Sound Analysis for Speed Estimation of Induction Motors Under Non-Stationary Conditions

This paper presents a novel methodology for the estimation of rotational speed of induction motors under non-stationary conditions, based on the analysis of acoustic signals captured by a low-cost microphone. The proposed approach employs multi-rate and time-frequency processing techniques to extract speed-dependent harmonic components from the sound emitted by the motor, effectively isolating them from noise and interfering harmonics. Experimental validation is performed under both steady-state and transient operating conditions, demonstrating the method's capability to accurately track dynamic speed variations caused by load changes and mechanical irregularities. The estimated instantaneous speed closely matches reference measurements obtained from a speed sensor. The results confirm that the proposed acoustic-based technique provides a reliable, non-invasive, and cost-effective solution for speed monitoring of induction motors.