In this paper, a dynamic modeling and signal processing framework integrating Doppler compensation is proposed for the accurate prediction of broadband noise characteristics in external-rotor permanent magnet synchronous motors (ERPMSM). First, a multiphysics coupling model incorporating tangential electromagnetic forces is established, with its spatiotemporal distribution validated through finite element analysis. Subsequently, multi-channel synchronous signal acquisition is implemented using a uniform circular microphone array, combined with a time-varying spatial filtering and reordering strategy to dynamically correct spectral distortion induced by time-varying sound propagation paths. Furthermore, a boundary element method modified by a dynamic Green's function is employed to precisely characterize the radiation behavior of moving acoustic sources. Experimental results demonstrate that the proposed method significantly reduces sound pressure level prediction errors, enhances the identification accuracy of high-frequency harmonics, and maintains stable spatial directivity prediction capability under all operating conditions.
30 October 2026 in International Journal of Structural Stability and Dynamics