Permanent Magnet Synchronous Machine for Electrical Vehicles: Optimized Electromagnetic Loads for Better Vibro-Acoustics Performance

摘要

New challenges arise for engineers with the emergence of Hybrid Electrical Vehicles (HEV) and Battery Electrical Vehicles (BEV). As an electric motor has a lower overall noise emission compared to an Internal Combustion Engine (ICE), the overall noise level inside an EM driven car’s cabin is more colored by road noise and windnoise. That being said, the EM can still be very audible because of its prominent high-pitched tonal content, linked to the machine design (number of slots, poles). Hence, in view of driver and passenger comfort, reduction of cabin noise also involves optimizing the noise mission of electric motors. The noise produced by an electrical machine in operation can be assessed early on in the design process using simulation methods. This paper presents currently available approaches for assessing during the ideation phase the noise induced by a brushless Permanent Magnet Synchronous Machine (PMSM) which are widely used in electrical vehicles. The presented multi-physics workflow combines methods based on Finite Volumes to get the electromagnetic forces, and methods based on Finite Elements to compute the structural and vibro-acoustic response. The method demonstrated is first applied to an isolated PMSM to compute its vibro-acoustic performance in acoustic free-field. Additionally, the motor is then also placed in an engine bay in order to quantify the installation effects due to the presence of surrounding reflective and absorbing surfaces. Optimization of the electromagnetic force loading is also investigated in this paper. A multiple Objective Trade-off Study is performed to maximize the output power of the shaft and efficiency whereas to minimize cost due to the active materials and magnet weight within the design constraints. Evaluating around 3000 designs in the Electrical Machine Design Optimization, an optimized PMSM was selected. The same workflow in assessing the noise is applied to the optimized model. Finally the vibro-acoustic performance of the two models are compared.