Abstract
Fractional pitch slot / pole arrangements are commonly used in permanent magnet synchronous motors (PMSMs) to significantly reduce cogging torque, however, maximum benefit depends on accurate stator and rotor manufacturing and alignment. This thesis presents a method of identifying the sources of manufacturing induced cogging torque.Decoupling of cogging caused by stator and rotor manufacturing variation is possible due to stator and rotor affected harmonics being independent of one another. A hybrid FEA / analytical method was developed to simulate cogging torque with multiple manufacturing faults induced which proved to be 6 orders of magnitude faster than FEA alone. The new method utilises a library of FEA derived pole transition over single stator slot waveforms which are then assembled in the correct order and phased corresponding to the slot / pole interactions for a given PMSM.
Angular and eccentric misalignment can be identified by investigating the presence of first and second order sidebands about the pole and slot harmonics. Static angular misalignment induces first order sidebands around the slot harmonics while dynamic angular misalignment induces sidebands about the pole harmonics. Static eccentricity induces both first and second order sidebands around the slot harmonics, while dynamic eccentricity has the same effect on the pole harmonics.
A ‘Fault Diagnosis Flow Chart’ is presented to identify the manufacturing sources of additional harmonics. To demonstrate the effectiveness of the proposed method, all combinations of 10 production stators and 10 rotors of 24 slots and 10 poles in an axial flux configuration were experimentally measured and analysed. The stator slot variation and pole misplacement were found to be the largest contributors to unexpected cogging torque. Pole strength variation and static angular misalignment had minor contributions while dynamic angular, static eccentricity and dynamic eccentricity were found to not have a significant impact on the motors tested.
Date of Award | Nov 2013 |
---|---|
Original language | English |
Supervisor | Friso De Boer (Supervisor) & Greg Heins (Supervisor) |