Control of Permanent Magnet Synchronous Motors for Automotive Applications

Due to its high power density and very high efficiency, the permanent magnet synchronous machine (PMSM) is the preferred motor topology in today's automotive applications. One major distinctive criterion is the location of the magnets. PMSMs with interior or buried magnets, so-called IPMSM, have an even higher power density than those with surface-mounted magnets (SPMSM), which represent the standard in industrial applications so far. This advantage can be ascribed to strongly asymmetric reluctances in IPMSM, which make the modeling and control design more complex, however.

The department's research focus in this area can be summarized as follows:

IPMSM Modeling

  • Electrical (Saturation, Iron Losses)
  • Thermal (Observer, LPTN)

Efficiency Optimization

  • Selection of Operating Point
  • DC-Link-Voltage Adaptation

Control Design (FOC/DTC)

  • Conventional (PI-Controllers
  • Model Predictive (MPC)
  • FPGA-based

A typical control structure for an electrical drive system is depicted in Fig. 2. With respect to the achievable control performance, it has been proven to be advantageous to control the IPMSM in rotor-fixed coordinates. The crucial components in this context are the current controllers, which guarantee that the actual (measured) currents in the motor follow their reference provided by the operating point selection block. The design of the controllers can be performed in various ways, strongly depending on the required dynamics, robustness and available computation power. In this context, precise machine models, which account for varying inductivities due to iron saturation, can be utilized in the control design to improve the overall control performance. Therefore, LEA developed a semi-automatic method to generate motor specific LUTs through offline measurements. A further challenge lies in the rather small number of sampling instants per electrical rotation, especially at high motor velocities. In this context, advanced model discretization techniques, <del cite="mailto:DEFIGGH" datetime="2011-08-11T14:19"> </del>along with measures such as regular sampling become crucial to maintain a good control performance.

The reference operating points are selected by taking into account the desired torque T*, the DC-link voltage uDC, the electrical angular velocity wel as well as the previously generated motor-specific LUTs. Fig. 3 displays such a LUT. Regarding the constant torque trajectories (green), it becomes obvious that mapping the torque to an operating point i*d, i*q is not unique. The degree of freedom can, however, be made use of to select an 'optimal' operating point within the reasonable operation range (yellow), while taking into account the given current and voltage limits (red and purple). Therefore, a target function representing the optimization criteria needs to be determined. Possible criteria are the minimization of ohmic losses (Maximum Torque per Current - MTPC) or total losses (incl. iron losses) in the motor (Maximum Efficiency - ME).

In order to protect the motor from heating up beyond its specified limits or to include temperature variations in the operation point selection, the acquisition of temperatures at spots of interest (e.g. magnets, stator windings) becomes necessary. In order to avoid costs related to the employment of temperature sensors, model-based approaches are preferred. Figure 4 and 5 show two examples for a model-based temperature acquisition. In Fig. 4, the rotor temperature is detected via an observer structure based on the voltage equations of the IPMSM. The idea is to utilize the temperature dependency of electrical variables, such as the resistance of the stator windings or the permanent magnet flux, to deduce temperatures from parameter variations. In Fig. 5, on the other hand, a generic Lumped Parameter Thermal Network (LPTN) is depicted. Those LPTNs can describe the actual thermal characteristics of the motor. Both approaches have unique advantages and drawbacks, for which reason they are both part of our research focus.

References:

[1] J. Böcker, T. Huber, A. Specht
Elektromobilität in Paderborn -  Kerntechnologien für Automobile von morgen
Forschungsforum Paderborn Ausgabe 14-2011
[2] Joachim Böcker, Tobias Huber, Michael Meyer, Wilhelm Peters, Daniel Pohlenz, and Andreas Specht
Control of Electric Propulsion Drives in Electric and Hybrid Vehicles
Automatisierungstechnik, October 2010, Vol. 58, No. 10: 599-608
[3] A. Specht, J. Böcker
Observer for the Rotor Temperature of IPMSM
EPE-PEMC 2010, Ohrid, Makedonien, Sept. 2010
[4] D. Pohlenz, J. Böcker
Efficiency Improvement of an IPMSM using Maximum Efficiency Operating Strategy
EPE-PEMC 2010, Ohrid, Makedonien, Sept. 2010
[5] T. Grote, M. Meyer, J. Böcker
Direct Torque Control für den verlustoptimalen Betrieb eines Permanentmagnet-Synchronmotors mit eingebetteten Magneten
Internationaler ETG-Kongress 2007 – Hybridantriebstechnik - Energieeffiziente elektrische Antriebe, Karlsruhe, Deutschland, 2007
[6] Meyer, M.; Böcker, J.
Optimum Control for Interior Permanent Magnet Synchronous Motors (IPMSM) in Constant Torque and Flux Weakening Range
12th International Power Electronics and Motion Control Conference (EPE-PEMC), August 2006, Portoroz, Slowenien
[7] Meyer, M.; Böcker, J.
Transient Peak Currents in Permanent Magnet Synchronous Motors for Symmetrical Short Circuits
Int. Symposium on Power Electronics, Electrical Drives, Automation and Motion (SPEEDAM), Mai 2006, Taormina, Italien
business-card image

Dr.-Ing. Frank Schafmeister

Power Electronics and Electrical Drives

Interim Professor Power Electronics and Electrical Drives

Write email +49 5251 60-3881