Predictive Energy Management for Battery Electric Vehicles with Hybrid Models

TL;DR

This work tackles BEV energy consumption prediction under variable conditions by coupling a physics-based BEV model with data-driven residual corrections. It assesses three corrective approaches—GAMMs, Random Forests, and Boosting—using leave-one-out cross-validation on 72 trips from the TUM dataset, and finds that hybrid models substantially outperform purely physics-based estimates. The boosting-based hybrid delivers the best performance, reducing average error from $0.379$ (physics only) to about $0.103$–$0.115$, demonstrating that external factors and driver heterogeneity can be effectively captured with data-driven corrections. The approach also separates vehicle dynamics from external influences, enabling easier transfer to different vehicles or driving styles and showing promise for robust energy forecasting under varied HVAC and environmental conditions.

Abstract

This paper addresses the problem of predicting the energy consumption for the drivers of Battery electric vehicles (BEVs). Several external factors (e.g., weather) are shown to have huge impacts on the energy consumption of a vehicle besides the vehicle or powertrain dynamics. Thus, it is challenging to take all of those influencing variables into consideration. The proposed approach is based on a hybrid model which improves the prediction accuracy of energy consumption of BEVs. The novelty of this approach is to combine a physics-based simulation model, which captures the basic vehicle and powertrain dynamics, with a data-driven model. The latter accounts for other external influencing factors neglected by the physical simulation model, using machine learning techniques, such as generalized additive mixed models, random forests and boosting. The hybrid modeling method is evaluated with a real data set from TUM and the hybrid models were shown that decrease the average prediction error from 40% of the pure physics model to 10%.

Figures (9)

• Figure 1: Schematic view of utilized physic-based simulation model.
• Figure 2: Energy prediction for a trip utilizing proposed hybrid model.
• Figure 3: Model diagnosis for GAMM with Student’s t family fitted on TUM data set.
• Figure 4: Estimated smooths from the GAMM with Student’s t family fitted on the TUM data set.
• Figure 5: Overall error distribution from LOOCV for GAMM with Student’s t family fitted on the TUM data set.