Condition monitoring of gear boxes in real time

Analysis of the vibration signals is one of the most powerful techniques available for determining the condition of operating machinery. Most serious faults will result in an increase of vibration level before actual breakdown, and thus allow shutdown before catastrophic failure. One such machinery is a gearbox, which finds its application in many areas. Spectrum analysis helps in converting the time-domain signal from the gearbox into the frequency-domain, thus enabling identification of the troubling frequency. But this tends to become difficult as the signal becomes complicated due to the multiple gear pairs in a gearbox. Therefore, in order to get a clearer picture of the condition of the gearbox, another mathematical tool called Cepstrum is needed which separates the side bands and makes the diagnosis easy. Here we present this tool, which was developed by us, and its results.