Optimal DC-link voltage adaptation for minimum switching losses in battery electric vehicles

Electrification is one of the macrotrends that mostly influenced research and innovation in road transportation across the last decade, aiming to provide solutions for the efficient development of EVs’ components and architectures. In this context, the European founded H2020 project FITGEN (Functionally Integrated E-axle Ready for Mass Market Third GENeration Electric Vehicles) aims at designing and testing an innovative and brand-independent e-axle for the next generation of BEVs, validated on an A-segment vehicle. The developed e-axle integrates a high speed 6-phase electrical motor and SiC inverter, a DC/DC converter boosting the battery voltage to the 800 V DC range and a single speed transmission gearbox. Among the other features of the FITGEN e-axle, the DC/DC converter permits to online adapt the inverter DC-link voltage for maximizing the efficiency of the power electronic converters. This thesis, developed at Politecnico di Torino, takes part to the FITGEN project, and in particular focuses on the evaluation of a variable DC-link control strategy for maximum efficiency of traction motor drive. In order to guarantee the controllability of the traction drive in the whole speed/torque operating range and during the whole propulsion system lifetime, particular attention was devoted to the computation of the optimal set-point of the controlled electrical variables. Based on the machine model, accurate look-up table (LUTs) were computed to define the current and flux vector minimizing the Joule loss in the entire operating range. This includes MTPA tracking and FW operation, both under variable DC-link. Dealing with the DC/DC converter, the computed LUTs also permit to define the optimal DC-link voltage on varying the motor torque and speed. At first, based on the computed LUTs, the torque and the optimal DC-link voltage references were directly imposed, and the system response was tested in the simulation platform. Anyway, since the LUTs are computed based on the motor model, accurate knowledge of its parameters, and particularly the magnetic saturation characteristic, is necessary. Therefore, two control strategies, namely “Control type 1” and “Control type 2”, were analyzed for automatically adapting the DC-link voltage to minimize the losses in the inverter and in the DC/DC. Particular effort was devoted to the optimal calibration procedure, requiring a tradeoff between fast motor control dynamic with reliable safety margin and power loss minimization at medium speed. Additionally, the calibration procedure needs to consider the relevant actuation delay of the DC/DC converter. One of the key features of these two solutions is that the optimal DC-link adaptation is almost independent by the adopted motor control strategy. To prove this assert, the global system, including the drive and the DC/DC, was tested in the Matlab-Simulink environment with two motor control algorithms, namely CVC-FOC and DFVC. Despite the reference FITGEN drive is a 6-phase system, the DC/DC control was initially tested for an equivalent 3-phase motor for debugging and calibration purposes, simulating the two three-phase sets connected in parallel. Then, the DC/DC control has been extended to the 6-phase case. Moreover, to confirm the expected performance of the proposed control strategy, the obtained aggregate losses reduction (DC/DC converter + inverter) was numerically evaluated based on a loss analysis of the two converters.