Assessing Digital Elevation Models for Battery Electric Buses Energy Prediction

Energy transition from combustion engines to electric vehicles in urban mobility is key to face climate change, especially in countries with clean, renewable electric matrices, such as Brazil. As battery electric bus (BEB) acquisition costs remain a barrier to the public transportation energy transition, energy prediction models are being developed to improve fleet management, route planning, and reduce operational costs. One of the main variables influencing energy consumption is the road slope; however, the accuracy of elevation data used to calculate the slope input in energy models is not being taken into consideration, and therefore, estimates diverge from actual consumption. Intending to improve energy prediction, this paper assesses seven Digital Elevation Models (DEMs), comparing their altitude values, calculated slopes, and resulting energy predicted values for a BEB route in the University of Campinas (Unicamp) campus. The energy prediction is performed using a physical energy model and compared with real consumption values acquired with the BEB’s Controller Area Network (CAN) module. Beyond the accuracy analysis, the DEMs’ cost-benefit is discussed. Among the evaluated elevation models, FathomDEM, an open-source dataset, showed the best overall performance in energy prediction, despite its lower spatial resolution. LiDAR- and photogrammetry-derived models also achieved high prediction accuracy, albeit with higher acquisition costs. Results confirm the significant impacts of the slope on energy consumption estimates, justifying the importance of this paper.