Open-Access Testbench Data for NVH Benchmarking of E-Machines under Electromagnetic Excitations 2019-01-1459

This paper presents an experimental setup dedicated to the analysis of noise and vibration due to Maxwell magnetic forces in electrical machines, a significant NVH source in hybrid and electric vehicles traction motors. Both electromagnetic excitations and structural response of the electrical machine are simplified to provide the first public benchmark of e-NVH phenomenon (electromagnetic Noise, Vibration, Harshness).

The paper first describes how the testbench is designed and tested in order to reduce as much as possible modelling and experimental uncertainties. A Permanent Magnet Synchronous Machine topology used in EV/HEV applications is used to illustrate tooth modulation effect and interaction between radial and tangential force-induced vibrations, and designed to generate the resonance of several stator structural modes with simplified electromagnetic loading (open-circuit case). A larger air gap allows the insertion of a fine search-coil network to measure time and space distribution of the air-gap flux density and resulting Maxwell stress harmonics. Accelerometers are placed on stator tooth tips to capture tooth bending motion, as well as on the outer yoke of the stator. Besides vibration measurements, sound pressure and sound power level measurements are carried.

Then, some of the key measurement results are presented including Experimental Modal Analysis, Operational Deflection Shapes, and Order Tracking Analysis, spectrograms and spatiograms, Sound Pressure Level and Sound Power Level measurements. The origin of the different NVH harmonics are analyzed and their physical origin is detailed.

All the benchmark data is available in open access and can be used to compare different multiphysic simulation strategies of e-NVH in terms of accuracy and computing time, such as analytic, semi-analytic, numerical and hybrid methods using during electromagnetic, structural mechanics and acoustic calculations. The benchmark will be used in further work to study the most common noise mitigation strategies used in EV/HEV electric motors such as skewing, notching, pole and slot shaping, and harmonic current injection.