Electro-thermo-mechanical frequency-decoupling for multi-modal NVH suppression in high-torque EV motors

Higher torque density in electric vehicle (EV) traction motors (>200 N·m) increases electro-thermo-mechanical coupling effects, which makes noise-vibration-harshness (NVH) a critical design constraint. Resonant excitation occurs when structural modes coincide with electromagnetic force harmonics. To address this, an Electro-Thermo-Mechanical Frequency-Decoupling (ETMFD) framework is proposed. The ETMFD framework integrates multi-physics modeling—combining Maxwell stress tensor analysis (0–10 kHz), modal strain energy mapping, and thermal conduction equations—with triple-domain decoupling. The ETMFD framework shifts the critical mode from 1045 to 1220 Hz through yoke thickening (+22%) and slot width reduction (–15%), and targeted damping applied to strain energy hotspots. An experimental test on a 70 kW PMSM shows 6.3 dB(A) noise attenuation in the 1–2 kHz band, 31% reduction in (2,0)-mode vibration, and maintained torque accuracy (230.8 N·m, error: +0.3%) with 94.2% efficiency at 3714 rpm. The proposed framework reduces manufacturing costs by $6.8 per unit and shortens development cycles by 30%, providing a systematic paradigm for the optimization of NVH in high-torque EV motors.