Dynamic Control, Parameter Identification and Field Current Estimation for Electrically Excited Synchronous Machines

Doktorsavhandling, 2024

Electric vehicles (EVs) have experienced substantial growth in recent years. Permanent magnet synchronous machines (PMSMs), due to their simple control and high torque density, are currently the predominant choice of electric machines (EMs) in EVs. However, the permanent magnets (PMs) used in PMSMs are crafted from non-recyclable rare-earth materials, whose extraction and refining processes may cause severe environmental pollution. Hence, research on EMs free from rare-earth materials becomes crucial.
Electrically excited synchronous machines (EESMs) have emerged as a promising candidate. In an EESM, in addition to stator windings, there is also a field winding that generates the field flux, thereby avoiding the need of PMs. However, compared to PMSMs, the structure of EESMs is more complex, which also introduces new challenges. The aim of this thesis is to address these challenges encountered by EESMs when employed as traction EMs in EVs. Several topics are studied in this thesis. First, a dynamic current controller is developed that accounts for the magnetic mutual coupling between the stator and field windings. This current controller can effectively mitigate the disturbances caused by the induced voltages and reduce the current spikes during transients by more than 50%. Secondly, an optimization method is designed to determine the optimal selection of the d-axis, q-axis, and field current references during the online operation, which considers the minimization of total copper losses, prevention of winding overheating, and satisfaction of various current and voltage constraints. Moreover, two EESM parameter estimation methods are proposed. In the first method, measurements taken during transient processes are analyzed. By extracting and processing current error signals from the current controller, EESM parameters are estimated without the need for an additional observer. In the second method, an enhanced particle swarm optimization (PSO) is designed to estimate EESM parameters by analyzing steady-state measurements. Implementing these two methods allows for precise estimation of both apparent and incremental values of EESM self and mutual inductances. Lastly, an extended Kalman filter (EKF) based field current estimation method is developed to overcome the challenge of accessing field current information in brushless EESMs. The methods proposed in this thesis have all been experimentally verified. By implementing these techniques, the performance of EESMs can be significantly improved.