Electric vehicle drivetrain optimisation

Abstract

This study provides a detailed analysis of an optimal drivetrain configuration, based on multi-cycles, for a plug-in electric vehicle (EV). The investigation aims to identify the best EV configuration according to the required power and the transmissible traction torque. The study focuses on an EV with four different combinations of drive systems among in-wheel motors and differential ones. To find out the best EV drive system configuration, it is adopted an optimisation process by means of a genetic algorithm that defines the electric motors (EMs) torque curves and powertrain transmission ratio in order to improve vehicle travel range and performance. The vehicle power demand is divided between the drive systems following rules established by the power management control which aims to reduce the lithium-ion battery discharges during the driving cycles: FTP-75 (urban driving), HWFET (highway driving) and US06 (high speeds and accelerations). After the simulations, the potential of each configuration is indicated according to its respective drive system and hence the best configurations are determined.

Energy storage and control optimization for an electric vehicle

“…The vehicle longitudinal dynamics was modeled by the equations proposed in Gillespie 18 that are based on the movement resistance forces. The equations were adapted as in Eckert et al 13 for an EV propelled by 4 in-wheel EMs, 2 at the front wheels, and other ones with different power/torque curves at the rear.…”

“…The EMs used in the simulations are similar to the ones used in Eckert et al 13 Figure 3 shows a generic EM torque curve and efficiency map that contains torque and speed points used by the optimization algorithm to define the most adequate EMs configuration for each EV propelling system. To define optimum EMs torque curves, the algorithm determines the EM maximum torque T max (Nm) and the speed Tc (rad/s) that characterizes the constant torque phase.…”

“…The battery is a lithium-ion model provided by the Simulink database. The battery state of charge SoC is limited to 40% as in Eckert et al, 13 to prevent battery damage caused by the complete battery discharge. 22 The battery mass M bat (kg) is defined by Equation 15 according to the lithium-ion specific energy S E = 120 (Wh/kg) and also by the battery voltage V bat (V) and capacity B c (Ah) that will be defined by the optimization algorithm.…”

“…The third optimization criterion is related to the EV performance as compared with the velocities profiles of the simulated cycles. In this paper, the vehicle performance is evaluated by the correlation coefficient R as in Eckert et al 13,27 The R value defined by Equation 24 provides a numerical value that measures the vehicle capacity to reproduce the target cycle speed profile, comparing the results performed by the current vehicle speed V of EV in the first cycle loop simulation (Figure 2) with the target speed values V c (considering discrete time intervals i of 0.1 s).…”

“…In this paper, the simulated EV is based on the optimized configuration defined in a previous work presented in Eckert et al 13 that showed the best trade-off between vehicle performance and battery discharge. This EV configuration is propelled by 2 pairs of in-wheel electric motors (EMs) (with different characteristics) connected directly to the vehicle frontal and rear wheels.…”