Thorough Analysis of the Reliability of Measurements on Chassis Roller Dynamometer and Accurate Energy Consumption Estimation for Electric Vehicles

Abstract

The energy consumption of the electric test vehicle on the roller dynamometer test bench is thoroughly investigated in this work. Initially, we extensively examined the relevant testing approaches on chassis dynamometer test benches. Then, we presented the main issues and limitations of this type of testing facility. After that, an experimental analysis was performed by recording the NEDC and the WLTP-class 2 standard driving cycle test data. It was found that the power consumption associated with specific test maneuvers could be accurately simulated via the roller dynamometer system. However, the experimental test results showed some disparities compared to the expected results from the simulation models. These findings supported identifying the dynamic behavior resulting from the tire and roller contact. The LuGre friction model provided the solution, incorporating friction-related physical models such as stick-slip and the Stribeck effect. In addition, adjustments were made to the LuGre model to consider the dynamic influences of the roller on the tire. Finally, we created a systematic estimation model for energy consumption by integrating the boundary conditions of the test bench, a generic model for energy estimation of electric vehicles, and the proposed tire–roller contact model. The overall model requires only the measured traction speed of the test vehicle and the auxiliary energy losses as physical inputs. The results proved that the proposed model enabled the estimation of the total power consumption with high accuracy.

Predicting the Torque Demand of a Battery Electric Vehicle for Real-World Driving Maneuvers Using the NARX Technique

“…J. 2024, 15, x FOR PEER REVIEW 3 of Developing torque demand predictive models became a helpful approach toward a curate energy consumption estimation during actual driving routes [14][15][16][17]. The behav of a separately excited DC motor in the paper in reference [4] is recognized using NAR neural networks.…”