Numerical Analysis and Design Optimization of Permanent Magnet Synchronous Motor for Noise and Vibration Reduction

Abstract

Due to recent active research on motors, the power performance of motors has begun to meet the requirements of various applications, increasing interest in noise and vibration issues. The primary causes of noise and vibration in electric motors are the radial electromagnetic force and the resonance caused by the stator that receives this force. Many studies have been conducted to reduce these issues. However, previous studies have mostly evaluated noise and vibration performance at specific rotational speeds, not across a wide range of rotational speeds. Therefore, this study performs an optimal design to reduce motor noise and vibration, reanalyzes it, and compares the noise and vibration performance over a wide range of rotational speeds with the initial model.

To avoid resonance, which is the main cause of motor noise and vibration, the stator of the permanent magnet synchronous motor was optimally designed with the objective of maximizing the natural frequency from a mechanical improvement perspective. The optimal design was performed based on a surrogate model, which was fitted using polynomial regression on the first natural frequency data obtained from all combinations of design variable levels. The fitted polynomial showed high predictive power for the first natural frequency, and sensitivity analysis identified the design variables that most significantly affect the natural frequency.

Subsequently, the noise and vibration numerical analysis of the permanent magnet synchronous motor with the initial stator model and the one with the optimized stator model was conducted to compare their performance. After calculating the periodically varying electromagnetic forces over time, these forces were mapped as nodal forces onto the stator for noise and vibration numerical analysis. The noise and vibration analysis results confirmed that the maximum sound pressure level decreased by 22% due to resonance avoidance of the stator, as evaluated through a waterfall diagram that assesses sound pressure levels over a wide range of rotational speeds. Additionally, an order tracking analysis verified that the overall sound pressure level across a wide range of rotational speeds decreased by an average of 20%. The frequency components of the radial electromagnetic forces, which cause the most noise and vibration issues, and the corresponding dynamic characteristics of the stator were also identified.

This study applied a surrogate model based design optimization to maximize the natural frequency of the stator in a permanent magnet synchronous motor. It reanalyzed the permanent magnet synchronous motor with the initial and optimized stator models to compare noise and vibration performance over a wide speed range. The noise and vibration analysis results showed that the maximum noise caused by resonance was reduced, achieving noise and vibration reduction over a wide range of rotational speeds. The study also identified the vibration modes of the electromagnetic forces and the natural modes of the stator that cause the most noise and vibration issues. This study suggests a direction for improving the noise and vibration performance of permanent magnet synchronous motors in the future.