Road Section Weather and Vehicle Energy Consumption Modeling

Abstract

Accurate vehicle energy consumption modeling for heavy-duty trucks requires precise characterization of aerodynamic drag and rolling resistance, both of which are highly sensitive to local weather and road conditions. Existing models often rely on standard atmospheric assumptions or dedicated on-board sensors, which limits their ability to capture real-world operating conditions. To address this, the present work proposes an integrated physics-based modeling framework that links road-segment level meteorological data to vehicle motion resistance without requiring additional sensing hardware. For the wind prediction model, a topographical model is used as a base and further corrected with a two-layer wind speed correction model developed to translate regional meteorological station measurements into local road-level wind estimates. The successive layers account for a global power law bias and geographic environment effects through clustering of road segments into sheltered and wind-amplified categories. Validated against over 200 repeated real-world journeys on RV (Riksväg) 40, Sweden, the model achieves a 48.4% reduction in mean absolute error over the regional baseline, consistently across seasons and travel directions, compared to using measurements from regional meteorological stations. A wind direction model in the form of a polynomial is incorporated to complete the wind characterization. For the rolling resistance component, a two-state tire thermodynamic observer is formulated, estimating tire bulk temperature from standard Tire Pressure Monitoring System (TPMS) measurements and an external temperature sensor via a Kalman filter. This enables model-based prediction of temperature-dependent rolling resistance as a function of ambient conditions and driving dynamics, without dedicated thermal sensors. The proposed framework provides a scalable, low-cost approach to weather-aware energy consumption forecasting, directly applicable to range prediction and route planning for heavy-duty trucks operating under dynamic real-world road section conditions.