper presents a method to predict the energy demand of a hybrid electrical vehicle. Within this method, data-based approaches, such as neural networks, Gaussian processes, and look-up tables, are applied and assessed regarding their ability to predict the behavior of separate powertrain parts. The applied approaches are trained using measured data of a test vehicle. As a result, for every separate powertrain part, the best-suited data-based approach is selected to obtain an optimal energy demand prediction method. On a validation data set, this method is able to predict the transmission ratio of the gearbox causing a rmse of 0.426. The combustion engine’s torque prediction results in an rmse of 19.01 Nm and the electric motor torque prediction to 19.11 Nm. The root mean square error of the motor voltage results to 1.211 V.
Organisation(s): Institute of Mechatronic Systems
External Organisation(s): IAV GmbH
Type: Article
Journal: SN Computer Science
Volume: 5
No. of pages: 7
ISSN: 2662-995X
Publication date: 11.01.2024
Publication status: Published
Peer reviewed: Yes
ASJC Scopus subject areas: General Computer Science, Computer Science Applications, Computer Networks and Communications, Computer Graphics and Computer-Aided Design, Computational Theory and Mathematics, Artificial Intelligence
Electronic version(s): https://doi.org/10.1007/s42979-023-02475-9 (Access: Open)

BibTeX

@article{7d48722d99fd43ad89385a9b6de218cc,
title = "Data-Based Energy Demand Prediction for Hybrid Electrical Vehicles",
abstract = "To achieve a resource-efficient automotive traffic, modern driver assistance systems minimize the vehicle{\textquoteright}s energy demand through speed optimization algorithms. Based on predictive route data, the required energy for upcoming operation points has to be determined. This paper presents a method to predict the energy demand of a hybrid electrical vehicle. Within this method, data-based approaches, such as neural networks, Gaussian processes, and look-up tables, are applied and assessed regarding their ability to predict the behavior of separate powertrain parts. The applied approaches are trained using measured data of a test vehicle. As a result, for every separate powertrain part, the best-suited data-based approach is selected to obtain an optimal energy demand prediction method. On a validation data set, this method is able to predict the transmission ratio of the gearbox causing a rmse of 0.426. The combustion engine{\textquoteright}s torque prediction results in an rmse of 19.01 Nm and the electric motor torque prediction to 19.11 Nm. The root mean square error of the motor voltage results to 1.211 V.",
keywords = "Energy demand prediction, Hybrid electrical vehicles, Systems modeling",
author = "Daniel Fink and Oliver Maas and Daniel Herda and Zygimantas Ziaukas and Christoph Schweers and Ahmed Trabelsi and Jacob, {Hans Georg}",
note = "Funding Information: Open Access funding enabled and organized by Projekt DEAL. The underlying project of this study was funded by IAV GmbH, Berlin, Germany. ",
year = "2024",
month = jan,
day = "11",
doi = "10.1007/s42979-023-02475-9",
language = "English",
volume = "5",
journal = "SN Computer Science",
issn = "2662-995X",
publisher = "Springer",
}