Vibration Analysis of Electrical Machine

Abstract

Recently industries that possess electrical machines mainly focuses on machine monitoring which involves many methods. Some of the methods are chemical, thermal and vibration monitoring. These methods require high accuracy sensors but in this case of vibration monitoring high accuracy sensors are not required which is emphasized in this work. The new approach is introduced to recognize the machine age based on vibration signal and the results are extracted by using signal processing techniques. Generally old machine creates huge vibration but in new machine vibrations are less observed. Our algorithm and techniques will easily recognize the machine type. In this paper, DWT (Discrete Wavelet Transform) and DWPT (Discrete Wavelet Packet Transform) used for recognizing old and new machines. Using these transform domain techniques, global threshold, threshold coefficient and statistical features like (entropy) were computed. From the results, it is convenient to recognize the machine's age and its lifetime.