Bearing fault detection using neural networks
Mayssa Hajar, Mohamad Ali Khalil · 2012
In this paper we present an application for the Artificial Neural Networks in industrial monitoring domain. Because it is frequently defected, we chose the bearing element to diagnose. Applying the pattern recognition principle, we used the bearing vibration signals which represent this element in its different states (normal, defected and severely defected) to extract their power spectral density parameters and classify them using the Feed Forward Neural Networks. The performance of the used networks reached 100% in some cases.