Vibroacoustic optimization of electric motors taking advantages of viscoelastic resins

Abstract

Electric drivelines are the source of tonal noise which can be particularly unpleasant in many industrial settings, and in particular in vehicles. Different approaches are investigated to minimize this phenomenon, based on the control of the electromagnetic source for example, or on the optimization of the structural design. This paper introduces a framework allowing time-efficient reduction of noise radiation without affecting electromagnetic performance of electric motors, based on the combination of reduced order model and optimization of viscoelastic properties of potting resins. Although viscoelastic resins are frequently used in electric motors, they are primarily used for insulation, protection, and mechanical stability. The originality of this work is to endow them with a new functionality by dimensioning them in such a way as to reduce acoustic radiation. Indeed, the elastodynamic properties of resins depend on the frequency and on the temperature, and can be optimized to reduce the vibrations and thus the radiated noise. The design is complex because it involves coupled multiphysical phenomena that cannot be considered separately, and because the numerical models considered are large-scale models. A methodology is proposed in this paper and applied to the case of a high-speed electric motor. To guarantee fast and accurate estimation of the acoustic power in the optimization process, a reduced order model of the engine is developed from a multi-model basis taking into account the thermal and frequency dependencies of the materials. The numerical optimization thus carried out allows to identify the optimal resin properties, and a motor is coated to experimentally validate the results. Experimental characterization thus confirms that the radiated noise can be considerably reduced by using a suitable resin. This work opens the way to a new design strategy for electric motors providing resin coating with an acoustic function.