Experimental investigation of the sound behavior of a three-phase asynchronous machine under variable frequency operation

Abstract

In industry, where acoustic characterization is becoming the preferred non-invasive technique, the diagnosis of problems in three-phase asynchronous machines is essential. Without the requirement for sensors installed directly on the machine, it makes it possible to identify fault-specific sound signatures, hence enabling defect anticipation. Acoustic analysis is notable for its speed and precision when compared to other methods like thermography. This is especially true when determining the influence of electrical harmonics and inverters, which have a significant impact on noise and vibration. These auditory effects are exacerbated by the usage of variable speed drives, which impact the machine as well as the comfort of the staff’s hearing. Although various fault types can be identified using intelligent sound classification techniques, databases still require enrichment for the best categorization. By combining acoustic and electrical data, this research investigates the acoustic characterization of a three-phase asynchronous machine with a frequency converter in an effort to identify failure indicators early and improve diagnosis accuracy while minimizing invasiveness. • Study sound signals from a Variable Frequency Drive, focusing on low frequencies, to assess the acoustic behavior of a three-phase asynchronous machine. • Measure and evaluate parameters like THD, SNR, and other statistical features to distinguish between healthy and faulty machine states. • Collect acoustic emissions and stator current signals simultaneously to explore their relationship, improving fault detection accuracy and enabling intelligent classification algorithms.