Loss-optimal control of electrically excited synchronous machines without position sensor in automotive traction drives
Project goal
Development of a loss-optimal torque operation strategy for electrically excited synchronous machines (EESM) in the traction environment, as well as a high-performance sensorless current control algorithm for its implementation.
Due to the high cost of high-voltage batteries in electric vehicles, increasing the efficiency of the electric powertrain is one of the key development drivers in the field of electromobility. In addition to various development trends in the area of frequency converters, the selection of the electric machine also represents a degree of freedom for optimizing the overall system.
Since the currently dominant permanent-magnet synchronous machines (PMSMs) are increasingly being operated outside their optimum range, especially at higher speeds, due to the need for field weakening, the development trend is toward the use of EESMs that can still be operated with maximum efficiency in the partial load range even at high speeds. A prerequisite for this, however, is a loss-optimized operating strategy that selects the electrical operating point in real time so that the torque demanded by the driver is implemented with minimum system losses, taking into account all relevant system states (including speed and DC link voltage).
However, the subsequent realization of the optimum current operating point is more challenging for EESMs than for PMSMs, since they represent a nonlinear multivariable system due to the saturation-related nonlinearities and the strong magnetic coupling between the stator and rotor circuits. In order to meet the high dynamic requirements common in the automotive sector, advanced concepts for controller decoupling and linearization, as well as for intelligent setpoint control, are therefore necessary.
In addition to the efficiency advantage over PMSMs, EESMs are also characterized by the fact that the rotor position required for field-oriented control can be determined from electrical variables independent of saturation, even at low speeds, and thus a position sensor can be dispensed with. However, the detection and observer algorithms used in this process must be robust in interaction with the field-oriented current control with respect to the highly dynamic operating point changes that occur due to the loss-optimal operating strategy.
The Ostbayerische Technische Hochschule Amberg-Weiden is working on the research project as part of a cooperative doctorate with the Technische Universität Darmstadt.
Project duration
July 2022 to August 2026 (50 months)
Project staff
Jan Herold