Measuring and modelling of core losses under non-sinusoidal waveforms and further development of the in-house post-processor for estimation of stator core losses in permanent magnet synchronous machines
- type:Masterarbeit
- time:20.04.2015
- tutor:
- Image:
Artem Rodionov
-
The precise estimation of the core losses in the electrical steels helps the efficiency optimization of the electric machines. But the phenomena like the rotational field vector and the non-sinusoidal flux density waveforms are not regarded seriously in the conventional finite element (FE) software which makes the precise estimation difficult. In this thesis different kind of electrical steels should be measured under sinusoidal as well as non-sinusoidal waveforms and a validated model should be integrated into the in-house post-processor to estimate the core losses in the stator of permanent magnet synchronous machines.
An extensive literature research about the core losses models under non-sinusoidal waveforms should be firstly carried out with regarding the charged hysteresis losses due to the minor-loops. To validate these models different kind of non-oriented electrical sheets should be measured under standard sinusoidal as well as with non-sinusoidal waveforms with help of the in-house Epstein and the toroidal test benches. The parameters of the core losses models should be determined for these electrical sheets and the quality of the models should be compared. In the next step the most suitable model (eventually with own modification) should be chosen for the further calculation.
Because of the fact that the non-sinusoidal waveforms of flux density appear in the stator due to the high saturation grad, the slotting harmonics and different combinations of the lead angle and the phase current, a detailed analysis of the simulated flux density waveforms in the given FE-model should be carried out with help of the in-house post-processor. Then the core losses model should be integrated into the post-processor and the losses in the stator should be calculated. The simulation’s results of the new model should be compared to those of the standard one. At the end the measurement results of a single tooth can be compared to the simulated losses.