AC_SCIM_002

Objective

Topology Squirrel Cage Induction Machine (SCIM)
Model Types Permeance / mmf electromagnetic model EM1.SCIM
Analytical structural model SM1
Analytical acoustic model AC1
Quantities Sound power level during run-up (SWL)
Geometry Internal Rotor _Semi closed slots
Winding Shorted pitch distributed winding

The aim of this validation case is to compare MANATEE results (obtained in a few seconds of computation with permeance/mmf model) with accurate experimental tests run in a semi-anechoic chamber using ISO 3744 standard to calculate the sound power level radiated by a traction induction motor.

Machine

The studied machine is a traction squirrel cage induction machine with p=2 pole pairs, Zr=38 and Zs=48 stator slots.

Machine SCIM_010 Topology
Machine SCIM_010 Topology

Results

The following graph favorably compares the calculation of natural frequencies with 2D FEA used in the article [R1]. For variable speed applications the accuracy on natural frequencies is less important, but if necessary MANATEE can be directly coupled to structural FEA for the computation of natural frequencies and Frequency Response Functions.

Natural frequencies comparison
Natural frequencies comparison
FEA (article) vs Analytical (MANATEE)

The next graph compares the calculation of the sound power level with MANATEE and during tests (run-up at no load). Measured noise including mechanical, aerodynamic and magnetic acoustic noise sources. At high speed, acoustic noise level is driven by aerodynamic noise due to fans. Aerodynamic noise is not calculated by MANATEE which explains the difference between MANATEE and tests. One can observe that two strong resonances occur close to 1250 rpm in tests, while only one resonance is observed in MANATEE simulation. This is probably due to the fact that the analytical model considers a cylindrical symmetrical stator, whereas the real machine is asymmetrical which leads to double structural modes. Another explanation could be the excitation of a secondary longitudinal mode which is not included in this MANATEE simulation based on 2D analytic models. However, the strong resonance is correctly predicted by MANATEE and the noise level given by MANATEE is close from experiments. This means that MANATEE can be efficiently use in the early electromagnetic design process to avoid this type of vibro-acoustic resonance. In that particular case the resonance is due to saturation effects which introduce a excitation force of wavenumber 2=Zs-Zr-4p, resonating with the stator stack elliptical mode m=2.

40. Comparison of sound power level (MANATEE vs Experimental measurements)
Comparison of sound power level (MANATEE vs Experimental measurements)

References

[R1] Nguyen, M. K., (2014). Predicting Electromagnetic Noise in Induction Motors, Master’s Thesis, Royal Institute of Technology of Stokholm, Sweden.

Previous Next