Data-Driven Modeling and Simulation Approach for Energy Demand Prediction of Off-Road Electric Truck

As the adoption of electrical transportation continues to rise, understanding and addressing performance degradation becomes increasingly critical to ensuring reliability and efficiency. This paper proposes a data-driven modeling and simulation approach for off-road electric trucks (e-truck). It employs a comprehensive simulation framework incorporating vehicle dynamics, terrain variability, and operational conditions to predict energy consumption with high precision. The model includes representations of the control system, electrical motor, regenerative energy mechanisms, and vehicle dynamics. This approach links the e-truck physical modeling to the realistic input data such as road conditions, truck loading, and driving behavior to provide valuable insights into the realistic operational performance of the e-truck. The developed approach addresses two main simulation models. The first one is the Terra-mechanics model that is developed in the Vortex environment to estimate the e-truck rolling resistance for different soil types. The second is the energy model, which is designed in Maplesim to estimate the instantaneous power demand and the energy required to perform a mining cycle. In the specific mining case study, rolling resistance and e-truck loading are the top sensitive variables in the energy prediction model. The fully loaded e-truck consumes 52.5% energy more than the empty truck, on average with 5% rolling resistance. The estimated energy demand increases by 55.9% for each percentage increase of rolling resistance, on average with the maximum e-truck loading. Detailed route simulation and sensitivity analysis were reported.