[Zurück]

M. Watscher:
"Beurteilung und Vergleich verschiedener Spannungsmodelle zur sensorlosen Regelung einer permanenterregten Synchronmaschine im Grunddrehzahl- und Feldschwachbereich";
Betreuer/in(nen): E. Robeischl; Institut für Elektrische Antriebe und Maschinen, 2002.

This diploma-thesis deals with the analysis of different models designed for sensorless control of industrial permanent magnet synchronous motors (PSM) in ground speed range and in the field-weakening region. In this context sensorless means the replacement of the mechanical sensor for detection of the rotor position. The omission of the mechanical sensor has the following positive impacts. First, the drive can be produced more cost-effectively and second, the system gets more robust.
The subject of the analysis includes only velocities higher than the 0.2-fold nominal speed. In this particular speed range the back EMF of the machine can be used in order to determine the position of the rotor without a sensor. In the present case the rotor position is evaluated according to two different models: the integrated voltage model and the short-circuit model.
The first part of this diploma-thesis deals with the operation in ground speed range and in the field weakening region on a sensor-controlled basis. The field-weakening region has to be specifically taken into account because the machine should be operated at highest speeds and with limited DC-link voltage of the voltage source inverter (VSI). The main emphasis is put on the appropriate guideline of the field weakening flux-parallel current-component of the machine. Following that, the above-mentioned models are realized in a step-by-step approach to replace the mechanical sensor. In this context the research focuses especially on the impact of field weakening current on the achieved results. Depending on the applied model and on the operating point of the machine, certain systematic errors occur relating to the sensorless determination of the rotor position. In the course of this diploma-thesis theses errors are discussed and corrected.
In addition to that, two different strategies are examined, to implement the control structure, which is entirely implemented in a digital 16-bit fixed point digital signal processor (DSP). First, a structure is demonstrated according at which all parts of the program are run through in one sampling cycle. The disadvantage of this structure is that due to the necessary calculations, the sampling time has to be determined relatively long, which causes problems with regard to the highest speed range. Therefore another structure is demonstrated, where certain parts of the program (current controller, extrapolation of sensorless estimated rotor position) are realized within faster interrupt-based tasks in order to achieve higher dynamics of the control structure and to reach a noise reduction and an extension of the maximum possible speed range with the short-circuit model.
Finally, this diploma-thesis compares and evaluates the models with regard to the accuracy of their position estimation with the help of statistic measurements.